r*:,. II ^f mM- ITEO STATES ."ITMENT OF AMERCE LIGATION i(- "^ ^H <:.:..■ >*^ ^^^1 Fishery Bulletin U.S. DEPARTMENT OF COMMERCE National Oceanic and Atmospheric Administration National Marine Fisheries Service J.i^-fn.^ Boicgical Uborcitory LIBRARY FEB 14 1372 Woods Hole, Mass. Vol. 70, No. 1 January 1972 MATSUMOTO, WALTER M., ELBERT H. AHLSTROM, S. JONES, WITOLD L. KLAWE, WILLIAM J. RICHARDS, and SHOJI UEYANAGI. On the clarification of larval tuna identification particularly in the genus Thunnus 1 PARSONS, T. R., K. STEPHENS, and M. TAKAHASHI. The fertilization of Great Central Lake. I. Effect of primary production 13 LeBRASSEUR, R. J., and O. D. KENNEDY. The fertilization of Great Central Lake. II. Zooplankton standing stock 25 BARRACLOUGH, W. E., and D. ROBINSON. The fertilization of Great Central Lake. III. Effect on juvenile sockeye salmon , 37 PERRIN, WILLIAM F., and JOHN R. HUNTER. Escape behavior of the Hawaiian spinner porpoise (Stenella of. S. longirostris) 49 EVANS, W. E., J. D. HALL, A. B. IRVINE, and J. S. LEATHERWOOD. Methods for tagging small cetaceans 61 DAVY, BRENT. A review of the lantemfish genus Taaningichthys (Family Myctophi- dae) with the description of a new species 67 DRUCKER, BENSON. Some life history characteristics of coho salmon of the Karluk River system, Kodiak Island, Alaska 79 PERKINS, HERBERT C. Developmental rates at variou-s temperatures of embryos of the northern lobster (Homarus americanus Milne-Edwards) 95 SICK, LOWELL V., JAMES W. ANDREWS, and DAVID B. WHITE. Preliminary studies of selected environmental and nutritional requirements for the culture of penaeid shrimp 101 KELLEY, CAROLYN E., and ANTHONY W. HARMON. Method of determining car- otenoid contents of Alaska pink shrimp and representative values for several shrimp products Ill LEWIS, ROBERT M., E. PETER H. WILKENS, and HERBERT R. GORDY. A de- scription of young Atlantic menhaden, Brevoortia tyrannus, in the White Oak River estuary. North Carolina Ill KEPSHIRE, BERNARD M., JR., and WILLIAM J. McNEIL. Growth of premigratory Chinook salmon in seawater 11£ MAJOR, RICHARD L., and GERALD J. PAULIK. Effect of encroachment of Wanapum Dam Reservoir on fish passage over Rock Island Dam, Columbia River 12S MOSHER, KENNETH H. Scale features of sockeye salmon from Asian and North American coastal regrions 14j HOUDE, EDWARD D. Development and early life history of the northern sennet, Sphy- raena borealis DeKay (Pisces: Sphyraenidae) reared in the laboratory 18{ CAHN, PHYLLIS H. Sensory factors in the side-to-side spacing and positional orienta- tion of the tuna, Euthynnus af finis, during schooling 19' ROSENTHAL, RICHARD J., and JAMES R. CHESS. A predator-prey relationship between the leather star, Dermasterias imbricata, and the purple urchin, Strongyloeeri' trotus purpuratus 20J JUDKINS, DAVID C, and ABRAHAM FLEMINGER. Comparison and foregut con- tents of Sergestes similis obtained from net collections and albacore stomachs 211 Note HOOPES, DAVID T., and JOHN F. KARINEN. Longevity and growth of tagged king crabs in the eastern Bering Sea . . 22! U.S. DEPARTMENT OF COMMERCE Maurice H. Stans, Secretary NATIONAL OCEANIC AND ATMOSPHERIC ADMINISTRATION Robert M. White, Administrator NATIONAL MARINE FISHERIES SERVICE Philip M. Roedel, Director Fishery Bulletin The Fishery Bulletin carries original research reports and technical notes on investigations in fishery science, engineering, and economics. The Bulletin of the United States Fish Commission was begun in 1881 ; it became the Bulletin of the Bureau of Fisheries in 1904 and the Fishery Bulletin of the Fish and Wildlife Service in 1941. Sep- arates were issued as documents through volume 46; the last document was No. 1103. Beginning with volume 47 in 1931 and continuing through volume 62 in 1963, each separate appeared as a numbered bulletin. A new system began in 1963 with volume 63 in which papers are bound together in a single issue of the bulletin instead of being issued individually. Beginning with volume 70, number 1, January 1972, the Fishery Bulletin became a periodical, issued quarterly. In this form, it is available by subscription from the Superintendent of Documents, U.S. Gov- ernment Printing Office, Washington, D.C. 20402. It is also available free in limited numbers to libraries, research institutions, State and Federal agencies, and in exchange for other scientific publications. EDITOR Dr. Reuben Lasker Scientific Editor, Fishery Bulletin National Marine Fisheries Service Southwest Fisheries Center La JoUa, California 92037 Editorial Committee Dr. Elbert H. Ahlstrom National Marine Fisheries Service Dr. William H. Bayliflf Inter-American Tropical Tuna Commission Dr. Daniel M. Cohen National Marine Fisheries Service Dr. Howard M. Feder University of Alaska Mr. John E. Fitch California Department of Fish and Game Dr. Marvin D. Grosslein National Marine Fisheries Service Dr. J. Frank Hebard National Marine Fisheries Service Dr. John R. Hunter National Marine Fisheries Service Mr. John C. Marr Food and Agriculture Organization of the United Nations Dr. Arthur S. Merrill National Marine Fisheries Service Dr. Virgil J. Norton University of Rhode Island Mr. Alonzo T. Pruter National Marine Fisheries Service Dr. Theodore R. Rice National Marine Fisheries Service Dr. Brian J. Rothschild National Marine Fisheries Service Dr. Oscar E. Sette National Marine Fisheries Service Mr. Maurice E. Stansby National Marine Fisheries Service Dr. Maynard A. Steinberg National Marine Fisheries Service Dr. Roland L. Wigley National Marine Fisheries Service ON THE CLARIFICATION OF LARVAL TUNA IDENTIFICATION PARTICULARLY IN THE GENUS Thunnus ^ r.A v Walter M. Matsumoto,^ Elbert H. Ahlstrom,'' S. Jones,^ WiTOLD L. Klawe,* William J. Richards/ and Shoji Ueyanagi" ABSTRACT A Larval Tuna Identification Workshop was held at the Bureau of Commercial Fisheries Biological Laboratory (now the National Marine Fisheries Service, Southwest Fisheries Center), Honolulu, Hawaii, on March 2-6, 1970, to resolve conflicting views on the identification of larvae of Thunnus alalunga and T. albacares and to clarify the status of larval identification of other Thunnus species. The identification of T. alalunga (Yabe and Ueyanagi, 1962), T. albacares (Matsumoto, 1958), T. obesus (Matsumoto, 1962), and T. thynnus (Yabe, Ueyanagi, and Watanabe, 1966) was agreed upon as correct, except that the description of T. albacares should include the appearance of black pigmen- tation at the tip of the lower jaw when the larva attains a length of about 4.5 mm SL and that the lower size limit of reliable identification of T. alalunga be set at about 4.5 mm SL until further studies indicate more precisely whether the black pigmentation at the tip of the lower jaw in T. albacares appears earlier. There was no difference in appearance of T. thynnus larvae from the Atlantic and Indo-Pacific Oceans. The identification of T. tonggol and T. maccoyii larvae was not conclusive. The larvae of T. atlanticus required further study. The workshop further concurred that juveniles (13-200 mm SL) of several species of Thunnus may be separated by internal and external characters: T. atlanticus by vertebral count, T. alalunga by shape of first elongate haemal spine and arrangement of pterygiophores of the second dorsal fin rel- ative to two adjacent neural spines, and T. thynnus by configuration of lateral line and arrangement of pterygiophores of the second dorsal fin; and that juveniles of T. obesus and T. albacares may be sep- arated from the previous three species by arrangement of pterygiophores of the second dorsal fin, but not from each other. The proper identification of larval tunas has been a perplexing and difficult problem for many years. Although progress in the past two dec- ades has resulted in agreement on the identifica- tion of larvae of a number of species {Katsu- womis pelamis, Euthyymus alletteratus, E. af- finis, E. lineatus, Thimnus thynnus, T. obesus, and Auxis spp.) , there is still some disagreement ^ National Marine Fisheries Service, Southwest Fish- eries Center, Honolulu, HI 96812. ^ National Marine Fisheries Service, Southwest Fish- eries Center, La Jolla, CA 92037. ^ Department of Zoology, University College, Trivan- dum-1, India. (Formerly: Central Marine Fisheries Research Institute, Mandapam Camp, South India.) * Inter-American Tropical Tuna Commission, La Jolla, CA 92037. ^ National Marine Fisheries Service, Southeast Fish- eries Center, Miami, FL 33149. * Far Seas Fisheries Research Laboratory, Shimizu, Japan. and confusion on the identity and description of T. alalunga and T. albacares. At the present time there are two diflferent descriptions given for T. alalunga (Matsumoto, 1962; Yabe and Ueyanagi, 1962). The identity of other tunas, such as T. tonggol, T. maccoyii, and T. atlanticus, has yet to be confirmed or resolved. One of the chief problems in larval tuna iden- tification is the difficulty in obtaining good series of larvae for study. Tuna larvae are seldom taken in sufficient numbers by the usual collect- ing methods, and individuals over 10 mm stan- dard length (SL) are taken rather infrequently. Additionally, although the young of a number of tuna species are found together in many parts of the ocean, some species are localized in certain areas. Consequently, it is extremely difficult for workers in diflferent parts of the world to have Manuscript accepted September 1971. FISHERY BULLETIN; VOL. 70, NO. I, 1972. FISHERY BULLETIN: VOL. 70, NO. 1 at hand a complete series of larvae of more than two to four species. In an attempt to resolve the conflicting views on T. alalunga and to review the identification of larvae of other species of tunas, a Larval Tuna Identification Workshop was held at the Bureau of Commercial Fisheries (BCF) Biological Lab- oratory (now the National Marine Fisheries Service, Southwest Fisheries Center) , Honolulu, Hawaii, on March 2-6, 1970. The workshop also afforded an opportunity to workers specializing on larval tuna identification to assemble speci- mens of the various species of tunas and to ex- amine these together. The procedure followed at the workshop was (1) to summarize the status of larval tuna identification to date by species and (2) to evaluate the identifying char- acters by examining larval specimens. As time permitted, the status of juvenile tuna identifi- cation was also examined. The participants included: Mr. Walter M. Matsumoto, Convenor Dr. Elbert H. Ahlstrom, Advisor Dr. Santhappan Jones Mr. Witold L. Klawe Dr. William J. Richards Dr. Shoji Ueyanagi Dr. Jean-Yves Le Gall of the Centre Ocean- ologique de Bretagne, Brest, France, attended the workshop as an observer. The sessions were conducted informally with a summary of the present status of larval tuna identification, including recent developments, followed by evaluation of the various characters that could be relied upon for positive identifica- tion. Most of the sessions were devoted to direct examination of larval specimens of the various species and discussions of unpublished data of- fered by participants. This report summarizes the proceedings and results of the workshop. RECENT DEVELOPMENTS IN THE IDENTITY OF Thunnus alalunga Two diff'ering versions of the identity and de- scription of T. alalunga had arisen from reliance on black pigmentation in diff"erent parts of the body. Matsumoto (1962) relied upon black pig- mentation on the dorsal and ventral edges of the trunk forward of the caudal fin base, where- as Yabe and Ueyanagi (1962) relied upon black pigmentation on the tips of the upper and lower jaws and the absence of pigmentation on the trunk. The lack of sufficient larvae fitting Matsu- moto's description from areas presumed to be spawning grounds on the basis of gonad studies casts some doubt on his identification. On the other hand, good correspondence in the occur- rence of larvae fitting Yabe and Ueyanagi's de- scription with catches of adult T. alalunga in various areas in the Pacific seemed to support the latter identification. A study of red pigment patterns in larvae prior to preservation (Ueya- nagi, 1966) reinforced Yabe and Ueyanagi's identification and description. Additional ob- servations on red pigmentation by Matsumoto (see later discussion) confirmed Ueyanagi's re- sults and also provided more data to enhance the reliability of red pigmentation as a supple- mentary character for identifying T. alalunga. IDENTIFICATION OF TUNA LARVAE With the problem of differences in the identity and description of T. alalunga larvae fairly well settled at the outset, there remained the tasks of evaluating the various identifying characters, not only for this species but for other tunas as well, and of describing the species at various size categories. DEFINITION OF LARVA In tunas, as in many other fish, it is difficult to clearly separate the larval from the juvenile stages because there is no marked metamor- phosis and the usual adult characters used for species identification develop gradually and sep- arately. It is generally accepted among workers in larval tunas that the larval stage ends when the larva has developed the full complement of spines and rays in all the fins, all the vertebrae have ossified, and the anal opening has moved back near the origin of the anal fin. For nearly all tuna species, these developments occur when the larva has attained 10 to 13 mm SL. We use this as our definition, also. MATSUMOTO ET AL.: LARVAL TUNA IDENTIFICATION EVALUATION OF CHARACTERS In identifying fish larvae collected in plankton nets, the easiest and perhaps the only recourse is to identify the largest stage and work down to the smallest. Unfortunately, very few tuna larvae above 9 mm SL are taken in plankton .j\et tows so that this process cannot be followed W all times and identification, therefore, must depend upon those nonadult characters that are the most distinctive and consistent throughout the size range. Characters that have been used in the past were reviewed and evaluated. A resume of the usefulness of the various characters follows. Meristic The number of myomeres is useful in sepa- rating Katsutvomcs pelamis (42-43) and Euthi/n- nus lineatus (38-39) from other tunas, including other species of Euthynnus, all of which have similar numbers of myomeres (40-41). The number of fin rays and spines are not useful for separation of Thunmis because all species are similar in this respect. Morphological Shape of first dorsal fin, when completely formed, is useful to distinguish late larval stages of K. pelamis, Euthynnus, and Auxis from those of Thunnus. Preopercular spines are unreliable . because they undergo rapid growth changes and position of eye relative to longitudinal axis of body needs to be determined more accurately. Distribution (number and position) of pterygio- phores in the second dorsal fin in relation to neural spines is useful in separating several Thunnus species, but only after these bones have ossified in larvae longer than 10 mm SL. Other characters of the axial skeleton useful in identification, such as the position of the first haemal arch and the position of the zygapophyses on the vertebrae, also form late. Measurements Morphometries have not been used extensively to date, although there may be some with good possibilities, such- as the relations of body depth to standard length, snout length to head length, and snout length to orbit diameter. Some of the reasons for not using measurement data are that the larvae not only shrink in preservatives, but the degree of shrinkage varies in different preservatives and with duration of preservation; the distortion of the body at the time of fixing cannot be controlled; and, more important, there are too few larvae in undistorted condition for reliable measurements. Added to these are other sources of variability such as rapid changes in body parts due to growth, changes which often occur in spurts, and distension of the abdomen, as well as stretching of the body at each feeding. Pigmentation For the most part black pigment patterns have been the most widely used and accepted character in identifying tuna larvae. There are variations and changes in black pigment patterns on tuna larvae due to growth, but in certain areas of the body these patterns have been found to be consistent enough for identification pur- poses. This is particularly true of pigment pat- terns on the first dorsal fin, posterior half of the trunk, forebrain, and tips of both jaws. The larval size at which black pigment cells appear in certain areas of the body, especially at the upper and lower jaw tips, may be useful in sep- arating T. albacares from T. alahinga. Red pig- ment patterns, although not species specific, have been useful in confirming the identification of T. alalunga when used in conjunction with black pigment patterns. Of all the characters reviewed and examined, pigment patterns, both black and red, were con- sidered to be the most reliable for identification of the larval stages, despite their known varia- bility, when supplemented by the use of certain morphological characters such as the distribu- tion of pterygiophores in the second dorsal fin and characteristics of the vertebral column, whenever these are developed. VERIFICATION OF RED PIGMENTATION Ueyanagi (1966) reported on the usefulness of red pigmentation in identifying tuna larvae. Up to then identification of tuna larvae by pig- mentation had been based on black pigment only. 3 FISHERY BULLETIN: VOL. 70. NO. 1 Ueyanagi examined 350 larvae and concluded that T. albacares and T. alalunga, which are difficult to identify by the usual diagnostic char- acters, could be distinguished by differences in red pigment patterns: larvae of T. alalunga consistently had more red pigment spots on the dorsal and ventral edges of the body and along the mid-lateral line forward of the caudal pe- duncle than larvae of T. albacares; red pigment patterns in larvae of T. thynnus and T. obesus were intermediate between those of T. alalunga and T. albacares; the red pigment pattern in Allothunnus fallal was similar to that in Thun- nus; and the pattern in K. pelaniis resembled that in Auxis spp. and E. affinis but differed from that in Thunnus. To confirm these results and to provide addi- tional information on red pigmentation in tuna larvae, the results of observations made on 432 larvae taken in Hawaiian waters during August and September 1967 were presented. Tables 1 and 2 give the number of red pigment cells along the dorsal, ventral, and lateral lines on the pos- terior half of the trunk and a summary of the number of larvae examined, the number of larvae observed with red pigmentation, and the mean numbers of red pigment cells at the three sites. In Table 2, the number of red pigment cells ob- served most frequently are given in bold face type and those observed occasionally or seldom are enclosed in parentheses. The pigment patterns agreed generally with those reported by Ueyanagi for the species listed in the tables. Differences in the patterns were noticeable mainly in the dorsal edge of the trunk and, to a lesser extent, in the mid-lateral line. There was no significant difference in the num- ber of pigment cells t)etween the left and right sides of the body. The appearance and extent of red pigment cells varied in larvae taken in night and day tows. In larvae taken at night the pigment cells were numerous, distinct, and brightly colored, whereas in larvae taken during the day the pig- ment cells were faintly colored, often not visible, and in many instances the pigment spots were united, forming single continuous lines. Of the species taken in both day and night tows {T. al- bacares, T. obesus, and K. pelamis) , red pigmen- tation was not visible in 41.5% of the larvae taken during the day, compared with only 3.6% of the larvae taken at night. Thus, observations of red pigment cells must be made largely on larvae taken at night to reduce the effects of diel variations. Despite the variations, red pigmentation is a useful supplementary character to either sepa- rate certain species or verify the identification made on the basis of other characters. That the red pigment pattern is not species specific is clearly seen in the similarity among K. pelamis, Auxis spp., and E. affinis and between T. obesus and T. albacares; however, it is useful in sepa- rating T. alalunga from the other kinds of Thimnus. EXAMINATION AND DISCUSSION OF SPECIES Thunnus alalunga and T. albacares Larvae of these two species were examined together because they are the only species lack- ing black pigment cells on the trunk, exclusive of the caudal fin and abdomen (Yabe and Ueya- nagi, 1962), Characters, including some that have not been used in the past, for separating the two species are summarized in Table 3. The larval stage was divided into two size categories, small larvae less than 10 mm SL and larger lar- vae 10 to 13 mm SL, because the characters used in differentiating small larvae became ineffective or obscured with growth. As mentioned earlier, pigmentation, particularly the presence of black pigment cells at the tips of the upper and lower jaws and the amount of red pigment cells on the trunk, was the most reliable character in sepa- rating larvae of the two species. In small larvae, black pigment cells on the lower jaw tip were first observed in larvae of T. albacares about 4.5 to 6.0 mm SL, and often as small as 3.8 mm SL; those on the upper jaw tip were first observed in larvae about 7.0 mm SL (Figures 1 and 2, reproduced from Matsumoto, 1958') . In T. alalunga these pigment cells were ' The difference in developmental stages per given size in the figures by Matsumoto (1958) and Ueyanagi (1969) is due to method of preservation: Matsumoto's figures are of larvae preserved in 10% Formalin; Ueya- nagi's figures are of larvae preserved in 70% alcohol. 4 MATSUMOTO ET AL.: LARVAL TUNA IDENTIFICATION US ft W I -t-> m •i-H w s 2 CO c3 C 01 o (V 13 cr BQ H «, t^ S3S i> o "' tX 11 V 0)12 ^:5 5 i; 0) S3^ ,2 oS; 0)3 PI o n 1> 0)73 0>3 J;^ U C - o r>T3 O t/i o »o to 1 I r-. r-^ -^ "1 CO o ^ »— o> lo CN CN O- CN •— •O CO ^ CM O ■* I lO -* — I O •— CN CO ■^ o "O so ' »o 'O o >o — ■» tx ^ — -O I I I i< d - d ' ' ' IT) — — I I I I CO I I I I r I tv -if — rx r ' ' K 00 so K ' CO ■>»• >o O; CO ■* -o — "S K d CO CM ■— CS CS — 00 CO CO CO LO ■— ' — -O — O K CO cs CO r-. <) ■^ "o vd — 00 SD CO CO d — CO cs — o — CN CO -^ "1 -o c 3 -6 is r rs. N. o hv o CO K r 1 1 N. -^ CO "^ rx N. 1 -^ 1 I 1 1 1 1 ' ' K lo d ^o is! i< "— CO •— > iri ' ' ' ' ' ' — CN — ' C(i ■>)■' K CN •- CO — — CM CN — I CM I I I I I I I I I I I I I I I ICM — COt^CSCO'O'* I I I I I I I I I I I I I •— I I I I I I I I I I I I lo I'O'Oixo-p'O'^rs.cssO'^'O'opio i ' lo d 'd->trv-10 mm SL): Array of ^Da pterygiophores between two adjacent neural spines Position of first haemal arch (vertebra number) 3-12 [mean = 7.0] 5-12 [mean 1.0] 1, 2, 1, 1 nth 2, 2, 3, 2, 2, 1 10th 1, 2, 2, 2, 3, 1 D2 refers to second dorsal fin. Other Thunnus species These species, which include T. thynnus {T. thynnus thynnus of Atlantic and T. thynnus or- ientalis of Pacific) , T. tonggol, T. maccoyii, and T. obesus, have been identified mainly by black pigmentation on the trunk other than that over the abdominal w^all. In small T. thynnus of both Atlantic and Pa- cific Oceans (larvae between 3 and 10 mm SL), one or two large black pigment cells are present on the dorsal edge of the trunk between the sec- ond dorsal and caudal fins (Table 4, Fig-ures 5 and 6) , the anterior one usually being the larger. There may also be one to four black pigment cells on the ventral edge of the trunk between the anus and the caudal fin. Black pigmentation in T. thynnus from both oceans agrees quite well, except that in 5 out of 10 Atlantic specimens one or two tiny black pigment cells were noted along the mid-lateral line of the body near the pectoral fin, and in two instances a single tiny black pigment cell was found on the mid-lateral line beneath the posterior end of the second dor- sal fin. These pigment cells were not considered reliable for identification purposes. Observation of red pigmentation on larvae of Atlantic and Pacific T. thynnus is incomplete. Only one Atlantic T. thynnus larvae was exam- ined for this character, but unfortunately the specimen was taken in a day tow so that the pigmentation appeared as a continuous streak on both the dorsal and ventral edges of the trunk as well as on the ventral surface of the lower jaw. In Pacific T. thynnus there were one to five red pigment cells, usually three, on the dorsal edge of the trunk. The number of red pigment cells on the mid-lateral line and ventral edge of the trunk has not been recorded, but according to the Illustration by Ueyanagi (1966), the pig- ment pattern may be similar to that of T. obesus. On the basis of black and red pigmentation, the Atlantic and Pacific T. thynnus were not sep- arable. The identification of T. tonggol, based on size series of 4.2 to 7.3 mm, has yet to be confirmed. Following the description of the species by Ma- tsumoto (1962), larvae similar to these having the anteriormost black pigment cell on the dorsal edge of the body ahead of the second dorsal fin origin have been found in 1963 in the mid-South Atlantic Ocean near Ascension Island. Confir- mation of the species description requires the finding of adults within this area and the finding of additional larvae to extend the identified size range. The identification of T. maccoyii, which was first described as having black pigment pattern similar to that of T. thynnus (Yabe, Ueyanagi, and Watanabe, 1966) and later as having the black pigment cells on the dorsal edge of the trunk reduced to pinpoints (Ueyanagi, 1969), also needs verification (see discussion on T. thyn- nus) . The correspondence of published descrip- tions based on eight specimens and observations of larvae identified as this species were not con- clusive. T. obesus was easily separated from T. thyn- nus by the absence of black pigmentation on the bases of the anterior dorsal finlets. Sometimes a single small black pigment cell was present along the ventral edge of the trunk near the caudal peduncle, but more often one to three pig- 8 MATSUMOTO ET AL.: LARVAL TUNA IDENTIFICATION Table 4. — Characters used to separate larvae of Thunnus species having black pigmentation on trunk. Characters Thunnus thynnus (Atlantic) Thunnus thynnus (Pacific) Thunnus tonggot Thunnus maccoyii Thunnus ohesus Small larvae (3-10 mm SL): Number of black pigment cells: Upper jaw tip Lower jaw tip Dorsal edge trunk Lateral lino Ventral edge trunk Number of red pigment cells: Dorsal edge trunk Lateral line Ventral edge trunk Lower jaw ventral view Large larvae (>10 mm SL): Array of ^D^ pterygio- phores between two adjacent neural spines No observation 2 on inner edge 1 or 2 0-2 near mid-trunk M Streak on caudal peduncle'^ Indistinct! Streak anus to caudal peduncle^ Streak along margin anterior half of jaw and midline^ 1, 2, 2, 3, 2, 2, 1, 1 Appears above 6 mm SL 2 on inner edge above 4 mm SL 1 or 2 None 2 or more 1-5, mostly 3 Number not available Number not available 2 well spaced on anterior half No observation No observation No observation 1, 2, or more None 2 or more No observation No observation No observation No observation No observation Appears above 5 mm Few spots above 5 mm SL SL Appears above 4 mm 0-2 on inner edge SL below 4 mm SL 1 or 2, very small None or 1 near mid-trunk None 1-3 1 or more No observation No observation No observation No observation No observation 0, 1, (2) 0, 1, 2, 3, 4 1-8 [mean = 5.3] 1 on each side near tip 1, 2, 2, 2, 3, 2, 1, 1 • Only one larva taken in a day tow was examined. 2 D2 refers to second dorsal fin. ment cells were present along the base of the posterior half of the anal fin. Red pigmentation did not differ from that in T. albacares. In larger larvae (10-13 mm SL) the array of pterygiophores of the second dorsal fin between two adjacent neural spines was sufficient to sep- arate T. thynnus from T. ohesus and both species from T. alalunga (Tables 2 and 3). In T. thyn- nus the greatest number of pterygiophores (3) between two adjacent neural spines appeared in the fourth position in the array, whereas in T. ohesus and T. alalunga it appeared in the fifth and sixth positions, respectively. T. ohesus was not distinguishable from T. alhacares by this character. The identification of T. atlanticus was not re- solved. No larvae from the distributional range of this species (tropical western Atlantic) have been found which are distinguishable from any of the species considered above. One of us (Richards) suspects that T. atlanticus larvae are very similar to larvae of T. ohesus. This suspicion is based on the great abundance of larvae resembling those of T. ohesus in this area, particularly at times and places where T. ohesus adults are rarely found or absent, studies are needed. Further SUMMARY OF LARVAL IDENTIFICATION On the basis of the examination and discussion above, the workshop agreed that: 1. The description of T. albacares by Matsu- moto (1958) was correct (see Figures 1 and 2), but that the "appearance of black pigmentation at the tip of the lower jaw at about 4.5 mm SL" should be included. 2. The description of T. alalunga by Yabe and Ueyanagi (1962) and illustrations by Ueyanagi (1969) were correct (see Figures 3 and 4), but that the lower size limit should be set at about 4.5 mm SL until further studies indicate more precisely the earlier appearance of black pig- mentation at the tip of the lower jaw in T. al- hacares. 3. It is not possible to separate larvae of T. alhacares from T. alalunga below 4.5 mm SL, prior to the appearance of black pigment cells at the tip of the lower jaw in T. alhacares. 4. The description of T. thynnus by Yabe, Ueyanagi, and Watanabe (1966) was correct FISHERY BULLETIN: VOL. 70, NO. 1 3.6 mm 6.6mm 9.6mm 4 4mm 5.3 mm 6.1mm Figure 5. — Larval stages of Thunnus thynnus, I. (From Yabe, Ueyanagi, and Watanabe, 1966. Lengths have been converted from total to standard.) (see Figures 5 and 6), and that there was no difference in T. thynnus from the Atlantic and Pacific Oceans. 5. The identification of T. tonggol was not substantiated by an adequate size series. 6. The description of T. maccoyii, based on tiny melanophores on the dorsal edge of the trunk, was not conclusive. 7. The description of T. ohesus by Matsumoto (1962) was correct, though it needed to be aug- mented by illustrations of a complete size series. 12.2 mm 16.8mm Figure 6. — Larval stages of Thunnus thynnus, IL (From Yabe, Ueyanagi, and Watanabe, 1966. Lengths have been converted from total to standard.) 8. The identity of T. atlanticus larvae is un- resolved. IDENTIFICATION OF JUVENILES In spite of the intention of the workshop to assemble as many specimens of juvenile tunas as possible, only a few juveniles of T. albacares and T. ohesus, not nearly enough to warrant their detailed examination, were available for study. The discussion on juvenile tuna identifi- cation, therefore, dealt mainly with published reports and with contributed data, resulting in a summary of identifying characters which the workshop considered useful and reliable. 10 MATSUMOTO ET AL. : LARVAL TUNA IDENTIFICATION Once the young tuna has acquired the full complement of spines and rays in all the fins, complete ossification of all the vertebrae, and the relocation of the anus near the origin of the anal fin, it is generally considered a juvenile of the species. Certain characters such as the full num- ber of gill rakers, however, develop much later, when the juvenile has attained a length of 40 or 45 mm SL. If we consider juveniles to in- clude all sizes up to the time of full gonad de- velopment signified by initial spawning, the size range of the juvenile stage would extend from about 13 mm SL to 700 mm FL (fork length) in T. albacares (Yuen and June, 1957) and to 860 mm FL in T. alalunga (Otsu and Hansen, 1962) . For the purpose of clarifying species identifi- cation of the young, however, individuals beyond 200 mm SL need not be included. The term ju- venile, as used here, thus refers to tunas between 13 and 200 mm SL. EVALUATION OF CHARACTERS The greatest difficulty in identifying juveniles of Thunnus is that the most useful characters are located internally. Except for the flattened first elongate haemal spine in T. alalunga, there is no single character that is peculiar to each of the species; but by using a combination of characters it should be possible to identify most of the other species. A summary of the most useful characters discussed is listed in Table 5. The size of juvenile at which each of the char- acters can be observed is listed also. Those char- acters whose usefulness in the early juvenile stages has not been shown conclusively are indi- cated by a question mark (?). The general formula of distribution of pterygiophores of the second dorsal fin has not been used before. The counts and descriptions given for those characters listed with a question mark gener- ally are those of the adults. These have not yet been substantiated for juveniles as well. Changes in the position of the first haemal arch with growth, for example, have been known to exist in other closely related fish such as the wahoo, Acanthocyhium solandri (Matsumoto, 1967). This could be true of the tunas also. Comparisons of body parts, particularly of or- bit diameter, body depth at origins of the first dorsal and anal fins, preanal and postanal dis- tances, and snout length, have not been investi- gated sufficiently in the past. The unavailability of specimens in sufficient numbers as well as the nonuniformity of body lengths (fork and stan- dard) used have contributed greatly to this ne- glect. Acceptance of standard length as the standard measure of body length and publishing of actual measurements in the future should help in the accumulation of sufficient data for anal- yses. This has to be done by all workers in this field of study, since the juveniles are not easily taken in large numbers. Table 5. — Characters for separating juveniles of Thnnnus species. Character Useful on juveniles obove Thunnus thynnus Thunnus alalunga Thunnus atlanticus Thunnus obesus Thunnus albacares First haemal arch ? 10 10 11 11 11 Ceratobranchial including angle 40 mm SL 17-20 15-16 12-13 15-16 15-16 Vertebrae 13 mm SL 18 + 21 18 + 21 19 -f 20 18 + 21 18 -f 21 Array of 1D2 pterygio- phores between two adjacent neural spines 10 mm SL 1, 2, 2, 3, 2, 2, 1, 1 1, 1, 2, 2, 2, 3, 2, 1 1, 2, 2, 2, 3, 2, 1, 1 1, 2, 2, 2, 3, 2, 1, 1 1, 2, 2, 2, 3, 2, 1, 1 First prezygapophysis and position on hoennal arch ? 15, 16, 17, high 15, 16, high 16, 17, low 15, 16, high 13, 14, low Postzygapophysis near first prezygapophysis ? Short, directed posterior Short, directed posterior Long, directed vertical or slightly anterior Short, directed posterior Long, directed vertical, some slightly anterior First haemal spine 30 mm SL Winglike at some stages Extremely wing- like Winglike at some stages — ~ Lateral line above base of pectoral fin 25 mm SL Acute, nearly 90° Obtuse Obtuse Obtuse Obtuse ^ Da refers to second dorsal fin. 11 FISHERY BULLETIN: VOL. 70. NO. 1 DISCUSSION AND SUMMARY T. thynnus below 25 mm SL can be separated from the other Thunnus species by the array of pterygiophores of the second dorsal fin, the last four positions containing 2, 2, 1, 1 pterygio- phores; in T. alabmga the sequence is 2, 3, 2, 1, and in T. atlanticus, T. obesus, and T. albacares it is 3, 2, 1, 1, T. thynnus above 25 mm SL can be separated from all other Thunnus by the sharp angle (nearly 90°) which the lateral line follows near the base of the pectoral fin; in all other species this angle is obtuse. In juveniles above 40 to 45 mm SL, T. thynnus has the highest number of gill rakers on the ceratobranchial, in- cluding that at the angle (Potthoff and Richards, 1970). T. alalunga below 30 mm SL can be separated from other Thunnus species by the distribution of pterygiophores of the second dorsal fin. Above 30 mm SL, T. alalunga is the only species whose first elongated haemal spine is flattened laterally and appears extremely winglike. T. atlanticus as small as 13 mm SL can be sep- arated from other Thunnus species by its dis- tinctive precaudal and caudal vertebral counts. It is the only species having 19 precaudal and 20 caudal vertebrae. Above 40 to 45 mm SL, this species can be separated from the others by the low (12-13) gill raker count on the cerato- branchial (Potthoff and Richards, 1970), in ad- dition to the vertebral formula. T. obesus and T. albacares are the only two species that cannot be distinguished from each other on the basis of internal characters. Com- parisons of body parts, i.e., orbit diameter, body depth or preanal and postanal distances, may have to be used. ACKNOWLEDGMENTS We thank John C. Marr, former Area Director, BCF Biological Laboratory, Honolulu, who orig- inated the idea of the workshop; Richard S. Shomura, former Acting Area Director of the same Laboratory, who continued with the orig- inal idea and organized the workshop ; and the BCF Biological Laboratory, Honolulu, Hawaii, for providing laboratory space and facilities. The workshop was supported entirely by the Bureau of Commercial Fisheries (now National Marine Fisheries Service). LITERATURE CITED Matsumoto, W. 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. 1962. Identification of larvae of four species of tuna from the Indo-Pacific region I. Dana Rep. Carlsberg Found. 55, 16 p. 1967. Morphology and distribution of larval wahoo Acanthocybimn solandri (Cuvier) in the central Pacific Ocean. U.S. Fish Wildl. Serv., Fish. Bull. 66: 299-322. Otsu, T., and R. J. Hansen. 1962. Sexual maturity and spawning of the alba- core in the central South Pacific Ocean. U.S. Fish Wildl. Serv., Fish. Bull. 62: 151-161. Potthoff, T., and W. J. Richards. 1970. Juvenile bluefin tuna, Thunnus thynnus (Lin- naeus), and other scombrids taken by terns in the Dry Tortugas, Florida. Bull. Mar. Sci. 20: 389-413. Ueyanagi, S. 1966. On the red pigmentation of larval tuna and its usefulness in species identification. [In Jap- anese, English summary.] Rep. Nankai Reg. Fish. Lab. 24: 41-48. 1969. Observations on the distribution of tuna lar- vae in the Indo-Pacific Ocean with emphasis on the delineation of the spawning areas of albacore, Thunnus alalunga. [In Japanese, English sum- mary.] Bull. Far Seas Fish. Res. Lab. 2: 177-256. Yabe, H., and S. Ueyanagi. 1962. Contributions to the study of the early life history of the tunas. Occas. Pap. Nankai Reg. Fish. Res. Lab. 1: 57-72. Yabe, H., S. Ueyanagi, and H. Watanabe. 1966. Studies on the early life history of bluefin tuna Thunnus thynnus and on the larva of the southern bluefin tuna T. maccoyii. [In Japanese, English summary.] Rep. Nankai Reg. Fish. Res. Lab. 23: 95-129. Yuen, H. S. H., and F. C. June. 1957. Yellowfin tuna spawning in the central equa- torial Pacific. U.S. Fish Wildl. Serv., Fish. Bull. 57: 251-264. 12 THE FERTILIZATION OF GREAT CENTRAL LAKE. I. EFFECT OF PRIMARY PRODUCTION T. R. Parsons/ K. Stephens,- and M. Takahashi^ ABSTRACT Commercial fertilizer was added at a rate of 5 tons per week to a lake (51 km^, mean depth 200 m) over a period of 5 months from May to October 1970. As a result of these additions, surface primary produc- tion was increased approximately tenfold while the primary production of the euphotic zone was doubled. The standing stock of primary producers and water clarity were substantially the same as in the pre- vious year when no fertilizer was added. The productive index (mg C/mg Chi a/hr) was increased, especially in the immediate area of nutrient enrichment. However, the principal phytoplankton species were very similar at locations near and distant from the area of fertilization. In conclusion, it appears that as a result of adding nutrients at a low but sustained level, primary productivity was increased without substantially changing the nature of the food chain at the primary level of production. In the Pacific northwest, an earlier study (Nel- son and Edmondson, 1955) on the fertilization of a small salmon-producing lake in Alaska showed that the addition of phosphate and nitrate fer- tilizer increased the production of sockeye salm- on (Oncorhynchus nerka); in more recent studies by Donaldson et al. (1968), an increase in the production of steelhead trout (Salmo gairdneri) was demonstrated in a small lake in the state of Washington. The natural fertiliza- tion of lakes from decomposing salmon carcasses has been discussed by Krokhin (1967), who has suggested that the potential deficit from salmon removed by the fishery should be replaced by ar- tificial fertilization. In the report presented here we have carried out a fertilization experi- ment which differs from the two previous reports (Nelson and Edmondson, 1955; Donaldson et al., 1968) in several respects. These include the size scale of the experiment which was very much larger than any previous experiments, the application of fertilizer as a solution, control of the N:P ratio, and, finally, sustained weekly nu- trient additions over a period of 5 months. ^ Fisheries Research Board of Canada, Biological Sta- tion, Nanaimo, B.C.; present address: Institute of Oceanography, University of British Columbia, Van- couver, B.C., Canada. ^ Fisheries Research Board of Canada, Biological Sta- tion, Nanaimo, B.C., Canada. Manuscript accepted September 1971. FISHERY BULLETIN: VOL. 70, NO. 1, 1972. Preliminary results of our experiment have been reported (Parsons et al., in press) together with our conclusion that lake production was increased by the addition of fertilizer and that this was achieved without causing a condition of eutrophication. The following account deals specifically with the effect of nutrient enrich- ment on the primary level of production. Inten- sive studies on the effect of nutrient additions were carried out during the period May to August 1970 while a more general monitor pro- gram has been maintained from 1969 to the present (March 1971). The first sustained nu- trient additions were made during the period June to October 1970 and further additions are planned for the next 5 years. The primary purpose in this study is to in- crease levels of production in an oligotrophic lake, but not to change the trophic relationships which lead to the production of young sockeye salmon. In this respect the ultimate desideratum of the experiment is to produce larger sockeye smolts at their time of seaward migration; earlier reports have demonstrated that there is a close positive relationship between smolt size and survival (Ruggles, 1965; Johnson, 1965). Since previous studies (Parsons et al., in press) have shown that the migrant smolts from Great Central Lake are small (63 ± 1 mm) and that the primary productivity is very low (ca. 5 g 13 FISHERY BULLETIN: VOL. 70, NO. I C/m^/year) , application of nutrients at a level that would increase primary and secondary pro- duction seemed reasonable. Data used in this presentation have been ob- tained from Stephens et al, (1969') and Kennedy et al. (1971^). METHODS ANALYTICAL PROCEDURES Chlorophyll a, nutrients, oxygen, and total CO2 were all measured as described previously (Strickland and Parsons, 1968); bacteria were enumerated from plate counts after 24 hr in- cubation at room temperature on Millipore uni- versal medium; major phytoplankton species were enumerated after settling preserved sam- ples; conductivity was measured using a Beck- man Solu Bridge (Cedar Grove, N.J.). Primary productivity was measured as the difference in uptake of '^COo in light and dark bottles; how- ever, on a few days the dark-bottle uptake was exceptionally high and this requires further in- vestigation. For the purpose of this presenta- tion, data have been used only for days when the dark-bottle uptake was less than 20 9f of the maximum light-bottle uptake. Radiation was measured with an Epply pyra- nometer and corrected to give photosynthetically available radiation (PAR) as described previ- ously (Parsons and Anderson, 1970). Light attenuation was routinely measured with a Secchi disc (SD), and an empirical relationship between SD depth (m) and the vertical extinc- tion coefficient was established using a Schiiler meter (maximum response at 430 nm). This relationship for light at 430 nm was: K4.30 _ in — 2.1 SD depth The (total) extinction coefficient for the water column was then found from Jerlov's (1957) ' Stephens, K., R. Neuman, and S. Sheehan. 1969. Chemical and physical limnological observations, Babine Lake, British Columbia, 1963 and 1969, and Great Cen- tral Lake, British Columbia, 1969. Fish. Res. Board Can., Manuscr. Rep. 1065: 41-52. * Kennedy, 0. D., K. Stephens, R. J. LeBrasseur, T. R. Parsons, and M. Takahashi. 1971. Primary and sec- ondary production data for Great Central Lake, B.C., 1970. Fish. Res. Board Can., Manuscr. Rep. 1127. 379p. light attenuation curves. Mean radiation (Im) for the water column of depth (dm) was deter- mined from the expression im — h (1- u^nK -e-^^m) where h was the surface radiation and k was the attenuation coefficient for light below the surface meter. The expression was also used to determine the light at various depths in re- lation to the photosynthetic activity at those depths. NUTRIENT ADDITIONS The choice of a suitable fertilizer for the waters of Great Central Lake has been discussed previously (Parsons et al, in press). The nu- trient addition consisted of a commercial grade of ammonium phosphate and ammonium nitrate which contained trace quantities of other ele- ments essential for plant growth. The mixture is known commercially as 27-14-0 (27% N; 14% P2O5; 0% K2O) and has an N:P ratio of 10:1. The ammonium nitrate and ammonium phos- phate were dissolved separately in 5-ton amounts (total) and the concentrated solutions mixed be- fore distribution. A small quantity of organic material was added to the dissolved inorganic fertilizer at a dilution of 6 liters of fish solubles (obtained from B.C. Packers Ltd.) for every 2 tons of nutrient solution. The dissolved ferti- lizer was distributed at 10 gal/min (38 liters/ min) in the wake of a vessel travelling at approx- imately 8 knots. The area of nutrient additions is shown in Figure 1, together with sampling stations 1, 2, and 3. Station 1 was sampled dur- ing 1969 and 1970, Station 2 was sampled during 1970, and Station 3 was sampled sporadically during 1970 in order to check on the flow of nutrients in a westerly direction; in addition, areal surveys for chlorophyll a, transparency, and bacteria were carried out over the whole lake in order to determine within-lake variation. The area over which nutrients were added represented ca. 3 sq mi (8 km-) of lake surface; however, from studies on lake circulation it was apparent that the material was transported east 14 PARSONS, STEPHENS, and TAKAHASHI: LAKE FERTILIZATION. I. Figure 1. — Great Central Lake showing principal sam- pling stations (1, 2, and 3) and area of nutrient en- richment ( crosshatched ) . and west at rates of up to 6 miles per day or ca. 10 km/day (Parsons et al., in press). Thus while Station 2 was generally under the most immediate influence of the nutrient additions, Stations 1 and 3 also received an accumulative enrichment. Fertilizer was added at the rate of 5 tons per week from June through to October 1970. During May 1970 approximately 2 tons of fertilizer were added in experiments to deter- mine the rate of mixing and distribution of nu- trients in the vessel's wake. RESULTS LAKE MORPHOMETRY Great Central Lake is located on Vancouver Island, B.C., at lat 49°20' N on an east/ west axis between long 125°00' W and 125°25' W (Figure 1). It is a long narrow lake (ca. 33 X 1.5 km) with steep sides and a mean depth of 200 m. The yearly mean discharge is approximately 6 X lO^mVday with a range from 0.4 x 10^ to 32 X lO^mVday. TEMPERATURE The temperature structure at Station 1 is shown in Figure 2. The results are representa- tive for the open waters of the whole lake, and it is apparent that the lake was isothermal dur- ing January and February; a thermocline started to form during March and was well established by May. Maximum surface tem- perature during July was 21.2° C; surface cool- ing started in September but a thermocline of Figure 2.— Temperature (°C) stratification at Station 1. 10° persisted through October and the lake did not become isothermal until January of the fol- lowing year. RADIATION Changes in photosynthetically active radiation (PAR) at the lake surface are shown in Figure 3 together with the mean radiation for the water column to 20 m, calculated -on a 24-hr basis. Figure 3. — (A) Surface photosynthetically active ra- diation (PAR) and (B) 15-day mean PAR in the first 20 m. From the latter results it is apparent that radi- ation in the water column increased by 50% from the beginning of May until the middle of June; the decrease in radiation during the second part of June was due to a combination of higher ex- tinction coefficients and lower surface radiation. The average radiation remained virtually con- stant during July and decreased by 20% during the latter half of August. CHLOROPHYLL A Surface chlorophyll a concentrations are pre- sented in Figure 5 in the same way and for the same stations and years as SD data in Figure 4. The two figures have some mirrored similarities ; 15 FISHERY BULLETIN: VOL. 70, NO. 1 14 i t- Q. UJ O 10 I o o 8 1 ii A lo o !,° i 1 ll'" I ,* B I STATION I A 1970 2 1970 STATION I o 1969 JUNE JULY AUGUST Figure 4. — Secchi disc depth at Stations 1 and 2, 1970, and Station 1, 1969. (Mean and standard deviation of values from areal surveys shown as bars.) ^ 0.6 0.4 - 0.2 0.0 STATION I A 1970 2 1970 STATION I o 1969 m I i ■'A fc i ' 1 A o JUNE ± JULY AUGUST Figure 5. — Surface chlorophyll a data at Stations 1 and 2, 1970, and Station 1, 1969. (Mean and standard de- viation of values from areal surveys shown as bars.) thus Station 2 chlorophyll a values from June to August were generally higher than either Sta- tion 1 data for 1969 or 1970; minimum mean SD data (8 to 10 m) occurred between June and July during a maximum in the mean chlorophyll a concentration. However, 1969 chlorophyll a data at Station 1 do not appear to be significantly different from 1970 chlorophyll a data at the same station. The depth distribution of chlorophyll a gen- erally showed a maximum between 10 and 20 m following stratification and nutrient depletion in the surface layers. pH, CALCIUM, TOTAL CO,, AND CONDUCTIVITY pH values were generally between 7,1 and 8.3 with some indication of a seasonal cycle towards higher pH values in summer. Several assays for calcium showed a concentration of 5 mg/liter while specific conductivity was very consistent at 33 ix mhos/cm, except in the immediate vi- cinity of small streams entering the lake; total carbon dioxide varied over a range from about 2.2 to 4.2 mg C/liter. OXYGEN Oxygen profiles to 200 m showed that surface oxygen concentrations were between 80 and 90% saturation during winter and up to 110% sat- uration during summer. Deepwater oxygen concentrations appeared constant at around 10 mg/liter or about 80% saturation. An oxy- gen maximum occurred at ca. 20 m during the summer. NITRATE, AMMONIA, PHOSPHATE, AND SILICATE Nitrate depth profiles at Station 1 during May to October, 1969 and 1970, are shown in Figure 6. The general form of the two profiles is similar; thus a depletion in the winter level of nitrate (1.0 to 2.0 fjig at./liter) becomes apparent to- wards the end of May and by the end of June about 1 iJLg at. NOnN/liter has been removed from the water column, to 10 m. During July and August nitrate in the first 10 to 15 m is close to the limit of detection, but there is a partial return to winter levels during September and October. Some difl["erence in the form of these events is apparent between 1969 and 1970; the utilization of nitrate was more rapid and apparently more complete during 1970; in addition, surface ni- trates did not increase in September-October 1970 as they did in 1969. Starting from a winter level of 2 yug at. NO.tN/ liter, the total utilization of nitrate in the water column has been determined for the periods Feb- ruary to May, June, and July-August using data shown in Figure 7. The accumulative amount 16 PARSONS. STEPHENS, and TAKAHASHI : LAKE FERTILIZATION. I. MAY I JUNE I JULY ' AUG I SEPT IQCT 197 Figure 6. — Nitrate (^g at./liter) profiles at Station 1, May to October 1969 and 1970. 1200 ^, 1000 - E °- 800 ■o 0) '£ 600 Z. 400 E 200 • N utilized in the woter column A N added as fertilizer O total N utilized F'MA'M'J'JASO Figure 7. — N utilization at Station 1. of inorganic nitrogen added as fertilizer (ex- pressed per m- for the entire 51 km- lake sur- face) is also shown; since this was utilized with- in hours following each addition, the total ni- trogen budget is represented as the sum of the natural and added inorganic nitrogen. Some mixing occurred during September and October, and the utilization of inorganic N during this period is shown as an indefinite extrapolation of the nitrogen utilized by the end of August. From these curves and Figure 5 it may be seen that the fertilizer was the principal source of new nitrogen during the period July-August when the lake nitrate was practically exhausted in the euphotic zone. Ammonia values tended to show sporadic in- creases during 1970, and at times ammonia may have been the principal inorganic form of ni- trogen in the lake, probably through being re- cycled as excretory products of the zooplankton (Beers, 1962). However, due to analytical dif- ficulties with this radical, further investigation of its seasonal behavior is required, especially with reference to the verification of high values. Phosphate showed similar variations to nitrate although the depletion of phosphate was less reg- ular. Seasonal concentrations ranged from <0.01 to 0.04 fxg at. P/liter with about 3% of the values falling in a much higher range of 0.1 to 0.6 /Jig at./liter. A determination of phos- phate utilized and phosphate added (similar to the inorganic N budget shown in Figure 7) was difficult to describe because of the unpredictable occurrence of phosphate throughout the summer; this may have been due to phosphate regener- ation. As an overall assessment, however, if a winter level of 0.03 fig at. P/liter were complete- ly utilized in the water column to 30 m, the addition of 100 tons of 27-14-0 would increase the supply of phosphate over the whole lake by a factor of about 450% compared with the increase in the inorganic nitrogen budget of ap- proximately 100% (Figure 7). From winter to summer, silicate concentra- tions ranged from about 1.8 to 3.0 mg silica/ liter. According to Lund (1965) silicate be- comes rate limiting for diatoms at about 0.5 mg/ liter, which is considerably lower than the sea- sonal range for Great Central Lake. BACTERIA Plate counts of bacterial colonies per 100 ml are shown in Figure 8, together with the range of counts obtained on several days when areal surveys were made. During May, the total num- ber of colonies per 100 ml was generally below the mean value of ca. 9,000 reported by Henrici (1940) for oligotrophic lakes; however, there is 17 FISHERY BULLETIN: VOL. 70, NO. 1 >30,000 - 10,000 - o o ac UJ Q. (O UJ o o -J < a: < ffi 1,000 < 100 I I (6) (10) O-r f •O I (3) (1 0) " (9) * • ' 4 • • • o • < ' (? )) o -' ) • - o o o o lOOA 1 1 MAY JUNE JULY AUG Figure 8. — Bacterial colonies per 100 ml surface lake water (O) Station 2; (•) Station 1; (5) number of samples in areal survey and range, I. some indication in the data that bacterial num- bers increased by one or two orders of magnitude during the latter part of June through to August. Summer increases in bacterial flora have been widely observed in lakes (e.g., Snow and Fred, 1926; Nauwerck, 1963) , and while nutrient level could have affected this increase (e.g., see Bosset, 1965), we have no previous data on which to judge the effect. PHYTOPLANKTON SPECIES Principal phytoplankton species from surface samples at Station 1 and 2 during 1970 are shown in Figure 9 on a relative scale. From these re- sults it is apparent that the predominant algae during May and early June were Dinobryon, Rhizosolenia, and Nitzschia. During June and July Gymnodinium, Cyclotella, and the euglenoid Phncus reached maximum numbers but tended to decline by August. Predominant algae of late summer and autumn were the chlorophyte Nan- nochloris and the cyanophyte Chroococcus. Sec- ond maxima in Dinobryon, Rhizosolenia, and Cyclotella occurred during the winter together with a maximum in Tabellaria. Two studies (May and June) on the depth distribution of the principal species showed that maxima in Rhizosolenia, Tabellaria, and Phaciis were found at the bottom of the thermocline (ca. 20 m) ; Cyclotella and Gymnodinium maxima occurred at the top of the thermocline (ca. 10 m) while Nannochloris, Dinobryon, Nitzschia, and Chroococcus showed maxima within the top to 10 m. PRIMARY PRODUCTION Surface primary production values at Station 1 during 1969 and 1970 and at Station 2 during 1970 are shown in Figure 10; the mean and co- efficient of variation of surface primary pro- duction for the months of June to August are also shown on each figure. The total average pri- mary production in the water column to 30 m at Stations 1 and 2 during 1970 was approxi- mately 12 g C/m^/year compared with approx- imately 6 g C/mVyear at Station 1 during 1969. Primary production per unit of chlorophyll a at different depths for Station 1, 1969 and 1970, and Station 2, 1970, are shown plotted against the light intensity at the same depths in Figure 11. A considerable amount of scatter is appar- ent in the data which is partly due to differences in environmental factors as well as to the lack of precision in attempting to establish photo- synthesis versus light intensity relationships on the basis of ecological rather than experimental data. Polynomial curves were fitted to each set of data using an IBM computer. The shape of these curves is consistent with P vs. / relation- ships obtained by physiologists under experi- mental laboratory conditions and differences in asymptotic values reflect differences in the nu- trient supply (Ichimura and Aruga, 1964). 18 PARSONS, STEPHENS, and TAKAHASHI: LAKE FERTILIZATION. I. Dinobryon (100%= 20 4 cells /ml) Cy dote II a (100% = 8 80 cells/ml) Nannochloris (100% =3880 cells/ml) Months M onths Figure 9.— Principal phytoplankton species at Station 1 (•), Station 2 (A), Station 3 (O) during 1970. 19 4.0 35 3.0 2.5 "£2.0 E 1.5- 1.0 0.5- Station I 1969 N 10 P 0.18 CF 72% Jl Station I 1970 N_ II P0.49 CF 69% Station 2 1970 N 10 PI.63 CF 76% T LJ, M ' J ' J ' A M ' J 'J' A I M ' j ' J ' A MONTHS Figure 10. — Surface primary production, May to August. (N = number of samples, P ^= mean surface produc- tion, and CF ^ coefficient of variation — all values for the period, June to August.) However the degree of scatter in the ecological data requires some expression of confidence limits. At Station 1 (1969), which was located at a considerable distance from the area of fertil- ization, 95 /f confidence limits for the asymptotic value of 1.03 mg C/mg Chi a/hr were 0.84 and 1.22; for 1970 at the same station the 95% con- fidence limits for the asymptotic value of 1.55 Figure 11. — Productivity indices plotted against light intensity at Station 1, 1969 and 1970, and Station 2, 1970 (O O computed best fitting polynomial curve). 2.0 1.5 1.0 0.5 FISHERY BULLETIN: VOL. 70, NO. 1 ST. I, 1969 3.0 2.5 -c 2.0 o M .»: » J L 1.5 o £ 1.0 05 14.0 12.0 10.0 8.0 6.0 4.0 2.0 ST. I, 1970 ST. 2,1970 J_ 0.1 0.2 0.3 0.4 I y / mifi L L_ a5 06^ 20 PARSONS, STEPHENS, and TAKAHASHI: LAKE FERTILIZATION. I. mg C/mg Chi a/hr were 1.12 and 1.98. At Sta- tion 2 in 1970, however, the scatter of points is so great that 95% confidence hmits become very large. The probable reason for this is that the station was sometimes in the area to which nu- trients were first added, and sometimes the move- ment of water containing freshly added nutrients was away from Station 2 (Figure 1) . If in fact it is assumed that there were only two alter- natives in such a narrow lake (i.e., movement of nutrients towards or away from Station 2) then the 50% confidence limits for the asymp- totic value of 4.17 mg C/mg Chi a/hr were 2.26 and 6.07. DISCUSSION The principal purpose of this report is to establish the effect of inorganic nutrient enrich- ment on the primary production of Great Central Lake. From data in Figure 10 it is quite ap- parent that primary productivity was increased in surface samples during 1970 compared with 1969, both at Station 1 and particularly at Sta- tion 2, which was very close to the area of re- peated enrichment. However, while the effect of nutrient enrichment was apparent to the extent of a tenfold increase in surface primary pro- ductivity, the integrated productivity for the water column only showed an approximate dou- bling in primary productivity during the first 3 months of nutrient enrichment (see Parsons et al, in press, for primary production depth pro- files) . This result is in keeping with the fact that the total inorganic nitrogen addition to the lake (Figure 7) was only sufficient to approx- imately double the natural reservoir of inorganic nitrogen in the upper 10 m, based on winter ni- trate levels. However, it does not take into ac- count nitrogen fixation by the blue-green alga, Chroococcus, which may have taken arvantage of the increased supply of phosphate to become one of the predominant summer plankters. The question is, whether some factor other than fertilization could have accounted for the increased primary productivity? Firstly, it is apparent that since the largest increase in pri- mary productivity occurred at the surface, it can- not be argued that the increased primary produc- tivity was due to greater enrichment of the water column from the hypolimnion, especially in view of the high degree of stratification (Figure 3) and apparent nitrate depletion in the epilimnion (Figure 6). It might be argued that the in- creased productivity was due to an increase in standing stock of primary producers and in- creased radiation. Data in Figure 5 indicate that the standing stock of primary producers at Station 2 was generally higher than at Station 1 during 1969, although the effect is within a 95% probability of being due to within-lake variations in standing stock of chlorophyll a. However, in order to examine this question in more detail, primary productivity data for Sta- tions 1 and 2 in 1970 and Station 1 in 1969 have been expressed as the production per unit chlor- ophyll ft and plotted against the calculated light intensity at various depths (Figure 11). This presentation of data has been used by Ichimura and Aruga (1964) to compare the productivity of oligotrophic, mesotrophic, and eutrophic lakes under conditions of different standing stocks of primary producers, light conditions, and photo- synthesis. From their findings it was concluded that oligotrophic lakes had a productive index of between 0.1 and 1.0 mg C/mg Chi a/hr, which is very similar to the range of values computed from the data in Figure 11 for Station 1 during 1969. The computed range for Station 1 during 1970 was appreciably higher, however, and en- ters the classification for mesotrophic lakes which have a photosynthetic index of up to 2 mg C/mg Chi a/hr; finally the asymptotic value (4.17) from Station 2 in 1970 is within Ichi- mura's and Aruga's (1964) range for eutrophic lakes, which the authors report as having photo- sjnithetic indices of up to 6 mg C/mg Chi a/hr. Since the only basis for this classification is the effect of nutrient enrichment in enhancing the photosynthetic response, it may be concluded that our observed increase in primary produc- tivity was determined by the artificial addition of fertilizer. Secondary effects of nutrient enrichment may also have influenced the primary formation of particulate material through a heterotrophic cycle. Unfortunately, our evidence for this is not substantial and rests mainly on the increase 21 FISHERY BULLETIN: VOL. 70, NO. I in bacterial numbers (Figure 8) and the fact that very high dark uptake of ^^C-bicarbonate (up to 50^; of the light bottle uptake) were encountered during the summer at some stations following fertilization. We are at present not sure of the accuracy of this result, however, and it will be reinvestigated during 1971. Nauwerck (1963) has concluded that the heterotrophic for- mation of particulate material is a principal mechanism for supplying food to particle feeders in some lakes and one might expect this mech- anism to be enhanced by the additional avail- ability of nitrogen and phosphorus. The most interesting aspect of changes in the species composition of the principal primary producers is that in spite of differences in sur- face primary productivity at Stations 1 and 2 during 1970 (Figure 10) the relative abundance of principal species at these two stations (and on several occasions at Station 3) was substan- tially the same (Figure 9) . This was important because it was intended that there should be no change in the species composition of organisms leading up the food chain to young salmon, but only an increase in their productivity. In ad- dition, the occurrence of the Cyclotella-Chroococ- cus association is characteristic of oligotrophic lakes (Hutchinson, 1967) which indicates that the general classification of the lake (based on species association) had not been changed by fertilization. However, some eutrophic species of phytoplankton, such as Ceratium, Peridwium, and Scenedesmtis, were also observed as minor constituents of the plankton, especially during the summer. In conclusion, it appears that the fertilization of Great Central Lake resulted in an increased primary production but did not substantially change the standing stock of primary producers, water clarity, or the principal phytoplankton species at locations near and distant from the site of nutrient enrichment. The effect of zoo- plankton on the primary producers was essen- tially to suppress the increase in standing stock of phytoplankton while the standing stock of zooplankton itself increased by almost an order of magnitude. Zooplankton growth and distri- bution are described in the second paper in this series (LeBrasseur and Kennedy, 1972). ACKNOWLEDGMENT The authors wish to acknowledge the assist- ance of S. Sheehan in carrying out analyses of water samples. LITERATURE CITED Beers, J. 1962. Ammonia and inorganic phosphorus excre- tion by the planktonic chaetognath Sagitta lispida Conant. In J. H. Ryther and D. W. Menzel (edi- tors), The biochemical circulation of elements in the Sargasso Sea. Append., Prog. Rep. 1, Sept. 1961 to 31 Mar. 1962, Bermuda Biol. Stn., 6 p. BOSSET, E. 1965. Incidences hygieniques de la vaccination des eaux de boisson au moyen de polyphosphates. MonatsbuU. Schweiz. Ver. Gas-u. Wasserfachm. 45: 146-148. Donaldson, L. R., S. M. Olsen, P. R. Olson, Z. F. Short, J. C. Olsen, H. E. Klassen, and R. W. Kiser. 1968. Fern Lake program. In Research in fish- eries . . . 1967, p. 38-39. Coll. Fish. Fish. Res. Inst., Univ. Wash. Contrib. 280. Henrici, a. T. 1940. The distribution of bacteria in lakes. Am.. Assoc. Adv. Sci., Publ. 10: 39-64. Hutchinson, G. E. 1967. A treatise on limnology. II. Introduction to lake biology and the limnoplankton. John Wiley and Sons Inc., p. 355-397. ICHIMURA, S., AND Y. ArUGA. 1964. Photosynthetic natures of natural algal com- munities in Japanese waters. In Y. Miyake (edi- tor) , Recent researches in the fields of the hydro- sphere, atmosphere and nuclear geochemistry, p. 13-37. Maruzen Co., Ltd., Tokyo, Jerlov, N. G. 1957. Optical studies of ocean waters. Rep. Sweden Deep Sea Exped. 3: 1-59. Johnson, W. E. 1965. On mechanisms of self-regulation of popu- lation abundance in Oncorhynchus nerka. Mitt. Int. Ver. Limnol. 13: 66-87. Krokhin, E. M. 1967. Influence of the intensity of passage of the sockeye salmon Oncorhynchus nerka (Wald.) on the phosphate content of spawning lakes. Izd. "Nauka", Leningrad 15(18): 26-31. (Fish. Res. Board Can. Trans. Ser. 1273.) LeBrasseur, R. J., and 0. D. Kennedy, 1972. The fertilization of Great Central Lake. II. Zooplankton standing stock. Fish. Bull., U.S. 70: 25-36, 22 PARSONS, STEPHENS, and TAKAHASHI: LAK£ FERTILIZATION. I. Lund, J. W. G. 1965. The ecology of the freshwater phytoplankton. Biol. Rev. (Cambridge) 40: 231-293. Nauwerck, a. 1963. Die Beziehungen zwischen Zooplankton und Phytoplankton im Zee Erken. Symb. Bot. Ups. 8(5) : 1-163. Nelson, P. R., and W. T. Edmondson. 1955. Limnological effects of fertilizing Bare Lake, Alaska. U.S. Fish Wildl. Serv., Fish. Bull. 56: 414-436. Parsons, T. R., and G. C. Anderson. 1970. Large scale studies of primary production in the North Pacific Ocean. Deep-Sea Res. 17: 765-776. Parsons, T. R., C. D. McAllister, R. J. LeBrasseur, AND W. E. BARRACLOUGH. In press. The use of nutrients in the enrichment of sockeye salmon nursery lakes - a preliminary re- port. FAO Technical Conference on Marine Pollution, Rome, Dec. 9-18, 1970. RUGGLES, C. p. 1965. Juvenile sockeye studies in Owikeno Lake, British Columbia. Can. Fish. Cult. 36: 3-21. Snow, L. M., and E. B. Fred. 1926. Some characteristics of the bacteria of Lake Mendota. Trans. Wis. Acad. Sci. Arts Lett. 22: 143-154. Strickland, J. D. H., and T.R. Parsons. 1968. A practical handbook of seawater analysis. Fish. Res. Board Can., Bull. 167, 311 p. 23 THE FERTILIZATION OF GREAT CENTRAL LAKE 11. ZOOPLANKTON STANDING STOCK R. J. LeBrasseur and 0. D. Kennedy^ ABSTRACT The regional, vertical, and seasonal abundance of the dominant zooplankton species were studied in con- junction with a series of nutrient additions to Great Central Lake. Two rotifer species, Kellicottia spp., Conochilus unicornis, three cladocera species, Bosmina coregoni, Holopedium gibherum, and Daphnia longiremis, and three copepod species, Cyclops bicuspidutns thomasi, Epischura nevadensis, and Diap- tomus oregonensis were the most numerically abundant zooplankton species. The introduction of the fertilizer and the consequent higher rate of primary production produced no changes in the species com- position. The zooplankton exhibited a relatively uniform horizontal distribution within the upper 20 m along the lake, a factor which was attributed to the lake circulation. All eight species were concentrated in the euphotic zone (upper 40 m), and five were most abundant in the upper 10 m. The center of abun- dance for the remaining three species was between 20 and 30 m depth. The respective depths of maximum abundance for the various species showed little variation between daylight and darkness. Seasonally, there were two periods, June to July and September to October, of maximum abundance for most spe- cies. The cause for somewhat lower levels of abundance in August is not known. The average zoo- plankton biomass showed a similar seasonal pattern with a maximum weight in July which exceeded 8 g/m2. The average biomass over a 6-month period, May through October, exceeded 5 g/m^ (more than 10 times greater than for the comparable period prior to fertilization in 1969). In contrast to the high standing stock of zooplankton, the estimated growth rate for underyearling sockeye salmon, the principal predator species in the lake, was only slightly improved over 1969 (1.2 vs. 0.99^ /day). In comparison with other lakes producing young salmon the growth rates appear low with respect to the zooplankton stock. It was suggested that the temperature structure of the lake, 14° to 23°C above the thermocline and 4° to 6°C below the thermocline, may reduce availability and prevent the efficient utilization of the zooplankton by the underyearling sockeye salmon. The following account is the second in a series of papers which report on the effects of sustained nutrient additions to an oligotrophic lake. In the first report, Parsons et al. (1972) showed that an increased primary productivity resulted from nutrient additions made to Great Central Lake, B.C.; the objective of this report is to de- termine if nutrient additions affected the stand- ing stock and diversity of secondary produc- ers. The overall purpose of these studies has been to determine if nutrient additions will increase sockeye salmon (Oncorhynchus nerka) produc- tion; zooplankton, as the principal food of under- yearling sockeye salmon, occupy a central posi- tion in the food chain of young sockeye during lake residence. Previous studies (Ivlev, 1961; ^ Fisheries Research Board of Canada, Biological Sta- tion, Nanaimo, B.C., Canada. Johnson, 1965; Brocksen et al., 1970) have sug- gested that prey density and availability may limit the predator biomass. The latter authors compiled data for several sockeye nursery lakes with which they were able to demonstrate a direct relationship between mean zooplankton biomass (prey) and the mean growth rate and biomass of underyearling sockeye salmon (pred- ator). Other studies (Ricker, 1962) have indi- cated that the ocean survival, i.e. the return to coastal waters of adult sockeye salmon, can be directly correlated in many instances with the size at which the sockeye as year-old migrants leave the nursery lakes to enter the ocean for 2 or more years. While the above studies rely heavily upon circumstantial data, as well as data which were collected for other purposes, they serve as a rational basis for attempting to increase the available zooplankton biomass for the enhancement of salmon growth. Manuscript accepted September 1971. FISHERY BULLETIN: VOL. 70, NO. 1, 1972. 25 FISHERY BULLETIN: VOL. 70, NO. 1 METHODS Sampling of zooplankton was initiated in mid- 1969, using a 0.25 m- mouth area cylinder-cone net with 100 micron mesh aperture, hauled ver- tically from 20 m or 50 m. The samples were collected at infrequent intervals during 1969 and the first 14 weeks in 1970; thereafter vertical net hauls were made at two locations at least once every 4 days until the first week of No- vember. Additional vertical net hauls were made once or twice each month from the lake bottom, 200 m, to the surface. Miller nets (Mil- ler, 1961) with 0.01 m- mouth area and 100 mi- cron mesh aperture were used at weekly intervals during the period June through August and thereafter at monthly intervals to determine the area] and vertical distribution of zooplankton. The areal sampling at 18 locations along the lake consisted of 5 min oblique tows from 20 m to the surface while underway at 2 m/sec. The daylight vertical distribution of zooplankton was monitored at 18 depths between the surface and 65 m by making three consecutive tows each with six Miller nets at 2 m/sec at one location. Ad- ditional tows were made to sample other depths and also at other periods of the day. Details of the sampling and sampling locations are re- ported elsewhere (Kennedy et al., 1971"). ' Kennedy, 0. D., K. Stephens, R. J. LeBrasseur, T. R. Parsons, and M. Takahashi. 1971. Primary and sec- ondary productivity data for Great Central Lake, B.C., 1970. Fish. Res. Board Can., Manuscr. Rep. No. 1127, 379 p. In the analyses of samples special effort was made to maintain up to date species counts and measurements for comparison with other events as they were occurring in the lake. The common zooplankton constituents were identified, mea- sured into size categories, and counted from an aliquot of the total sample ; fractions of 1/50 or 1/100 using a Stempel pipette were used de- pending upon the sample size. The size cate- gories (length in microns) reported in Table 1 were based upon individual measurements for different stages of development of the respective species. It is to be noted that the lengths refer to mean sizes of organisms during the spring and summer growing period. Individual length measurements for the different species and for diflferent times of the year may be found in the MS data report (see footnote 2). Species counts in vertical net hauls are re- ported as number per m^, in oblique and hori- zontal tows as number per m^ In this report, unless otherwise indicated, counts all refer to numbers of individuals which fall within the size range occupied by mature (Stage VI) copepods and egg-bearing cladocera; these were usually the two largest size groups for the species re- ported in Table 1. RESULTS SPECIES Table 2 lists the species of zooplankton which have been found in Great Central Lake. Addi- tional species may be present as minor constitu- Species Table 1. — Zooplankton size ranges for species sorting in Great Central Lake. Group III IV VI VII Cyclops bicuspidatui Egg 125-275 C. vernalii Egg 275-375 Epischura nevadensis Egg 275-450 Diaplomus orfgoneniis Egg 125-275 D. kenai Egg Bosmina Egg 125-225 llolopedium Egg 375-450 Daphnia longirtmii Egg 450-750 D. pultx Egg 450-650 Kellicottia Egg ca. 80 Conochilus Egg ca. 80 Keratella Egg ca. 80 Size range y. 375-550 550- 750 750- 850 850- 950 950-1,100 750- 900 900-1,100 1.100-1,350 450-650 650- 900 900-1,100 1,100-1,350 1,350-2,250 450-750 750- 900 900-1,100 1,100-1,350 650-1,100 1,100-1,750 1,750-2,500 225-325 325- 450 450- 650 650- 600 450-750 750-1,100 1,100-1,750 450-750 750-1,100 1,100-1,750 450-750 750-1,100 1,100-1,750 1,750-2,500 26 LeBRASSEUR and KENNEDY; LAKE FERTILIZATION. II. Table 2. — Zooplankton species found in Great Central Lake, 1970. Rotifera *Kellicottia spp. Ktratella cochlearis K. quadrata *Conockilus unicornis Cladocera *Bosmina coregoni *Holopedium gibberum "Daphnia longiremis D. pulex Scapholeberis kingi Polyphemus pediculus Alona affinis Copepoda *Cyclops bicuspidatus thomasi C. vernalis *Epischura nevadensis *Diaptomus oregonensis D. kenai Unknown Actinopoda Pollen Egg clusters Arachnoidea (mites— 2 spp.) Chironomid larvae Fish larvae (cottid) * Indicates the most common species. ents of the zooplankton and it is also possible that new species are being introduced into the lake through a hydroelectric installation which discharges water from an adjacent watershed into the lake. It will be noted from Table 2 that the common zooplankton constituents con- sisted of two rotifer species, three species of cladocera, and three species of copepods; these species are identified throughout the text by their generic names. There has been no change in the species composition during the course of the experiment, i.e. the common species have re- mained numerically abundant while the rare spe- cies have continued to occupy a minor role. PATCHINESS It was anticipated that the zooplankton would exhibit contagious distributions reflecting local circulation patterns, species preferences, and predation. Accordingly, oblique samples from 20 m were collected at weekly intervals at 18 positions along the lake, both near the shore and in midlake. In general, with the exception of the area near the inlet and outlet of the lake where the abundance of organisms was some- times low, there was greater variability found with respect to the date of sampling than the lo- cation of sampling. Weekly means and standard deviations computed for each species showed that Cyclops was the only species in which the standard deviation exceeded the weekly count for more than half the surveys (11 out of 17). The apparent variability in Cyclops abundance might be due contagion or, more likely, to the fact that sampling was limited to depths (20 m) where Cyclops were seldom abundant (see sec- tion on vertical distribution) . The major source of variability in weekly mean counts appears to be associated with the number of organisms counted, i.e. the number of organisms in a sample and the size of the aliquot counted. The weekly mean number of organisms for each species were grouped together with their respective standard deviations as follows: 1-50, 51-250, 251-500, 501-1,000, 1,001-2,000, 2,001-5,000. The mean coefficient of variation (C.V.), the range, the number of means present in each group, and the number of times a standard deviation ex- ceeded its respective mean are shown in Figure 1 (e.g. for 50 or fewer organisms counted, the standard deviation in 17 out of 23 samples ex- ceeded the mean). The magnitude of C.V., or the relative variation about a mean, is closely associated with the number of organisms counted. The high degree of variability about a mean of 50 or fewer organisms reflects counting errors due to the subsampling technique used in the initial analyses of the samples. However, counts of organisms of 250 or more per m^ tend 5100 c 50 o o 1 1 r 2 000 Meon (No/m3) Figure 1. — Coefficient of variation computed for mean counts of species sampled in oblique (20 m to surface) tows, where N is the number of weekly means, m and cr is the standard deviation. 27 FISHERY BULLETIN: VOL. 70, NO. 1 to be relatively uniform, suggesting that hori- zontal patchiness or local contagion is not a ty- pical feature of Great Central Lake zooplankton. The observations of lake circulation (McAllister, personal communication) (Parsons et al., in press), and the chlorophyll a distribution (Par- sons et al, 1972) confirm that the epilimnion is well mixed, thus assuring a nearly uniform dis- persal of planktonic organisms along the lake. The 50-m vertical hauls made at Stations 1 and 2 provide further opportunity for examining the variability with respect to different sampling lo- cations. Here, the comparisons of mean counts indicated a high degree of similarity between the two locations with respect to species comj^osition, stage of development, and abundance. However, examination of samples collected on the same day indicated a high degree of variability. For- ty-nine samples were collected from Stations 1 and 2 during May through December; species counts for Station 1 were plotted against the re- spective count for Station 2. Values which fell outside of a mean ±- half the expected mean (where the expected mean equals half the counts for Stations 1 and 2 combined) are tabulated in Table 3. On half the sampling dates the counts for a particular species tended to be si- milar at both locations (e.g. in the first column of Table 3, the number of samples with a mean ± m/2 is generally greater than half the total number of samples, N/2). Greater numbers of four species were found at Station 2 than at Station 1. It is noteworthy that three of the four species, Kellicottia, Cyclops, and Daphnia, have their greatest abundance below the epilimnion at depths greater than that sampled on the areal surveys. However, there was no apparent cor- Table 3. — Comparison of counts of zooplankton from 50-m vertical hauls at Stations 1 and 2, N = 49. Species Counts = (m : 2 Station I >(m + m) Station 2 <(m — m) 2 Cyclops 33 5 11 Epischura 31 10 8 Diaptomus 24 10 15 Bosmina 26 17 6 Ilolopfdium 25 15 9 Daphnia 22 6 21 Kellicottia 24 9 16 Conochilus 29 14 6 relation in the relative abundance of any of these three species with respect to each other or to other species at either station. There were, however, periods when three or more species would be more abundant at one po- sition than at the other. For example, from July 3 to July 21 (six sets of samples) three to six species were most abundant at Station 1 while during the period August 21 to September 8 (six sets of samples) three to seven species were most numerous at Station 2. Similarly, for other periods of 4 to 12 days, one or another species was in greater abundance at one station than the other. These data have not been examined further to show if variation in species abundance be- tween sampling positions can be correlated with variations in the lake circulation or other envi- ronmental factors such as fertilization or pre- dation by underyearling sockeye. However, it is apparent that all seasonal changes in species composition and abundance were reflected throughout the near-surface waters of the lake and that no local area of high or low zooplankton concentration could be clearly defined within the main body of the lake. VERTICAL DISTRIBUTION Horizontal tows made within the upper 60 m revealed marked differences in species compo- sition and abundance with depth during the pe- riod of thermal stratification. As an example the weekly tows made during July were com- bined and the average concentration of each spe- cies at each of 17 depths sampled during daylight is shown in Figure 2. The inset associated with each species distribution shows the relative dis- tribution (25% quartile intervals) of the respec- tive populations sampled during a 24-hr period in August. Five of the eight species shown in Figure 2 have their maximum concentration within the upper 10 m, while the maximum concentration of the other three species was below 20 m. Thus the species maxima fall either above or below the thermocline. However, it should be noted that the number of organisms per m^ decreased from a maximum of greater than 7,000/m^ be- 28 LeBRASSEUR and KENNEDY: LAKE FERTILIZATION. II. CofiOcMji colonel 400 'III CfClops (.9001 No/m^ 200 400 D''omus ()900) No/m' Epischura (i 900 1 No/m^ 300 600 J \ 1_ Boiiofw al thetmoclnw *6*C Figure 2. — Vertical distribution of common zooplankton species in Great Central Lake mean no./m^ for July 1970. (Horizontal lines indicates the top and bottom of the thermocline, McAllister (personal communication). Inset shows the vertical distribution at Ih'/V quartile intervals over a 24-hr period. Note: the scale indicating the quantity of organisms varies for each species.) tween 3 m and 5 m to a minimum of approxi- mately SOO/m'' at depths below 40 m. The max- imum concentration of individual species ranged from 180/m^ for Daphnia to greater than 3,000/m^ for Holopedium. The tendency for some species to show an increase in abundance in deep samples was likely due to contamination from shallower depths since in the process of setting and hauling with nonclosing nets the deeper nets actually sample for a slightly longer time than the shallower nets. Variations in abun- dance with respect to time of sampling was noted for all species (Figure 2 inset) . The effect was generally most pronounced just after sunset when the maximum concentration per m^ of a species might be increased by 30' Rotifers, which were presumably the least motile of the zooplankton, exhibited the largest shift in abundance towards the surface with the onset of darkness. Some species, notably Holopedium and Eplschura, returned to their daylight depth of maximum abundance within 2.5 hr after sun- set. Other species, such as rotifers, Bosm'ma, and Daphhia exhibited relatively little movement during darkness. It is apparent from Figure 2 that the shift in species abundance were all with- in the daylight range occupied by the bulk of the respective populations. Furthermore, more than 75 Sr of the zooplankton populations were at all times within the euphotic zone (i.e. surface to 30-40 m). SEASONAL ABUNDANCE In Figure 3 the mean monthly numbers of zooplankton are shown for the 50-m vertical haul samples. Three species, Cyclops, Bosmina, and KelUcottia, were relatively abundant through- out the year, whereas the other species were present in numbers which exceeded 1,000/m' for periods of 4 to 5 months. Co7iochilus were the only species present during 1970 to appear subsequent to the initiation of nutrient addition. (They were present in 1969 samples.) Cyclops ranged from a winter minimum of 2,000/m' to a maximum in September and Oc- tober of 30,000/m-. Eplschura were never nu- merically dominant but ranged in numbers from 2,000 to 4,000/m- from May through September. Counts of Diaptomus did not exceed l.OOO/m^ until August, but by September there was a 29 FISHERY BULLETIN: VOL. 70. NO. I No/m ahousoKlsl 10 20 30 40 SO La_l I I 1 I 'mMui&ii^^ J4N FEB MSH iPR MAY JUN JUL AUG SEP OCT NOV DEC JAN FEB MAR Figure 3.-Monthly mean zooplankton abundance (no./m2 in the upper 50 m data from Stations 1 and 2 combined). 7-fold increase which in turn doubled to ca. 26,000/m2 by October. The November-December catches of Diaptomus exceeded lO.OOO/m^, which was approximately two orders of magnitude greater than their standing stock 12 months earlier. Substantial numbers of Diaptomus, 3,000 to 4,000/m-, were carried through into 1971. Bosmina were the most abundant species collected throughout the year. Their numbers ranged from ca. 8,000/m2 in January to ca. 60,000/m- in June and again in October. The December concentrations of Bosmina were twice that of the preceding January. However, by January of 1971 Bosmina had virtually disap- peared from the water column, to 50 m. Holo- pedium attained their maximum abundance in July, approximately 3 months after they began appearing in the samples in significant quan- tities, i.e. greater than l.OOO/m-. Following a secondary maximum in October, Holopedium were virtually absent from samples collected from December through March. Daphnia were the least numerous of the zooplankton spe- cies routinely sampled. They occurred in num- bers of 1,000 to 3,000/m2 from June through September. Kellicottia exceeded 10,000/m2 from May through August and again in October and November. Nearly twice as many Kellicot- tia were present in December as were present at the beginning of 1970. The maximum abun- dance of Conochilus colonies (2,000/m") was during July; no colonies were found prior to June and by December the number of colonies had declined to approximately 500/m-. In toto there were two to three times more zooplankton present in December of 1970 than there were the preceding January. It is of in- terest to note that the greater abundance of zoo- plankton at the end of 1970 was not maintained through the first 3 months of 1971 and further, that Bosmina had been apparently supplanted by Diaptomus in 1971. On a monthly basis there were fewer than 22,000 organisms/m^ , in January while in October, where the max- imum concentration was observed, there were nearly 10 times as many organisms present. Zooplankton counts exceeded 100,000/m- in June, July, September, and October. The de- crease in zooplankton abundance in August was approximately 15 Sr lower than that in either July or September; this decline was attributable mainly to fewer numbers of Bosmina. Individual species counts (4-day running mean number/m-) in vertical hauls have been pre- sented in Figure 4 in order to show the seasonal variations in abundance in greater detail than is shown in Figure 3. The general features of both figures are the same but in Figure 4 the rapid increase and decrease in numbers of some species are shown more clearly, e.g. Holopedium and Epischura. From Figure 4 it is possible to infer some relationship between the addition of nutrients and the appearance of Conochilus or the sustained increase in the abundance of Diap- tomus. It is noteworthy that all species, with the exception of Epischura and possibly Daphnia, went through a secondary maximum in October which was nearly as great as or greater than their level of abundance earlier in the summer. SEX RATIO Adult stages of Cyclops and Epischura showed marked imbalances from an expected 50: 50 ratio of females to males through the year (Table 4). Males of these two species were clearly pre- dominant during the late winter and early spring months. Cyclops females were predominant among the adults taken in June through August whereas Epischura females were never numer- ically dominant for more than two or three sam- pling periods, i.e. July 21 to 31, August 28 to 30 LeBRASSEUR and KENNEDY: LAKE FERTILIZATION. II. H 1 1- -I 1 \ 1 r- r'^^ -4 1 h H ♦- ■MT JUNC JULY AUft SEPT OCT OCC HAT JUNC SePT OCT MOV 0€e •**» JUNE JUL* »U0 SEPT OCT NOV Qt? AUfi «PT OCT MOV OCC Figure 4.— Species counts (no./m^ in 50-m vertical hauls at Stations 1 and 2. (Points represent a 4-day running mean, solid triangle indicates monthly mean, circle with dot indicates more than one sample with the same count. Note: the numbers of organisms are shown on a logarithmic scale.) Table 4. — Copepod sex ratios. Species Month Cycl ops Diapt 7mus Episch F/M ura F/M F2/F1 F/M F2/F1 F2/F1 Jan. -April 0.6 1.2 0.5 May 0.6 2.8 1.5 0.7 0.7 1.3 June 2.6 1.1 1.1 1.9 0.7 1.4 July 1.9 I.O 1.5 1.1 1.0 1.1 Aug. 1.9 1.6 1,3 1.0 1.1 0.8 Sept. 0.8 2.2 1.2 0.9 1.4 0.9 Oct. -Dec. 1.0 2.0 1.3 0.9 1.0 0.5 F/M Number of adult females/number adult males. Fi'/Fi Number of adult females at Station 2/number of adult females at Station I. September 14. In contrast there was a tendency for Diaptomus females to be slightly more abun- dant than males throughout the year. The only period for which Diaptomus males were con- sistently more numerous than females was from June 22 to July 10. Included in Table 4 is the ratio of the number of female copepods at Sta- tion 2 to the corresponding number at Station 1. Cyclops was the only species in which the fe- males were as numerous or more numerous at Station 2 than at Station 1. EGG PRODUCTION Counts were made of all readily identifiable eggs; these consisted of eggs in the brood pouch of cladocera and the egg sacks of Cyclops and Diaptomus. Rotifer species were not examined for eggs, while Epischura eggs were positively identified on only one occasion from a horizontal tow made at 1 m depth in August. It is possible that Epischura eggs develop close to the surface at depths of less than 1 m since they were not found at other standard depths sampled between 1 m and 65 m. Also other data, not presented here, indicate that the smaller size groups of Epischura were found closer to the surface than the adult stages. The data presented in Figure 5 show the ratio of eggs per female for vertical samples collected at Station 1 and Station 2. It was noted in Table 4 that maximum numbers of copepod females oc- curred during the summer, June through Sep- tember; Cyclops females were more numerous at Station 2 than Station 1 and Diaptomus 31 FISHERY BULLETIN: VOL. 70, NO. 1 o E a; UJ O 0} o E JAN ' FEB ' MAR' APR ' MAY ' JUN JUL ' AUG' ' SEP 'OCT ' NOV ' DEC 32 LeBRASSEUR and KENNEDY: LAKE FERTILIZATION. II. females were in about equal numbers at both stations. In Figure 5 the eggs per Cyclops were about equally numerous at both stations. There were three and possibly four periods of maximum egg production for Cyclops females, i.e. June, mid-July through to the third week in August, and the last week of September through to the first week of November; the lat- ter time interval could possibly be interpreted as consisting of two separate periods of egg pro- duction (late September and late October). Diaptomus females at Station 1 had two major periods of egg production, mid-June through mid-July and mid-August through the first week of September, with a period of relatively low egg production from mid-July to the end of August. At Station 2 there was no clear cessation of Diaptomiis egg production from mid-June through to the first week of September. For a major part of this period there were more than 10 eggs per female being produced. There was also a brief period of Diaptomus egg production in mid-May. The production of Bosmlna eggs ranged be- tween and 0.5 per individual. From May through mid-August more eggs were produced at Station 2 than at Station 1 and thereafter the egg production was nearly equal at both sta- tions. The summer minimum which occurred in the first 2 weeks of August was followed by a rise in the number of eggs in the first week of September continuing until the end of the third week of September. The summer max- imum of adult Bosmina shown in Figure 4 oc- curred approximately 1 week after that of the eggs while the maximum standing stock of Bos- mina (which occurred in mid-October) was pre- ceded by the production of eggs 3 to 5 weeks earlier. Holopedium exhibited two clearly de- fined peaks in the production of eggs, from the first to the third week of June and again from the first to the third week of September. The corresponding maximum in the standing stock of Holopedium shown in Figure 4 occurred Figure 5. — Ratio of the number of eggs to the number of adult females. The data from 50-m vertical samples at Stations 1 and 2 have been averaged to give a 4-day running mean ratio. Note: scale changes for different species. from the second week of July through to August 25 and from September 27 to about October 20; the summer minimum occurred between the two peaks. The production of Daphnia eggs took place from June to mid-September with a second brief rise in egg production during mid-October at Station 2. The numbers of eggs produced per female at Station 2 by all species of clado- cera was generally greater or equal to that at Station 1. It should be noted that the latter was found for both the prefertilization period in May as well as during the period of nutrient additions. ZOOPLANKTON BIOMASS The wet weights for 1970 50-m vertical hauls at Stations 1 and 2 were combined and expressed as a monthly mean wet weight (g/m-) together with the range about the mean weight (Figure 6) . Included in Figure 6 (below) are individual weights for the 1969 sampling. The maximum wet weight in 1969 never exceeded 1 g/m- where- as in 1970 the weights ranged as high as 15 g/m^. The average wet weight of zooplankton during the period May through October was approxi- mately 0.5 g during 1969; for the same period in 1970 the average weight was 10 times larger, i.e. 5.3 g. The sample weights increased at a rate of 3'^r per day May through July to a max- imum average wet weight of 8.6 g/m-; there- JON FEB M4HCH IPBIL MAY JUNE JULY 4UG SEPT OCT NOV DEC Figure 6. — Zooplankton wet weight (g/m^) for 50-m vertical hauls. In the lower part of the figure, the points marked "x" indicate individual weights (g/m2) for 1969 samples. 33 FISHERY BULLETIN; VOL. 70, NO. 1 after the weights declined at an average rate of 1.5^'c per day to the December biomass of 0.9 g/m-. The actual rate at which the mean weight of zooplankton declined in any one month fol- lowing July was greater than the average com- puted above due to the increase in biomass in October. Inspection of Figures 3 and 4 indicates that the decline in biomass seen in August was a result of fewer numbers of Epischura, Bosmi- na, and Holopedium. Epischura never reached its earlier level of abundance after August and was virtually absent from the samples by Octo- ber whereas most other species, Daphnia excep- ted, showed an increase in abundance in October which gave rise to the October increase in bio- mass. Dry weights of Great Central Lake zooplank- ton (determined by the freeze-dry method) ranged from 14 '/f to 26 % of the wet weight M'ith a mean of 19 '^r . The variation in the percentage dry weight was directly attributable to the spe- cies composition of a sample. For example, the dry weight of Holopedium was 14 '"r of their wet weight, whereas the dry weight of Cyclops was approximately 26 ^V of the wet weight. The av- erage dry weights of the summer zooplankton (May through October) integrated over a 25-m column, i.e. the depth range in which most zoo- plankton were concentrated, for 1969 and 1970 was 4 mg and 40 mg/m^ respectively (from Fig- ures 2 and 6). Table 5. Species -Length- weight measurements of adult crustaceans. Mean length (Microns) Wet weight (Micrograms) Cyclops Diaptomus Epischura Bosmina Holopedium Daphnia 960 6 1,100 11 1,500 66 300 4 900 \7 900 10 Length-wet weight determinations were made for different sizes and stages of the common crustacean species and the data are summarized in Table 5. The data in Table 6 were obtained by multiplying the maximum concentration (no./m^) of a species within particular depth intervals (Figure 2) by their respective weight from Table 5, thereby providing a measure of biomass with depth. Included in Table 6 are the mean July temperatures within the respective depth intervals. Nearly 60 ^r of the total bio- mass occurs in the upper 10 m where the mean temperature was about 18°C. In the thermo- cline, from 10 to 20 m, with a temperature range from 12° to 6°C (mean temperature, 9°C) the biomass was about 50 mg/m^ or approxi- mately 30 Sr of the total. From 20 to 30 m depth the biomass was about 89r of the total. The re- maining 3 to A*^ of the total biomass occurred below 30 m (30 to 60 m). While these data were derived from July sampling it should be noted that the general distribution of the biomass with depth was similar throughout the period of thermal stratification, i.e. June to October. DISCUSSION The zooplankton standing stock in 1970 shows a phenomenal increase over 1969. This can be largely attributed to the affect of the nutrient additions upon the rate of primary production. The results of Parsons et al. (1972) demonstrate a marked increase in the rate of primary pro- duction within the upper 5 m; at the same time there was little or no change in the standing stock of primary producers. While experiments and observations of a direct relationship between particular species of primary and secondary pro- ducers have not been attempted, the obvious in- ference is that the zooplankton through increas- Table 6.- — Relative biomass of July crustacean zooplankton in various depth interval 3 (from Figure 2). Depth range (m) mT °C Mean maximum b iomass (mg/m' ) Cyclops Diaptomus Epischura Bosmina Holopedium Daphnia Total 0-10 10-20 20-30 >30 18 9 6 <5 1.2 3.1 .6 1.3 .3 .3 39.6 19.8 4.9 2.3 7.2 2.8 1.2 .7 51.0 28.9 3.4 1.7 .1 .5 1.8 .2 99.2 53.5 14.7 5.6 34 LeBRASSEUR and KENNEDY: LAKE FERTILIZATION. II. ing stock size were able to utilize the higher rates of primary production. It is also apparent that the higher biomass in 1970 cannot be entirely attributed to fertiliza- tion since the biomass in May (prefertilization) was also higher than any of the 1969 values. However, the nearly continuous production of eggs by most species and the maintenance of an increased standing stock over a 6-month period are indicative of a direct relation between zoo- plankton and nutrients. It is also noteworthy that there was no change in species diversity. The techniques employed for wet weight de- terminations in this study have produced weights which are apparently lighter than would be ob- tained by other investigators. Wet to dry ratios in the literature suggest that the dry weight is 5% to 10% of the wet weight. Schindler and Noven (1971) employed a ratio of 6%, although their reason for using this particular value is not given; the present results indicate that the dry weight is 19% of the wet weight. Conse- quently, the present weights could be increased approximately three times for comparison with other studies. Thus in the lakes which range from oligotrophic to eutrophic, listed by the above authors. Great Central Lake has, in terms of its mean summer zooplankton biomass, changed from oligotrophic to oligotrophic-meso- trophic, i.e. 12 mg in 1969 to 120 mg dry weight/m^ in 1970. In lakes producing sockeye salmon the mean abundance of zooplankton ranges from values which are less than 5 mg dry weight/m^ to greater than 1 g dry weight/m^ (Johnson, 1965). The mean concentrations in Great Central Lake have increased from the very low end of the range to values which are com- monly reported for some of the larger sockeye producing lakes, e.g., Babine Lake. Johnson (1965) concluded that there was a general relationship between the rate of growth of underyearling sockeye and zooplankton abundance. However, he also suggested that with increasing fish density food abundance was supplanted by a space effect as a limiting factor. In Great Central Lake the underyearling sock- eye in October of 1970 were ca. 30% heavier than fish caught in October of 1969 (Parsons et al., in press; Barraclough and Robinson, 1972). In addition to the increase in weight these authors report (on the basis of the number of adult salm- on spawning) that the number of sockeye fry in the lake were from two to five times more numerous than in the previous year. Assuming an initial weight of 120 mg for individual fry of each year the respective rate of growth over their first 200 days of lake residence was 0.9% and 1.2% per day for 1969 and 1970 respec- tively. The increased growth rate of sockeye in 1970 is less than might be anticipated from the 10-fold increase in zooplankton abundance. Johnson's data (1965) indicated that a pop- ulation density of 1 fish per m- might be the point at which space becomes a factor limiting growth. The maximum estimate of 1 x 10'' sockeye in Great Central Lake during 1970 is approximately 1 fish in every 5 m^. Conse- quently it appears unlikely that the density of the fish population in Great Central Lake limited their growth. Among other factors which limit growth of sockeye, Foerster (1968, Figure 45) indicates that temperature has a major affect upon growth and the efficiency with which food is utilized. The optimum temperature for food conversion for sockeye lies between 10° and 15°C. At higher or lower temperatures the efficiency of food conversion decreases, especially at temper- atures in excess of 20°C or less than 6°C. The laboratory studies of Brett et al. (1969) with fingerling sockeye support the findings reported above. In their experiments 15°C was found to be the optimum temperature for growth at high rations; however, maximum efficiencies with which a ration was utilized occurred at lower temperatures, e.g. the maximum food conversion efliciency of 40% with a 0.2% increase in fish weight per day occurred at a temperature range of 8° to 10°C and a ration of 1.5% of the fish weight/day. Temperatures between 5° and 17°C were found to provide the laboratory fish the optimum conditions for conversion efficien- cies and growth. In Great Central Lake, during their first 200 days of lake residence, the under- yearling sockeye concentrate at depths of 50 m or greater during daylight; with the approach of sunset the fish move to shallower depths and by nightfall the major portion of the population 35 FISHERY BULLETIN: VOL. 70, NO. 1 is at depths between 10 and 20 m while some fraction of the population occur in the upper 10 m. This pattern of vertical migrations ap- pears to be repeated daily (Barraclough and Robinson, 1972). Narver (1970) has reported similar vertical movements for sockeye popula- tions in Babine Lake. For the greater part of the day the salmon in Great Central Lake are at temperatures of 4° to 5°C, a somewhat shorter period (ca. 6 hr) is spent at temperatures of 6° to 12°C (10 to 20 m depth) while a rel- atively brief period (ca. 1 hr) may be spent at temperatures ranging from 14° to 23 °C (0- 10 m depth) . Details of the time actually spent at different depths by the sockeye are reported by Barraclough and Robinson (1972). It is apparent that the fish are utilizing the maximum concentrations of prey which occur at above op- timum temperatures in the upper 10 m for very short intervals. Consequently in assessing the relationship between the increased abundance of prey brought about through fertilization and the sockeye it should be noted that possibly 60 Sr of the total biomass, i.e. the portion in the upper 10 m, may be only partially available to the fish (Table 6) . Furthermore, some prey species be- cause of their size (rotifers) or structure (Holo- pedium) may not be a particularly useful food source for the salmon. Holopedium, for ex- ample, was among the largest and most numer- ous species of crustaceans in the lake; however, a large fraction of their biomass is comprised of a gelatinous material of dubious food value. The difference in the wet to dry weight ratio between Holopedium and other zooplankton (14% to ca. 26^/ -respectively) attests to the water composition of Holopedium. The quality of prey together with the observations of Foers- ter (1968) and Brett et al. (1969) empha- size the need for caution in interpreting preda- tor-prey relations. In the present instance, the benefits of the fertilization appear to have been only partially transferred to the sockeye salmon. Since the thermal structure of the lake is a factor beyond immediate control, it would be in- teresting to consider possible benefits from the addition or deletion of some prey species and to attempt to shift the level of primary and sec- ondary production to depths and temperatures favoring sockeye salmon growth. LITERATURE CITED Barraclough, W. E., and D. G. Robinson. 1972. The fertilization of Great Central Lake. III. Effect on sockeye salmon. Fish. Bull., U.S. 70: 37-48. Brett, J. R., J. E. Shelbourn, and C. T. Shoop. 1969. Growth rate and body composition of finger- ling sockeye salmon, Oncorhynchiis nerka, in re- lation to temperature and ration size. J. Fish. Res. Board Can., 26: 2363-2394. Brocksen, R. W., G. E. Davis, and C. E. Warren. 1970. Analysis of trophic processes on the basis of density-dependent functions. In J. A. Steele (editor), Marine food chains, p. 468-498. Oliver and Boyd, Edinburgh. Foerster, R. E. 1968. The sockeye salmon, Ovcorhynchus nerka. Fish. Res. Board Can. Bull. 162, 422 p. Ivlev, V. S. 1961. Experimental ecology of feeding of fishes (Transl. by D. Scott). Yale Univ. Press, New Haven, 302 p. Johnson, W. E. 1965. On mechanisms of self-regulation of popula- tion abundance in Oncorhynchiis nerka. Mitt. Int. Ver. Limnol. 13: 66-87. Miller, D. 1961. A modification of the small Hardy Plankton Indicator for simultaneous high speed plankton hauls. Bull. Mar. Ecol. 5: 165-172, Narver, D. 1970. Diel vertical movements and feeding of underyearling sockeye salmon and the limnetic zooplankton in Babine Lake, British Columbia. J. Fish. Res. Board Can. 27: 281-316. Parsons, T. R., C. D. McAllister, R. J.LeBrasseur, AND W. E. Barraclough. In press. The use of nutrients in the enrichment of sockeye salmon nursery lakes — a preliminary report. FAO Technical Conference on Marine Pollution, Rome, 9-18 December 1970. Parsons, T. R., K. Stephens, and M. Takahashi. 1972. The fertilization of Great Central Lake. I. Effect on primary production. Fish. Bull. U.S. 70: 13-23. Richer, W. E. 1962. Comparison of ocean growth and mortality of sockeye salmon during their last two years. J. Fish. Res. Board Can. 19: 531-560. Schindler, D. W., and B. Noven. 1971. Vertical distribution and seasonal abundance of zooplankton in two shallow lakes of the exper- imental lakes area, northwestern Ontario. J. Fish. Res. Board Can. 28: 245-256. 36 THE FERTILIZATION OF GREAT CENTRAL LAKE III. EFFECT ON JUVENILE SOCKEYE SALMON W. E. Barraclough and D. Robinson^ ABSTRACT Nutrient levels and rates of primary production in nursery lakes are factors which may limit production of sockeye salmon. This paper describes the effect of artificial fertilization on feeding behavior and growth of juvenile sockeye salmon in Great Central Lake, Vancouver Island, British Columbia. Under- yearling sockeye salmon grew 30% larger in 1970 than in 1969 as a result of adding 100 tons of fertilizer to Great Central Lake. The growth pattern for the whole population was complex, however, and the increase in size of juvenile sockeye was not as much as had been expected from the increase in quantity of their food organisms. The fact that the sockeye did not appear to appreciably crop the high epilem- netic concentrations of zooplankton during July and August 1970 may have been partly due to avoid- ance of high temperatures by the fish. Decomposing carcasses of anadromous fish, such as the sockeye salmon {Oncorhynchiis nerka) , contribute to the fertilization of nursery lakes following spawning in the lake. In most instan- ces the extent of this fertilization is not known but the removal of maturing sockeye by a com- mercial fishery may deny lake waters of their essential nutrients and contribute to lowered productivity. Particular attention has been fo- cussed on the imbalance of phosphate in the na- tural fertilization of lakes from decomposing salmon carcasses (Krokhin, 1959) and the sug- gestion has been made (Krokhin, 1967) that a positive balance should be maintained by the ar- tificial replacement of the phosphate with inor- ganic fertilizers. Early studies carried out in a small unstrati- fied lake in Alaska (Nelson and Edmondson, 1955; Nelson, 1958) showed that the addition of phosphate and nitrate fertilizer resulted in in- creased length and weight of sockeye smolts leaving the lake. The potential role of a natural imbalance of phosphate in nursery lakes on sock- eye salmon is emphasized in the following quo- tation from Foerster (1968): One wonders whether sufficient significance has been given to this feature of the phosphate balance. With ^ Fisheries Research Board of Canada, Biological Station, Nanaimo, B.C., Canada. Manuscript accepted September 1971. FISHERY BULLETIN: VOL. 70, NO. 1, 1972. sockeye populations in all areas showing such evident declines, despite legislation on regulation and limita- tion of fishing, it might well be that some basic factor such as this may be having a much more limiting effect on productivity than seems apparent. In addi- tion to the smaller amounts of phosphorus introduced into a lake in the carcasses of fewer sockeye spawners, there may also be occurring a steady decline in the phosphate content of the runoff waters as the phos- phates of the soil and rock become leached out over the years. Future studies of the phosphate balance of sockeye-producing waters and the direction of its trend may prove most enlightening. Addition of suit- able fertilizers may be found advantageous. In recent years it has become more evident that suitable fertilizers should not only include phosphates but also other nutrients, including trace elements, in order to increase aquatic pro- ductivity (Goldman, 1960, 1964). The theory and application of adding natural fertilizers to aquatic environments has been practiced in fish farming for many centuries. Parsons et al. (in press; 1972) have presented data on various aspects of lake fertilization stud- ies carried out by others. In summary of these findings, there is much evidence to show that the larger the sockeye smolts at the time of seaward migration, the higher the percentage return from the sea (Burgner, 1962; Ricker, 1962). Since food supply is one of the important factors 37 FISHERY BULLETIN: VOL. 70, NO. I governing growth, the effect of increasing the food supply to underyearling sockeye salmon through artificial fertilization of Great Central Lake, B.C., is presented here. It has already been established (Parsons et al., in press; 1972) that the waters of Great Central Lake are relatively unproductive of sockeye salmon, the average size of yearling smolts at the time of seaward migra- tion being much smaller than in Babine Lake, B.C. (63 mm versus 79 mm) (McDonald, 1969). The average size of yearling smolts from 14 other lakes in Washington, British Columbia, Alaska, and Kamchatka is larger than the yearling smolts from Great Central Lake (Foerster, 1968). In the following account particular attention is given to changes in size of juvenile sockeye salm- on in Great Central Lake associated with changes in their food supply prior to and after the addi- tion of inorganic nutrients (see Parsons et al, 1972; LeBrasseur and Kennedy, 1972). STUDY AREA Great Central Lake (Figure 1) is located in central Vancouver Island, British Columbia. The lake is about 33 km long and varies between 1 and 2.5 km in width. The shoreline length is 72 km and the surface area is ca. 51 km^. Ele- vation of the lake surface is 83 m above sea level and the mean depth is 200 m, with a maximum depth of about 280 m. The outlet of the lake runs into the Stamp River. Most of the shore- line slopes very abruptly into deep water. This feature is an important factor in regulating hor- izontal distribution of juvenile salmon in the lake, by providing a maximum amount of the lake surface available to juvenile sockeye. LAKE SPAWNING A brief account of the spawning sites of the sockeye salmon is presented here because the location of the in-lake spawning grounds is an important factor in the emergence of the alevin and dispersal of the fry at the time of their initial intake of food. Little or no published informa- tion is available on the migration and spawning of adults in the lake. Mr. F. C. Boyd of the De- partment of the Environment has kindly granted permission to refer to his internal manuscript reports on the subject. Adult sockeye salmon bound for Great Central Lake first enter the Stamp River as early as the first week in June. This migration up the Stamp River continues through June, peaks in July, and in most years, ends in early August. The peak I2^°25 20' 49''25 49° i^ 05 1250OO' I25°25' GREAT CENTRAL LAKE FISHING STATIONS 20' 15' 10' 49°25 49°20' 05' I25''00 Figure 1. — Great Central Lake showing the six fishing stations and depth contours in meters. 38 BARRACLOUGH and ROBINSON: LAKE FERTILIZATION. III. of migration may not occur until the first week in September. It takes between 2 and 5 days for the sockeye to migrate up the Stamp River into Great Central Lake, depending on the water levels in the river. The fish remain in the lake but do not commence to spawn until the latter part of September. Great Central Lake is one of the few lakes in British Columbia where most of the spawning occurs in areas along the lake- shore rather than in tributary streams. Only a few hundred sockeye spawn in tributary streams a short distance away from the lake. Drinkwater, Lindsay, and Fawn Creeks (Fig- ure 1) receive most of the stream spawners. Lakeshore spawning commences in the last week of September, reaches a series of peaks in three principal locations during October, and ends in November. About 50% of the spawning occurs along 4.63 km of lakeshore between Lind- say Creek and Forestry Creek, 30 '^r along 1.1 km of shoreline west of Fawn Point, and 20 '^'r along 1.6 to 4.8 km of lakeshore off" North Creek. Redds were found at depths between 0.6 and 24 m but most were between 12 to 15 m. Spawnings were observed by scuba divers at depths as great as 41 m. It is now realized that the location of separate major in-lake spawning areas is im- portant in providing the potential basis for the immediate and rapid distribution of juvenile sockeye throughout the lake, shortly after the fry emerge from the gravel and commence to feed. Two spawning areas are adjacent to the lake area where fertilizer was applied (see Par- sons et al., 1972). METHODS LOCATION AND DISTRIBUTION OF JUVENILE SOCKEYE SALMON A high frequency (200 kHz) moist paper re- cording echo sounder (Furuno model No. FNV- 3000)' was used to locate the young sockeye in ^ Reference to trade names in the publication does not imply endorsement of commercial products by the National Marine Fisheries Service. the lake and monitor their horizontal and vertical distribution. During the day, young sockeye are generally distributed throughout the lake at depths between 45 and 90 m, but are most abund- ant at about 65 m. They commence to migrate toward the surface about half an hour before sunset. In the summer months at civil twilight, when the sun is 96° from the zenith (or 6° be- low the horizon) they are distributed irregularly in density between 5 and 30 m. At nautical twi- light when the sun is 102° from the zenith (or 12° below the horizon) the juvenile sockeye form a layer between the depths of 10 and 20 m, with a maximum density of about 14 m. At night during the winter months they are distributed more uniformly between 20 and 60 m. In sum- mer the downward migration commences shortly before sunrise and is usually complete 15 to 30 min after sunrise. The young sockeye were sampled with mid- water trawls. Sampling commenced at night when the fish were in a layer between 10 and 20 m. Samples were also collected during day- light at diff"erent depths throughout the depth range of the young sockeye. The depth of trawling was adjusted to coincide with the depth of maximum fish concentration as shown by the echo sounder traces. FISHING GEAR A trawl net with a mouth opening 3 m wide, 6.1 m deep, and 17.7 m long was towed at 2.7 to 3.2 km/hr by a single vessel, the Decihar, to sample the sockeye between the depths 5 and 25 m. Three mesh sizes of knotless nylon netting were used in the construction of the net: 5 cm and 2.5 cm stretched mesh in the body and 1.3 cm in the cod end. The cod end measured 1.2 m wide by 1.8 m deep at the mouth and it tapered to a blunt end about 76 cm in diameter. A standard Henson plankton net (350/x mesh) 76 cm in diameter at the mouth, was secured to the blunt aft end of the cod end to retain the smallest juvenile sockeye and minimize the loss 39 FISHERY BULLETIN: VOL. 70, NO. 1 of their minute scales by abrasion against net- ting in the main cod end of the trawl. An Isaacs-Kidd midwater trawl was towed from the Decibar to samjile the juvenile sockeye at depths greater than 25 m and to evaluate the fishing capabilities of the large midwater trawl towed at the same depths. The mouth opening of this trawl was 1.9 m^ and the net was con- structed of 6.3 mm stretched mesh knotless netting. FISHING STATIONS Juvenile sockeye were sampled with trawls taken at intervals of about 3 weeks at 6 different stations (Figure 1). Most of the tows were of 30 min duration but some tows wei-e shorter, when the echo sounder traces indicated that young sockeye were especially abundant between 12 and 14 m at night. ANALYSES OF SAMPLES The length of all fish was measured to the near- est millimeter from the snout to the end of the central rays of the caudal fin. This measurement is referred to as the fork length. Lengths of smaller fish were measured in a graduate tray under a binocular microscope; calipers were used for larger individuals. All fish were weighed by fork length groups using a center-loading milligram balance (KERN Model No. T1226-1) . Weights recorded are from "blot-dried" specimens. Moisture was blotted from the exterior of the fish, and gentle pressure was applied to the buccal cavity and branchial chamber to remove moisture from these spaces. Age was determined from scales using y 254 projections of thermoplastic impres- sions. Stomach analyses for food were done on fish selected to represent proportionally as complete a size range as possible. The food weight was measured by subtracting weight of stomach shell from weight of stomach plus food. The number of all species of food organisms were counted according to size and state of condition of each stomach examined. RESULTS FOOD OF UNDERYEARLING (AGE 0) SOCKEYE During the latter part of March and up to mid-April, 1970, pre-mature fry" (24 to 28 mm fork length) with a small portion of the yolk sac remaining were caught at night at a depth of 14 m in midlake positions off the 3 major spawning areas. A few fry (28-30 mm) with empty stomachs were caught during the day at depths between 35 and 100 m in late March, and the first actively feeding fry (28 to 33 mm) were caught at depths between 12 and 55 m at night during the latter part of April. The number of fry caught at midwater depths increased in May at Stations 3, 4, 5, and 6 and reached a maximum in June at all stations. Fry continued to be caught in July and were still being caught in trawl nets at night in late August and early September. The fry and larger underyearling sockeye ate the same food organisms throughout the year, but the larger juvenile fish had more food in their stomachs. Figure 2 shows the number and weight of all species of food organisms per underyearling sockeye (Age 0) from August, 1969, when in- lake sampling began to April, 1970, when about 85 '^r of the fish migrating were yearling smolts.' The percentage of the total number of the six major food categories from all the fish sampled for stomach contents through the same period is shown in Figure 3. A list of the different gen- era of food organisms found in the stomachs of juvenile sockeye from 1969 to 1971 is given in Table 1; the smallest is listed at the top of the column and the largest at the bottom. Epischura was the predominant form (60%) in the stomachs in August, 1969 but was almost replaced by /fo/o?9edmm (60-80%) from Septem- ber to December (Figure 3). The incidence of ^ "Embryo" is defined as a larva minus its yolk-sac. An "alevin" is a larva of an age following hatching but prior to yolk absorption. Following this stage the fish becomes a "fry" (cf. Bams, 1969). ' In 1969 ca. 86% of migrant smolts were yearling, 10% were 2 year old, and 4% were 3 year old. In 1970 ca. 85% of smolts were yearling and 15% were 2 year old. 40 BARRACLOUGH and ROBINSON: LAKE FERTILIZATION. III. 500,- -i50 AUG SEPT OCT 1969 NOV DEC JAN FEB MAR 1970 APR Figure 2. — Average number and weight of all food or- ganisms (all species combined) per fish for underyearling sockeye salmon in Great Central Lake from August, 1969 to April, 1970. lOO AUG I SEP I OCT I NOV DEC JAN FEB 'MAR APR 1969 1970 Figure 3. — Food of underyearling sockeye salmon ex- pressed as a percentage of the total number of organisms from August, 1969 to April, 1970. Table 1. — List of organisms found in juvenile sockeye stomachs. Size range Organism mm 0.3-0.6 Bosmina, usually B. coregoni 0.6-1.1 Cyclops, usually C. biscuipidatus thomasi and C. vernalis 0.9-1.2 Holopedium gibberum 0.9-1.5 Daphnia, usually D. longiremis 0.8-1.3 Diaptomus, usually D. oregonensis 2.0-2.5 D. kenai 1.1-1.9 Epischura nevaiemii 3 Insects of the order Diptera (other than Chironomidae) 3-5 Insects of the family Chironomidae— larvae 8-11 Insects of the family Chironomidae— larvae 6-11 Larvae of the sculpin, Cottus asper Bosmina and Cyclops increased gradually from less than 5 Sf in August to a peak of 30 to 50 % in January-February, 1970. Chironomid larvae were the only organisms eaten from February to early March and in turn were replaced by Bos- mvna (SO^r) in late March and April. The per- centages of Epischura and Holopedium in the stomachs by number (Figure 3) and by weight (Figure 4) were similar from August to De- cember. There was a pronounced difference between the percentages by number and by weight of Cyclops and Bosmina per fish. Al- though the percentage by numbers of both organisms per fish increased markedly between December, 1969 and February, 1970, the per- centage weight per fish remained less than 7% for Bosmina and never exceeded 20% for Cy- clops. The importance of the chironomid larvae in the diet of underyearling sockeye from Feb- ruary through March to early April, 1970 is more indicative when expressed as a percentage by weight (Figure 4) than by number of or- ganisms (Figure 3). AUG I SEP ' OCT I NOV ' DEC ' JAN ' FEB ' MAR 1969 197 Figure 4. — Food of underyearling sockeye salmon ex- pressed as a percentage of the total weight of organisms from August, 1969 to April, 1970. The fry which emerged in late March and April, 1970 commenced to feed actively by mid- April and, as juvenile sockeye, they continued to increase their intake in number and weight of all food organisms throughout the summer, 41 FISHERY BULLETIN: VOL. 70, NO. I reaching a peak in September-October (Figure 5). In August, 1970, 3 months after lake fertilization began, the underyearling sockeye had twice the number of organisms per stomach as in August, 1969 and contained about 60% more food by weight. The high consumption in September-October, 1970 represents an increase of about 45^ r in number of organisms, and 40% by weight, compared to the stomach contents per fish in the same period in 1969. A slight decline in number and a significant decline in weight of food organisms per fish was shown from No- vember, 1970 to February, 1971; an abrupt in- crease occurred to a second high in March, which represented an increase manyfold over March, 1970. This increased food consumption occurred I to 2 months prior to their emigration from the lake as yearling smolts. Five species of food organisms contributed chiefly to the diet of underyearling sockeye in Great Central Lake in 1970. In April and May Bosviina contributed about 50 '^r of both the total number (Figure 6) and total weight (Figure 7) of all organisms found in their stomachs. The numbers of Bosmhia consumed were insignificant throughout the rest of 1970 and the first 3 months of 1971. Epischura was the most im- portant food organism from May to July (Fig- ures 6 and 7) and was probably the principal source of energy for the rajiid growth of the underyearlings during this period (see Figures II and 12). There was a transition in late July and August when Cyclops, Holopedium, and Daphnia gradually became more abundant in the stomach samples. In the 3 months which followed, September to November, Cyclops and Holopedium were the predominant genera. Cyclops continued in importance and formed about 50% of the number of food organisms to the end of January, 1971. However, during this period of 6 months Cyclops formed only 15 to 30% of the food by weight whereas Holopedium constituted 30 to 80% by weight. Diaptomus was first observed in the stomachs in the latter part of October, increased markedly in Decem- ber and January, and was the predominant food organism by number in February and March, 1971. Thus Diaptomus was the most numerous food organism in the stomach samples just before SOOr 200 I APR I MAY I JUN I JUL I AUG I SEP IQCT I NOV I DEC I JAN 1 FEB I MAR I 1970 19 7 1 Figure 5. — Number and weight of all food organisms per fish, for underyearling sockeye salmon from April, 1970 to March, 1971. Figure 6. — Food of underyearling sockeye salmon ex- pressed as a percentage of the total number of organisms from April, 1970 to March, 1971. 'APR 'MAY 'JUN ' JUL ' AUG ' SEP ' OCT ' NOV 'DEC ' JAN ' FEB ' mar' 1970 1971 Figure 7. — Food of underyearling sockeye salmon ex- pressed as a percentage of the total weight, from April, 1969 to March, 1971. 42 BARRACLOUGH and ROBINSON: LAKE FERTILIZATION. III. smelt migration in 1971. Bosmina was the most numerous in the previous year. The few chiro- nomid larvae (Figure 6) in the diet of juvenile soekeye from December, 1970 to March, 1971, formed 30 to 70 7^- by weight of all the food or- ganisms (Figure 7). The importance of chiro- nomid larvae during the winter months was ob- served also in the previous winter (Figures 3 and 4). FOOD OF YEARLING (AGE 1) SOCKEYE Those underyearling soekeye in Great Central Lake in 1969 which did not migrate to sea as yearling smolts in 1970, but remained in the lake for a second year, attained a mean length of only 51 mm and weight of 1.1 g between the latter part of April and May, 1970, whereas the migrating smolts had a mean length of 70 mm and weighed 3.5 g. The yearling soekeye which remained in the lake were collected from samples taken at all six stations and not from the end of the lake where smolts were schooling and heading seaward. Reference will be made later to the fact that the smallest size smolt, caught in the Robertson Creek weir (Figure 1) or in the nets set to capture smolts in the Stamp River, measured 55 mm and weighed 1.5 g. Food organisms found in the stomachs of the yearling soekeye were similar to those eaten by the underyearlings during most of the year in 1970, but the yearling soekeye were more se- lective in cropping the larger forms of zooplank- ton (Figure 9). Both the underyearling and yearling soekeye fed heavily upon Epischura from May to July (Figures 6 and 9) , but it was evident from the large numbers and weight of food organisms per fish (Figure 8) that the year- ling soekeye elected to feed or were able to prey more heavily upon Epischura (Figure 9) than the underyearlings during September and Oc- tober. Few yearling soekeye were caught in the trawls during the winter months of 1970-1971 prior to their migration as 2-year-old smolts. Diaptomus, Holopedium, and Cyclops were in the stomachs of these fish. ° 500- 400 100 TBO -70 60 A-A, 50 30 ^ 20 * -10 °l APR : MAY I JUN I JUL UUG I SEP I OCT I NOV I DEC I JAN I FEB I MAR P 1970 1971 Figure 8. — Number and weight of food organisms of yearling soekeye salmon from April, 1970 to February, 1971. APR MAY JUN JUL I AUG I SEP I OCT I NOV I DEC I JAN I FEB 1970 1971 Figure 9. — Food of yearling soekeye salmon expressed as a percentage of the total number of organisms from April, 1970 to February, 1971. DIEL FEEDING OF JUVENILE SOCKEYE From midafternoon on June 17 to midday on June 18 a series of 11 tows, each of 15 min dur- ation, were made with an Isaacs-Kidd midwater trawl. The trawl was towed through the middle of the densest portion of the stock during their 43 FISHERY BULLETIN: VOL. 70. NO. I diel migration. The tows were made to deter- mine what portion of different food organisms contributed to the salmon's ration during the day and night, as well as during the period of their diel migration. Data on depth and time of each tow, the number, size range, and mean length of sockeye caught, together with those sampled for the number and species of food organisms, and the weight of the food as a per- centage of the body weight are given in Table 2. The degree of freshness of food organisms was arbitrarily determined as fresh, fragmented, or largely unidentifiable. Fresh food was des- ignated when no indication of digestion had oc- curred. The percentage of fresh zooplankton in the stomachs is given in Table 2. Depth of the densest portion of the layer of juvenile sockeye at different times of the day and night is indi- cated in Figure 10a by a broken line. The depth and time of each trawl tow relative to the depth of the fish is also indicated in Figure 10a. The 24-hr data collection shows that in the day the densest poi'tion of the layer of juvenile sockeye was formed at 75 m where the temper- ature of the water was 4° C; of the fish exam- ined for stomach contents (Table 2) from tow No. 1, few food organisms per fish (Figure 10b) were noted and only 5Vf of the species were in fresh condition (Bosmina) . The remaining spe- cies, Epischura, Cyclops, and Daphnia, were digested. Tow No. 2, through a less dense sec- ondary layer at 105 m, indicated the same feed- ing pattern. Young sockeye commenced to mi- grate upward from 75 m between 1700 and 1800 hr. No differentiation in migration between underyearling or yearling sockeye could be de- tected at any level in the layer, either by net sampling or from high frequency echograms. A tow just after sunset at a depth of 35 m revealed that the fish were eating Bosmina and Cyclops (Figure 10c) as they moved upward and 22% of the contents were in fresh condition. At 2200 hr the sockeye had passed 25 m where the heaviest concentration of Cyclops and Daphnia was located (LeBrasseur and Kennedy, 1972) ; in passing they had eaten Cyclops (Fig- urelOc) . It should be recognized, however, that there is a natural time lag between feeding at any depth and the time the fish was captured by the trawl at a shallower depth, as they migrated toward the surface. At nautical twilight, most of the fish had com- pleted their upward migration and were distrib- uted in a layer between 10 and 20 m where tem- peratures ranged from 6° to 12°C (Figure 10a), Echograms indicated many of the juvenile sock- eye salmon appeared to spend brief periods be- tween and 10 m at temperatures ranging from 14° to 23° C, during which time the young fish fed heavily upon Epischura (Figure 10c). In the 4 hr between the beginning and end of nau- tical twilight no feeding occurred (Table 2). Table 2.— Die! feeding of juvenile sockeye; tows made with an Isaacs -Kidd midwater trawl, each of 15 min dur- ation, over a 24-hr period from June 17 to 18, 1970 at Station 4 in Great Central Lake. Tow No. Depth of tow Time (PST) at start of 15 min Number fish caught Size range underyearling Mean length Mean weight Size range sampled for food Mean length No. samp for food le Total no. food organisms No. of organisms per fish % fresh Weighl food as % body weight (m) mm mm mg mm mm 1 75 1449 47 23-40 32 292 28-40 33 10 190 19 5.5 1.1 2 105 1538 17 27-36 30 212 27-36 31 9 112 12 1.0 3 55 1820 21 27-38 32 272 27-38 33 10(1)1 67 7 1.0 4 35 2018 11 27-36 31 263 27-36 32 10 164 16 22 I.O 5 18 2142 6 26-36 31 282 26-36 31 6(2) 132 22 46 1.6 6 14 0023 71 26-41 32 315 26-41 33 11 537 49 6 1.7 7 19 0254 50 26-39 32 296 28-39 34 10(1) 257 26 1.5 8 62 0517 14 28-39 32 271 28-39 32 14(4) 586 41 64 2.1 9 68 0738 10 29-37 33 320 29-37 33 9 447 50 84 1.6 10 70 0945 9 29-39 33 278 29-39 33 9 543 60 34 1.6 n 75 1159 6 31-39 35 353 31-39 35 6 324 54 2 2.1 Time of sunset 2011 Time of sunrise 0350 Time of nautical twilight 2201 Time of nautica 1 sunrise 0200 1 Number in parentheses is number of items In sample which contained no food in stomachs. 44 BARRACLOUGH and ROBINSON: LAKE FERTILIZATION. III. 1400 1600 1800 2000 2200 2400 0200 0400 0600 OBOO lOOO '2°° _, (500 1700 SOO 2100 2300 OiOO 0300 0500 0700 0900 "OO 'JTO Figure 10. — (a) Depth of the densest portion of the layer of juvenile sockeye salmon at different times of the day and night from June 17 to June 18, 1970 is indicated by a broken line. A secondary layer is shown at 105 m for a 2-hr period. The depth of each tow with a midwater trawl is shown relative to the depth of the fish, (b) Number of all food organisms of underyearling sockeye. (c) Food species of underyearling sockeye salmon expressed as a percentage of the total number of organisms. A second feeding period was noted at the time of the diel migration downward. Stomach sam- ples from juvenile sockeye collected during this period contained many fresh Daphnia and Cy- clops in tows 8 to 10. Only 2^/r of the zoo- plankton in the stomachs of sockeye caught at midday were in a fresh condition (Table 2), which indicates a marked reduction in feeding activity. GROWTH OF UNDERYEARLING SOCKEYE The average size of underyearling sockeye (Age 0) in 1969 and 1970 in Great Central Lake is shown in Figures 11 and 12. A total of 1,760 underyearling sockeye were caught in 1969 and 20,783 fish in 1970 from all six stations. A com- plete record of all data on which this analysis is based has been reported by Barraclough and 70 60 E t 50 ,o 40 30 Or 1969 Underyeorlings a -. 1970 Underyeorlings ^^IwARlflPRlMflYljUNljUL I AUG I SEP'OCTInOV I DECl JAN I FEBI MAR Figure 11. — Average length of underyearling sockeye salmon in each month, 1969 and 1970. 2 50 - 2 00 E 1 50 100 050 o = 1969 Underyeorlings O = 1970 Underyeorlings D — a-o-16°C, and this reduced their feeding effi- ciency (Foerster, 1968) . Thus increases in tem- perature of the epilimnion through the long period of fry emergence decreased the apparent benefit to late-hatching fish. However, in spite of this, most of the fry hatching later in the year achieved a length greater than 55 mm and mi- grated from the lake in June 1971; under normal conditions it is believed that these fry would not have reached 55 mm and would have migrated the following year as 2-year-olds. Prelimi- nary examination of the scales from juvenile sockeye in 1971 reveals an absence of a winter check on many scales which suggests that the high concentrations of zooplankton persisting through the winter enabled many fish to smoltify and leave the lake. The combination of an early run of very large 1-year-old smolts, combined with this later run of much smaller smolts, tended to reduce the overall apparent effective- ness of lake fertilization. Thus the real effects of fertilization seem likely to be greater than would be judged from considering only changes in overall mean size of smolts. LITERATURE CITED Bams, R. A. 1969. Adaptations of sockeye salmon associated with incubation in stream gravels. In Symposium on salmon and trout in streams, p. 71-87. H. R. MacMillan Lectures in Fisheries, Univ. B. C, Inst. Fish., Vancouver, B.C. Brett, J. R., J. E. Shelbourn, and C. T. Shoop. 1969. Growth rate and body composition of finger- ling sockeye salmon, Oncorhynchus nerka, in re- lation to temperature and ration size. J. Fish. Res. Board Can. 26: 2363-2394. Burgner, R. L. 1962. Studies of red salmon smolts from the Wood River Lakes, Alaska. Univ. Wash. Publ. Fish., New Ser. 1: 247-314. Foerster, R. E. 1968. The sockeye salmon, Oncorhynchus nerka. Fish. Res. Board Can. Bull. 162, 422 p. Goldman, C. R. 1960. Primary productivity and limiting factors in three lakes of the Alaska Peninsula. Ecol. Monogr. 30: 207-230. 1964. Primary productivity and micro-nutrient lim- iting factors in some North American and New Zealand Lakes. Int. Ver. Theor. Angew. Limnol., Verhandl. 15:365-374. Johnson, W. E. ^ 1961. Aspects of the ecology of a pelagic, zooplank- ton-eating fish. Verh. int. Ver. Limnol. 14:727- 731. Krokhin, E. M. 1959. (On the effect of the number of spawned-out sockeye salmon (Oncorhynchus nerka) in a lake on its supply of biogenic elements.) Dokl. Akad. Nauk SSST 128(3) :626-627. (Fish. Res. Board Can. Transl. Ser. 417.) 1967. Influence on the intensity of passage of the sockeye salmon (Oncorhynchus nerka Walb.) on the phosphate content of spawning lakes. Izdatel'- stvo "Nauka" Leningrad 15:26-31. (Fish. Res. Board Can. Transl. Ser. 1273.) LeBrasseur, R. J., W. E. Barraclough, 0. D. Kennedy AND T. R. Parsons. 1969. Production studies in the Strait of Georgia. Part III. Observations on the food of larval and juvenile fish in the Eraser River plume, February to May, 1967. J. Exp. Mar. Biol. Ecol. 3:51-61. 47 FISHERY BULLETIN: VOL. 70. NO. 1 LeBrassei'r, R. J., AND 0. D. Kennedy. 1972. The fertilization of Great Central Lake. II. Zooplankton standing stock. Fish. Bull., U.S. 70: 25-36. McDonald, J. G. 1969. Distribution, groN^-th and survival of sockeye fry (OncorhyncliHs nerka) produced in natural and artificial stream environments. J. Fish. Res. Board Can. 26:229-267. Nelson, P. R. 1958. Relationship between rate of photosjTithesis and growth of juvenile red salmon. Science 128: 205-206. Nelson, P. R., and W. T. Edmondson. 1955. Limnological effects of fertilizing Bare Lake, Alaska. U.S. Fish Wildl. Serv., Fish. Bull. 56: 414-436. Parsons, T. R., C. D. McAllister, R. J. LeBrasseur, and W. E. Barraclough. In press. The use of nutrients in the enrichment of sockeye salmon nursery lakes — a preliminary report. FAO Technical Conference on Marine Pollution, Rome, December 9-18, 1970. Parsons, T. R., K. Stephens, and M. Takahashl 1972. The fertilization of Great Central Lake. I. Effect on primary production. Fish. Bull., U.S. 70:13-23. RiCKER, W. E. 1962. Comparison of ocean growth and mortality of sockeye salmon during their last two years. J. Fish. Res. Board Can. 19:531-560 48 ESCAPE BEHAVIOR OF THE HAWAIIAN SPINNER PORPOISE {Stenella cf. 5". longirostris) William F. Perrin and John R. Hunter^ ABSTRACT Incidental mortality of porpoise (Cetacea, Delphinidae) occurs in the tropical tuna seine fishery. Ex- periments were carried out in a crowding chamber to determine behavioral responses of trained and naive Hawaiian spinner porpoise (Stenella cf. S. longirostris) to barriers of purse-seine netting, monofilament webbing, pol>winyl sheeting, rows of floats, and openings of various dimensions in a net wall. The object of the experiments was to generate information to be used in development of rescue gear and methods for the fishery. Openings of less than 1.5 m in width and/or 1 m in depth markedly inhibited escape. Negative effect of a line of floats aci'oss an opening at the surface was pronounced. Barriers of visually and acoustically relatively transparent monofilament webbing and polyvinyl sheeting were not appar- ently detected by porpoise prior to physical contact. Recommendations pertaining to potential design of rescue gear are presented. Incidental mortality of porpoise occurs in the American purse-seine fishery for tropical tunas (Perrin, 1970). In 1970, the National Marine Fisheries Service began a program of research to develop improved gear and methods to reduce the porpoise mortality due to tuna seining. This paper reports the results of experiments on the responses to netting and other barriers by the Hawaiian spinner porpoise {Stenella cf. S. longi- rostris) , a form closely related to one of the species involved in the tuna fishery." We studied the response of the spinner porpoise to barriers of net, transparent monofilament nylon webbing, transparent polyvinyl sheeting, rows of floats, ' National Marine P^isheries Service, Southwest Fish- eries Center, La Jolla, CA 92037. ^ Taxonomic note: The spinner porpoise of Hawaii has been variously referred to Stenella longirostris Grav 1828, by Nishiwaki (1967) and Tomich (1970), and to S. roseivcntris Wagner 1846, by Fraser (in Morris and Mowbray, 1966) and Rice and Scheff"er (1968). The spinner porpoise of the tuna grounds of the far eastern Pacific has been referred to .S'. microps Gray 1846 (Miller and Kellogg, 1955; Handlev, in Hester, Hunter, and Whitney, 1963; Nishiwaki, 'l967; Pilson and Waller, 1970) and to .S. longiros^tris (Rice and Scheffer, 1968; Harrison, Boice, and Brownell, 1969). No critical re- view of the genus has been accomplished since True's work on the Delphinidae in 1889. The usage here of S. longirostris for the Hawaiian spinner is provisional pending the I'esults of taxonomic studies underway at the Southwest Fisheries Center and elsewhere. and to openings of different dimensions in a net wall. The results of these studies will be ap- plied in the design of an escape opening in the tuna purse seine. The experiments were carried out at Oceanic Institute, Oahu, Hawaii, in May, June, and July 1970. METHODS AND MATERIALS THE ANIMALS Three of the five porpoise (Table 1) used in the experiments had been in captivity at Oceanic Institute and Sea Life Park for various lengths of time and are referred to below as the "trained porpoise"; the remaining two, referred to below Table 1. — Hawaiian spinner porpoise (Stenella cf. S. longirostris) used in behavioral experiments. Name Date of capture Sex Weight at time of capture ke Trained porpoise Waimea Mar. 6, 1969 Male 50.0 Nani Dec. 4, 1969 Female 59.2 Nohea Dec. 4, 1969 Male 65.9 Naive porpoise Westward June 11, 1970 Female 72.7 Moana July 9, 1970 Female 51.3 Manuscript accepted September 1971. FISHERY BULLETIN: VOL. 70, NO. I, 1972. 49 FISHERY BULLETIN: VOL. 70, NO. 1 as the "naive porpoise," were freshly captured, were not exposed to any training: procedures prior to the experiments, and were tested imme- diately upon arrival at the Institute. THE APPARATUS The crowdinjr chamber (Fig-ure 1) was con- structed in a large pool at the Oceanic Institute. The pool, known as "Bateson's Bay," is roughly circular, 24.7 m across at its greatest diameter, and approximately 1 m deep at its center. A smaller holding tank communicates with the jwol through a wooden gate. Three hemisjiher- ical underwater viewing ports allow surveillance of the entire pool. Net barriers were placed at various points along the pool wall to construct a circular en- closure or crowding chamber about 20 m in di- ameter in which j^orpoise were tested. The crowding chamber had two radial walls of net- ting that extended from the outer edge of the chamber to a central aluminum mast. One of the walls was stationary and was provided with escape openings of various dimensions. The other wall was movable and was used to drive the animals through the opening in the sta- tionary wall. The movable wall pivoted on the central mast and was supported along the leading edge by an aluminum beam and on the distal end by a plastic float. The edge of the pool was marked at 1° intervals. The walls were made of tuna purse-seine web- bing (4Vi.-iiich stretched mesh [10.8 cm] *42 thread knotted nylon). Flotation was provided by purse-seine-type corkline constructed of 6-inch diameter x 3V^-inch (15 >( 9 cm) sponge- plastic floats. The basic escape opening was 18 ft (5.5 m) wide and 6 ft (1.8 m) deep. Flaps of purse- seine webbing were laced in, to variously de- crease width to 10, 5, or 2V2 ft (approximately 3.0, 1.5, or 0.8 m) and/or depth to 31/2, 3, 2, 1, or 1/2 ft (approximately 1.1, 0.9, 0.6, or 0.2 m) . For tests of response to a barrier across the opening at the water surface, a corkline constructed of hollow plastic floats (5 y 9 inch [13 X 23 cm], 4 per m) was strung across the top of the open- ing. In tests of response to barriers of acous- tically low-reflective materials, a panel of 3%- inch (stretched) mesh (8.6 cm) *12 monofila- ment webbing, a panel of 0.38-mm-thick poly- vinyl sheeting, or a panel of 1.04-mm-thick poly- vinyl sheeting, was laced into the opening. STATIONARY CONTAINING NET WALL P STATIONARY CONTAINING NET WALL Figure 1. — Crowding chamber. Largest diameter of pool is 80 feet, to scale. Sketch not drawn 50 PERRIN and HUNTER: ESCAPE BEHAVIOR OF PORPOISE Acoustical tests carried out by the Naval Under- sea Research and Development Center, San Di- ego, Calif., on plastic sheeting of similar thick- nesses indicated effective acoustic transparency in the range of porpoise emanations (personal communication from W. E. Evans). PROCEDURES The moving wall was rotated around the cen- ter mast at a rate sufficient to completely close the chamber in about 4 min. An attempt was made to maintain a constant rate of rotation. Time required for an animal to escape from the chamber was recorded in seconds with a stop- watch, and position of the moving wall at time of escape was recorded in degrees. The reading in degrees was later used to calculate surface area remaining in the crowding chamber at time of escape. The wall was rotated alternately in clockwise and counterclockwise directions. After escape of an animal, the moving net wall was rotated until it was against the stationary wall, and the two radial nets remained together until the beginning of the next trial. Trails were spaced initially at 15-min intervals, to allow time for changing the escape opening. After our pro- ficiency in altering the opening increased, the in- terval was decreased to 10 min. Two major types of experimental design were used : ( 1 ) a long series of trials alternating two treatments and (2) a series of blocks of consec- utive trials of various treatments. In some ex- periments, the two approaches were combined to yield a factorial design testing simultaneously the effects of variation in two or three of the factors of width, depth, and presence or absence of corkline, monofilament, or polyvinyl barriers. In some tests of the monofilament and polyvinyl panels, the animal was subjected to a single trial with the panel after a series of learning trials without the jianel or at the beginning or conclu- sion of an experiment involving other variables. The design of these experiments is referred to below as "single trial." The results of the first series of experiments (Waimea I, II, and III; see Table 2) using the alternating trials design indicated a probable influence by the direction of rotation of the net wall or by stage of practice effect. The small number of trials in each ex- ]-»eriment precluded complete randomization, but the treatments in subsequent experiments were staggered to offset the effect of direction of ro- tation. A typical sequence of trials was: a, h, a, b, a, a, h, a, b, b; where a and b were different treatments, and rotation in the first trial was clockwise, in the second counterclockwise, and so on in alternating fashion. In this manner, an equal number of clockwise and counterclockwise trials was assured for each treatment. Table 2. — Preliminary experiments with trained porpoise. Porpoise Experiment Variables tested Design Number of trials Waimea 1 Width Alternating trials 20 II Width Alternating trials 20 III Depth, corkline Factorial 26 IV Monofilament panel Single trial 14 Nan! 1 Depth Alternating trials 20 11: trialsl-19 trial 20 Width, depth Corkline Factorial Single trial 19 1 III: trial 1 Corkline Single trial 1 trials 2-16 Depth Block 16 trials 17-30 Depth, monofilament panel Factorial 14 Nohea 1 II Depth Depth Alternating Block trials 22 36 III: trials 1-36 Width Block 36 trials 37-44 Depth Alternating trials 8 trial 45 Monofilament panel Single trial 1 IV Width, depth, corkline Factorial 32 V: trials 1-8 trial 9 Depth Thin polyvinyl panel Alternating Single trial trials 8 1 VI: trials 1-8 trial 9 Depth Thick polyvinyl panel Alternating Single trial trials 8 1 51 The remaining surface area between the ad- vancing net wall and the stationary wall at the time of escape was used as a criterion of the animals' readiness to esca])e. This index was the inverse of latency, since the smaller the area that remained when the animal escaped the longer would be the latency. Although we measured latency in seconds, we felt the net po- sition was the preferable measurement because the rate of net movement was imi)recise, where- as the actual stimulus for escaj^e, the reduction in the swimming area, could be measured rel- atively accurately. In presentation of the data, the logarithm (to base 10) of the remaining area at the time of escape is plotted on trial number. Because procedures and plans were modified during the course of the experiments, results and interpretation are combined in the presen- tation of the results. PRELIMINARY EXPERIMENTS WITH TRAINED PORPOISE We anticipated that the behavior of the por- poise would change rapidly during the course of the experiments; thus, to avoid wastage of the naivete of the limited and expensive supply of untrained animals, we conducted a series of pre- liminary exi:)eriments with three trained por- poise (Table 2, Figures 2-4). WIDTH OF OPENING The effect of the width of the opening on the escape l)ehavior of the three trained porpoise was first tested by presenting on alternate trials an escape route of standard width, 5.5 m, and one either 3.1 or 3.8 m wide. In Waimea (Figures 2-1 and II) and Nohea (Figure 4-1 V) there was some evidence that the porpoise escai)ed sooner when the wider e-scajie route was used but not in Nani (Figure 3-II). We felt this was i)robably an artifact of experimental design as described above and consequently we considered only the data from the block experiments for evaluating effects of width on the trained pori)oise. To de- termine the width of opening that would influ- ence performance, Nohea was tested over six FISHERY BULLETIN: VOL. 70, NO. I WAIMEA FAILED TO ESCAPE 092m DEEP, WIDTH VARIED 092m DEEP. WIDTH VARIED TRIAL NUMBER Figure 2. — Results of experiments with trained porpoise Waimea. Each plot summarizes one day's continuous experimentation, as follows: I. May 21, II. May 22, III. May 23, IV. May 25, 1970. NANI 55m WIDE, DEPTH VARIED / I Im DEEP, WIDTH VARIED / / o 55m WIDE, „ DEPTH VARIED OI5m DEEP O O O FAILED TO ESCAPE 55m WIDE. DEPTH VARIED TRIAL NUMBER Figure 3. — Results of experiments with trained porpoise Nani. I. May 26, II. May 27, III. May 28, 1970. 52 PERRIN and HUNTER: ESCAPE BEHAVIOR OF PORPOISE I • < /^ • / OlSmOEE 1 5.5. n WIDE. DEPTH VAH EO ZO S5 30 35 Vv^ .^V*^^ l-Jm *iOt ; Wm WIDE / 1 DEEP, WIDTH VARIED FAILED TO ESCAPE O 55m WIDE; DEPTH VARIED Sim WIDE; DEPTH VARIED TRIAL NUMBER Figure 4. — Results of experiments with trained porpoise Nohea. I. May 30, II. May 31, III. June 1, IV. June 21, V. June 28, VI. June 29, 1970. blocks of six trials each (Figure 4-III). The width of the escape route was 3.1 m in the sec- ond, 1.5 m in the fourth, and 0.76 m in the sixth block of experimental trials, with interspersed blocks of trials at 5.5 m. A significant decrease in performance occurred only when the escape opening was narrowed to 0.76 m. In one trial at 0.76 m the animal refused to leave the crowd- ing chamber and had to be extricated from the webbing. In others, the porpoise exhaled air and sank passively to the bottom of the pool and did not move even when the chamber was com- pletely closed. Exhalation of air and sinking to the bottom was a pattern that appeared in other porpoise in other experiments and was accom- panied by failure to escape. Our tentative conclusion was that the width of the opening w^as not a significant variable in the block experiments if it exceeded about 1.5 m. DEPTH OF OPENING We determined the effect of the depth of the escape route by varying depth of the hole from 1.8 m to 0.15 m while maintaining the standard hole width of 5.5 m. We will describe the results for each porpoise separately since the experi- ments w^ere diflFerent for each animal. Waimea failed in the first eight trials to escape through an opening 0.92 m deep (Figure 2-1). Performance improved thereafter to a plateau that was maintained throughout a subsequent identical experiment the next day, throughout a series of alternating trials with 0.92-m- and 0.61-m-deep openings on the third day of the ex- periments, and in a fourth experiment (Waimea IV) 2 days later with a 0.92-m-deep opening. Nani showed no difference in response after the first two trials with openings 1.8 m and 1.1m deep (Figure 3-1). High performance contin- ued through a series of trials with a 1.1-m-deep opening, but dropped in blocks of trials of 0.61-m- and 0.30-m-deep openings (II). In sub- sequent experiments Nani failed to escape twice when openings 0.15 m and 0.30 m deep were used and performed erratically in blocks of trials with openings 0.61 m and 0.30 m deep (III). After seven trials with the 0.30-m opening, no failures was experienced in seven trials with a 0.15-m opening, but the animal escaped consistently earlier (larger remaining area) when the 1.1-m- deep opening was used. Nohea escaped earlier when the hole was 1.1m than when it was 0.61 m deep in 14 of the first 16 trials of an alternating series (Figure 4-1), In a subsequent series of blocks of trials at decreas- ing depths (II), a pronounced drop in perform- ance occurred at depths of 1 ft and 5 ft. The following day's performance remained at a high level except when a corkline was strung across the opening (IV). Our tentative interpretation of the above re- sults was that a critical depth of opening lay near 1 m: 11 failures to escape occurred at 0.92 m or shallower; none occurred with openings 1.1 m deep or deeper; and performance was even more adversely aflfected by further decreasing the depth of the opening. We also concluded that the results of the first few trials for each ani- 53 FISHERY BULLETIN: VOL. 70, NO. 1 mal were of most importance in predicting the probable response of naive wild porpoise, as the animals were able quickly to achieve high levels of performance even at very shallow depths. CORKLINE A corkline across the toj) of the ojiening caused Waimea (III) and Nani (II) to fail on initial trials, and greatly affected the performance of Nohea (IV) . Waimea, after four failures, over- came reluctance to pass through an oi:)ennig with a corkline at the surface and reachieved a high level of performance. An interaction between the corkline and depth of opening was apparent in the factorial experiment with Nohea. Initial trials with the corkline (second block) produced a temporary drop in performance with a 1.1-m- deep opening. In the fourth block, the corkline was again inserted, and performance dropped at 0.61 m depth but not at 1.1 m. MONOFILAMENT PANEL When the panel of nylon monofilament web- bing was inserted into the opening (1.1 m deep) after a series of trials in which performance was consistently high, Waimea (IV), Nani (III), and Nohea (III) swam into the webbing as if it did not exist. Performance in subsequent trials without the i^anel was not affected (Nani III). Upon hitting the webbing, the porpoise became entangled and had to be extricated by a diver. POLYVINYL PANEL In the two single trials with a panel of clear polyvinyl sheeting inserted in the opening, Nohea (V and VI) hit the ])anel and slid over the top as it buckled. No difference was noted in be- havior in these trials from that in trials in which the panel was absent. During these experiments Nohea in several trials passed back and forth through the opening two or three times after the initial escape, while the net wall was being closed. The values for the surface area index shown in the figure are for the first passage. The incidence of such be- havior throughout the course of all the experi- ments occurred only after considerable expe- rience with a particular net configuration. In most cases, only one or two double "escapes" occurred during an experiment. EXPERIMENTS WITH NAIVE PORPOISE Eleven experiments were conducted with the naive porpoise (Table 3, Figures 5 and 6). The first naive animal. Westward, was captured on June 12, 1970, and after a relatively short han- dling period was placed in Bateson's Bay. Her swimming behavior during the first 5 days of captivity was unlike that of the trained porpoise. The trained porpoise continually swam about the tank during and between experiments, diving and "porpoising," and spinning. Westward, on Table 3. — Experiments with naive porpoise. Porpoise Experiment Variables tested Design Number of trials Westward 1 Depth Alternating trials 20 II Depth Alternating trials 20 III Depth Block 36 IV: triols 1-40 Depth, width, corkline Factorial 40 trial 41 Monofilament panel Single trial 41 V V\'idth Block 36 VI: trials I-IO Depth, thin polyvinyl panel Factorial 10 trials 11-15 Thick polyvinyl panel Block 5 Moana 1 Depth Alternating trials 11 II Depth Alternating trials 20 III Width Block 36 IV Depth Block 24 V: trials 1-12 Depth Block 12 trials 13-25 Width, corkline Factorial 13 54 PERRIN and HUNTER: ESCAPE BEHAVIOR OF PORPOISE WESTWARD MOANA 55m WIDE: DEPTH VARIED N v\// fAILEO ,T9 ESCAPE, 5.5m WIDE; DEPTH VARIED n DEEP :a6lmOEEP : l.lm DEEP | Q30m DEEP 55m WIDE, DEPTH VARIED ■t] r\ < 5 10 IS 20 25 30 35 TRIAL NUMBER 500 f m 1 Mm DEEP, , w V^^i . S^^ llmOEEP^ . VNmDEEP H- lOO ■'^^im DEEP ;'!'" deep/ \ ,/^lm WtOE / /06lm DEEP m \ '\ 5.5m WIDEi DEPTH VARIED FAILED >l \ TO ESCAPE© a DEPTH VARIE, OI5m DEEP ,V^ O TO ESCAPE /55m WiDf Urn DEEP, WIDTH VARIED : &5m WiOE I DEEP; WIDTH VARIED FAILED TO ESCAPE O 20 25 50 35 TRIAL NUMBER Figure 6. — Results of experiments with naive porpoise Moana. I. July 10, II. July 11, III. July 12, IV. July 13, V. July 15, 1970. carried out the day after capture. Subsequently Westward's behavior slowly changed, until 5 days later on June 17 it was indistinguishable from that of the trained porpoise. The re- mainder of the Westward experiments were car- ried out after June 17. Moana, the second naive porpoise, was cap- tured on July 9, 1970. When placed in Bateson's Bay, she exhibited the same behavior as West- ward, but to a lesser extent. Periods of surface swimming in a semiupright position, but without head-bobbing, were interspersed with periods of normal porpoising and diving. During the first experiment she swam slowly at the surface in the diagonal posture but during the second and subsequent experiments, her behavior was sim- ilar to that of the trained animals. 55 FISHERY BULLETIN: VOL. 70, NO. 1 DEPTH OF OPENING In the first few trials (Fioure 5-1) a marked difference existed in the response of Westward to an opening: 1.1 m deep and one 0.61 m deep. The animal swam slowly at the surface, circling or moving back and forth in the chamber. When the opening was 1.1 m deep, she moved slowly through the opening .iust as the moving wall closed. When the opening was 0.61 m deep, she moved past or circled slowly in front of the open- ing and then dove and entangled herself in the webbing of the moving wall. In the sixth trial, her behavior became more varied; she swam in tight circles beneath the surface and attempted to pass between the end of the moving wall and the periphery of the chamber before passing through the opening, after which she slapped her tail against the water surface. Behavior in subsequent trials became increasingly erratic. In trial 13 she darted through the opening rather than moving through slowly as in the previous trials. In trial 14, she tried again to squeeze past the moving wall and became lodged in the narrow opening. In trial 15, she assumed a position across the corkline of the moving wall, half in and half out of the chamber, and remained there until removed. In the remaining trials, she moved rapidly through the opening, and in the last two, she assumed a horizontal attitude sim- ilar to that usually taken by the trained ])or- poi.se and .stop])ed bobbing her head but still kept her blowhole above the surface. An identical experiment (II) of alternating trials was carried out 5 days later, after all traces of the slow surface-swimming and head-bobbing l)ehavior had disappeared. Performance was consistently higher with the 1.1-m opening. The effect of depth is clearly seen in the results of a block-design experiment for Westward (III). The second naive porpoise, Moana, a smaller and presumably younger animal than Westward, achieved a higher rate of successful passage in the first depth experiment (Figure 6-1). She failed only once, with the 1.1-m-deep opening. In the .second depth experiment (II) her per- foi-mance was extremely variable compared to that of Westward, and no relation between depth and success rate existed. In two trials (18 and 20) while swimming in tight circles near the apex of the chamber, she snagged her flipper in the webbing and had to be extricated. In later block-design experiments (V and IV) the effect of depth was evident as it was for Westward. WIDTH OF OPENING Results of block-design experiments testing the effect of width of opening for Westward (Figure 5-V) and for Moana (III) were similar to those for the trained porpoise (Nohea), but an effect was discernible at widths of 1.5 m. As with the other experiments, performance of Westward was higher and more stable than that of Moana. Westward began to pass through the opening two or three times during a single trial. The frequency of multiple "escapes" was higher for the 5.5-m-wide opening than for the narrower openings (Table 4). CORKLINE Insertion of a corkline at the surface across the top of the opening sharply affected the per- formance of Westward (IV) and Moana (V). The performance of Moana showed the greatest effect. After a series of preparatory trials, Moana failed to pass through the opening in five straight trials with the corkline. In each trial she laid the anterior part of her body across the corkline and remained there until removed. In the block-design experiments with Westward (IV), the second block of trials with a corkline produced a smaller drop in performance than did the first, with the 0.61-m-deep opening only, dem- onstrating as for the trained porpoise (Nohea IV, Figure 4) an interaction between depth and presence or absence of a barrier at the surface. Table 4. — Multiple escapes of Westward. Width of open in block of six trials ng Number of double escapes Number of triple escapes m 5.5 3.1 1 55 2 2 1,5 1 5.5 2 0.76 56 PERRIN and HUNTER: ESCAPE BEHAVIOR OF PORPOISE MONOFILAMENT AND POLYVINYL PANELS When the monofilament panel was inserted in- to the 1.1-m-deep openino- at the end of a series of depth trials, Westward (IV) "s'ot u]i a full head of steam and plowed into the monofilament" (extracted from field notes of W. Wasden) and became entangled. Insertion of polyvinyl panels produced similar results in multiple trials; West- ward (VI) in each trial hit the panel and slid over it and out of the chamber. There was nothing in the behavior of the porpoise to indi- cate that they recognized the presence of the panels. DISCUSSION AND CONCLUSIONS The swimming behavior of the naive porpoise Westward and, to a lesser extent, of Moana, the first few days after capture was very similar to that of porpoise {Stenella spp.) in tuna purse seines as ol)served by one of us (Perrin) off Cen- tral America. A typical "failure to escape" epi- sode is illustrated for Moana in Figure 7. Im- mediately after a purse-seine net has been set, when the diameter of the encircled area is great- est (approximately 250 m), the porpoise swim about quite rapidly in small tight groups of a dozen or so individuals, the mem])ers of a group diving and surfacing together (Figure 8). As the net is hauled and the area enclosed becomes smaller, especially after the backing down oper- ation (see Perrin, 1969) , the porpoise congregate and raft near the center of the enclosure and mill very slowly, holding their bodies in a semiup- right position with blowhole exposed and ros- trum at or slightly below the surface (Figure 9) . At this point, individual animals can be seen to leave the grou]) and dive. When the net has been completely hauled, animals are often found with their snouts entangled in the webbing several meters below the corkline. Although the head bobbing exhibited by West- ward was not observed in the purse-seine situ- ation, the similarities in Ijehavior between freshly captured animals and those captured in a purse seine were striking. In both cases the animals did not display normal motor patterns; they rested or swam at abnormally slow speeds, and this behavior was often ended l^y a rapid dive beneath the surface with no noticeable change in behavior preceding the act. The prin- cipal characteristics of this behavior, the inhi- bition of activity in a fear-inducing environ- ment, resembled fear responses described for many other vertebrates and frequently classified as an immobility or freezing response (Ratner and Thompson, 1960; Hinde, 1970). Hogan (1965, 1966) suggested that withdrawal and im- mobility are separate, mutually inhibitory sys- tems. If this view is correct, then driving por- poise through an escape route in the luirse seine would not be successful once the animals began to show the immobility response, because with- drawal would be inhibited. Under these circum- stances the additional fear stimulus associated with driving might be the catalyst for the rapid dive to escape, which results in entanglement. Driving may have to be carried out before im- mobility begins. Once the animals became im- mobile the only strategy may be to ])ull the net out from beneath them as is currently done dur- ing the "backing down operation" (Perrin, 1969). That the behavior of Westward and Moana evolved into more typical behavior during the course of a single ex])eriment also supports the notion that their unusual behavior was caused by the circumstance of captivity rather than ill health. Our conclusions with respect to projected de- sign of a rescue gate for removing porpoise from a purse seine during fishing operations were: 1. The gate should be sufficiently wide so that when the perimeter of the net circle buckles after pursing, the width does not become less than 1.5 m. Considering the equivocal results of the ex])eriments for o])enings wider than 1.5 m, the opening should be as wide as prac- tically possible. 2. Depth of the opening should be not less than 1 m and as deep as it is possible to make it with- out causing loss of the fish in the net. 3. There should be no line, corkline, or other barrier across the oi)ening at the surface. 4. A self-actuating release port that will open when struck by a porpoise swimming into it 57 FISHERY BULLETIN: VOL. 70, NO. I 58 PERRIN and HUNTER: ESCAPE BEHAVIOR OF PORPOISE Figure 7. — Typical "failure to escape" episode. Moana (1) pa- trols moving wall at beginning of trial, then (2) takes up position at apex of chamber and remains there for most of trial, in vertical attitude. As chamber nears clo- sure Moana dives (3) , orients to- ward opening (4) , and turns and swims into moving wall (5), be- coming entangled (6). %fe - -^*^ Figure 8. — Porpoise (Stevella grnffmnni) in tuna purse seine at beginning of set, when net is at near-maximum diameter. An- imals are circling and diving in groups of a dozen or so individu- als. Figure 9. — Porpoise in purse seine, after most of net has been taken aboard. Animals are "rafting" in compact group, each maintaining approximately vertical attitude, with blowhole exposed and dorsal fin submerged. Large fish underwater in foreground are yellowfin tuna. 59 FISHl.RY BULLETIN: VOL. 70, NO. 1 miorht be feasiV)le if constructed of acoustically transi)arent materials, providing: that it were so constructed that the fish in the net would not also use it. 5. It is to be exjiected that grreat difficulty will be encountered in inducing: wild porpoise to pass throuo:h an opening in the perimeter of a purse- seine enclosure. ACKNOWLEDGMENTS We thank the Oceanic Institute and its Di- rector. Dr. Kenneth Norris, for providino- the facilities and the porpoise used in this study. William Wasden, National Marine Fisheries Service, Honolulu, Hawaii, assisted in all phases of the study, and por]wise trainers Tng-rid Kano-, Sea Life Park, and Scott Rutherford. Oceanic Institute, provided as.sistance and advice during the course of the experiments. LITERATURE CITED Harrison, R. J., R. C. Roice, .wd R. I.. Rrownrll, Jr. 1969. Rpproduction in wild and captive dolphins. Nature (London) 222: 1143-1147. HESTKR, F. J., J. R. Ht'NTKR, AND R. R. WHITNEY. 196.'5. Jumping and spinning hohavior in the spin- ner porpoi.se. J. Mammal. 44: 586-588. HiNDE, R. A. 1970. Animal behaviour. A synthesis of ethology and comparative psychology. McGraw-Hill Book Company, 876 p. HOG.AN, J. A. 1965. An experimental study of conflict and fear: An analysis of behavior of young chicks toward a mealworm. Part I. The behavior of chicks which do not eat the mealworm. Behaviour 25: 45-97. 1966. An experimental study of conflict and fear: An analysis of behavior of young chicks toward a mealworm. Part K. The behavior of chicks which eat the mealworm. Behaviour 27: 273-289. Miller, G. S., Jr., and R. Kellogg. 1955. List of North American recent mammals. U.S. Natl. Mus. Bull. 205: 1-954. Morris, R. A., and L. S. Mowbray. 1966. An unusual barnacle attachment on the teeth of the Hawaiian spinning dolphin. Nor. Hval- fangsttid. 55: 15-16. NlSHIWAKI, M. 1967. Distribution and migration of marine mam- mals in the North Pacific area. Bull. Ocean Res. Tnst., Univ. Tokyo 1: 1-64. Perrin, W. F. 1969. Using porpoise to catch tuna. World Fish- ing 18(6): 42-45. 1970. The prolilem of porpoise mortality in the U.S. tropical tuna fishery. Proc. Sixth Annu. Conf. Biol. Sonar and Diving Mammals, Stanford Res. Inst., Menlo Park, Calif., p. 45-48. Pilson, M. E. Q., AND D. W. Waller. 1970. Composition of milk from spotted and spin- ner porpoises. J. Mammal. 51: 74-79. Ratner, S. C, AND R. W. Thompson. 1960. Immobility reactions (fear) of domestic fowl as a function of age and prior experience. Be- haviour 8: 186-191. Rice, D. W., and V. B. Scheffer. 1968. A list of the marine mammals of the world. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 579, 16 p. ToMicii, P. Q. 1969. Mammals in Hawaii. Bernice P. Bishop Mus. Spec. Publ. 57, 238 p. True, F. W. 1889. Contributions to the natural history of the cetaceans. A review of the family Delphinidae. U.S. Natl. Mus. Bull. 36, 191 p. + pi. I-XLVIL 60 METHODS FOR TAGGING SMALL CETACEANS W. E. Evans,'- ^ J. D. Hall,- A. B. Irvine,^ and J. S. Leatherwood^ ABSTRACT Four types of tags have been used on four species of delphinids. These include a circular plastic button tag that is attached to the dorsal fin by a nylon bolt, a highly visible dart-type spaghetti tag that is placed near the base of the dorsal fin, a radio transmitter tag, and a freeze brand. Use of button tags has been discontinued due to high shedding rate. The dart-type spaghetti tag has proved best for tagging large numbers of animals without capturing them. The radio tag provides very detailed information on behavior and movements, while freeze branding provides a permanent mark, though both require capturing the animal. The importance of marking commercially valu- able species of whales (primarily the larger baleen whales and the sperm whale) has long been recognized. Since their development in the mid-1920's, "Discovery-type" tags have been used to mark large numbers of these animals (Rayner, 1940; Brown, 1962; Clark 1962). Re- turns from these tags have provided valuable in- formation on the species' distribution, migration, and abundance and on such basic aspects of their biology as relative growth rates and the timing of the events in their lives (Mackintosh, 1965). The relationship of several small delphinid species to commercial fish populations and the potential of these cetaceans as a major economic resource has renewed interest in their stocks during the last decade (Perrin, 1970). Early attempts to study these populations in the wild have been hampered by the difficulty of posi- tively identifying an animal or a population from one encounter to the next. Therefore, develop- ment of a reasonable method for marking these animals for identification would facilitate studies of their life histories. Although several investigators have tried tag- ging small cetaceans, only three have had even moderate success. In a program conducted by ^ Authors are listed in alphabetical order. '^ Marine Life Sciences Laboratory, Naval Undersea Research and Development Center, San Diego, CA 92132. ' Mote Marine Laboratory, Sarasota, FL. 33581. the Oceanic Institute, Oahu, Hawaii, plastic cat- tle eartags were placed on two Steno hredanensis and one Stenella attennata (Evans, 1967) . This program was continued by Norris and Pryor (1970), and at least one of the tags was still on a Stenella attennata when it was resighted after 3i/4 years. Sergeant and Brodie (1969) tagged 812 be- lugas, Delphi napteriis leucas, in Hudson Bay, Canada, over a 2-year period. Six hundred and ninety-four of these animals were tagged with a spaghetti tag originally designed by Mather ( 1963) for use in tagging pelagic fishes and man- ufactured by Floy Tag Company,' Seattle, Wash. The remaining 118 belugas were tagged with Petersen disc tags, similar to the button tags we used. Of the 812 animals tagged, 2 with spaghetti tags were recovered by the beluga fish- ery. A third spaghetti tag was observed in a live animal temporarily stranded by the ebbing tide 1 year after the original tagging. Perrin and Orange (1971) tagged 21S, Ste7iella spp. in 1969 and approximately 1,000 in 1970 in the eastern tropical Pacific with spaghetti-type dart tags. Five tags have been recovered; max- imum time at liberty was 138 days (916 km net movement) , Since 1968, personnel of the Naval Undersea Research and Development Center's Marine Bio- science Division at San Diego, Calif., have been * Reference to commercial products does not imply endorsement by the National Marine Fisheries Service. Manuscript accepted September 1971. FISHERY BULLETIN: VOL. 70, NO. 1, 1972. 61 FISHERY BULLETIN: VOL. 70, NO. 1 investigating the distribution and biology of sev- eral odontocete cetaceans off the southern Cali- fornia coast. In order to delineate migration routes and to keep track of local herds of the com- mon dolphin, Delphinns delphis auctt., a tagging program was initiated. During the same period, a tagging program was also initiated for Tur- siops tnoicatus on the west coast of Florida. The special problems associated with tagging odontocete cetaceans required the modification of old and the development of new tagging tech- niques. This paper discusses the relative merits of the four marking methods used by our labora- tory. In addition, it presents some preliminary results of the program in order to substantiate the utilitv of the various methods. METHODS AND RESULTS We have used modified dart-type vinyl spa- ghetti tags (Floy Manufacturing Company) on four species of Eastern Pacific delphinids in an area from Point Conception, Calif., to Cabo San Lucas, Baja California, Mexico, and throughout the Gulf of California. Our original spaghetti tags were 5 mm in diameter by 17 cm long. In order to increase visibility and flow character- istics of the tag, we increased the length to 30 cm (Figure 1). Using the modified tag, w^e have marked 240 D. delphis, 10 Lagenorhynchm obliquidens, 8 Tursiops pilli auctt., and 13 Stenel- la gmffmani to date (July 1971). The animals were all tagged at the anterior insertion of the dorsal fin while they were surfing on the bow pressure wave. Several dolphins were observed to continue riding the bow pressure wave after being tagged, so the tagging process apparently did not affect their normal behavior. A T. gilli auctt., tagged on 27 October 1970, off Magdalena Bay, Baja California, was recov- ered by an American tuna boat off Manzainillo, Mexico, on 22 January 1971. The animal had covered at least 816 km between the time of tag- ging and the time of capture, a period of just less than 3 months. Three D. delphis bearing spaghetti tags have been observed swimming in the vicinity of the Coronado Islands near San Diego, Calif., and at least one spaghetti-tagged D. delphis has been sighted off Magdalena Bay, Baja California. Each of these animals was known to have been carrying the tag for from 2 weeks to several months. Circular plastic "button" tags (10 cm diam) (Figure 2) were through-bolted to the dorsal fins of 46 D. delphis and 6 L. obliquidens between 1967 and 1970. These tags are similar to those employed by Norris and Pryor (1970) in Hawaii, but are larger to make them more easily spotted. Button tags were attached to animals captured off the southern California coast, or near Cedros ' V 1 'J r ■■™ffTraeAi.«i»««u»«i! Figure 1. — The dart-type spaghetti tag in place on the tagging apparatus. 62 EVANS ET AL. ; TAGGING SMALL CETACEANS To meet this need, a new lightweight radio tag (170 g) with a 9-12 month transmitter life was developed. This tag combines the advantages of a radio beacon and a button tag in that it contin- ues to serve as a color coded marked even after it no longer transmits (Figure 3), Further- more, the new radio tag is available commer- cially at less than 10 ^r of the cost of the 900/gm transmitters. Figure 2. — Three Lagenorhynchus obliqiddens with the plastic button tag, just prior to release. Island, Baja California. Two of the L. obliqui- dens tagged in 1969 were resighted almost 1 year later, and a D. delphis tagged in 1968 was re- sighted 21 months later. Twenty-four T. truncatus were tagged with the button tags near Sarasota, Fla., from August 1970 through September 1971. Animals bearing tags have been resighted several times. The third and most successful short-term tag is the radio transmitter tag with which at least four species of small cetaceans have been success- fully marked to date (Evans, in press, Martin, Evans, and Bowers, 1971) . The original package used in these studies was a 27 mHz (11m) trans- mitter and antenna housed in a waterproof envelope which is attached to the dorsal fin of a dolphin or a small whale by means of a spring- loaded corrosible link. The link dissolves and releases in 30 days, allowing the package to slip off the animal. These early radio beacons, designed for short- term transmission (30-60 days), weighed up to 900 g, and though they proved especially useful in studying the detailed movements of D. delphis in the waters off San Diego, Calif., their size, cost, and relatively short transmission time made them unacceptable for long-term monitoring of herd movements. Figure 3. — The lightweight (170 g) radio tag. 63 FISHERY BULLETIN: VOL. 70. NO. 1 The fourth method, freeze branding, consists of applying a supercooled branding iron, usually copper, to the epidermal surface of the dolphin for 5-30 sec. Evidence from freeze branding cattle indicates that the branding process is pain- less to the animal and has no lasting effect other than leaving a permanent mark (Farrell, Lais- ner. and Russell, 1969) . Though evidence of the branding usually becomes indistinct shortly after application, after about 2 months the animal will display a highly legible brand (Figure 4). We have used this method on eight wild T. ti-uncatus near Sarasota in conjunction with either a but- ton tag or a spaghetti tag. The number on a freeze branded animal was clearly visible, from a distance of 40 yards, when the animal was resighted 10 weeks after tagging. FiGi.RE 4. — A Tumiops truncatus with the freeze brand on the dorsal fin. DISCUSSION We have discontinued use of the button tag in favor of the spaghetti and radio tags. In- cidence of loss of button tags from animals has been exceptionally high among the T. truncatus around Sarasota, and the few resightings of but- ton-tagged dolphins off southern California lead us to believe that button tag loss is high in this area also. A major disadvantage of the button tag is that the animal must be captured in order to be tagged. The spaghetti tag, on the other hand, is normally placed in the animal while it is free swimming and thus does not require cap- ture. Using this method we have placed over 50 spaghetti tags in one herd of D. delphis in less than 2 hr. When spaghetti tags are placed in the fibrous tissue at the insertion of the dorsal fin, incidence of tag loss appears to be lower for spaghetti tags than for the button tags (Nishi- waki, Nakajima, and Tobayama, 1966). In either case, the numbered information on the tag is so small that it cannot be read on a moving animal at sea. Unless the spaghetti tags are color-coded, resighting at sea can give no in- formation on the original tagging location. Spa- ghetti tags may also be placed in an animal that has been captured. The radio tags can be placed only on captured animals but provide very detailed information concerning exact movement and diving patterns of the animal. While freeze branding involves capture of the animal, it appears to provide permanent and highly legible identification of cetaceans. Tom- ilin (1962) reported taking a Black Sea D. delphis in 1953 which bore a brand posterior to the eye. The brand was quite legible and contained numbered information. The source and nature of the brand were not known. In the future, we plan to freeze brand all the dolphins we capture for radio tagging and to continue to use the spaghetti tags for free-swim- ming delphinids. An advertisement was placed in the July issue of National Fisherman requesting that any in- formation on sightings of tagged delphinids in the Eastern Pacific be forwarded to the Marine Bioscience Division of the Naval Undersea R&D Center, San Diego, Calif. (Evans, Leatherwood, and Hall, 1971). Copies of this advertisement have been placed at sportfish landings and com- mercial docks from Santa Barbara to San Diego, Calif. LITERATURE CITED Brown, S. G. 1962. A note on migration in fin whales. Nor. Hvalfangst-Tid. (Norw. Whaling Gaz.) 51(1): 13-16. 64 EVANS ET AL.: TAGGING SMALL CETACEANS Clarke, R. 1962. Whale observation and whale marking off the coast of Chile in 1958, and from Ecuador to- wards and beyond the Galapagos Islands in 1959. Nor. Hvalfangst-Tid. (Norw. Whaling Gaz.) 51(7): 265-287. Evans, W. E. 1967. Vocalization among marine mammals. In W. B. Tavolga (editor), Marine Bio-Acoustics, Vol. 2, p. 159-186. Pergamon Press, New York. In press. Orientation behavior of delphinids: radio telemetric studies, presented at the Conference of Animal Orientation : Sensory Basis, sponsored by New York Academy of Sciences, Feb. 8-10, 1971. Ann. N.Y. Acad. Sci. Evans, W. E., J. S. Leatherwood, and J. D. Hall. 1971. Request for information on tagged porpoises on the eastern Pacific. Natl. Fisherman 52(3): 15A. Farrell, R. K., G. a. Laisner, and T. S. Russell. 1969. An international freeze-mark animal identi- fication system. J. Am. Vet. Med. Assoc. 154 : 1561-1572. Mackintosh, N. A. 1965. The stocks of whales. Fishing News (Books) Ltd., London, 232 p. Martin, H., W. E. Evans, and C. A. Bowers. 1971. Methods for radio tracking marine mammals in the open sea. Transactions of the IEEE Con- Terence on Engineering in the Ocean Environment, September 1971, San Diego, Calif. Mather, F. J., III. 1963. Tags and tagging techniques for large pe- lagic fishes. Int. Comm. Northwest Atl. Fish., Spec. Publ. 4: 2. NiSHiwAKi, :m., M. Nakajima, and T. Tobayama. 1966. Preliminary experiments for dolphin mark- ing. Sci. Rep. Whales Res. Inst. 20: 101-107. NORRis, K. S., ANT) K. W. Pryor. 1970. A tagging method for small cetaceans. J. Mammal. 51: 609-610. Perrin, W. F. 1970. The problem of porpoise mortality in the U.S. tropical tuna fishery. Proc. 6th Annu. Conf. Biol. Sonar Diving Mammals. Stanford Res. Inst, Menlo Park, Calif., p. 45-48. Perrin, W. F., ant) C. J. Orange. 1971. Porpoise tagging in the eastern tropical Pa- cific. Proc. 21st Tuna Conf., Lake Arrowhead, Calif., October 1970, p. 5. Rayner, G. W. 1940. W^hale marking, progress and results to De- cember 1939. Discovery Rep. 19: 245-284. Sergeant, D. E., and P. F. Brodie. 1969. Tagging white whales in the Canadian Arctic. J. Fish. Res. Board Can. 25: 2201-2205. TOMILIN, A. G. 1962. The migrations, geographical races, the thermo-regulation and the effect of the tempera- ture of the environment upon the distribution of the cetaceans. Fish. Res. Board Can., Transl. Ser. 385: 1-24. 65 A REVIEW OF THE LANTERNFISH GENUS Taaningkhthys (FAMILY MYCTOPHIDAE) WITH THE DESCRIPTION OF A NEW SPECIES Brent Davy^ ABSTRACT The genus Taaningichthys includes three known species, one of which is here described as new. The species of the genus Taaningichthys do not appear to perform daily vertical migrations. Evidence indicates vertical stratification of juveniles and adults. Although photophores and lateral line are reduced, the species of Taaningichthys possess very large eyes which may be related to capture of luminescent prey. Otoliths of all three species have been examined and found to be taxonomically important. Bolin (1959) erected the genus Taaningichthys to include two species, T. hathyphilus and T. mi- nimus, previously placed in the genus Lampade- na by Taning (1928). The main characters which distinguish Taaningichthys from Lampa- dena are: (1) the origin of the dorsal fin in Taaningichthys is clearly behind the base of the pelvic fins; (2) the development in Taaningich- thys of a crescent of white tissue' on the pos- terior half of the iris, although a similar white (luminous?) crescent is present on the dorsal portion of the iris in Lampadena chavesi (Naf- paktitis and Paxton, 1968); (3) the presence of a single SAO, or none, in Taaningichthys (always three SAO in Lampadena) ; (4) re- duced dentition and lateral line in Taaningich- thys. Taaningichthys may be distinguished from all other myctophid genera by the combination of the white crescent of tissue on the posterior half of the iris, the undivided luminescent caudal glands, and the single or altogether absent SAO. Berry and Perkins (1966) reported what they thought to be a third form of Taaningichthys apparently without photophores. Following the capture of a number of specimens of this form ^ Department of Biological Sciences, Allan Hancock Foundation, University of Southern California, Los An- geles, Calif. 90007. "This tissue is not visible until some time after preser- vation and is hardly distinguishable in specimens initially frozen and then preserved. by the RV Velero IV of the University of Southern California and the examination of con- siderable material made available to me by nu- merous institutions around the world, I felt that a review of the genus was appropriate. MATERIALS AND METHODS Members of the genus Taaningichthys are deep-dwelling, fragile myctophids, easily dam- aged by the net. Scales are readily lost, and damage to the bones of the snout, upper jaw, and operculum is very common. Consequently, measurement of jaw, head, and snout length is often very difficult if at all possible. The follow- ing measurements were taken on the best pre- served specimens: Eye diameter (ED) — hori- zontal distance across the orbit; jaw length ( JL) — length of premaxillary; predorsal (Pre D) — anterior tip of premaxillary to base of anterior- most ray of dorsal fin ; pre ventral (Pre V) — an- terior tip of premaxillary to base of anteriormost ray of ventral fin; preanal (Pre A) — anterior tip of premaxillary to base of anteriormost ray of anal fin; prepectoral (Pre P) — anterior tip of premaxillary to base of anteriormost ray of pec- toral fin; preadipose (Pre Ad) — anterior tip of premaxillary to posterior end of base of adipose fin; length of supra- and infracaudal luminous glands — length of exposed luminous tissue only; anal-infracaudal distance — anterior tip of Manuscript accepted July 1971. FISHERY BULLETIN: VOL. 70, NO. 1, 1972. 67 FISHERY BULLETIN': VOL. 70, NO. I infracaudal gland to end of base of anal fin. Sizes of specimens are given in standard lengths (SL) only. Terminology of body photophores follows that of Bolin (1939). Unless otherwise specified, the term photophore refers to the primary body l^hotophore. Otoliths were measured with an eyepiece mi- crometer as follows: Length (OL)— the great- est length parallel to the sulcus; height (OH) — greatest height iierpendicular to the sulcus. Fol- lowing measurements, otoliths were lightly smeared with graphite to bring out detail and then photographed. Otolith terminology fol- lows that of Frizzell and Dante (1965). Female specimens were considered gravid when eggs included oil globules and completely filled the oviduct. Most specimens examined were captured with open nets and depth sampled is here considered as the maximum depth reached l\v the net (appendix). Counts of procurrent caudal rays are given as dorsal + ventral. KEY TO THE SPECIES OF THE GENUS Taatihigichthys la. VO 8-10; AO 5-7 + 4-6, total 9-13; Pol directly below or anterior to base of adipose fin; Prci-Prc2 interspace equal to or greater than two photophore diameters; as many as five pairs of broad-based, hooklike teeth on dentary near symphysis T. minimus (Taning, 1928) lb. VO, if present, 3-5; AO, if present, 1-4 + 1-2, total 2-5; Pol, if present, clearly behind base of adipose fin; Prci-Prc2, if present, interspace equal to, or less than, one photophore diameter; no broad-based, hooklike teeth on dentary near sympiiysis 2 2a. Photophores present as in lb above; anal-infracaudal distance half as long as length of infracaudal gland, or longer T. hathyphilns (Taning, 1928) 2b. Photophores absent; anal-infracaudal distance less than half length of infracaudal gland T. j)au)'olychnus n. sp. ' /y. ■^^:.- .iAS W^^;- Figure 1. — Taaningichthys minimus (Taning) ; 46 mm, Ocean Acre stn. 7-21. 68 DAVY; REVIEW OF LANTERNFISH GENUS TaaningUhthys GENUS Taaningichthys Bolin Taaningichthys minimus (Taning, 1928) Figures 1 and 2 Lampadena minima Taning, 1928: 63; Parr, 1928: 154, Figure 37. Lampadena (Lampadena) minima Fraser-Brun- ner, 1949: 1078, Figure. Taaningichthys miriimus Bolin, 19.59: 26. D 11-13; A 12-13 (11-14); P 15-17; V 8; gill rakers 4-5 + 1 + 11 (10-13), total 16-17 (15- 18); VO 8-10; AO 5-7 + 4-6, total 9-13; Pre 2 + 1; vertebrae 40-41; procurrent caudal rays 8-10 + 8. Mouth terminal, moderately large, JL about 1.5 in Pre P; maxillary slightly expanded pos- teriorly. Eye large, ED 2.2-3.4 in Pre P. Pter- otic spine long and directed posteriorly. Oper- cular margin concave posterodorsally, slightly convex posteriorly. Pectoral fin long, reaching VOe or VO7; its base about midway between ventral body margin and horizontal septum. Pre V 2-2.4 in SL. Pre D 1.9-2.3 in SL; end of base of dorsal fin clearly in advance of vent. Pre A 1.4-1.6 in SL. Anterior end of base of adipose fin on vertical through posterior end of base of anal fin; Pre Ad 1.2-1.3 in SL. A band of dark pigment along anteroventral margin of orbit containing a series of light gray, triangular patches of tissue not present in the other two species, Dn absent; Vn present between anterior mar- gin of orbit and posteroventral margin of nasal rosette. PVOi on or behind vertical through upper end of base of pectoral fin and about mid- way between it and ventral margin of body; PVO2 in front of middle of base of pectoral fin; a straight line through PVOi and PVO2 passing in front of PLO. PLO about halfway between upper end of base of pectoral fin and horizontal septum. Five PO. VLO above base of pelvic fin, usually closer to horizontal septum than to ventral margin of body. Last VO usually slightly elevated. SAO 1-2 photophore diameters be- low horizontal septum, directly above vent. AO level. AO series overlaps anterior end of in- fracaudal gland. Pol directly below or in ad- vance of base of adipose fin, 1-2 photophore ^ '"^j B «::>'"'"'# X«« Figure 2. — Taayiingichthys minimus. A. Side view, sex- ually dimorphic supracaudal gland of male, 53 mm. B. Top view, sexually dimorphic supracaudal gland of same male. C. Side view, sexually dimorphic supracaudal gland of female, 54 mm. D. Top view, sexually dimorphic supracaudal gland of same female. 69 FISHERY BULLETIN: VOL. 70. NO. 1 diameters loelow horizontal septum. Prci and PrC'j level, behind infracaudal g-land; Prc:i at horizontal septum. Caudal luminous glands undivided, the infra- caudal larg-er than the sujjracaudal and both covered by scales. Sexual dimorphism is evi- dent in supracaudal gland of adults (specimens about 40 mm and larger) ; in males this gland is about twice as large as in females (Figure 2) . Mesopterygoid teeth in narrow oval patches. Narrow band of needlelike teeth on palatine. No vomerine teeth. Both jaws with needlelike teeth which bend medially. A single row of broad-based, anteiiorly hooked teeth occupying posterior two-thirds of media! surface of den- tary. As many as five pairs of similar teeth, most often directed posteriorly, on medial sur- face of dentary near symphysis, and another two to three pairs projecting forward and lat- erally on symphysial area of iiremaxillary; be- low these, on anterior part of premaxillary, sev- eral posteriorly curved teeth (longer than rest of premaxillary teeth) . Tann'myichthys minimus is the shallowest dwelling, most firm-bodied, and smallest of the three sj^ecies, the largest examined specimen measuring 65 mm. Gravid females (about 40 mm and larger) were captured in August-September. Horizontal distribution — T. minimus occurs circumglo))ally between about lat 35° N and 30° S (Figure 5). It has been taken less frequently than T. bathyphilus. Vertical distribution — Closing-net data from the Project "Ocean Acre" in the north Atlantic Ocean suggest vertical stratification of juveniles and adults. Juveniles appear to inhabit depths of 140 to 250 m, the smallest specimen (21 mm) having been captured at 140 m. Adults occur ]-»redominantly in depths between 450 and 500 m. T. minimus does not appear to perform daily vertical migrations. Taanhigichthys bathyphilus (Taning, 1928) Figure 3 Lampadena bafhyphila Taning, 192S: 63; Parr, 1928: 151, Figure 36. Lampadena (Lampadena) bathyphila Fraser- Brunner, 1949: 1078, Figure. Taaningichthys bathyphilus Bolin, 1959: 26, Figure 6. D 12-13 (11-14); A 13 (12-14); P 12-14; V 8; gill rakers 3 + 1+7-8 (5-9), total 11-12 (9-13); VO 4 (3-5); AO 3 (1-4) +1(2), total 4(2-5); Pre 2 + 1; vertebrae 34-36; procurrent caudal rays 7 + 6. Mouth terminal, moderately large, JL about 1.5 in Pre P; maxillary slightly expanded poster- iorly. Eye large, ED about 2.5 in Pre P. Pterotic spine inconsi)icuous. Opercular margin as in T. mininuis. Pectoral rays reaching VOi; base of pectoral fin nearer to horizontal septum than to ventral margin of body. Pre V 2.1-2.5 in SL. Pre D 1.9-2.2 in SL; end of base of dorsal fin on, or slightly in advance of, vertical through SAO. Pre A 1.5-1.7 in SL. Base of adipose fin above end of base of anal fin; Pre Ad 1.2-1.4 in SL. V -O-jU^-a--** ' Fic.rRK ^.—Tamiingiehthys bathyphili(S (Taning) ; 02 mm, RV Velcro stn. 11733, LACM 30034-1. 70 DAVY: REVIEW OF LANTERNFISH GENUS Taamnguhthys i t ■1/ ^^ ' :'^. Figure 4. — Taaningichthys pmirolychnus, new species, holotype, 67 mm, SIO 70-19. Dn absent; a very small oval Vn, visible in young individuals and masked by darkly pig- mented tissue in adults. Position of PVOi far- ther forward than in T. minimus, a line through PVOi and PVO2 passing behind PLO; PVO, mid- way between upper end of base of pectoral fin and ventral margin of body; PVO2 midway between PVOi and upper end of base of pectoral fin. PLO varying in position, usually closer to hor- izontal septum than upper end of base of pec- toral fin. PO 5-6. VLO above base of pelvic fins, closer to horizontal septum than ventral margin of body. VO level. SAO 1-2 photophore dia- meters below horizontal septum, directly above or slightly behind urogenital papilla. AO level; AOp over anterior end of infracaudal gland. Pol position variable, generally midway be- tween anterior end of infracaudal gland and end of base of anal fin (always well behind base of adipose fin), and one photophore diameter or less below horizontal septum. Prci and Prc2 level; Prcs at horizontal septum. Secondary photophores present on snout and rays of caudal fin. Length of supracaudal luminous gland 1.5-2 in length of infracaudal; sexual dimorphism not apparent ; both glands undivided and surround- ed by dark pigment. Mesopterygoid teeth rather sparsely distrib- uted. Single row of needlelike teeth on palatines. No vomerine teeth. Both jaws with needlelike teeth which bend medially (those on the anterior- most premaxillary somewhat longer). Several broad-based, anteriorly hooked teeth on posterior medial surface of dentary (not as many as in T. minimus). Two to three pairs of similar teeth projecting forward and laterally on sym- physial area of premaxillary. Taaningichthys bathyphilus is the intermedi- ate of the three species in terms of depth of occurrence, photophore development and size. It does not seem to grow larger than about 80 mm. Of the specimens examined, only one gravid female (57 mm) was found which had been cap- tured in late June. Horizontal distribution — T. bathyphilus oc- curs circumglobally within a broad zone between lat 41° N and 67°31' S (Figure 5). It appears to be more common or, perhaps, more easily cap- tured than its two congeners. Vertical distribution — The shallowest depth of capture for T. bathyphilus is 580 m (a ju- venile male, 32 mm). An adult female, 65-mm long, was captured at a depth of 675 m. Mem- bers of this species have not been taken above these depths. The maximum depth of occur- rence is not yet known. T. bathyphilus does not appear to perform daily vertical migrations. Taaningichthys paurolychnus, NEW SPECIES Figure 4 Holotype: 1 (67 mm), 17 Dec. 1969, 31° N, 119° W, Scripps Institution of Oceanography. Paratypes: 1 (68 mm), 22 Nov. 1969, 17° 47' N, 25°22' W, National Institute of Ocean- ography; 1 (87 mm), 13 Sept. 1968, 17° S, 71 FISHERY BULLETIN: VOL. 70, NO. I 86° W. Institute of Oceanolog-y, Academy of Sci- ences of the USSR, Moscow; 1 (49 mm), 20 Sept. 1961, 33° N, 17° W, Museu Municipal do Fun- chal; 1 (57 mm), 29 Jan. 1922, 19° N, 79° W, Zooloofical :\Iuseum, University of Copenhagen; 2 (79-95 mm), 17 Dec. 1969, 31° N, 119° W, 2 (65-71 mm), 10 June 1967, 35° N, 123° W, Scripps Institution of Oceanography; 1 (75 mm) , 17 Sept. 1966, 1° N, 81° W, 1 (80 mm), 15 Jan. 1969. 32° X, 120° W, Smithsonian Oceanographic Sorting Center; 1 (82 mm) , 13 Apr. 1962, 30° N, 120° W, 1 (77 mm) , 29 Mar. 1962, 35° N, 129° W, National Marine Fisheries Service. D 12-13 (11); A 13 (11-14); P 14 (13-15); V 8; gill rakers 3-4 + 1 -^ 9-10 (8-11) . total 13-15 (12-16); vertebrae 35-36; procurrent caudal rays 7 + 6-7. Mouth terminal, moderately large, JL about 1.5 in Pre P. Eye large, ED 2.2-3.2 in Pre P. A short pterotic spine directed posterolaterally. Oi)ercular margin slightly concave posterodor- sally to a level above upper end of base of pectoral fin, slightly convex posteriorly. Pectoral rays reaching base of pelvic fins; base of pectoral fin midway between ventral margin of body and hor- izontal septum. Pre V 2.1-2.3 in SL. Pre D 1.9-2.1 in SL; end of base of dorsal fin in ad- vance of origin of anal fin. Pre A 1.5-1.7 in SL. Base of adipose fin directly above, or somewhat behind end of base of anal fin; Pre Ad 1.2-1.4 in SL. Vn apparently absent. Head and body photo- phores absent. Secondary photophores present on snout and interradial membranes of caudal fin. Length of supracaudal luminous gland 1.5-2 in length of infracaudal gland; sexual dimor- phism not apparent; both glands undivided and surrounded by dark pigment. Mesopterygoid teeth rather sparsely distrib- uted. Single row of needlelike teeth on palatine. No vomerine teeth. Both jaws with needlelike teeth which bend medially (those on the anterior part of premaxillary somewhat longer). Se\^- eral broad-based, anteriorly hooked teeth on pos- terior medial surface of dentary as in T. bathy- philus. Two to three pairs of similar teeth pro- jecting forw^ard and laterally on symphysial area of premaxillary. Taaningichthys paurolychnus is the largest of the three species, the longest specimen examined measuring 95 mm. It has apparently lost its °T. minimus »T. bathvphilus *T. paurol y chnus W Figure 5. — Catch localities of Tanningichihys viinimuft, T. bathyphilns, and T. paurolychnus. Capture locality for specimen taken from 67°31' S/9()°26' W not .shown. 72 DAVY: REVIEW OF LANTERNFISH GENUS Taaningichthys primary photophores, retaining only the caudal glands and the simple, presumably secondary, photophores, on the head and caudal fin. It is also the deepest dwelling of the three species. Gravid females have been captured in March- April and August-September. The smallest gravid female examined was 65 mm. Horizontal distribution — T. pam'ohjchnus is distributed circumglobally between about lat 40° N and 20° S (Figure 5). It does not appear to be as common as T. bathyphilus. Vertical distribution — T. paurolychnus has not been taken above 900 m (an adult female, 77 mm). The lower limits of its distribution are not yet known. T. paurolychnus does not appear to perform daily vertical migrations. Etymology — The name paurolychnus refers to the absence of primary photophores and the presence of limited, presumably secondary pho- tophores. It is derived from the Greek pauros meaning few, small, and lychnus meaning light. OTOLITHS OL as a percentage of SL ranges from 4.4 to 5.5% in T. minimus, 3.9 to 4.6% in T. bathy- philus, and 2.8 to 3.6% in T. paurolychnus. OH as a percentage of OL ranges from 66.7 to 77.7% in T. minimus, 72.1 to 77.9% in T. bathyphilus, and 78.3 to 91.7% in T. paurolychnus. The sulcus is more pronounced in the otolith of T. bathyphilus than it is that of T. paurolych- nus, but less so than in that of T. minimus (Fig- ure 6). The otolith of T. paurolychnus has al- most no antirostrum, and the antirostrum in T. bathyphilus is less pronounced than that in T. minimus. The posterior margin of the otolith in T. paurolychnus is nearly straight vertically, making the general outline almost square, where- as the otoliths of its two congeners are smoothly rounded posteroventrally, so that the general outline is oval. The otolith (Figure 6) of a single specimen of T. bathyphilus from the north Atlantic is differently shaped and very large for a specimen of its size (60 mm). However, no other differences in the fish were found and more material from the north Atlantic must be examined before anything further can be stated. DISCUSSION The various hypotheses and ideas regarding the function, or functions, of luminous organs of midwater fish are reviewed by Nicol (1969). The photophores within the genus Taaningich- thys show drastic reduction in terms of numbers and development. T. minimus has, relatively, the best developed photophores as well as the greatest number; these organs are seldom rubbed ofl[" unless the specimen is damaged. T. bathyphilus has fewer and less well-developed photophores which are easily rubbed ofl". T. paurolychnus has lost all primary photophores but retains simple, presumably secondary, photo- phores on the snout and caudal fin. As already mentioned, there are no indications that the members of the genus Taaningichthys undertake diel vertical migrations as most myctophids do. It may therefore be that i:)hotophores of the myc- tophid type are not selected for in a deep-water, nonmigratory fish, which would account for the reduction of these organs and, eventually, their loss. Unlike photophores, eyes are very well de- veloped in Taaningichthys, regardless of depth of occurrence. Even the deepest of the species has large, nearly binocular eyes. This may be corre- lated with the food habits of these fish. Mycto- phids, in general, feed on zooplankters, many, if not most, of which are bioluminescent. It is pos- sible therefore that Taaningichthys strongly de- pends on large, presumably highly eflFective eyes for locating and capturing its prey, which is probably not very abundant in those dark mid- water depths. Furthermore, retention of this energetically expensive visual equipment may account for the very poorly developed lateral line system. ACKNOWLEDGMENTS I thank Basil G. Nafpaktitis of the University of Southern California, Robert J. Lavenberg of the Los Angeles County Museum of Natural History, and Theodore W. Pietsch of the Uni- versity of Southern California for their critical review of the manuscript and helpful sugges- tions. Thanks are also due to John E. Fitch 73 FISHERY BULLETIN: VOL. 70, NO. I Figure 6. — Medial views of right otoliths, anterior end to the left. A. Tarnihif/iclithys miniimts, otolith 2.0 mm long, specimen 49 mm. B. T. bathyphilidi, otolith 1.8 mm long, specimen 64 mm. C. T. ixtiirolycIiriKS, otolith 1.7 mm long, specimen 87 mm. D. T. batluiphUiis, otolith 2.8 mm long, specimen (from north Atlantic) 60 mm. of the State of California Department of Fish and Game for comments concerning otoliths. I am indebted to several jieople and their in- stitutions for making si)ecimens available: Rob- ert L. Wisner, Scripps Institution of Oceanogra- phy (SIO); Julian Badcock, National Institute of Oceanography, Surrey, England; Roljert J. Lavenberg; Jorgen Niel.sen, Zoological Museum, University of Copenhagen; V. E. Becker, Insti- tute of Oceanology, Academy of Sciences of the USSR, Moscow ; Robert H. Gibbs, Jr., U.S. Na- tional Museum; Leslie W. Knapp, Smithsonian Oceanographic Sorting Center; G. E. Maul, Museu Municipal do Funchal, Madeira; Thomas Clarke, University of Hawaii; E. H. Ahlstrom and II. Geoffrey Moser, National Marine Fish- eries Service; Richard H. Backus and James E. Craddock, Woods Hole Oceanographic Institu- 74 DAVY: REVIEW OF LANTERNFISH GENUS Taamnguhthys tion; G. Palmer, British Museum (Natural History); Michel Le Gand, Office de la Re- cherche Scientifique et Technique Outre-Mer, Noumea, New Caledonia (material presented to the Los Angeles County Museum of Natural History) . Illustrations were made by Sharon Calloway, Los Angeles County Museum of Natural History, and otoliths were photographed by the Photogra- phy Department, Los Angeles County Museum of Natural History. I am grateful to my friend Lu Duffy for having typed the manuscript. LITERATURE CITED Berry, F. H., and H. C. Perkins. 1966. Survey of pelagic fishes of the California Current area. U.S. Fish Wildl. Serv., Fish. Bull. 65: 625-682. BOLIN, R. L. 1939. A review of the myctophid fishes of the Pa- cific coast of the United States and of lower Cal- ifornia. Stanford Ichthyol. Bull. 1: 89-156. 1959. Iniomi. Myctophidae from the "Michael Sars" North Atl. Deep-Sea Expedition 1910. Rep. Sci. Results "Michael Sars" North Atl. Deep-Sea Exped. 1910 Vol. 4, Part 2, No. 7, 45 p. Fraser-Brunner, a. 1949. A classification of the fishes of the family Myctophidae. Proc. Zool. Soc. London 118: 1019- 1106. Frizzell, D. L., and J. H. Dante. 1965. Otoliths of some early Cenozoic fishes of the Gulf Coast. J. Paleontol. 39: 687-718. Nafpaktitis, B., and J. R. Paxton. 1968. Review of the lanternfish genus Lampadena with the description of a new species. Los An- geles County Mus. Nat. Hist. Contrib. Sci. 138. NicOL, J. A. 1969. Bioluminescence. In W. S. Hoar and D. J. Randall (editors), Fish physiology, Vol. 3, p. 355- 400. Academic Press, New York. Parr, A. E. 1928. Deepsea fishes of the order Iniomi from the waters around the Bahama and Bermuda Islands. Bull. Bingham Oceanogr. Collect. Yale Univ. Vol. 3, Artie. 3, 193 p. Taning, a. V. 1928. Sjnopsis of the scopelids in the North At- lantic. Vidensk. Medd. Dan. Naturhist. Foren. Kjobenhaven 86: 49-69. APPENDIX Material examined Taaningichthys minimus University of Southern California, RV Velero IV- Stn. 11168, 31 July 1966, 32° N/120° W, 350 m, 10-ft IKMT, 1 (36 mm), LACM 9705. Stn. 11185, 2 Aug. 1966, 29° N/118° W, 375 m, 10-ft IKMT, 1 (47 mm), LACM 9650. International Indian Ocean Expedition, RV Anion Briiuu, Cruise III: Stn. 156, 6 Sept. 1963, 29° S/60° E, 122 m, 10-ft IKMT, 1 (45 mm), LACM 31320. University of Hawaii, Institute of Marine Biology: LACM 31574, 11 Sept. 1969, Hawaiian waters, 380 m, 6-ft IKMT, 2 (55-57 mm). LACM 31575, 30 Oct. 1969, Hawaiian waters 780 m, 6-ft IKMT, 2 (63-65 mm). LACM 31576, 13 Nov. 1969, Hawaiian waters, 575 m, 10-ft IKMT, 2 (50-60 mm). Scripps Institution of Oceanography: SIO 57-86, 12 May 1955, 29° N/125° W, 700 m, 10-ft IKMT, 1 (42 mm). SIO 62-430, 24 Aug. 1962, 29° N/130° W, 600 m, 10-ft IKMT, 1 (52 mm). SIO 68-490, 22 Sept. 1968, 29° N/178° W, no depth, 10-ft IKMT, 1 (46 mm). SIO 69-341, 27 Mar. 1969, 13° N/110° W, 1,100 m, 10-ft IKMT, 1 (42 mm). Woods Hole Oceanographic Institution: RHB stn. 1112, 17 June 1965, 22° N/70° W, 200 m, 10-ft IKMT, 1 (28 mm). RHB stn. 1735, 8 July 1968, 28° N/67° W, 870 m, 10-ft IKMT, 1 (40 mm). Zoological Institute, Academy of Sciences of the USSR, Leningrad: RV Vitiaz stn. 4885, 20 Dec. 1960, 17° S/71° E, 2,700 m, RT, 1 (27 mm). RV Vitiaz stn. 5127, 28 Oct. 1961, 13° N/154 ° W, 1,000 m, CN, 1 (48 mm). National Marine Fisheries Service, RV Horizon: Cruise H6204, stn. 100-140, 15 Apr. 1962, 28° N/ 124° W, 1,676 m, 10-ft IKMT, 1 (53.5 mm). Cruise H6204, stn. 110-160, 17 Apr. 1962, 35° N/ 124° W, 1,676 m, 10-ft IKMT. 1 (50 mm). U.S. National Museum, Ocean Acre material: Stn. 2-2N, 6 Mar. 1967, 32°26' N/63°44' W, 140 m, 6-ft IKMT, 1 (21 mm). Stn. 3-3N, 4 July 1967, 33°4' N/64°37' W, 1,060 m, 10-ft IKMT, 1 (52.5 mm). Stn. 3-4N, 4 July 1967, 33°10' N/64°45' W, 480 m, 10-ft IKMT, 1 (37 mm). Stn. 3-6N, 5 July 1967, 33°9' N/64°33'W, 250 m, 10-ft IKMT, 3 (36-45.2 mm). Stn. 3-13N, 6 July 1967, 32°54' N/64°45' W, 161 m, 10-ft IKMT, 1 (20.5 mm). Stn. 4-9A, 4 Sept. 1967, 31°52' N/63°58' W, 479 m, 10-ft IKMT, 1 (49 mm). Stn. 4-9B, 4 Sept. 1967, 31°52' N/63°58' W, 479 m, 10-ft IKMT, 1 (42 mm). Stn. 4-16C, 6 Sept. 1967, 32° N/64° 17' W, 500 m, 10-ft IKMT, 1 (41 mm). Stn. 6-7B, 26 Apr. 1967, 31°47' N/63°53' W, 155 m, 10-ft IKMT, 1 (33.5 mm). 75 FISHERY BULLETIN: VOL. 70, NO. 1 Stn. 6-15B, 28 Apr. 1967, 32°13' 10-ft IKMT, 1 (33 mm). Stn. 6-1.5P, 28 Apr. 1967, 32°13' 10-ft IKMT, 1 (34.5 mm). Stn. 6-18P. 29 Apr. 1967, 32° 14' 10-ft IKMT, 1 (30 mm). Stn 6-26.\, 30 Apr. 1967, 32°18' 10-ft IKMT, 1 (29 mm). Stn. 7-14X, 8 Sept. 1967, 32°12' 10-ft IKMT, 1 (45 mm). Stn. 7-15N, 8 Sept. 1967, 32°21' 10-ft IKMT, 1 (42 mm). N/63°51' W, 160 m, N/63°51' W, 160 m, N/63°46' W, 750 m, N/63°55' W, 200 m, N/63°25' W. 250 m, N/63°29' W, 450 m, Taatihigichthys hathyphilm University of Southern California, RV Eltanin: Stn. 947, 27 Jan. 1964, 67°31' S/90°26' W, 2,690 m, 3-m IKMT, 1 (67 mm), LACM 10424. Stn 1724, 18 July 1966, 40°06' S/149°55' W, 1,180 m, 3-m IKMT, 1 (60 mm), LACM 11247. Universitv of Southern California, RV Velero IV: Stn. 8959, 17 Oct. 1963, 33° N/119° W, 900 m, 10-ft IKMT, 1 (69 mm), LACM 6435. Stn. 10607, 10 June 1965, 33° N/119° W, 900 m, 10-ft IKMT, 1 (61 mm), LACM 6723. Stn. 966, 15 May 1964, 33° N/118° W, 750 m, 10-ft IKMT, 1 (72 mm), LACM 8525. Stn. 8238 25 Oct. 1962, 33° N/118° W, 10-ft IKMT, 1 (64 mm), LACM 9036. Stn. 9860, 25 June 1964, 33° N/118° W, 750 m, 10-ft IKMT, 1 (60 mm), LACM 9089. Stn. 11538. 21 June 1967, 32° N/118° W, 1,300 m, 10-ft IKMT, 1 (66 mm), LACM 9676. Stn. 11539, 21 June 1967, 33° N/118° W, 950 m, 10-ft IKMT, 1 (63 mm), LACM 9677. Stn. 11617, 16 Aug. 1967, 31° N/118° W, 1,130 m, 10-ft IKMT, 1 (68 mm), LACM 9682. Stn. 11312, 25 Jan. 1967, 28° N/116° W, 1,325 m, 10-ft IK.MT. 1 (58 mm), LACM 9708. Stn. 10373, 23 Feb. 1965, 33° N/118° W, 10-ft IKMT, 2 (61-68 mm), LACM 9764. Stn. 11696, 12 Oct. 1967, 32° N/118° W, 860 m, 10-ft IKMT, 1 (67 mm), LACM 9796. Stn. 11733, 8 Nov. 1967, 20° N/106° W. 1,400 m, 10-ft IKMT, 1 (62 mm), LACM 300.34. Stn. 11767, 16 Nov. 1967, 24° N/109° W, 1,500 m, 10-ft IKMT, 1 (57 mm), LACM 30045. Stn. 12066, 12 Apr. 1968, 26° N/114° W, 1,300 m, 10-ft IKMT, 1 (61 mm), LACM 30075. Stn. 12072, 14 Apr. 1968, 29° N/118° W, 750 m, 10-ft IKMT, 1 (65 mm), LACM 30079. Stn. 12184, 24 July 1968, 31° N/119° W, 820 m 10-ft IK.MT, 1 (50 mm), LACM 30271. Stn. 12597, 17 Jan. 1969, 32° N/120° W, 770 m, 10-ft IKMT, 1 (67 mm), LACM 30348. Stn. 12392, 11 Oct. 1968, 32° N/118° W, 1,110 m, 10-ft IKMT, 1 (61 mm), LACM 30403. Stn. 12349, 12 Sept. 1968, 32° N/118° W, 1,400 m, 10-ft IK.MT, 1 (67 mm), LACM 20598. Stn. 12491, 20 Nov. 1968, 29° N/118° W 910 m, 10-ft IKMT, 2 (61-65 mm), LACM 30609. Stn. 13385, 28 Oct. 1969, 28° N/118° W, 780 m, 10-ft IKMT, 1 (68 mm), LACM 30886. Smithsonian Oceanographic Sorting Center, RV Anton Brunn, Cruise III and VI: Label no. 7033, 18 Aug. 1963, 4° N/60° E, 2,120 m, 10-ft IKMT, 1 (34 mm), LACM 31292. Label no. 7057, 23 Aug. 1963, 5° S/60° E, 2,030 m, 10-ft IKMT, 1 (26 mm), LACM 31303. Label no. 7083, 6 Sept. 1963, 29° S/60° E, 1,150 m, 10-ft IKMT, 7 (44-55 mm), LACM 31320. Label no. 7173, 23 May 1964, 8° N/65° E, 2,850 m, 10-ft IKMT, 1 (44 mm), LACM 31344. Label no. 7177, 23 May 1964, 7° N/65° E, 940 m, 10-ft IKMT, 1 (36 mm), LACM 31345. Label no. 7204, 27 May 1964, 2° N/65° E, 1,250 m, 10-ft IKMT, 1 (46 mm), LACM 31358. Label no. 7217, 28 May 1964, 14° S/65° E, 2,250 m, 10-ft IKMT, 1 (53 mm), LACM 31361. Label no. 7265, 4 June 1964, 12° S/64° E, 1,930 m, 10-ft IKMT, 1 (38 mm), LACM 31375. Label no. 7273, 6 May 1964, 14° S/65° E, 880 m, 10-ft IKMT, 1 (47 mm), LACM 31376. Label no. 7305, 24 June 1964, 24° S/65° E, 3,500 m, 10-ft IKMT, 1 (57 mm), LACM 31401. Label no. 7312, 25 June 1964, 24° S/65° E, 1,100 m, 10-ft IKMT, 1 (56 mm), LACM 31404. Office de la Recherche Scientifique et Technique Outre- Mer, Noumea, Ne\v Caledonia : RV Cnridc, Cruise I, stn. 36A, 23 Sept. 1968, 0°2' N/ 137°51' W, 950 m, 10-ft IKMT, 3 (51-60 mm) , LACM 31439. RV Cnride, Cruise I, stn. 39A, 24 Sept. 1968, 0°14' N/ 138°17' W, 1,130 m, 10-ft IKMT, 1 (46 mm), LACM 31440. RV Cnride, Cruise I, stn. 69A, 29 Sept. 1968, 0°5' N/ 144°41' W, 580 m, 10-ft IKMT, 1 (32 mm), LACM 31446. RV Caride, Cruise I, stn. 74 A, 29 Sept. 1968, 0°/ 145°41' W, 820 m, 10-ft IKMT, 1 (58 mm), LACM 31448. RV Cnride, Cruise I, stn. 77A, 30 Sept. 1968, 0°25' N/ 146° 17' W, 1,110 m, 10-ft IKMT, 1 (48 mm), LACM 31450. RV Cnride, Cruise I, stn. 78A, 30 Sept. 1968, 0°2' S/ 146°29' W, 1,280 m, 10-ft IKMT, 1 (42 mm), LACM 31451. RV Caride, Cruise III, stn. 17, 7 Feb. 1969, 11°17' S/ 142°47' W, 1,050 m, 10-ft IKMT, 1 (60 mm), LACM 31459. RV Cnride, Cruise III, stn. 18, 8 Feb. 1969, 11'='7' S/ 142°35' W, 1,050 m, 10-ft IKMT, 1 (53 mm), LACM 31460. RV Cnride, Cruise III, stn. 60, 18 Feb. 1968, 0°12' S/ 139°19' W, 850 m, 10-ft IKMT, 1 (55 mm), LACM 31464. RV Caride, Cruise III, stn. 64, 19 Feb. 1969, 0°/ 140°9' W, 900 m, 10-ft IKMT, 1 (48 mm), LACM 31467. RC Caride, Cruise III, stn. 68, 10 Feb. 1969, 0°/ 140°42' W, 1,080 m, 10-ft IKMT, 1 (29 mm), LACM 31469. RV Caride, Cruise III, stn. 122, 24 Feb. 1970, 0°3'N/ 147°2' W, 1,100 m, 10-ft IKMT, 2 (45-51 mm), LACM 31478. RV Caride, Crui.se III, stn. 200, 2 March 1970, 0°1' N/ 154°14' W, 930 m, 10-ft IKMT, 1 (46 mm), LACM 31491. RV Caride, Crui.se III, stn. 204, 2 March 1970, 0°/ 154°25' W, 1,160 m, 10-ft IKMT, 1 (34 mm), LACM 31492. RV Cyclone, Cruise III, stn. 8, 4 May 1967, 2°13' S/ 169°47' E, 1,125 m, 10-ft IKMT, 1 (37 mm), LACM 31501. RV Cyclone, Cruise III, stn. 17, 5 May 1967, 4°23' S/ 169°52' E, 1,090 m, 10-ft IKMT, 1 (51 mm), LACM 31505. RV Santo, Cruise 68, stn. 6, 20 July 1968, 16°17' S/ 166°40' E, 1,395 m, 10-ft IKMT, 1 (52 mm), LACM 31528. 76 DAVY: REVIEW OF LANTERNFISH GENUS Taaningichthys U.S. National Mu.seum: Ocean Acre material: Stn. 3-2N, 4 July 1967, 33° N/64°45' W, 1,425 m, 10-ft IKMT, 1 (44 mm). Stn. 3-llN, 5 July 1967, 33° N/64°40' W, 1,920 m, 10-ft IKMT, 2 (53-58 mm). Stn. 6-lOB, 27 Apr. 1967, 31°59' N/63°43' W, 900 m, 10-ft IKMT, 1 (55 mm). Stn. 6-24N, 30 Apr. 1967, 32°13' N, 63°40' W, 750 m, 10-ft IKMT, 1 (67 mm). Stn. 7-13N, 8 Sept. 1967, 32°18' N/63°30' W, 1,500 m, 10-ft IKMT, 1 (36 mm). N/21° W, N/24° W, N/24° W, Carlsberg Foundation, Dana collections: Dana stn. 1156 VII, 25 Oct. 1921, 25° 2,000 m, S 150, 1 (37 mm). Dana stn. 1159 II, 29 Oct. 1921, 18° 2,000 m, S 150, 1 (43 mm). Dana stn. 1159 III, 29 Oct. 1921, 18° 1,500 m, S 150, 1 (35 mm). Dana stn. 1181 III, 21 Nov. 1921, 13° N/57° W, 1,500 m, S 150, 1 (28 mm). Dana stn. 1217 III, 29 Jan. 1922, 19° N/79° W, 1,500 m, S 150, 1 (45 mm). Dana stn. 1342 I, 15 May 1922, 34° N/70° W, 2,250 m, E 300, 1 (43 mm). Dayia stn. 1365 IX, 8 June 1922, 32° N/42° W, 2,500 m, E 300, 1 (55 mm). Zoological Institute, Academy of Sciences of the USSR, Leningrad: RV Lyra stn. 50, 25 Mar. 1966, 3° N/120° W, 1,000 m, CN, 1 (43 mm). RV Lyra stn. 3717, 9 Jan. 1957, 3° N/128° E, 1,250 m, CN, 1 (35 mm). RV Vitiaz stn. 4183, 6 Dec. 1958, 40° N/127° W, 675 m, CN, 1 (65 mm). RV Vitiaz stn. 4189, 7 Dec. 1958, 40° N/133° W, 1,000 m, CN, 1 (49 mm). RV Viiiaz stn. 4939, 4 Feb. 1961. 9° N/87° E, 1,000 m, CN, 1 (39 mm). National Institute of Oceanography, Surrey, England: NIO stn. 4687, 30 Aug. 1961, 29°57' N/32°3' W, 800 m, IKMT, 1 (50 mm). NIO stn. 4746, 30 Sept. 1961, 29°59' N/22°56' W, 1,100 m, IKMT, 2 (38-44 mm). NIO stn. 5799, 19 Oct. 1965, 28°9' N/14°9' W, 675 m, IKMT, 1 (48 mm). NIO stn. 5810, 7 Nov. 1965, 28°4' N/13°51' W, 800 m, IKMT, 1 (28 mm). NIO stn. 5813, 10 Nov. 1965, 28°5' N/14°ll' W, 950 m, IKMT, 1 (56 mm). NIO stn. 6687, 7 Mar. 1968, 20°37' N/22°56' W, 1,000 m, RMT8, 1 (56 mm). NIO stn. 7072, 30 Oct. 1969, 20°27' N/25°32' W, 1,000 m, RMT8, 2 (32-55 mm). NIO stn. 7079, 3 Nov. 1969, 17°40' N/27°6' W, 1,000 m, RMT8, 1 (24 mm). NIO stn. 7089 #54, 22 Nov. 1969, 17°47' N/25°22' W, 1,000 m, RMT8, 2 (54-77 mm). NIO stn. 7089 #55, 22 Nov. 1969, 17°47'N/25°22' W, 2,000 m, RMT8, 1 (46 mm). British Museum of Natural History: Rosaura collection, 26 June 1969, 17° N/86° W, 1,100 m, S 200, 1 (57 mm). Rosaura collection, 26 June 1969, 11° N/76° W, 1,200 m, S 200, 1 (46 mm). National Marine Fisheries Service, RV Horizon: Cruise H6204 stn. 120 • 70, 23 Apr. 1962, 26° N/117° W, 1,676 m, 10-ft IKMT, 1 (65 mm). Woods Hole Oceanographic Institution : RHB stn. 977, 26 Feb. 1963, 1° S/27° W, 10-ft IKMT, 1,100 m, 1 (65 mm). RHB stn. 979, 28 Feb. 1963, 3° S/29° W, 10-ft IKMT, 1,100 m, 1 (57 mm). RHB stn. 1603, 6 Oct. 1967, 39°46' N/70°30' W, 10-ft IKMT, 1,000 m, 1 (65 mm). Taauingichthys paurolychniis University of Southern California, RV Velero IV: Stn. 10675, 28 Aug. 1965, 29° N/118° W, 1,625 m, 10-ft IKMT, 1 (44 mm), LACM 9350. Stn. 11187, 2 Aug. 1966, 29° N/118° W, 1,720 m, 10-ft IKMT, 1 (87 mm), LACM 9567. Stn. 11257, 21 Oct. 1966, 29° N/118° W, 940 m, 10-ft IKMT, 1 (77 mm), LACM 9408. Stn. 11628, 18 Aug. 1967, 32° N/119° W, 1,300 m, 10-ft IKMT, 1 (33 mm), LACM 9693. Stn. 12331, 24 Aug. 1968, 29° N/118° W, 1,100 m, 10-ft IKMT, 1 (55 mm), LACM 30284. Stn. 12340, 26 Aug. 1968, 32° N/118° W, 1,130 m, 10-ft IKMT, 1 (17 mm), LACM 30591. Stn. 12475, 18 Nov. 1968, 28° N/119° W, 900 m, 10-ft IKMT, 1 (42 mm), LACM 30606. Stn. 12483, 19 Nov. 1968, 28° N/119° W, 2,080 m, 10-ft IKMT, 3 (52-81 mm), LACM 30382. Stn. 12592, 15 Jan. 1969, 32° N/120° W, 1,950 m, 10-ft IKMT, 3 (23-82 mm), LACM 30429. Stn. 12593, 16 Jan. 1969, 32° N/120° W, 1,920 m, 10-ft IKMT, 1 (91 mm), LACM 30430. Stn. 12594, 16 Jan. 1969, 32° N/120° W, 1,250 m, 10-ft IKMT, 1 (79 mm), LACM 30431. Stn. 12786, 16 Mar. 1969, 32° N/118° W, 1,350 m, 10-ft IKMT, 1 (66 mm), LACM 30423. Stn. 12791, 17 Mar. 1969, 32° N/118° W, 1,200 m, 10-ft IKMT, 1 (64 mm), LACM 30428. Smithsonian Oceanographic Sorting Center, RV Anto7i Briuni, Cruise III and VI: Label no. 7057, 23 Aug. 1963. 4° S/60° E, 2,030 m, 10-ft IKMT, 1 (51 mm), LACM 31303. National Marine Fisheries Service, RV Horizon: Cruise H6204 stn. 60.60, 26 Mar. 1962, 37° N/123° W, 1,863 m, IKMT, 1 (86 mm). Cruise H6204 .stn. 60.140, 29 Mar. 1962, 35° N/129° W, 1,863 m, IKMT, 1 (77 mm). Cruise H6204 stn. 80.90, 18 Mar. 1962, 33° N/123° W, 2,234 m, IKMT, 1 (29 mm). Cruise H6204 .stn. 100.60, 13 Apr. 1962, 31° N/119° W, 1,676 m, IKMT, 1 (68 mm). Cruise H6204 stn. 100.80, 13 Apr. 1962. 30° N/120° W, 1,676 m, IKMT, 1 (82 mm). Scripps Institution of Oceanography: SIO 54-95, 23 June 1954, 23° N/119° W, 2,500 m, 10-ft IKMT, 1 (49 mm). SIO 60-283, 12 Aug. 1960. 28° N/135° W, 3,000 m, 10-ft IKMT, 1 (44 mm). SIO 60-284, 13 Aug. 1960, 29° N/132° W, 3,000 m, 10-ft IKMT, 1 (71 mm). SIO 64-11, 3;) Jan. 1964, 24° N/113° W. 5,300 m, 10-ft IKMT, 1 (78 mm). SIO 66-31, 5 Apr. 1966, 29° N/117° W, 4,000 m, 10-ft IKMT, 1 (84 mm). SIO 67-52, 22 Apr. 1967, 30° N/117° W, 4,000 m, 10-ft IKMT, 1 (80 mm). 77 FISHERY BULLETIN: \0L. 70, NO. I SIO r>7-102, 10 June 1967, 35° N/123° W, 2,200 m, 10-ft Zoological Institute, Academy of Sciences of the USSR, IK.MT, 2 (65-71 mm). Leningrad: SIO 70-19, 17 Dec. 1969. 31° N/119° W, 4,000 m, 10-ft RV Akademik Kurchntov stn. 233, 13 Sept. 1968, IKMT, 1 (67 mm). 17° S/86° W, 2,000 m, CN, 1 (87 mm), 39908. SIO 70-20, 17 Dec. 1969:31° N/119° W, 4,000 m, 10-ft ^ , . , ,, IKMT '^ (79-95 mm) Zoological Museum, University of Copenhagen: ' " Dana stn. 1217 I, 29 Jan. 1922, 19° N/79° W, 2,000 m, E 300, 1 (57 mm), P2330669. National Institute of Oceanography, Surrey, England: Museu Municipal do Funchal, Madeira: NIO .sta. 7089 #55, 22 Nov. 1969. 17°47' N/25°22' W, RV Discovery 4742, 20 Sept. 1961, 32°42' N/16°32' W, RMT8, 3 (45-73). 1,700 m, IKMT, 1 (49 mm), MMF 22115. 78 SOME LIFE HISTORY CHARACTERISTICS OF COHO SALMON OF THE KARLUK RIVER SYSTEM, KODIAK ISLAND, ALASKA Benson Drucker^ ABSTRACT This paper contains data on some life history characteristics of the coho salmon of the Karluk River system, Kodiak Island, Alaska: age, fecundity, length, and egg size of adults; and migration charac- teristics, age, and size of smolts. The greater age at maturity of Karluk coho salmon (4 and 5 years) because of the longer freshwater residence of the juveniles is unique among reported North American stocks and may result in greater freshwater mortality but less marine mortality because the smolts are larger when they enter the ocean. Fecundity of Karluk coho salmon also differs from that reported for other North American stocks in that they are extremely fecund — more similar to Asiatic stocks of the Kamchatka Peninsula. Coho salmon, Oncoi^hynchus kisutch, are widely distributed along the Pacific coast of North America and occur in commercially harvestable quantities from northern California to north- western Alaska. About one-third of the total North American commercial catch comes from Alaska waters, where from 1960 to 1968 the average annual catch of 16 million pounds was valued at almost $3.5 million to the fishermen." The amount of biological research on coho salmon in Alaska is small, and published scientific re- ports on Alaska coho salmon stocks are very few. In this paper I present data on some life his- tory characteristics of the coho salmon of the Karluk River system. This system is located on the southwest side of Kodiak Island, Alaska, at approximately lat 57° N and long 154° W and includes Karluk Lake, tributaries to the lake. Thumb and O'Malley Lakes, and Karluk River (Figure 1). Information is presented on age, fecundity, length, and egg size of coho salmon adults; and migration characteristics, age, and ^ National Marine Fisheries Service, Auke Bay Fish- eries Laboratory, Auke Bay, Alaska 99821 ; present ad- dress: National Marine Fisheries Service, Technical Advisory Division, Interior Building, Washington, D.C. 20235. " Nelson, Richard C. 1968. Alaska catch and pro- duction, commercial fisheries statistics. Alaska Dep. Fish Game, Stat. Leafl. 17. 29 p. (Unpublished.) Figure 1. — The Karluk River system, Kodiak Island, Alaska. Manuscript accepted July 1971. FISHERY BULLETIN: VOL. 70, NO. 1, 1972. 79 FISHERY BULLETIN: VOL. 70, NO. I size of coho salmon smolts. These life history features of Karluk coho salmon are compared with those reported for various coho salmon stocks from other areas on both the Asiatic and North American sides of the Pacific Ocean. In addition, the effect of a prolonged juvenile fresh- water residence, a feature unique to the Karluk system, on freshwater and marine survival is discussed. All of the data on Karluk coho salmon were collected incidentally during studies of sockeye salmon. 0. nerka, the dominant salmon species in the Karluk system; much of the in- formation on other Alaska stocks is from un- published administrative and progress reports. Because I intend to discuss differences between coho salmon in the Karluk system and those in other areas, a description of general features of the life cycle of coho salmon stocks is appropri- ate. Typically, the adults enter streams and rivers from late summer to November and spawn in late fall and early winter. Some Asiatic stocks, however, spawn as late as mid-March (Smirnov, 1960). The progeny emerge as fry in the spring following spawning and reside in rivers or lakes for 1 or 2 years before going to sea as smolts. In some areas the seaward smolt migration begins in late winter (Chapman, 1961; Smoker, 1953), but in most areas it takes place from April to August (Godfrey, 1965). The salmon grow rapidly in the ocean, and the adults return to the streams and rivers to spawn 12 to 18 months later. However, a significant per- centage of male coho salmon, particularly in their southern range of distribution in North America (California), mature precociously (6 to 9 months after they enter salt water) and return to spawn the same year that they mi- grated to sea (Shapovalov and Taft, 1954). These fish are known as jack salmon. METHODS The data for adult coho salmon of the Karluk system were obtained from fish from the 1966 escapement that were captured at the adult counting weir or caught by sport fishermen at the outlet of Karluk Lake about 300 yards up- stream from the weir site. All fish were mea- sured for length (mideye to fork of tail) with a caliper to the nearest millimeter. Mideye-fork length was used because of morphological changes that occur as the fish matures, partic- ularly the elongation of the snout. Ovaries for fecundity samples were removed from all females and were preserved in 10% Formalin solution for at least 48 hr. The eggs were then hand-counted to get total egg counts. The diameters of some eggs from the fecundity samples were measured. These eggs were removed directly from the ovary, water hardened, and placed in Stockard's solution. The diameters were then measured with a vernier measuring microscope calibrated to 0.01 mm. The ages of adult fish were deter- mined by reading scales that had been taken halfway between the lateral line and the poster- ior insertion of the dorsal fin. The data for smolts were obtained from fish captured in 1956, 1965, and 1968 in fyke nets fished on the downstream side of the adult counting weir. Fork lengths were taken to the nearest millimeter with a steel millimeter ruler, and weights were taken to the nearest tenth of a gram on a triple-beam balance. As with the adults, the ages of smolts were determined from scales taken halfway between the lateral line and the posterior insertion of the dorsal fin. AGE OF COHO SALMON The average age composition of coho salmon for several systems in northern and southern latitudes of North America and Asia is shown in Table 1. The differences from the northern to southern latitudes in age composition is sim- ilar to that noted by Marr (1943) and possibly represent a geographic cline. The Karluk system had three freshwater age classes," two of which were decidedly predom- inant (Table 1). The three age classes, 43, 54, and 65, designate fish that went to sea in their third, fourth, and fifth years of life and returned to spawn after being at sea for about 1 year. "Age classes are designated according to the system developed by Gilbert and Rich (1927). A 43 coho salmon is in its fourth year of life. It went to sea as a smolt at the beginning of its third year, having spent two growing seasons in fresh water. 80 DRUCKER: COHO SALMON OF KARLUK RIVER SYSTEM Table 1. — Age class composition of stocks of coho salmon from North America and Asia, arranged geographically from north to south. Area Percent coho salmon in age class— 2i 2a 33 33 4i 4a 43 53 65 Reference North America Nome River, Alaska Unalakleet River, Alaska Yukon River, Alaska Yukon River, Alaska Cook Inlet River, Alaska Resurrection Boy, Alaska Bear Creek, Alaska Dairy Creek, Alaska Mendenhall River, Alaska Hood Bay Creek, Alaska Karluk River, Alaska Sashin Creek, Alaska Port Herbert, Alaska Stikine River, Alaska Chignik River, Alaska Ketchikan River, Alaska Quatsino Bay, British Columbia Fraser River, British Columbia Georgia Strait, British Columbia West coast, Vancouver Island, British Columbia Langara Island, Georgia Strait, British Columbia Columbia River, Wash. Waddell Creek, Calif. Asia East coast of Kamchatka, USSR: Kamchatka River Lake Ushki Kyrganik River Paratunko River Avachin Gulf, Solevarko Bay Kalyger River West coast of Kamchatka, USSR: Kikhchik River Bolshaya River Ozernaia River Kukhtui River (Okhotsk) 2.6 1.5 1.3 0.1 2.0 0.6 O.I 6.1 18.4 1.2 Tr. Tr. Tr. Tr. Tr. Tr. 29.4 37.9 55.6 38.7 40.0 30.3 27.1 383.3 12.0 46.5 8.0 0.7 0.4 0.6 18.0 20.0 45.2 23.2 70.8 95.0 96.5 97.1 97.5 0.2 97.9 83.9 81.6 55.7 4.3 93.6 57.4 80.2 27.5 Tr. 0.3 Tr. Tr. 100.0 69.3 100.0 32.8 Tr. Tr. Tr. Tr. Tr. Tr. Tr. Tr. 1.0 70.6 62.1 44.4 58.1 60.0 68.8 71.1 16.7 80.0 47.5 56.9 77.0 76.0 51.9 72.4 29.2 0.1 3.0 0.2 0.4 0.4 0.7 0.2 0.9 9.7 43.1 92.9 6.4 42.6 19.8 72.5 30.1 67.2 3.2 0.9 1.8 1.9 6.0 41.7 5.0 4.0 4.4 1.4 2.8 Godfrey (1965) Godfrey (1965) Godfrey (1965) Gilbert (1922) Godfrey (1965) Logan (1963,i 1964=) Logan (1964) = Logan (1964) = (') Armstrong (1970) Present study Crone (1968) Crone (1968) Godfrey (1965) Israel (1933) Godfrey (1965) Godfrey (1965) Godfrey (1965) Godfrey (1965) Godfrey (1965) Pritchard (1940) Marr (1943) Shapovalov and Toft (1954) Godfrey (1965); Gribanov (1948) Gribanov (1948) Gribanov (1948) Gribanov (1948) Gribanov (1948) Gribanov (1948) Gribanov (1948) Godfrey (1965); Gribanov (1948); Semko (1954) Gribanov (1948) Godfrey (1965) In Dingell-Johnson project report, 1962-63, Vol. 4: In Dingell-Johnson project report, 1963-64, Vol. 5: 175-194, Alaska 133-151, Alaska 1 Logan, Sidney M. 1963. Silver salmon studies in the Resurrection Bay area Dep. Fish Gome, Sport Fish Div., Juneau, Alaska. (Unpublished.) = Logan, Sidney M. 1964, Silver salmon studies in the Resurrection Boy area Dep. Fish Game, Sport Fish Div., Juneau, Alaska. (Unpublished.) 3 The high percentage of age 3^ fish is atypical and not representative of Resurrection Bay streams. Dairy Creek juveniles ore reared in a brackish water lagoon rather than in the stream itself, resulting in 1-year smolts. (Personal communication, Sidney Logan, Area Management Biologist, Alaska Department of Fish and Game, Soldotna, Alaska, March 15, 1971.) * Collected by author in 1966. Fifty-seven percent of the Karluk fish had mi- grated in their third year and 42 9f in their fourth year; only l^c had migrated in their fifth year (Table 1). Although the freshwater residence of fish in the Karluk escapement var- ied from 2 to 4 years, Karluk coho salmon, like those in all other systems, returned to spawn after being at sea for 12 to 18 months. The presence of large numbers of fish( 42%) that had spent 3 years in freshwater residence (age class 64) is unique to the Karluk system. Fish of age class 64 have been found in other Alaska river systems, i.e., the Yukon River (Gil- bert, 1922), Resurrection Bay and Bear Creek (see footnotes 1 and 2, Table 1) , Hood Bay Creek (Armstrong, 1970), Sashin Creek and Port Herbert (Crone, 1968), and Chignik River (Israel, 1933) ; but the proportion of 64 fish in the total runs to these systems is small — usually less than 5% (Table 1). The age composition of stocks of coho salmon from systems on the Kamchatka Peninsula, 81 FISHERY BULLETIN: VOL. 70. NO. 1 USSR, is similar to that of coho salmon in the northern areas of the west coast of North Amer- ica. The main age classes are 82 and 43 (Table 1 ) . The ratio of one age class to the other varies, however, from year to year and from area to area (Gribanov, 1948; Semko, 1954). An additional comparison of the age compo- sition of coho salmon from northern to southern latitudes is shown in Figure 2, which gives the percent age composition of the major age classes from five geographical areas along the west coast of North America. In California, the southern limit of the range of coho salmon, the major age class is 82, but jack salmon (age 22) contribute significantly to the runs. The 82 age class is still dominant in Washington, but the number of jack salmon is less and 43 fish are starting to appear. North of Washington to central British Colum- bia, more than 95 ^r of the fish are age 82, and there are only traces of other age classes, mainly the 43 class. From central British Columbia and northward through Alaska, the primary age class is 43; 82 fish are the secondary class and 54 fish are found in small numbers. In Alaska, the increase in total age is the result of juvenile coho salmon residing an additional year in fresh water before migrating to sea. Possible exceptions to the dominance of the 43 age class in Alaska are the Ketchikan River, Dairy Creek, Yukon River, and Karluk River systems (Table 1). In the first three river systems, 82 fish are the dominant age class and 43 fish the secondary class. The sizes of the samples from these systems were small, however (less than 25 fish). In the Kar- luk system, although 43 fish were dominant, 64 fish rather than 82 fish were the secondary age class (Figure 2). 100 "50 CALIFORNIA WASHINGTON "n 11^ I ^11 BRITISH COLUMBIA ALASKA KARLUK RIVER i h h \ \ S h S h S ^ S ^ AGE CLASS Figure 2. — Average age composition of coho salmon runs along the west coast of North America by geographical area (minor age classes omitted). The presence of older fish (bi) in northern latitudes may be a result of the juveniles being reared in lakes rather than rivers. Typically, coho salmon spawn in rivers or tributaries to rivers and the emerging fry reside in these areas until they migrate to sea. In contrast, in some Alaska river systems where 54 fish are part of the run (Table 1 and Figure 2), the juveniles migrate from spawning grounds to lakes before migrating to sea. It appears that some of the juveniles that reside in lakes (lake type) go to sea at an older age than those that reside in rivers (river type)/ NUMBER AND SIZE OF EGGS In this section, information is presented on fecundity (number of eggs contained in a fe- male) as a function of latitude and length, the relative numbers of eggs in right and left ovaries, and egg size in relation to length and fecundity. Fecundity and factors related to it form the basis for determining the reproductive potential of a spawning stock and subsequent survival from egg to young. Knowledge of variations in fe- cundity and egg size is of increasing importance in fish stocking and fish rehabilitation programs. Size of egg may be useful in predicting the con- dition, or hardiness, of developing fry. Because the fecundity of fish differs among geographic areas, the reproductive potential must be deter- mined for each stock. FECUNDITY AS A FUNCTION OF LATITUDE The average fecundity for both North Amer- ican and Asiatic stocks of coho salmon is con- siderably higher in fish from northern latitudes than in those from southern latitudes (Table 2 and Figure 8) , Coho salmon from Alaska river systems (with the exception of the two small samples from Port Herbert and Sashin Creek) * Personal communication, 1969, Charles J. DiCos- tanzo, Chief, Salmon Investigations, National Marine Fisheries Service, Auke Bay Fisheries Laboratory, Auke Bay, Alaska 99821. 82 DRUCKER: COHO SALMON OF KARLUK RIVER SYSTEM Table 2. — Average fecundity of coho salmon stocks from North American and Asiatic river systems, arranged geographically from north to south. Area Lati- tuda Average number of eggs Reference North America Swonson River, Alaska Bear Creek, Alaska Dairy Creek, Alaska Karluk River, Alaska Pasogshak River, Alaska Scshin Creek, Alaska Port Herbert, Alaska Namu River, British Columbia Fraser River, British Columbia Nile Creek, British Columbia Cultus Lake Hatchery, British Columbia Port John, British Columbia Cowichan River, British Columbia Oliver Creek, British Columbia Beadnell Creek, British Columbia Seattle, Wash. Winter Creek, Wash. Fall Creek, Alsea River, Ore. Scott Creek, Calif. Asia East coast of Kamchatka, USSR: Ushki Hatchery Kamchatka River Paratunka River West coast of Kamchatka, USSR: Bolshoya River Sakhalin Island, USSR: Tymi River 61° N 3,378 Engel (1966)i 60° N 4,115 Lawler (1963,= 1964^) 60° N 4,177 Engel (1965),*; Lawler (1963) 57° N 4,706 Present study 57° N 4,510 Marriott (1968) = 56° N 2,868 Crone (1968) 56° N 2,565 Crone (1968) 54° N 3,002 Foerster and Pritchard (1936) 53° N 3,152 Foerster and Pritchard (1936) 49° N 2,310 Wickett (1951) 49° N 2,300 Foerster and Ricker (1953) 49° N 2,313 Hunter (1948) 48° N 2,329 Neave (1948) 48° N 2,267 Foerster (1944) 48° N 2,789 Foerster (1944) 47° N 3,141 Alien (1958) 47° N 2,447 Salo and Bayliff (1958) 44° N 1,983 Koski (1966) 37° N 2,336 Shapovalov and Toft (1954) 56° N 5,282 Gribanov (1948) 56° N 4,883 Gribanov (1948) 53° N 4,350 Gribanov (1948) 53° N 4,638 Semko (1954) 52° N 4,570 Smirnov (I960) 1 Engel, Larry J. 1966. Egg-take investigations in Cook Inlet drainage and Prince William Sound. In Federal aid in fish restoration, 1965-66 progress report. Vol. 7: 109-116, Alaska Dep. Fish Game, Sport Fish Div., Juneau, Alaska. (Unpublished.) - Lav^ler, Robert E. 1963. Silver salmon egg taking investigations in Cook Inlet drainage. In Ding- ell-Johnson project report, 1962-63, Vol. 4: 161-173, Alaska Dep. Fish Game, Sport Fish DIv., Juneau, Alaska. (Unpublished.) ^ Lawler, Robert E. 1964. Egg take investigations in Cook Inlet and Prince William Sound. In Dingell-Johnson project report, 1963-64, Vol. 5: 123-132, Alaska Dep, Fish Game, Sport Fish Div., Juneau, Alaska. (Unpublished.) * Engel, Larry J. 1965. Egg take investigations in Cook Inlet drainage and Prince William Sound. In Dingell-Johnson project report, 1964-65, Vol. 6: 155-163, Alaska Dep. Fish Game, Sport Fish Div., Ju- neau, Alaska. (Unpublished.) = Marriott, Richard A. 1968. Inventory and cataloging of the sport fish waters In southwest Alaska. In Federal aid in fish restoration, 1967-68 progress reporf. Vol. 9: 81-93. Alaska Dep. Fish Game, Sport Fish Div., Juneau, Alaska. (Unpublished.) are more fecund than coho salmon from more southerly areas in North America (Figure 3). Stocks of coho salmon from Asiatic river systems are extremely fecund, even more so than North American stocks in more northerly latitudes. The high fecundity of Karluk River coho salmon more closely resembles the fecundity of Asiatic stocks than North American ones. Contrary to these findings for coho salmon, Rounsefell (1957) suggests that for the genus Oncorhynchus, salmon in southern latitudes may be more fecund than those in northern latitudes because of ". . . the higher age at maturity, and therefore slower growth rates, from south to north." Rounsefell found that the amount of time juvenile sockeye salmon spent in fresh water had no effect on fecundity, but the amount of time the adults spent at sea did have an effect: adult sockeye salmon that spent 2 years at sea had higher fecundity counts than fish of the same size that spent 3 years at sea. With coho salmon, however, the greater age at maturity is not due to increased time in the ocean but to increased time in fresh water. 83 FISHERY BULLETIN: VOL. 70, NO. 1 63' 59' 57' 55' 53' ui5|. o 3 <49' -I I a 47' o z 45' 43"- 39' 37' 35' 0' © \F © © ®® ® © ® @ ® ® ® ®^A ® ® ® 3 4 5 6 7 8 9 10 I I 12 13 14 15 16 17 IB 19 20 21 22 23 24 _1 SWftNSON RIVER, ftLASKA BEAR CREEK. ALASKA DAIRY CREEK .ALASKA KARLUK RIVER, ALASKA PASA6SMAK RIVER, ALASKA PORT HERBERT, ALASKA SASHIN CREEK, ALASKA KAMCHATKA RIVER, USSR USHKI HATCHERY, USSR NAMU RIVER. BRITISH COLUMBIA ERASER RIVER, BRITISH COLUMBIA PARATUNKA RIVER. USSR BOLSHAYA RIVER. USSR TYMI RIVER, USSR CULTUS LAKE, BRITISH COLUMBIA NILE CREEK. BRITISH COLUMBIA PORT JOHN. BRITISH COLUMBIA OLIVER CREEK, BRITISH COLUMBIA COWICHAN CREEK, BRITISH COLUMBIA BEADNELL CREEK, BRITISH COLUMBIA MINTER CREEK, WASHINGTON SEATTLE, WASHINGTON ALSEA RIVER, WASHINGTON SCOTT CREEK, CALIFORNIA _1_ _1_ 1,000 2,000 3,000 4,000 NUMBER OF EGGS 5,000 6,000 Figure 3. — Average fecundity of various stocks of echo salmon from North America and Asia. fork length by the method of least squares. The result may be expressed by the equation y = —7,503.55 + 195.51X, where Y is the esti- mate of number of eggs and X is the mideye- fork length of female salmon (Figure 4). The mean number of eggs for the sample was 4,706 (range 1,724 to 6,906); the mean length was 62.1 cm (range 46.6 to 69.8 cm). 2 3,000 Figure 4. — Relation of fecundity to length of coho salmon sampled at Karluk v^reir, 1966. FECUNDITY AS A FUNCTION OF LENGTH The presence of a positive relation between fecundity and length in the genus Oncorhynchus is well known (Gilbert and Rich, 1927; Foerster and Pritchard, 1941; Allen, 1958; Hartman and Conkle, 1960). For fish in general, the relation of fecundity to length is logarithmic (Y = aX^) over a wide range of lengths. For salmon, how- ever, the narrow range in length at maturity permits this relation to be described adequately by a straight line of the foi'm Y = a + bX (Foer- ster and Pritchard, 1941; Rounsefell, 1957). I counted the total number of eggs in 49 coho salmon from the Karluk River and calculated the relation between number of eggs and mideye- It is difficult to determine if the high fecun- dity of coho salmon of the Karluk system (Fig- ure 3) is due to greater fecundity per unit length or simply to the fact that coho salmon from Kar- luk are very large. The average lengths of fe- male coho salmon from various spawning streams along the Pacific coast of North Amer- ica are quite variable and do not seem to follow any set geographic pattern (Table 3). More- over, Karluk fish were measured from mideye to fork of tail, and direct comparisons of lengths with coho salmon from other areas are difficult to make because of variability in the types of measurements used. For instance lengths re- ported from areas other than Karluk include tip of snout to fork of tail (fork length), tip of snout to tip of tail (total length), and tip of snout to base of tail (standard length). 84 DRUCKER: COHO SALMON OF KARLUK RIVER SYSTEM Table 3.— Average lengths of female coho salmon from river systems along the Pacific coast of North America and Asia, arranged geographically from north to south. • Area North America Yukon River, Alaska Swanson River, Alaska Resurrection Boy, Alaska Dairy Creek, Alaska Brooks River, Alaska Karluk River, Alaska Sashin Creek, Alaska Port Herbert, Alaska Nomu River, Britisli Columbia Fraser River, British Columbia Seattle, Wash. Minter Creek, Wash. Columbia River, Wash. Deer Creek, Oreg. Flynn Creek, Oreg. Needle Branch, Oreg. Scott Creek, Calif. WadcJell Creek, Calif. Asia East coast of Kamchatka, USSR: Kamchatka River Kyrganik River Kalyger River Avachin Gulf Paratunka River West coast of Kamchatka, USSR: Kikhchik River Bolshaya River Ozernaia River Average length Fork Total Standard Mideye- fork Refer 63.4 62.0 67.2 __ 72.8 — 58.8 70.5 " 67.8 __ _. 69.0 64.0 63.4 ~ 74.6 70.7 __ 69.4 67.6 66.3 — _ 63.9 60.9 68.1 61.0 __ 55.4 __ 59.4 — 58.6 57.4 __ 62.6 __ 62.1 62.1 Gilbert (1922) Engel (I966)i Logon (1965) = Engel (1965)^ (') Present study Crone (1968) Crone (1968) Foerster and Pritchard (1936) Foerster and Pritchard (1936) Allen (1958) Salo (1955) Marr (1943) Koski (1966) Koski (1966) Koski (1966) Shapovalov and Toft (1954) Shapovalov and Taft (1954) Gribanov (1948) Gribanov (1948) Gribanov (1948) Gribanov (1948) Gribanov (1948) Gribanov (1948) Gribanov (1948) Gribanov (1948) ' See footnote 1, Table 2. 2 Logan, Sidney M. 1965. Silver salmon studies in the Resurrection Boy area. In Dingell-Johnson project report, 1964-65, Vol. 6: 129-145, Alaska Dep. Fish Game, Sport Fish Div., Juneau, Alaska. (Unpublished.) ■' Lowler, Robert E. 1964. Egg take investigations in Cook Inlet and Prince William Sound. In Dingell-Johnson project report, 1963-64, Vol. 5: 123-132, Alaska Dep. Fish Gome, Sport Fish Div., Juneau, Alaska. (Unpublished.) * Eicher, George J., Jr. The effects of laddering a falls in a salmon stream. National Marine Fisheries Service, Auke Bay Fisheries Laboratory, Auke Bay, Alaska, 5 p. (Unpublished.) NUMBER OF EGGS IN RIGHT AND LEFT OVARIES The numbers of eggs in the right and left ovaries of the genus Oncorhynchtis are usually quite variable. Rounsefell (1957) noted that although the rate of maturation of eggs from Karluk Lake sockeye salmon was the same in both ovaries of the same fish, the number of eggs in each ovary varied. Eguchi and his co-workers (Rounsefell, 1957) found no significant diflFer- ences in the numbers of eggs in the two ovaries in chum salmon, 0. keta, in Japanese waters. Helle (1970) found the same lack of a significant difference in a sample of pink salmon, 0. gor- buscha, from Olsen Bay, Alaska, in 1963. Sock- eye salmon from Brooks Lake, Alaska, in 1957 and 1958 and from Karluk Lake in 1958 had more eggs in the left ovary than in the right (Hartman and Conkle, 1960). At Bare Lake, Alaska, sockeye salmon had more eggs in the right ovary than in the left (Nelson, 1959). I compared the numbers of eggs from the right and left ovaries of Karluk River coho salmon (Table 4) by means of a ^ test for paired ob- servations. The differences between the num- bers of eggs in the right and left ovaries were significant (t = 2.60; df = 31; P = 0.05). In 31 fecundity samples, 71% had more eggs in the right ovary than the left. I could not find comparable information on comparisons between the numbers of eggs in the ovaries of coho salmon from other areas. 85 FISHERY BULLETIN: VOL. 70, NO. 1 Table 4. — Numbers of eggs in right and left ovaries from coho salmon collected at the outlet to Karluk Lake, 1966. Sample number 1 2 3 4 5 6 7 8 9 10 11 12 13 U 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 Average Mideye-fork length (cm) Number of eggs Righl ovary 50.9 61.4 65.0 56.9 62.7 65.6 69.8 64.7 62.6 60.2 59.3 60.2 62.3 61.2 60.1 63.5 64.5 67.2 60.2 65.2 64.0 65.1 66.0 62.9 65.8 64.6 63.6 66.7 63.1 61.9 64.8 1,640 2,265 3.005 2,213 2,322 3,001 3,559 2,258 2,546 2,433 2,243 2,331 2,481 2,620 2,067 2,581 2,473 2,604 2,044 2,824 2,608 2,491 2,926 2,266 3,047 2,726 2,721 3,104 2,981 2,176 2,340 Left ovary 1,403 1,918 2,876 2,083 2,337 3,147 3,347 1,884 2,501 2,220 2,225 2,161 2,283 2,539 2,000 2,221 2,546 3,233 1,813 2,697 2,659 2,501 3,174 2,280 2,878 2,579 2,563 2,997 2,843 2,125 2,521 Total 3,043 4,183 5,881 4,296 4,659 6,148 6,906 4,142 5,047 4,653 4,468 4,492 4,764 5,159 4,067 4,802 5,019 5,837 3,857 5,521 5,267 4,992 6,100 4,546 5,925 5,305 5,284 6,101 5,824 4,301 4,861 r.oo • 680 ' 660 - . • • • 6.40 is.zo ;;;6.oo - • • • • • • • • • • • UJ Z 5 5.80 o - • / • 13 £5.60 - • 25.40 • S 5.20 - 500 - • 4.60 - 0.0 U — 1— 1 1 1 1 _l fill III 63.0 2,545 2,469 5,015 48 50 52 54 56 58 60 62 64 66 68 70 72 74 76 MIOEyE-FORK length Icm) Figure 5. — Relation of mean egg diameter to mideye- fork length of female coho salmon, Karluk River, 1966. length of the fish. Allen (1958) in his studies of coho salmon in Green River, Wash., also found no relation. The average diameters of eggs were plotted against number of eggs in individual fish to de- termine if a relation existed between the fecun- dity of a female and the size of her eggs (Figure 6). For 24 females, the egg diameter ranged RELATION OF EGG SIZE TO LENGTH AND TO FECUNDITY The average diameter of eggs obtained from the fecundity samples from Karluk River was plotted against the length of the female coho salmon from which the samples were taken ( Fig- ure 5) to determine if there was a relation be- tween the size of a female and the size of her eggs. The size of eggs increases as they mature, and so the eggs used had to be in the same stage of maturation. I therefore selected only females beginning to show secondary sexual character- istics and containing eggs that could not be readily expressed from the body cavity. For 25 females the eggs varied in size from 4.92 to 6.88 mm. (mean 6.11 mm) ; lengths varied from 50.4 to 69.8 cm (mean 62.0 cm). No relation was found between the size of the egg and the • 6.80 - 6.60 - • . 640 - • • • • §6.20 a: - • • • • • ;^6.00 - • • • UJ Z < 580 _ • o • • o • S5 60 - z • uj5.40 - Z 5.20 - 5.00 - • 4.80 - 0.0 U — 1 — 1,- 1 _l L. 1111 1 1 1 1 1,600 2;600 3,600 4;600 5^00 6,600 NUMBER OF ECGS 7,600 Figure 6. — Relation of mean egg diameter to number of eggs in female coho salmon, Karluk River, 1966. 86 DRUCKER: COHO SALMON OF KARLUK RIVER SYSTEM from 4.91 to 6.87 mm (mean 6.11 mm); fecun- dity ranged from 2,855 to 6,906 (mean 4,766) . No relation was found: eggs from a female with low fecundity were not necessarily large, nor were those from a female with high fecundity necessarily small. Allen (1958) reported similar findings. Unlike the general relation that exists between fecundity and length (i.e., larger fish are more fecund), the size or fecundity of the females I sampled apparently had no relation to the size of the eggs. Large, fecund females had a wide range of egg sizes (Figure 5) . Thus, the larger number of eggs in large females may be due to a larger body cavity that allows more eggs to develop rather than to the fish having smaller eggs. COHO SALMON SMOLTS Smolts of coho salmon, like those of other salmon that live for a while in fresh water be- fore migrating to sea, migrate seaward at a par- ticular season and under particular light inten- sities. This migration and the associated envi- ronmental factors and information on age and size of migrating smolts are discussed in this section. SEASONAL MIGRATION Coho salmon juveniles reside in Karluk Lake for 1 to 4 years before they migrate to sea as smolts. From 1961 to 1967 the migration began in mid-May and was usually over by early July (Figure 7). Although most coho salmon migrate in the spring (Hamilton and Andrew, 1954; Taft, 1934; Gharrett and Hodges, 1950; Semko, 1954) , several exceptions do exist. In the Paratunka River, Kamchatka Peninsula, the migration ex- tends from the end of May to the end of August (Gribanov, 1948) ; in several streams in Oregon it extends from late winter to May (Chapman, 1961); in some streams in western Washington it runs from early winter to late spring ( Smoker, 1953) ; and at Waddell Creek, Calif., small num- bers of atypical migrants migrate in the fall and early winter (Shapovalov and Taft, 1954) . The 100 -| 1 1 T — I 1 1 1 1 rr 15 20 25 30 4 9 14 19 24 29! 4 9 14 19 MAY ' JUNE ' JULY Figure 7. — Cumulative seasonal migration of coho salm- on smolts from Karluk Lake, 1961-67. number of coho salmon smolts involved in the early or late parts of these migrations, however, represents only a small percentage of the total number of smolts in each migration. The warming of the water after the ice breaks up is of major importance in initiating the sea- ward migration of smolts. Hartman, Heard, and Drucker (1967) found this to be a major factor in the migration of sockeye salmon in lakes of southwestern Alaska; and Logan (see footnote 2, Table 1) found that the coho salmon smolt migration in Bear Lake, Alaska, did not start until the ice cover on the lake was gone and the water temperature had risen to 4.2° C. Ninety percent of the Bear Lake coho salmon smolts had migrated to sea when water temperatures ranged between 5° and 13.3° C. Coho salmon smolts apparently migrate over a greater 87 FISHERY BULLETIN: VOL. 70, NO. I temperature range than sockeye salmon, whose migration generally ends when water tempera- tures reach about 10° C (see footnote 2, Table 1) . At Karluk Lake, for each year from 1961 to 1968 (excluding 1964) the date by which 50' > of the coho salmon smolts had migrated was later than the comparable date for sockeye salm- on smolts (Table 5). The difference in time of the two migrations ran from 6 to 19 days (aver- age 11 days). Not only did more of the coho salmon smolts migrate later than the sockeye salmon smolts, but the coho salmon smolts usu- ally migrated during a period of relatively warm- er water, when the abundance of migrating sockeye salmon smolts had greatly diminished. Similarly, Foerster and Ricker (1953) found that the coho salmon smolt migration in Cultus Lake and Sweltzer Creek, British Columbia, al- ways followed the sockeye salmon smolt migra- tion by about 10 days. Although the seasonal timing of the outmigra- tion of coho salmon smolts may vary from system to system, it is relatively consistent within a particular system. When time of migration is plotted against latitude, a definite south to north cline in time of migration becomes evident (Fig- ure 8). Coho salmon smolts migrate later in the season in northerly systems than in more southerly ones. More than a month separates the midpoint of smolt migration from the central coast of California (lat 37° N) to the Gulf of Alaska (lat 60° N). This relation also applies for the Asiatic side of the Pacific Ocean. DIEL PATTERN OF MIGRATION The transformation of juvenile coho salmon from either lake- or stream-type residents to Table 5. — Dates by which 50% of the coho and sockeye salmon smolts migrated from Karluk Lake, 1961-68. Year Sampling period 50% migration date Coho salmon Sockeye salmon 1961 May 25 to June 29 June 10 June 2 1962 May 17 to June 21 June 10 May 29 1963 May 18 to July 6 June 12 June 6 1964 May 17 to July 6 June 2 June 3 1965 May 16 to July 16 June 18 June 6 1966 May 18 to July 2 June 15 June 4 1967 May 18 to June 29 Juno 5 May 27 1968 May 17 to June 26 June 12 May 24 60' 55' HOOD BAY CREEK, ALASKA* SASHIN CREEK, ALASKA* LAKELSE LAKE, BRITISH COLUMBIA* 50* o z UJ ^< (E (J 40" 35° 0' BEAR CREEK, ALASKA* *LAKE EVA, ALASKA •karluk LAKE, ALASKA *BOLSHAYA RIVER, KAMCHATKA, USSR *CULTUS LAKE, BRITISH COLUMBIA *MINTER CREEK, WASHINGTON *WADOELL CREEK. CALIFORNIA 10 15 20 MAY 25 31 /I 10 15 20 25 JUNE Figure 8. — Average date when 50% of the coho salmon smolts had migrated from river and lake systems along the Pacific coast of North America and Asia. smolts is associated with avoidance of light and increasing nocturnal activity (Hoar, Keenley- side, and Goodall, 1957; Hoar, 1958; Smirnov, 1960). Although most of the migration of smolts to salt water occurs during the darkest hours of the night, some occurs during the daytime. At Karluk Lake, for instance, during some years almost 40% of the coho salmon smolts migrated in the daytime — between 0600 and 1800 hr (Fig- ure 9). In other coho salmon rivers, the per- centage of smolts that migrate seaward during daylight is quite variable. In the Bolshaya River in Kamchatka, during the years 1944-47, 6.3 to 50.0% of the age 1 smolts"^ and 8.8 to 73.2% Fish that go to sea in their second year. 88 DRUCKER: COHO SALMON OF KARLUK RIVER SYSTEM 67 MEAN JUNE 22) 22 ) TIME PERIOD Figure 9. — Migration of echo salmon smolts by time period from Karluk Lake for the years 1961-67, of the age 2 smolts migrated in daylight (Semko, 1954.) At Bear Creek, Alaska, 49.8^r of the smolts migrated between 0400 and 2000 hr in 1962 (see footnote 1, Table 1), and W/c mi- grated between 0900 and 1700 hr in 1964 (see footnote 2, Table 3). AGE In 1956, 1965, and 1968, scale samples taken from seaward-migrating coho salmon smolts at the Karluk Lake weir revealed that the dom- inant ages were 2 and 3 and that the age com- position was similar between years (Table 6). The freshwater age composition determined from the scales of adults collected in 1966 was strikingly similar to the age composition of sea- Table 6. — Freshwater age composition of Karluk Lake coho salmon as determined from smolt and adult scale samples. ter Perce nf compos tlon Freshwa age From smolt sea es From adult scales 1956 1965 1968 1966 1 1.4 3.0 3.0 2 44.5 51.5 48.5 56.9 3 49.1 43.9 42.5 41.7 4 4.9 1.5 6.0 1.4 ward-migrating smolts in 1965 — smolts that pro- duced the adults in 1966 (Table 6). The only group missing from the adult scale sample but present in small numbers in the smolt scale sam- ples was age 1 (fish that went to sea in their second year of life). Fish from this age class could have been missing in the returning adults 89 FISHERY BULLETIN: VOL. 70. NO. 1 because (1) they had poor ocean survival be- cause of their small size; (2) the young over- wintered in the river and migrated as age 2 smolts the following year; or (3) the young never migrated at all. SIZE Only lengths were measured in 1956, and lengths and weights were taken in 1965 and 1968. A summary of average size data by age class is presented in Table 7. Average lengths for comparable age classes were greater in 1968 than in 1956 and 1965. Average weights, with the exception of age 3 fish, were less in 1968 than 1965. Differences in lengths and weights bet\N'een smolts for the two comparable years (1965 and 1968) are reflected in the condition factor (K), or coefficient of condition, which indicates the relative well-being of the fish. In 1965 all age groups had K values greater than 1.0000; the range was 1.0544 to 1.3695. In 1968 all K values were under 1.0000; the range was 0.9187 to 0.9600. Information on the size of coho salmon smolts from other spawning systems is presented in Table 8. This table gives information for natural or "wild" populations and not for artificially hatched or reared stocks. Karluk Lake coho salmon smolts were generally as large as smolts of the same age from other areas or larger. POSSIBLE EFFECTS OF INCREASED FRESHWATER RESIDENCE ON SURVIVAL OF COHO SALMON The extended period of freshwater residence resulting in coho salmon smolts of age 3 occurs in many systems but seems to be significant only at Karluk. It is interesting to hypothesize what effect a prolonged freshwater residence has on the an- nual return of adult coho salmon at Karluk Lake. Is an increased freshwater residence advanta- geous or disadvantageous to survival of each year class? What effect is there on marine sur- vival of coho salmon if they take up ocean resi- dence at an older age and consequently a larger size? One means of answering these questions is to examine freshwater and marine survival rates for coho salmon from other areas. Survival from egg to smolt (fresh water) and smolt to returning adult (marine) are shown in Table 9 for some areas in California, Oregon, Washing- ton, and British Columbia. Both freshwater and marine survival for age 1 smolts from these areas are quite variable: 0.13 to 12.00% and 3.77 to 11.79% respectively. The survival data in Table 9 pertain to stocks in which the smolts were primarily age 1 when they migrated to sea, and the application of these data to more northern stocks in which the smolts are mostly older and larger when they migrate must be done with caution. The small population of age 2 smolts from Sweltzer Creek in British Columbia (Table 9) is of interest because these fish are more compar- able to Karluk smolts, in that they may possibly have had a period of lake residence. Marine survival of these older, larger fish was high. Of 72 fin-clipped migrating age 2 smolts, 19 (26%) returned 5 or 6 months later as 83 fish (Foerster and Ricker, 1953). Although marine survival for these age 2 smolts might have been lower if they had spent another year in the ocean, it Table 7. — Average length, weight, and condition factor of coho salmon smolts by age from Karluk Lake, 1956, 1965, and 1968. Age Length 1956 1965 Weight Condition factor Length Weight Condition factor Length 1968 Weight Condition factor mm 1 106.8 2 139.7 3 151.1 4 165.4 112.5 136.3 141.7 177.0 g 19.5 28.2 30.7 63.9 1.3695 1 .0544 1 .0790 1.1523 114.8 140.1 160.4 181.8 e 13.9 26.4 38.5 56.2 0.9187 0.9600 0.9329 0.9353 90 DRUCKER: COHO SALMON OF KARLUK RIVER SYSTEM Table 8. — Average fork lengths of coho salmon smolts of ages 1 to 4 from river and lake systems along the Pacific coast of North America and Asia, arranged geographically from north to south. Area and year Age Reference North America Hood Bay Creek, Alaska 1968 83.0 1969 79.0 Karluk River (Karluk Lake), Alaska 1956 106.8 ^ 1965 112.5 1968 114.8 Bear Creek, Alaska 1962 106.3 Sweltzer Creek (Cultus Lake), British Columbia _. 110-120 1939 - Minter Creek, Wash. 1940 296.3 1953 !'99.7 Deer Creek, Oreg. 1960 88,7 Flynn Creek, Oreg. 1960 88.1 Waddell Creek, Calif. 1933 113.5 1934 113.3 1935 113.1 1936 116.6 1937 114.8 1938 112.4 1939 112.4 1940 109.5 1941 103.1 Asia Bolshoya River, Kamchatka, USSR 85.0 96.0 91.0 139.7 136.3 140.1 118.7 291.6 130.0 151.1 141.7 160.4 150.8 165.4 177.0 181.8 Armstrong (1970) Armstrong (1970) Present study Present study Present study Logan^ Foerster and Ricker (1953) Foerster and Ricker (1953) Sab and Bayliff (1958) Salo and Bayliff (1958) Chapman (1961) Chapman (1961) Shapovalov Shapovalov Shapovalov Shapovalov Shapovalov Shapovalov Shapovalov Shapovalov Shapovalov and Toft and Taft and Taft and Toft and Taft and Taft and Taft and Taft and Taft (1954) (1954) (1954) (1954) (1954) (1954) (1954) (1954) (1954) Semko (1954) 1 Personal communication, Sidney M. Logan, Fishery Biologist, Alaska Department of Fish and Game, May 3, 1967. ^ Primarily age 1 fish. Table 9. — Average freshwater and marine survival for coho salmon from various streams along the Pacific coast of North America, arranged geographically from north to south. Streams Nile Creek, British Columbia Hooknose Creek, British Columbia Sweltzer Creek, British Columbia Sweltzer Creek, British Columbia Sweltzer Creek, British Columbia Minter Creek, Wash. Deer Creek, Oreg. Waddell Creek, Calif. Percent survival Fresh water Marine Reference Egg to Aga 1 smolt aga 1 smolt to adult 1.40 6.00 Wicketf (1951) 1 1.30 11.79 Godfrey (1965) I 10.13 8.07 Foerster and Ricker (1953) I 2,^0.33 _. Foerster and Ricker (1953) 1 *26.39 Foerster and Ricker (1953) 3.22 3.77 Salo and Bayliff (1958) 12.00 _^ Chapman (1961) 1.35 4.95 Shapovalov and Taft (1954) 1 Before piscivorous fishes were controlled. " After piscivorous fishes were controlled. ^ Geometric mean. * Age 2 fish only. 91 FISHERY BULLETIN: VOL. 70, NO. 1 nevertheless was considerably higher than for any of the age 1 smolts. In the absence of knowledge of survival rates for the more northern populations of coho salm- on, an examination of the effect of increased freshwater residence on sockeye salmon, the dominant species of salmon in the Karluk system, is of value. Sockeye salmon juveniles at Karluk Lake have long been known to reside in the lake a year or more longer than do sockeye salmon in other areas (Gilbert and Rich, 1927) . In most Alaska systems, sockeye salmon smolts migrate at the beginning of their second or third year of life, but at Karluk Lake most sockeye salmon smolts migrate at the beginning of their third or fourth years. Possibly the factor (s) respon- sible for the 1-year holdover of juvenile sockeye salmon in the lake may also be responsible for the holdover of juvenile coho salmon. Freshwater sui'vival of sockeye salmon at Kar- luk Lake is extremely poor, but marine survival is good. During the late 1920's and early 1930's, freshwater survival was less than 1 % and ocean survival was about 21% (Barnaby, 1944). In recent years, freshwater survival has dropped to less than 0.5% and ocean survival has in- creased to about 40%.' Ricker (1962) modi- fied Barnaby's data by applying a marking mor- tality factor derived from his Cultus Lake studies and determined that the older, larger smolts have greater marine survival and that Barnaby's ori- ginal estimate of 21% survival was too low. Average marine survival by freshwater age for the years 1926 and 1929-33 were as follows: age 1 smolts, 18.3%; age 2, 27.4%; age 3, 34.2%,, and age 4, 33.3%. Ricker attributed the high ocean survival to the large size of the smolts when they entered salt water. The larger size of the sockeye salmon smolts at the time of sea- ward migration, however, is offset by a greater total freshwater mortality due to their prolonged stay in the lake. I have shown that in the more northern lat- itudes coho salmon usually reside a minimum of one extra year in fresh water before they migrate ' Unpublished data on file at National Marine Fish- eries Service Auke Bay Fisheries Laboratory, Auke Bav. Ala.ska 99821. to sea. Generally, a longer period of freshwater residence will result in greater freshwater mor- tality but lower marine mortality because the fish are larger when they enter the ocean. Most likely, as with Karluk Lake juvenile sockeye salmon, an extra year in the lake for juvenile coho salmon probably results in a greater total freshwater mortality. Total marine mortality, however, may be less for coho salmon than for sockeye salmon because the coho salmon gener- ally spend less time at sea before returning to spawn (12 to 18 months rather than 24 to 30 months). SUMMARY AND CONCLUSIONS Both the freshwater and total ages of adult coho salmon increase from southern to northern latitudes. In California, the southern portion of the coho salmon's range, fish of ages 32 and 22 are in the majority, but in the northern areas, ages 43 and 32 predominate. Karluk coho salmon, however, are unique, in that although age 43 fish are still the primary age class, the age 32 fish are replaced by age 54, so that age 54 fish account for 42% of the run. In no other North Amer- ican or Asiatic coho salmon stock for which in- formation is available is such a large percentage of the run composed of 54 fish. The increase in total age of coho salmon from south to north is associated with the increased time the juveniles spend in fresh water. The small numbers of age 54 fish in several Alaska stocks may represent juveniles that live in lakes rather than rivers. Fecundity generally increases from south to north, and Karluk coho salmon are the most fecund of any North American stock and closely parallel the highly fecund Asiatic stocks from the Kamchatka Peninsula. In Karluk coho salmon, there is a relation between number of eggs and length but no relation between egg size and length or egg size and fecundity. Egg counts are significantly higher in the right ovary than in the left. Coho salmon smolts generally migrate after the ice breaks up and the water warms. In North America, the migration is earlier in southern latitudes than northern ones. The coho salmon 92 DRUCKER: COHO SALMON OF KARLUK RIVER SYSTEM migration at Karluk Lake is primarily nocturnal, although some daytime migration does occur. A prolonged freshwater residence by juvenile coho salmon in Karluk Lake should result in a greater total freshwater mortality, but the re- sulting larger smolts should have a lower total marine mortality. Coho salmon at Karluk may have an even lower marine mortality than sock- eye salmon, in part because the coho salmon spend less time at sea. LITERATURE CITED Allen, G. H. 1958. Notes on the fecundity of silver salmon {Oncorhynchus kisutch). Prog. Fish-Cult. 20: 163-169. Armstrong, R. H. 1970. Age, food, and migration of Dolly Varden smolts in southeastern Alaska. J. Fish. Res. Board Can. 27: 991-1004. Barnaby, J. T. 1944. Fluctuations in abundance of red salmon, Oncorhynchus nerka (Walbaum), of the Karluk River, Alaska. U.S. Fish Wildl. Serv., Fish. Bull. 50: 237-295. Chapman, D. W. 1961. Factors determining production of coho salm- on, Oncorhynchus kisutch, in three Oregon streams. Ph.D. Thesis, Oreg. State Univ., Cor- vallis, 214 p. Crone, R. A. 1968. Behavior and survival of coho salmon, On- corhynchus kisutch (Walbaum), in Sashin Creek, southeastern Alaska. M.S. Thesis, Oreg. State Univ., Corvallis, 79 p. FOERSTER, R. E. 1944. Appendix IV. Report for 1943 of the Pa- cific Biological Station, Nanaimo, B.C. Annu. Rep. Fish. Res. Board Can., 1943: 22-26. FoERSTER, R. E., and A. L. Pritchard. 1936. The egg content of Pacific salmon. Biol. Board Can., Prog. Rep. Pac. Biol. Stn. Pac. Fish. Exper. Stn. 28: 3-5. 1941. Observations on the relation of egg content to total length and weight in the sockeye salmon {Oncorhynchus nerka) and the pink salmon (O. gorbuscha) . Trans. R. Soc. Can., Ser. 3, Sec. 5, 35: 51-60. FOERSTER, R. E., AND W. E. RiCKER. 1953. The coho salmon of Cultus Lake and Sweltzer Creek. J. Fish. Res. Board Can. 10: 293-319. Gharrett, J. T., AND J. I. Hodges. 1950. Salmon fisheries of the coastal rivers of Ore- gon, south of the Columbia. Oreg. Fish Comm., Contrib. 13, 31 p. Gilbert, C. H. 1922. The salmon of the Yukon River. Bull. U.S. Bur. Fish. 38: 317-332. Gilbert, C. H., and W. H. Rich. 1927. Investigations concerning the red-salmon runs to the Karluk River, Alaska. Bull. U.S. Bur. Fish. 43, Part II: 1-69. Godfrey, H. 1965. Salmon of the North Pacific Ocean— Part IX, Coho, chinook and masu salmon in offshore waters. 1. Coho salmon in offshore waters. Int. North Pac. Fish. Comm., Bull. 16: 1-39. Gribanov, V. I. 1948. Kizuch [Oncorhynchus kisutch (Walb.)]: (biologichestii ocherk). (The coho salmon (On- corhynchus kisutch (Walbaum) — A biological sketch.) Izv. Tikhookean. Nauchn.-issled. Inst. Rybn. Khoz. Okeanogr. 28: 43-101. (Fish. Res. Board Can., Transl. Ser. 370.) Hamilton, J. A. R., and F. J. Andrew. 1954. An investigation of the effect of Baker Dam on downstream-migrant salmon. Int. Pac. Salmon Fish. Comm., Bull. 6, 73 p. Hartman, W. L., and C. Y. Conkle. 1960. Fecundity of red salmon at Brooks and Kar- luk Lakes, Alaska. U.S. Fish Wildl. Serv., Fish. Bull. 61: 53-60. Hartman, W. L., W. R. Heard, and B. Drucker. 1967. Migratory behavior of sockeye salmon fry and smolts. J. Fish. Res. Board Can. 24: 2069- 2099. Helle, J. H. 1970. Biological characteristics of intertidal and fresh-water spawning pink salmon at Olsen Creek, Prince William Sound, Alaska, 1962-63. U.S. Fish Wildl. Serv., Spec. Sci. Rep. Fish. 602, iv + 19 p. Hoar, W. S. 1958. The evolution of migratory behaviour among juvenile salmon of the genus Oncorhynchus. J. Fish. Res. Board Can. 15: 391-428. Hoar, W. S., M. H. A. Keenleyside, and R. G. Goodall. 1957. Reactions of juvenile Pacific salmon to light. J. Fish. Res. Board Can. 14: 815-830. Hunter, J. G. 1948. Natural propagation of salmon in the central coastal area of British Columbia. Fish. Res. Board Can., Prog. Rep. Pac. Coast Stn. 77: 105- 106. Israel, H. R. 1933. On the life history of the silver salmon On- corhynchus kisutch (Walbaum) of Chignik River, Alaska. M.S. Thesis, Stanford Univ., Palo Alto, Calif., 58 p. KosKi, K. V. 1966. The survival of coho salmon (Oncorhynchus kisutch) from egg deposition to emergence in three Oregon coastal streams. M.S. Thesis, Oreg. State Univ., Corvallis, 84 p. 93 FISHERY BULLETIN: VOL. 70, NO. I Mark, J. C. 1943. Age, length, and weight studies of three species of Columbia River salmon (Oncorhynchus keta, O. gorbuscha, and O. kisutch). Stanford Ichthyol. Bull. 2: 157-197. Nea\'E, F. 1948. Fecundity and mortality in Pacific salmon. Trans. R. Soc. Can., Ser. 3, Sect. 5, 42: 97-105. Nelson, P. R. 1959. Effects of fertilizing Bare Lake, Alaska, on growth and production of red salmon (0. nerka). U.S. Fish Wildl. Serv., Fish. Bull. 60: 59-86. PRITCHARD, A. L. 1940. Studies on the age of the coho salmon (O71- corhynchus kisutch) and the spring salmon (Oncorhynchus tshuwytscha) in British Columbia. Trans. R. Soc. Can., Ser. 3, Sect. 5, 34: 99-120. RiCKER, W. E. 1962. Comparison of ocean growth and mortality of sockeye salmon during their last two years. J. Fish. Res. Board Can. 19: 531-560. ROUNSEFELL, G. A. 1957. Fecundity of North American Salmonidae. U.S. Fish Wildl. Ser\'., Fish. Bull. 57: 451-468. Salo, E. 0. 1955. Silver salmon, Oncorhynchus kisutch, sur- vival studies at M inter Creek, Washington. Ph.D. Thesis, Univ. Wash., Seattle, 183 p. Salo, E. 0., and W. H. Bayliff. 1958. Artificial and natural production of silver salmon {Oncorhynchus kisutch) at Minter Creek, Washington. Wash. Dep. Fish., Res. Bull. 4, 76 p. Semko, R. S. 1954. Zapasy zapadnokamchatskikh lososei i ikh promyslovoe ispolzovanie. (The stocks of west Kamchatka salmon and their commercial utili- zation.) Izv. Tikhookean. Nauchn.-issled. Inst. Rybn. Khoz. Okeanogr. 41: 3-109. (Fish Res. Board Can., Transl. Ser. 30.) Shapovalov, L., and A. C. Taft. 1954. The life histories of the steelhead rainbow trout (Salmo gairdneri gairdneri) and silver salmon (Oncorhynchus kisiitch) with special ref- erence to Waddell Creek, California, and rec- ommendations regarding their management. Cal- if. Fish Game, Fish. Bull. 98, 375 p. Smirnov, a. I. 1960. K. kharakteristike biologii razmnozheniia i razuitiia kizhucha — Oncorhynchus kisutch (Wal- baum). (The characteristics of the biology of reproduction and development of the coho On- corhynchus kisutch (Walbaum).) Vestn. Mosk. Univ., Ser. VI Biol. Pochvoved., p. 9-19. (Fish. Res. Board Can., Transl. Ser. 287) Smoker, W. A. 1953. Stream flow and silver salmon production in western Washington. Wash. Dep. Fish., Fish. Res. Pap. 1: 5-12. Taft, A. C. 1934. California steelhead experiments. Trans. Am. Fish. Soc. 64: 248-251. WiCKETT, W. P. 1951. The coho salmon population of Nile Creek. Fish. Res. Board Can., Prog. Rep. Pac. Coast Stn. 89: 88-89. 94 DEVELOPMENTAL RATES AT VARIOUS TEMPERATURES OF EMBRYOS OF THE NORTHERN LOBSTER (Homarus americanus MILNE-EDWARDS) Herbert C. Perkins' ABSTRACT The rates of development, time from extrusion to hatching at various temperatures, and differential de- velopmental rates at the same temperature of lobster embryos are presented. The eyes of the embryos were measured to monitor the rates and degree of embryo development. Herrick (1890, 1896) discussed developmental rates for lobster embryos in the early stages at 20° to 22° C. Templeman (1940) determined the times required at various temperatures for lobster eggs to reach the 16-cell stage, and up to the formation of eye pigment. The information from these studies is valuable for determining the rates of early development in lobster egg- embryos but is not adequate for the accurate assessment of developing embryos once eye pig- ment has been formed. By monitoring the rate of development of lobster embryos throughout the embryonic period at various temperatures one can predict hatching times of larvae and con- trol hatching times by manipulating the water temperature in tanks holding egg-bearing fe- males, so that larvae can be available over a wide period of time for use in experiments. This paper presents the rates of development and time required to complete the embryonic period by lobster embryos at various temperatures and a method of continually monitoring that develop- ment. The work was conducted at the National Marine Fisheries Service, Biological Laboratory, Boothbay Harbor, Maine, as part of the Labora- tory's investigation of the early life history of the lobster. METHODS AND MATERIALS Most of the egg-bearing lobsters used in this study came from the offshore canyons of the ^ National Marine Fisheries Service, Northeast ' Fish- eries Center, Boothbay Harbor Laboratory, W. Boothbay Harbor, ME 04575. continental shelf, south and east of New England. A few came from the Boothbay Harbor area and are so noted. All egg-bearing females were kept in tanks at seasonal water temperatures or in water warmed to various constant temperatures. Water from the laboratory's seawater system was piped to the heated tanks at rates consistent with the capacity of the heaters. Salinity aver- aged 31'/( and ranged from 29 to Z2%c throughout the study period. Five egg-bearing females were kept in a tank through which natural seawater at seasonal tem- perature was circulated during the development- al period of their eggs. The purpose of holding these females at seasonal temperatures was to determine the rates of development of their em- bryos in a natural temperature regime. Four- teen female lobsters from the offshore canyons, with recently extruded eggs (eggs in prenaupliar condition), were kept at constant temperatures from 6.9° to 24.6° C. The primary objective at the higher temperatures (20°-24.6° C) was to force the eggs to hatch before the time they would do so at seasonal temperatures. Of fur- ther interest was the rate of development of the embryos at constant, rather than fluctuating, temperatures, and the time required for the eggs to hatch at these temperatures from a given point in their development. The rates and extent of development of the embryos were determined by measuring the size of their eyes. Measurements were made to the nearest micron with an ocular micrometer in a dissecting microscope at a magnification of 50 x . When measuring an eye, I took its greatest width and greatest length, combined these figures and Manuscript accepted July, 1971. FISHERY BULLETIN: VOL. 70, NO. 1, 1972. 95 FISHERY BULLETIN: VOL. 70, NO. 1 divided by two. The resulting figure was used as an index of development only and was not meant to represent the actual increase or growth of the eye. Samples of 15 to 20 eggs were re- moved from the periphery of an egg mass when- ever a measurement was desired (usually once a week) . From these samples five eggs were se- lected randomly, and one eye of each embryo measured. A mean of these five measurements was used as a working figure or index. Eggs were taken from the peripheral layer of the mass as these are the furthest advanced in develop- ment and are the first to hatch. Variation in the eye measurements of eggs from this layer ranges from zero to o^C When the eyes first appear and are large enough to measure, they are but thin crescents, darkly pigmented and surrounded by a halo of lighter material. The dark crescents only were measured. The eyes are very distinct for most of the developmental period and are easily measured, the crescents gradually becoming tear-drop in shape. The in- dex of the eye is about 70/x when it is first mea- surable; the index is about 560/i at hatching. All eyes were measured after the eggs had been preserved in a 5/f solution of Formalin in sea- water. Preservation in Formalin caused sig- nificant swelling in the eggs themselves but had no determinable effect on the size of the eyes. year, water temperatures were no higher than 6° C. Embryonic development during this pe- riod ceased in some of the egg masses and was barely discernible in others, at least by the method of eye index measurement. Squires (1970) , using the amount of yolk material in the eggs as a criterion, reported a standstill in devel- opment during the winter in embryos of New- foundland lobsters. tfi 600 I 500 y 400 Z 500 S 200 o 100 Eye Pigment Stofis 600 500 400 xa 200 100 JUL AUG SEP OCT NOV DEC JAN FEB MAR APR MAY JUN JUL AUG Figure 1. — Trends of development of the embryos of five female lobsters held in water of seasonal temperature at the Boothbay Harbor Laboratory and the temperature cycle during the period. The lines showing develop- mental trends are derived from plots of periodic measure- ments of the embryos' eye indices. The lines are num- bered in accordance with the age of the egg mass; the lower the number, the older the egg mass. RESULTS AND DISCUSSION TRENDS OF EMBRYONIC DEVELOPMENT IN WATER OF SEASONAL TEMPERATURES The developmental patterns, from onset of eye pigment to hatching, of the embryos of the five females held at seasonal temperatures are shown in Figure 1, as is the cycle of water temperature for the same period. The trends of embryonic development were plotted by using periodic (usu- ally weekly) measurements of the eye index of each egg mass. These five lobsters had extruded their eggs in the laboratory tanks so the age of each egg mass was known. From the latter part of November to the first of May of the following DIFFERENTIAL RATES OF DEVELOPMENT The developmental rates of lobster embryos appear to be governed not only by their thermal environment, but by the age or extent of de- velopment at which they are subjected to that environment. During the experiments I con- ducted the older or more advanced embryos de- veloped at slower rates than those less advanced, though all were maintained in the same tank. The oldest egg mass of the five females held under seasonal conditions was extruded 7 weeks before the youngest, but the total time for development of the younger egg mass was 414 weeks less than the older; the younger embryos had developed considerably faster. Measurements of the eye index of embryos in all five egg masses were made for the first time on November 7. In Fig- 96 PERKINS: DEVELOPMENTAL RATES OF NORTHERN LOBSTER EMBRYOS LlI $ a: UJ 0. z o a: <_) X UJ Q LlI O UJ to < UJ a: o 60 50 40 30 20 10 - (r=.952,P<.OI) _L. J_ _1_ (r= .995, P <.0I) _L J 50 100 150 200 250 300 350 400 450 500 EYE INDEX AT START (MICRONS) Figure 2. — Line A represents the different rates of in- crease of the eye index, in microns per week, of the em- bryos of the five lobsters held in the same tank, under seasonal conditions, from November 7, until hatching. Line B represents the different rates of eye increase of the embryos of seven females held at a constant tem- perature of 22.6° C. Size of eye index at the starting time is plotted against the corresponding rate of increase of the eye index up to the time of hatching. Table 1. — Lobster number, carapace length of female, age of eggs 10 January, increase of eye index of embryos from 10 January to 26 March, and the total develop- mental time for the embryos of the five female lobsters held under seasonal water conditions at the Boothbay Harbor Laboratory. Lobster Carapace number 'f"9\'i (mm) Area of capture Age of eggs Increase of Total 10 January , eye index weeks (weeks) 7J";''°"'/.T/^' u /?• ' 10 Jan. -26 Mar. hotchmg 1 97 Boothbay Harbor 29 0.00 51.4 2 94 Boothbay Harbor 26 0.46 50.6 3 147 Veatch Canyon 24 0.65 50.0 4 124 Hudson Canyon 23 1.49 49.4 5 94 Boothbay Harbor 22 2.52 47.0 surable increase in development during this time, whereas some development was noted in the em- bryos of the other egg masses. In fact, the em- bryos in the oldest egg mass showed no measur- able increase from the second week of December to the middle of the following April. The num- ber of weeks during the winter in which no de- velopment could be measured, for each egg mass, was as follows (as in Table 1, lobsters are num- bered according to the age of their egg mass) : weeks 'Number of "dormant' Lobster number during winter 1 18 2 14 3 8 4 6 5 ure 2(A) the rate of eye increase (microns per week) of the embryos in each egg mass for the remainder of the developmental period is plotted against the corresponding eye index taken on November 7. The same result is obtained if the age of an egg mass is substituted for its eye index on the abscissa. Rates of increase were calculated by dividing the total increase of the eye index by the total number of weeks the em- bryos took to complete development after No- vember 7. The increase in eye index (microns per week) of embryos in each egg mass, for the period January 10 to March 26, is presented in Table 1, During this time the water temper- ature ranged between 0.1° and 1.5° C; the mean was 1.0°. This was the coldest 10-week interval of the developmental period. The embryos in the oldest egg mass exhibited no noticeable or mea- The rates of increase (microns per week) in eye index for the embryos of seven lobsters held at a constant temperature of 22.6° C are indicated in Figure 2(B). These females came from offshore canyons of the continental shelf, off New England, The ages of these egg masses were not known, but the eye index of the embryos in each was measured before the females were placed in the warm water. The increase in eye index of the embryos in each egg mass was mon- itored weekly until hatching. Times and rates pertain only to the time spent at 22.6° C. Al- though one might expect that in a given time in- terval, at the same temperatures, younger em- bryos would develop faster than older ones, it might also be expected that the younger embryos would assume the slower growth rate of the older when they eventually reached the same age. Of 97 FISHERY BULLETIN: VOL. 70, NO. 1 the five females held in water of seasonal tem- perature, the youngest eggs were at lower tem- peratures than the older at the same age, making assessment of differential developmental rate difficult. However, of the seven egg masses held at 22.6° C the embryos of least development at the start developed faster at comparable levels of development than the more advanced embryos. The differential rate of development of lobster embryos, at the same temperature, seems to im- ply that in a given population where extrusion of eggs may be somewhat staggered in time among the females, hatching of the eggs would occur during a more limited period, providing the egg-bearing females occupied the same ther- mal environment. SOr 70- X u t- X 60 50 40 » 30 20 - 10 • = OBSERVED VALUES O « CALCULATED VALUES (•) 10 12 14 16 le 20 22 TEMPERATURE OF WATER (°C ) 24 26 Figure 3. — The number of weeks for lobster eggs to complete the embryonic period at various temperatures. Line A represents the time required from onset of eye pigment in the embryos; line B represents the time re- quired from extrusion to hatching. Points in paren- theses indicate times required at the mean temperature of a fluctuating thermal environment. RATES OF DEVELOPMENT AT VARIOUS TEMPERATURES The times required for the embryos to hatch at various temperatures are shown in Figure 3. Line A represents the time required for the em- bryos to hatch after the formation of eye pig- ment; line B represents the time required from extrusion to hatching. Most of the points in each line indicate the time required to complete de- velopment at constant temperatures. A few (points in parentheses), representing the time required for total development at the mean tem- perature of a fluctuating thermal environment, have been included as well. For example, the average time required for total development of the eggs of the five females held under seasonal conditions was 49.7 weeks. The mean water temperature during the period was 8.1° C. These values are virtually the same as would be expect- ed if the water was held constantly at 8° C. All values showing time from onset of eye pig- ment to hatching were observed. The times re- quired from extrusion to hatching at five tem- peratures were also observed. To find the unknown time required from extrusion to hatch- ing, at other temperatures, I used the following equation: At = Az Xi X2 where Ai was the observed time from onset of eye pigment to hatching at 20° C; A2 was the observed time from extrusion to hatching at 20° C; Xi was the observed time required from Table 2. — Number of weeks required from extrusion to onset of eye pigment, onset of eye pigment to hatching, and to hatching at certain temperatures, at salinities near 31%c. Weeks required from Water temperature C C) Extrusion to onset of eye pigment Onset of eye pigment to hatching Extrusion to hatching 5 10 15 20 25 40 9 S A 3 120 30 18 12 9 160 39 23 16 12 98 PERKINS: DEVELOPMENTAL RATES OF NORTHERN LOBSTER EMBRYOS onset of eye pigment to hatching at a given tem- perature; and A'2 is the unknown time required from extrusion to hatching at the same temper- ature as Xi. Templeman (1940, p. 74) used a similar method to find unknown developmental rates. The requisite times for development of lobster embryos at certain temperatures are summarized in Table 2. The relationship be- tween water temperature and the average in- crease in eye index of lobster embryos, in mi- crons per week, is linear at temperatures be- tween 5° and 25° C. The index of the embryonic eye must increase to approximately 560^t at hatching. If eggs are encountered with eyed embryos, their eye index may be subtracted from 560 and the difference divided by the value calculated from the following equation: Y = —8.3151 + 2.6019 (X) where Y is the increase of the eye index in mi- crons per week, and X is the developmental tem- perature. The resulting quotient is the average number of remaining weeks required for the em- bryos to hatch, depending on genetic variation and the differential rate of development noted earlier. SUMMARY 1. Once eye pigment has been formed, the course and rate of development of lobster em- bryos may be monitored by the periodic mea- suring of the eye of the embryos. 2. Lobster embryos develop differentially, under the same thermal conditions, depending on their age or extent of development when they are subjected to a given thermal environment. 3. As water temperature has a direct effect on the developmental rate of lobster embryos, that rate may be manipulated by adjusting the water temperature of holding tanks to insure periodic hatches of larvae throughout the year. LITERATURE CITED Herrick, F. H. 1890. The development of the American lobster, Homarus americarms. Johns Hopkins Univ. Circ. 9: 67-68. 1896. The American lobster : A study of its habits and development. Bull. U.S. Fish Comm. 15: 1-252. Squires, H. J. 1970. Lobster (Homarus americanus) fishery and ecology in Port au Port Bay, Newfoundland, 1960- 65. Proc. Natl. Shellfish. Assoc. 60: 22-39. Templeman, W. 1940. Embryonic developmental rates and egg-lay- ing of Canadian lobsters. J. Fish. Res. Board Can. 5: 71-83. 99 PRELIMINARY STUDIES OF SELECTED ENVIRONMENTAL AND NUTRITIONAL REQUIREMENTS FOR THE CULTURE OF PENAEID SHRIMP' Lowell V. Sick, James W. Andrews, and David B. White" ABSTRACT Types of substrate, type of aeration, and stocking density were compared as prerequisities for high-den- sity culture studies with penaeid shrimps. Neither sand-shell substrate nor brick subdivisions of cul- ture tank bottoms produced significantly higher survival rates than bare fiber glass tanks. Forced air supplied via airstones proved to be a more suitable form of aeration than did physical agitation of the water column in culture tanks by high-pressure nozzles. Survival rates of 80 to 90% were achieved when biomass densities did not exceed 40 g/m^. Semipurified pelleted diets (i.e., containing defined chemical ingredients plus one or more natural products) having a complement of nutrients including minerals and vitamins, various ratios of shrimp to fish meal, protein hydrolysates, and such diets fed at three percentages of total biomass daily were compared for their ability to produce increases in growth. Diets without fish or shrimp meal sustained biomass while those diets having the highest proportion of shrimp to fish meal in addition to added vitamins produced over 60% increase in total biomass over a 3-month period. Animals fed a combina- tion of yeast, soy, and casein hydrolysates increased 39% in biomass over the same period of time while those fed each of the above hydrolysates during the 3-month period separately showed only an average of 18% increase in weight. Feeding shrimp with a fish-shrimp base with added vitamins at a rate of 15% daily of the total biomass produced a 164% increase in weight with 95 to 100% survival during the 3-month period. Using semipurified pelleted diets, a food conversion ratio of 5.5 was obtained. Establishing selected preliminary environmental and nutritional requirements for penaeid shrimp re- sulted in the successful and reproducible production of major biomass increases with relatively high sur- vival rates and low food conversion ratios. The harvest of commercial shrimp suffers great seasonal variability and has failed to keep pace with ever-increasing domestic and export de- mands (Surdi and Whitaker, 1971). In order to supplement the natural harvest and provide a year-round supply of shrimp, several attempts have been made to culture shrimp in natural ponds, restricted portions of bays and estuaries, and laboratory tanks. In general, these efforts have had limited success and have explicitly il- lustrated the need for more accurately defining the nutritional and environmental requirements ^ This work is a result of research sponsored by NOAA Office of Sea Grant, U.S. Department of Commerce, under Grant #GH-73. ^ Skidaway Institute of Oceanography, 55 West Bluff Road, Savannah, GA 31406. necessary for culturing these species. Although pond culture has produced annual crops of shrimps (Villadolid and Villaluz, 1951; Lunz, 1967; Wheeler, 1967; Broom, 1969; Moore and Elan, 1970') , production has been minimal and highly variable. Attempts to obtain commercial quantities of shrimp by stocking enclosed por- tions of estuaries have to date not yielded pro- duction results (American Fish Farmer & World Aquaculture News, 1970) . During recent labora- tory studies, Subrahmanyam and Oppenheimer (1969) were able to maintain shrimp in labora- tory tanks using a pelleted diet consisting of fish meal, stickwater, and vitamins. However, the Manuscript accepted July 1971. FISHERY BULLETIN: VOL. 70, NO. 1, 1972. " Moore, W. R., and L. L. Elan. 1970. Salt water pond research. Tex. Parks Wildl. Dep., Austin. (Processed.) 101 FISHERY BULLETIN: VOL. 70. NO. 1 total biomass increase of shrimp fed their best diet for 6 weeks was only approximately 50% higher than initial biomass. Better results (on an individual weight basis) using Penaeus duo- rarum were obtained from animals grown on a sand substrate than those grown in bare tanks. The greatest promise for economical shrimp production lies in determining the exact nutri- tional requirements and developing an inexpen- sive artificial diet from feedstuffs for these spe- cies. Current commercial practices in Japan employ chopped clam (predominantly Tapes semidecussata, Reeve) as a diet for rearing shrimp. Despite the high market price for cul- tured shrimp in Japan (the retail price of cul- tured shrimp ranged from $4 to $10 per pound in 1970), shrimp farming there tends to be a marginal enterprise because of the high cost of a clam diet. However, in other parts of the world where shrimp does not command such a luxury price, the use of a high-value product such as clam for shrimp feed is prohibited. Pelleted diets (i.e., pellets containing all the chemical ingredients thought to be important to animal growth) have been designed consisting of purified soybean meal, glucose, sucrose, starch, glucosamine, chitin, cellulose, soybean oil, citric acid, succinic acid, amino acid, minerals, vita- mins, and cholesterol (Kanazawa et al., 1970). After growing penaeids on such diets, the ani- mals were in excellent physiological condition, but in the best group, total biomass increase was only 72% of the control group fed chopped clam. Thus, little progress has been made toward estab- lishing nutritional and environmental require- ments that will yield optimum growth (total biomass increase) and survival of penaeid shrimp. In the present study, an attempt was made to develop a suitable experimental culture system which could serve as a model for future nutri- tional and environmental studies. Several en- vironmental factors were examined, and, as a result, environmental conditions were created which would allow acceptable growth and sur- vival. Having first established suitable culture conditions, several diets were evaluated in pre- liminary studies of the nutritional requirements of shrimp. MATERIALS AND METHODS Both environmental and nutritional studies were conducted in round fiber glass culture tanks measuring approximately 1 m deep by 1 m in diameter and equipped with a venturi type cen- ter drain which maintained a water depth of 0.75 m. Three replicates were maintained for all treatments. Water (ranging in salinity from 26.8 to 29.3^r) from the Skidaway River was filtered through an oyster shell and sand filter to re- move major food particles. Filtered water was heated to 30° C in a stainless steel heat exchang- er and jetted into each tank at a rate of 1.9 liters/min through flow-control nozzles which were aimed so that the agitation of the water column in each tank was minimal. Temper- ature ranged from 25° to 28°C in each tank throughout the experimental period. White shrimp (P. setiferus) obtained from the Savannah, Ga., river and tributary systems were used in all environmental experiments, and brown shrimp (P. aztecus) obtained from the Tampa Bay, Fla., area were used in the nutri- tional studies. Shrimp weighing 4 db 0.8 g (mean and standard deviation based on 480 weighed shrimp) were selected from the above stock and used in all environmental studies (10 animals per tank) and fed pelleted diets (Table 1, Diet 1) at a rate of 5% of their biomass daily, on a dry weight basis. Shrimp were weighed each week and the percent increase or loss re- corded on a wet weight basis. SUBSTRATE STUDY Sand-shell substrates suitable for burrowing, subdivisions of tank bottoms, and bare fiber glass tank bottoms were provided for replicate groups of shrimp, and relative survival rates among the treatment groups were compared over a 5-week period. Sand-shell substrates were placed di- rectly onto the tank bottom in one group, and in another group, the same substrate was placed on a perforated platform 10 cm above the tank bot- tom, allowing a flow of water through the drain below the sand surface. Such an arrangement was designed to test the eflfect of decreasing the 102 SICK, ANDREWS, and WHITE: REQUIREMENTS FOR SHRIMP CULTURE S U p4 0) a 3 '•B O 3 6 -t-> CO »-H OS C _o '+3 •rH U -*-> ;3 C >H s o «M s 10 o "S b tn -£ -M "s a 03 O (D ^ M O ■o c cS CO ,2 3 OS •E cS > ,__, c3 -t^ C o 1 p lO I O lO I I I I p CO r O ' CO I lo to ' CO 00 CM o p lO I CN CO CN to 1 o >o ■"I- CNl 1 CO CN 1 1 p lO * 1 CO CN 1 1 1 r 1 1 1 1 >0 1 I 1 1 1 1 1 1 1 1 1 1 1 1 p 1 1 1 1 1 lO 1 1 1 1 1 1 o t 1 1 1 1 1 1 1 1 1 1 1 1 O lO 1 oS ' ' ' CM ' ' ' CO ' CO CM CM >o ' ' ' CJ CM ' to 1 1 1 rill 1 >o ■o 1 1 O >0 1 CO ' ' ' cs 1 1 1 1 CO ^ ' ' -^^ ci ' O I 1 1 1 1 1 o 1 >o o 1 1 O vo 1 CO ' ' ' ' ' ' 00 MD CO ' ' Tf CM ' to 1 1 1 1 1 T P 1 lO >o 1 1 O lO 1 00 ' ' ' CM 1 1 ' CC) ' oo CN od CM ' ' ^ CM ' 1 1 1 1 CN O 1 1 I 1 1 1 1 1 1 p 1 1 1 1 1 1 1 1 lO I I 1 (III 1 1 1 1 1 1 1 p 1 1 1 1 UO 00 o ^ 1 1 1 1 CO d > ' ' 1 1 1 Tf CN CO to CM CN O ^_ CN CO CO p lO 1 1 1 1 t 1 o >o O- C3 — d 6 O o CO 6 ' ' ' '''■*■ CM CN o CO 6 ' ' ' 111^ CN IX 2 S i 3 5 "^ E ^ E 2 — o p o p ^ i) o o -D >- *(U ^ if "^ *= c o^ ■- o^ O .± — p — a -^ c a> O (U c "^ D> p ^ .5 a. "' — n — USO«Uco5lEu^<^VUooU>U o E o O E 5 >^ 9 o- o co o- 00 o 6 CO u CO o s? CO CO cs o c o- o o X o o "O o> lO — o o D E o „ 2 s? *- lA O^ ( 5° ^.ro? 05 c £ cmB^ >2 ai — ^ O E'^ o o E o^ O u ^ CO 0) ■^ O w 0>>CN 5. . o 103 FISHERY BULLETIN: VOL. 70, NO. 1 buildup of anaerobic conditions. Division of tank bottoms into sundry tunnels and levels was created by specific placement of bricks and clay drain tiles. AERATION STUDY Aeration provided by jetting streams of fil- tered seawater into respective tanks was com- pared to aeration supplied by bubbling air through airstones into tanks in which water was continuously added with no agitation of the water column for an 8-week period. Two airstones were placed in each tank and valve-regulated air lines controlled the pressure at approximately 4 psi. Oxygen levels were monitored periodically and used along with survival rates as a basis for evaluation of replicate groups aerated by each method. STOCKING DENSITY STUDY Survival data were compared among triplicate tanks stocked at 10, 20, and 40 shrimp per m^ for an 8-week period. These densities of ap- proximately 40, 80, and 160 g/m- were chosen on the basis of data provided in pond and lab- oratory culture of penaeid shrimp (Broom, 1969; Subrahmanyam and Oppenheimer, 1969). PRELIMINARY NUTRITIONAL STUDY Triplicate groups of ten 4 g brown shrimp (P. aztecus) were fed a series of pelleted diets. Growth data (biomass increase) was used as a means of evaluation. Diets examined consisted of those patterned after Japanese purified diets (i.e., diets containing only chemical ingredients) (Table 1, Group I) (Diet 1 was conducted for 5 weeks and Diets 2, 3, and 4 for 11 weeks each) ; a second group of semipurified diets (i.e., con- taining defined chemical ingredients but contain- ing one or more natural products) providing four combinations of levels of protein, fat, shrimp, and fish meal (Group II) (conducted for 11 weeks) ; and a third group designed to compare the nutritional value of casein, yeast, and soy hydroly.sates (Group III) (conducted for 6 weeks). All of these groups were fed at 5% Table 2. — Percent of pellet dissolved over time and at at three concentrations of binder. (Values are means and standard deviation on two replicates with Diet 1.) Percent b inder added Hours (collagen) 6 12 24 1 3 5 10 13 ± 1.2 11 ±0.8 10 ±0.6 14 ±0.9 10 ± 0.6 10± 1.1 18 ± 1.7 10 ±0.6 10 ± 1.0 of their respective biomass daily. In addition, Diet 6 was fed at 5, 10, and 15% of biomass (Group IV) (conducted for 6 weeks). Combined environmental factors which pro- duced best survival in each of the environmental experiments (i.e., culture conditions consisting of bare fiber glass tank bottoms, supplied aera- tion, and a stocking density of approximately 40 g/m-) were used in all nutritional studies. This combination oflfered a maximum potential for an increase in biomass and therefore allowed accurate evaluation of diflferences among diets tested. Although survival in bare fiber glass tanks was not significantly different from sand substrates, the fact that bare tanks were simpler to maintain dictated that they be used for the nutritional studies. Prior to starting nutritional studies, the phys- ical properties of pelleted diets were evaluated for acceptability as shrimp food. Pellet consis- tency was determined according to its ability to resist dissolution over a given period of time, and texture and size were chosen according to animal performance when presented several choices. Collagen' proved to be a suitable bind- ing agent. Using an experimental design with time and collagen levels as variables, a pellet with 5% collagen added as a binder was found to offer optimum consistency over a 24-hr im- mersion in salt water (Table 2). Percent dis- solution was measured by taking dry weights after 6, 12, and 24 hr of immersion (no shatter- ing of pellets was observed, and all loss of weight was therefore assumed to be from dissolution). Animals were observed to feed most readily on * Supplied on an experimental basis by the Hides and Leather Division of the U.S. Department of Agriculture Eastern Utilization Laboratory in Philadelphia, Pa. 104 SICK, ANDREWS, and WHITE: REQUIREMENTS FOR SHRIMP CULTURE pellets 0.3 cm in diameter by approximately 1.5 cm in length and which sink in water, and hence, pellets having these characteristics were used in both environmental and nutritional experiments. RESULTS AND DISCUSSION SUBSTRATE STUDY A survival rate of 80% was obtained after 5 weeks in tanks without substrate, 80 to 90% survival was maintained over much of the dur- ation of the experiment among both treatments having sand-shell substrates, and less than 60% survival occurred among tanks having brick sub- divisions (Figure 1). Although P. setiferus is reported to burrow less than either P. duorarum or P. aztecus (Anderson, 1966; Perez Farfante, 1969) , it apparently was able to avoid predation, especially during the highly vulnerable moulting period, quite successfully with or without a sand substrate, since 5-week survival data among the two sand-shell treatments and the bare tank bot- tom treatment were not significantly different (P < 0.05) (Duncan, 1955). If the type of shelter is a factor in increased survival for penaeids maintained under culture conditions, the brick subdivisions should have enhanced sur- vival. However, the markedly high mortality 100 80 c 60 40 20 SiJND-SHELL SUBSTRATE ON PLATFORM SAND-SHELL ON TANK BOTTOM NO SUBSTRATE BRICK SUBSTRATE Figure 1. — Mean and standard error for percentage of animals surviving after 5 weeks of growth on four dif- ferent substrates. rate among this group, significantly different from the other three treatments (P < 0.05), may have resulted from either failure of the shrimp to behaviorally segregate and thus fully utilize this protection or from physical abrasion against the sharp and coarse brick surface. Al- though there may have been toxic substances in the brick materials, the bricks were carefully washed and assumed to be otherwise inert in any chemical effect they may have had on the animals. Although differences in volume of water caused by placing various substrates in their respective treatments was not controlled for, it was felt that these differences in a running water system were not critical to the survival of shrimp. Dif- ferences in bottom area among the treatments caused by placement of different types of sub- strate were neither controlled for nor measured but were also thought to be negligible compared to differences found among treatment groups. The high degree of cannibalism noted by Subrah- manyam and Oppenheimer (1969) in tanks with- out substrate was not observed in any groups. AERATION STUDY The group having oxygen supplied by injecting air through airstones had significantly higher survival rates (P < 0.05) when compared with a treatment aerated by agitation of the water column (Figure 2) . Although the average oxy- gen levels were similar between the two treat- ments (3.4-6.8 ppm), such levels in tanks aer- ated by high-pressure nozzles often dropped for short intervals due to clogging of the nozzles with silt and biological debris. Electrical power failures which affected water flow but not the compressed air supply (equipped with stand-by DC power) also caused intermittent drops in oxygen levels. Such short-term irregularities may have been more critical to shrimp toler- ances than is indicated from reference to average oxygen level values, per se. Also, at the rela- tively high temperatures maintained throughout the study, short drops in oxygen levels could have been very critical. Decreased survival in tanks with agitation of the 0.75-m water column may also have resulted from physical agitation of the animals. lor FISHERY BULLETIN: VOL. 70, NO. 1 100 Figure 2. — Mean and standard error for percentage of animals surviving after 8 weeks of growth with two types of aeration. STOCKING DENSITY STUDY Stocking densities higher than 40 g/m^ pro- duced proportionally higher mortalities indi- cating an appi'oximate carrying capacity for this particular culture system (Figure 3). If shrimp were stocked at 40 g/m-, a population of 32 g/m^ remained after 8 weeks. Similarly, when shrimp 100 80 a 60 3 40 20 '^^■a — -A ^ >" L ^^7 T K^^^^ ^^^ '' 'H 1 o— 80 •— • 160 1111 4 WEEKS Figure 3. — Mean and standard error for percentage of animals surviving after 8 weeks of growth at three stocking densities. were stocked at 80 g/m^, a relatively stable pop- ulation of approximately 52 g/m^ existed during the final 2 weeks of study. Similarly, mortality during the first 8 weeks among a population orig- inally stocked at 160 g/m^ created a population of approximately 80 g/m^, but in this case the survival rate was still declining after 8 weeks of growth. Therefore, a carrying capacity (max- imum biomass obtainable) for this culture sys- tem may have been somewhere between 32 and 80 g/m2. Although such a carrying capacity would de- pend on the particular culture system, applicable calculations, utilizing data from a laboratory culture study (Subrahmanyam and Oppenhei- mer, 1969) and a pond culture experiment (Broom, 1969), indicate a similar carrying capacity for populations of other systems (ponds, embayments, and laboratory tanks). In the case of the laboratory study, best survival was obtained when shrimp were stocked initially at 34 g/m^, yielding a biomass of 27 g/m^ at the termination of the experiment. Likewise, best survival and an increase in biomass occurred in the pond culture study when initial stocking den- sities were below 20 to 30 g/m^. Recent data from a commercial operator in Central America indicates that, regardless of stocked biomass, the carrying capacity ranged from 5.5 to 7.3 g/m^ (Smitherman and Moss, 1970). Such evidence suggested that final production expectations should be considered in choosing initial stocking densities. PRELIMINARY NUTRITIONAL EXPERIMENTS A comparison of several groups of diets (Table 1) revealed that semipurified diets with casein as the major source of protein (Group I), only produced an average of 18% increase in biomass above stocked biomass levels. Group II, having fish and shrimp meal as additional sources of protein, produced approximately 63% growth on the best diet. Group III diets comparing hydro- lyzed proteins yielded only 39% growth on the best diet, and animals fed at a rate of 15% of their total biomass (Group IV) increased their initial biomass 164%. In addition to increased 106 SICK, ANDREWS, and WHITE: REQUIREMENTS FOR SHRIMP CULTURE growth above that obtained in the environmental studies, survival was increased from 80 to 90% to 90 to 100% through the information acquired from the above nutritional comparisons. Shrimp obtained little if any sustenance from organic or settled detritus since the continuous flow of filtered seawater through the tanks kept the system relatively free of siltation and ex- traneous growth. Furthermore, starved animals were not able to sustain their initial biomass level beyond 2 weeks, and the populations in all three replicate tanks had died after 7 weeks. Cannibalism appeared to be prevalent among starved organisms, and the decline in weight was undoubtedly moderated due to growth of animals preying upon dead shrimp. Diets in Group I with casein as the major pro- tein source produced little growth above susten- ance. Diet 1 with an added mineral mix yielded a significantly higher biomass increase above ini- tial weights at the 5% feeding level than either Diets 2 or 3 which lacked the mix. In addition to the mineral mix. Diet 3 lacked sodium gluta- mate, glycine, citric acid, and succinic acid and correspondingly caused a loss in biomass over the 3-month growth period. Although the above results showed Diet 1 to be significantly different (P < 0.05) from the other diets after the first month of study, results are somewhat confounded with initial differences in stocked biomasses and poor response in general. These sustenance biomass levels represented far less growth increase than the 72% of control obtained by Kanazawa et al. (1970) and may be due to a lack of cholesterol in our study. Since many crustaceans are not able to synthesize cholesterol (Van Den Oord, 1964; Dall, 1965; Zandee, 1967), including recent evidence for shrimp (Kanazawa et al., 1971) , the lack of this entity undoubtedly was attributed to the poor performance of these diets. Shrimp fed on Group II diets averaged a 37 to 63% increase in growth. Although results from Diets 5, 6, and 7 were not significantly dif- ferent (P < 0.05), Diet 6, which consisted of a high ratio of shrimp to fish meal and a low level of casein, yielded greatest biomass increases. Total biomass decreased in diets having a de- crease in percentage of shrimp meal. Growth from Diet 8, which contained blended shrimp muscle and lower levels of shrimp and fish meal, was statistically less (P < 0.05) than the other three diets. Again, the control group of starved shrimp was not able to sustain its initial weight and declined in biomass after the first 2 weeks. Group III, consisting of yeast, casein, and soy protein hydrolysate diets, produced an average biomass increase of 18 to 39% (Table 1). The combination of diets containing casein, soy, and yeast hydrolysates produced significantly better (P < 0.05) growth than individual hydrolysates. Since results from this group were not better than results after 6 weeks from Diet 5 which was similar except it contained only intact pro- tein, these data indicate that hydrolyzed proteins are not utilized more efficiently than intact pro- teins. Comparing food supplied at 0, 5, 10, and 15% of total biomass using Diet 6 illustrated that growth was directly proportional to an increase in feeding rate (Group IV), and may reflect the natural feeding habit of the species. While the population of starved animals disappeared after 8 weeks, the treatments fed at 5% of their biomass increased 58% over their initial weight; those fed at 10% of their biomass increased 109%; and those fed at 15% biomass gained 164%. The above results indicate that penaeids are capable of consuming large amounts of food. This may be a reflection of their natural tendency to continuously graze upon large quantities of benthic material rather than feed periodically as would a strict carnivore. Although pellets used in all experiments were textured to main- tain consistency in solution for 24 hr, some shat- tering may occur as shrimp gnaw at them and thus some food may be lost through flushing, thus decreasing the efficiency of ingestion as feed levels are increased. Although growth was directly proportional to an increase in feeding rate, feeding at low levels was still justified in attempting to determine nu- tritional requirements of shrimp. The benthic material normally grazed upon is low in energy content and is often of relatively poor nutritional content. Feeding at lower fed levels but with food of proper nutritional value could conceiv- ably produce growth comparable to higher fed 107 FISHERY BULLETIN: VOL. 70, NO. 1 levels of natural or formula diets presently- known. Food conversion ratios (FCR) (weight of food fed for 6 weeks/ weight increase) were cal- culated from results in Group IV (calculated on a dry weight basis). Feeding at 109r biomass yielded an FCR of 6.7 and growth increase of 109 ^r. On the other hand, feeding at the 15 ^f level produced a 164 ''r growth increase and an FCR of 5.5. Such FCR, although not compar- able to those obtained for vertebrates such as the 1.6 or less for catfish (Andrews, in press), nonetheless represent a significant decrease over the FCR of 10 or greater reported for shrimp fed on natural foods (Fujinaga, 1963) . Further refinement of the FCR for penaeids can un- doubtedly be obtained through procurement of a more suitable pellet, better understanding of exact nutritional requirements of specific nu- trients, and more information on ingestion and assimilation phenomena. SUMMARY 1. Environmental conditions yielding 80 to 90% survival in the intensive tank culture of penaeid shrimp encompassed a combination of either no substrate or sand substrate on elevated platforms, air supplied externally by an aeration system, and population density of 40 g/m^. 2. Diets having balanced complements of pro- teins, lipids, carbohydrates, amino acids, fatty acids, minerals, and vitamins produced only sus- tained biomass levels. 3. Diets having 69.5^; of the total diet as shrimp meal produced growth increases of 63%. 4. Examination of soy, casein, and yeast hy- drolysates revealed that a combination of each produced 39 Sr growth increase while an average of 18% resulted from feeding each hydrolysate separately. Hydrolyzed proteins did not yield better growth than intact proteins. 5. Feeding at 5, 10, and 15 s; of the animals' biomass daily yielded directly proportional growth. A growth increase of 164% was achieved with a fish meal and shrimp meal diet fed at 15% of biomass daily. 6. Using semipurified pelleted diets, food con- version ratios were reduced by nearly half of that reported for penaeids feeding on clam and other natural foods. 7. Establishing selected preliminary environ- mental and nutritional requirements for penaeid shrimp resulted in reproducible production of major biomass increase with relatively high sur- vival and low food conversion ratios. 8. Results from these studies have allowed us to design facilities and experiments for future work with environmental and nutritional factors. Development of an inexpensive diet which will yield rapid and maximum growth will be an es- sential requirement for economical production of penaeid shrimp, ACKNOWLEDGMENTS The authors wish to sincerely thank Lee H. Knight and his engineering crew for their night and day effort to establish and maintain the fa- cilities and auxiliary power units that were es- sential for this study. In addition, we are grate- ful to Harry Carpenter and his crew for their efforts in general construction and maintenance of our mariculture facilities. LITERATURE CITED American Fish Farmer & World Aquaculture News. 1970. First cultured shrimp harvested at Florida farm. Am. Fish Farmer World Aquacult. News 2(1): 7. Anderson, W. W. 1966. The shrimp fishery of the southern United States. U.S. Fish Wildl. Serv., Fish Leafl. 589, 8 p. Andrews, J. W. In press. The stocking density and water require- ments for the culture of channel catfish in intens- ively stocked tanks. Foodstuffs. Broom, J. G. 1969. Pond culture of shrimp on Grand Terre Island, Louisiana, 1962-1968. Gulf Caribb. Fish. Inst. Proc. 21st Annu. Sess., p. 137-151. Dall, W. 1965. Studies on the physiology of a shrimp, Meta- penaetis sp. (Crustacea: Decapoda: Penaeidae). IV. Carbohydrate metabolism. Aust. J. Mar. Freshwater Res. 16: 163-180. 108 SICK, ANDREWS, and WHITE: REQUIREMENTS FOR SHRIMP CULTURE Duncan, D. B. 1955. Multiple range and multiple F tests. Bio- metrics 11: 1-42. FUJINAGA, M. 1963. Culture of Kuruma-shrimp (Penaeus japon- icus). Curr. Aff. Bull. Indo-Pac. Fish. Counc. 36: 10-11. Kanazawa, a., M. Shimaya, M. Kawasaki, and K. Kashiwada. 1970. Nutritional requirements of prawn — I. Feed- ing on artificial diet. Bull. Jap. Soc. Sci. Fish. 36: 949-954. Kanazawa, A., N. Tanaka, S. Teshima, and K. Kashiwada. 1971. Nutritional requirements of prawn — II. Re- quirements for sterols. Bull. Jap. Soc. Sci. Fish. 37: 211-215. Lunz, G. R. 1967. Farming the salt water marshes. Proceed- ings of the marsh and estuary management sym- posium, p. 172-177. Louisiana State Univ., Baton Rouge. Perez Farfante, I. 1969. Western Atlantic shrimp of the genus Peyi- aeiis. U.S. Fish Wildl. Serv., Fish. Bull. 67: 461-591. Smitherman, R. 0., AND D. D. Moss. 1970. Fish culture survey report for Panama. Ala. Agric. Exp. Stn., Auburn, 71 p. PB 195 912. SUBRAHMANYAN, C. B., AND C. H. OPPENHEIMER. 1969. Food preferences and growth of grooved penaeid shrimp. In H. W. Youngken, Jr. (editor) Food-drugs from the sea, proceedings 1969, p. 65- 75. Mar. Technol. Soc, Wash., D.C. SuRDi, R. W., AND D. R. Whitaker 1971. Shellfish situation and outlook 1970 annual review. Natl. Oceanic Atmos. Adm. Natl. Mar. Fish. Serv., Curr. Econ. Anal. S-20, 39 p. Van Den Oord, A. 1964. The absence of cholesterol synthesis in crab, Cancer pagurus L. Comp. Biochem. Physiol. 13 : 461-467. ViLLADOLID, D. v., AND D. K. ViLLALUZ, 1951. The cultivation of Sugpo {Penaeus monodin Fabricus) in the Philippines. Philipp. J. Fish. 1(1): 16-28. Wheeler, R. S. 1967. Experimental rearing of postlarval brown shrimp to marketable size in ponds. Commer. Fish. Rev. 29(3) : 49-52. Zandee, D. I. 1967. Absence of cholesterol synthesis as contrasted with the presence of fatty acid synthesis in some arthropods. Comp. Biochem. Physiol. 20: 811-822. 109 METHOD OF DETERMINING CAROTENOID CONTENTS OF ALASKA PINK SHRIMP AND REPRESENTATIVE VALUES FOR SEVERAL SHRIMP PRODUCTS Carolyn E. Kelley' and Anthony W. Harmon^ ABSTRACT An extraction method is described for estimating the amount of carotenoids in pink shrimp. The carot- enoid index is useful as a measure of quality and as an indicator of changes during storage. Values for several shrimp products are reported. The carotenoid content of Alaska pink shrimp is affected by many conditions and can be used as an index of the general quality of canned shrimp and of the changes in quality of frozen shrimp during storage. It has been also used as a factor in determining optimum peeling charac- teristics of shrimp (Collins and Kelley, 1969) and in selecting desirable retorting conditions (Kelley, 1971'). Color differences in shrimp at different seasons and in different areas may be important in harvesting and marketingpractices. The carotenoid in Alaska pink shrimp is pri- marily astaxanthin. Both total astaxanthin and astacin, its oxidation product, are measured by the method to be described, which wa'fe developed for use with frozen Alaska king crab (Ravesi, 1965') and adapted to Alaska pink shrimp which contain more interfering protein and moisture than crab. ' National Marine Fisheries Service, Fishery Products Technology Laboratory, Kodiak, AK; present address: 609 Schoenbar, Ketchikan, AK 99916. ^ Formerly, National Marine Fisheries Service, PMsh- ery Products Technology Laboratory, Kodiak, AK; pre- sent address: Department of Chemistry, Oklahoma State University, Stillwater, OK 74074. \ Kelley, C. E. 1971. Carotenoid content of pink shrimp: Effect of retorting conditions. National Ma- rine Fisheries Service, Fishery Products Technology Lab- oratory, Kodiak, Alaska. (Unpublished manuscript.) * Elinor Ravesi. 1965. Effect of processing and fro- zen storage on the carotenoid pigments of Alaska king crab. Unpublished manuscript filed at NMFS, Kodiak, Alaska. Manuscript accepted August 1971. FISHERY BULLETIN: VOL. 70, NO. I, 1972. METHOD OF DETERMINING CAROTENOID CONTENTS To 50 g of blended meat add approximately 10 g of silica gel and 100 ml of the proper ace- tone solution: 1. 759r acetone for canned shrimp with liquor. 2. 65% acetone for frozen cooked or raw meats. 3. 50 ^r acetone for raw shrimp with shells on. The silica gel, which serves as a filter aid, is not essential but makes subsequent extraction and filtration easier. Blend just enough to ensure complete mixing and filter through a medium porosity fritted glass funnel, maintaining suc- tion until dripping ceases. Rinse container and filter as needed with 50 "^r acetone. Discard colorless filtrate and blend residue about 2 min with 15 to 20 g anhydrous sodium sulfate and 100 ml of 1:1 2-propanol: chloroform. Filter and re-extract with 50 ml solvent one or two times as needed to get a colorless meat. Use 2- propanol: chloroform as rinse solution during these extractions. Transfer filtrate to 500 ml round bottom flask and strip the solvent, using a rotating vacuum evaporator. Add 5 to 10 ml chloroform and evaporate to dryness. Dissolve residue in enough pure cyclohexane to wash sides of flask and add 10 to 15 g aniiydrous sodium sulfate. Let set for a few minutes and filter through sodium sulfate on a fine porosity fritted 111 FISHERY BULLETIN; VOL. 70, NO. I glass funnel, washing: sodium sulfate with cyclo- hexane to remove all traces of color. If filtrate is clear, dilute to 100 ml. If it appears hazy, repeat the filtration, allowing solution to remain in fresh sodium sulfate for a brief time. Read at 474 nm on spectrophotometer, using cyclo- hexane as a blank. The precision of the method was determined by analyzing 11 identical samples in quadrupli- cate on 11 different days. Twenty-two cans of the same code of canned shrimp were blended in a Waring blendor, and the homogeneous mix- ture was sealed in cans and frozen at — 60° F. For each day's sampling, two cans of the frozen mixture were thawed and blended together. Different lots of solvents were used at inter- vals to determine the sensitivity to slight vari- ations in solvents. The solvent lot was not criti- cal but the cyclohexane used in the spectropho- tometer should be carefully distilled within a few days of use. We used a Gilford modification of a Beckman DU spectrophotometer' which gives readings with three place accuracy. The range of absorbance readings was 0.420 to 0.452, the average was 0.436, and the standard devi- ation was 0.008. The carotenoid content, expressed as the carot- enoid index, is a calculated value based on dry weight. The solids content of the shrimp was determined by the Association of Official Agri- cultural Chemists method (Horwitz, 1965: 346), using 5 to 10 g of the blended meat sample and heating at 105° C for 18 to 24 hr. The carotenoid index represents the absorbance (A) in 100 ml of solvent of the carotenoids from 1 g of dry sample, measured in a 1-cm cell. It is calculated as follows: C, = (A^ 474 nm in 100 ml cyclohe xane ) (100) (50-g wet sample) {% dry weight) The absorbance reading of a shrimp sami)le with average moisture content is roughly 10 times larger than the carotenoid index; there- fore the carotenoid index equivalent of the stan- dard deviation is slightly less than 0.001. The amount of carotenoid can also be ex- pressed as grams of pigTnent/gram tissue by using the extinction coefficient of 2150, as re- ported by Kanemitsu and Aoe (1958). The amount of astacin present in fresh shrimp is small and since the extinction coefficients of astacin and astaxanthin for calculation purposes do not introduce significant error for routine analytical work, the calculation would be: grams pigment/1 g tissue = (A aj^47^nm) (100 ml) 100 (50 g) (d) (2150) wiiere d is the cell width in centimeters. This could be converted to dry tissue weight by multi- plying by the percent of solids in the sample. CAROTENOID CONTENTS OF VARIOUS TYPES OF SHRIMP SAMPLES Table 1 gives carotenoid values of various types of shrimp samples described. Most of the data were collected as part of some other project so these samples are from several lots of shrimp caught at different times of the year. Only those grouped together in the table can be accurately compared with each other. All data, however, represent an average figure for the given sample and may be used to compare types of sample products or processing methods. Tablk 1. — Carotenoid values for 11 shrimp samples. Sampling conditions Carotenoid index Cause of color differences indicated by data ° The use of trade names is merely to facilitate de- scription and does not imply endorsement of a product. 1. Raw tails, shells on 0.237 Raw meats 0.086 2. Whole cooked, hand-peeled meats, frozen 0.112 3. Precooked, machine-peeled meats, frozen 0.086 After 6 months' storage , After 12 months' storage 0.070 0.062 Storage time 4. Precooked, machine-peeled canned 2-day iced, machine-peeled, canned 0.073 0.039 Precook versus Ice held conditioning 5. 1-day not iced, precooked, machine-peeled, canned 0.080 2-day iced, precooked, machine-peeled, canned 3-day iced, precooked, machine-peeled, canned 0.066 0.064 Time of ice holding 112 KELLEY and HARMON: CAROTENOID CONTENTS EXPERIMENTAL PROCEDURE 1. The shrimp were frozen whole as soon as possible after being caught, then were shipped to the laboratory. They were partially thawed, weighed, and separated according to weights. The shrimp used were about 80 count. All were headed and some were hand peeled to obtain meats. The tails with shells on and the peeled meats were refrozen as needed until analyses could be made. 2. Whole cooked, hand-peeled frozen shrimp meats were obtained from a commercial proces- sor. This is the conventional, cocktail style product. 3. Precooked, machine-peeled shrimp were produced under experimental conditions in a commercial plant. Shrimp were landed within 24 hr of catching, held overnight without ice, and precooked at 165° F for 10 sec, 110° F for 2 min, and machine peeled. The meats were col- lected at the end of the inspection belt and frozen in cans without vacuum. Analyses were made within a few days, after 6 months, and after 12 months of 0° F storage. 4. Precooked, machine-peeled canned shrimp were produced as described above except they were routinely retorted. The 2-day iced, ma- chine-peeled shrimp are a standard commercial pack from the same lot of shrimp. 5. The 1-day not iced; 2-day iced ; and 3-day iced, precooked machine-peeled, canned shrimp were also experimentally produced in a commer- cial plant. The 1-day not iced shrimp were held in the wooden boxes in which they were landed. The 2- and 3-day iced shrimp were held in large tanks and ice added as needed to keep them cool. All of these shrimp were precooked at 165° F for 10 sec, 110° F for 2 min, and then routinely peeled and canned. All samples were analyzed using the previ- ously described method of determining carote- noid contents. The averages reported in Table 1 represent 3 to 12 determinations under the given s,ampling conditions. ' Some of the factors which cause differences in the carotenoid content of shrimp are shown in Table 1. These include several processing variations which can be controlled by processors and fishermen. CONCLUSIONS The method of determining carotenoid content described is simple and precise and may be used on a variety of shrimp product forms. The carotenoid index for Alaska pink shrimp varies from 0.267 in raw tails to 0.059 in ice held, machine-peeled canned shrimp. Correla- tion with subjective color rating is quite good (Collins and Kelley, 1969) . At the higher color levels found in raw, hand-picked, or precooked shrimp, small differences are difficult to detect visually and the determination of the carotenoid index becomes even more useful in evaluating samples. Since the carotenoid content is usually closely correlated with other quality characteristics, the carotenoid determination may be useful in making decisions about the best ways to process or handle shrimp. LITERATURE CITED HORWITZ, W. (chairman and editor). 1965. Official methods of analysis of the Associ- ation of Official Agricultural Chemists. 10th ed. Association of Official Agricultural Chemists, Washington, D.C., xx + 957 p. Collins, J., and C. Kelley. 1969. Alaska pink shrimp, Pandalus borealis: Ef- fects of heat treatment on color and machine peel- ability. U.S. Fish Wildl. Serv., Fish. Ind. Res. 5: 181-189. Kanemitsu, T., and H. Aoe. 1958. Studies on the carotenoids of salmon. I. Identification of the muscle pigments. Bull. Jap. Soc. Sci. Fish. 24: 209-215. 113 A DESCRIPTION OF YOUNG ATLANTIC MENHADEN, Brevoortia tyrannus, IN THE WHITE OAK RIVER ESTUARY, NORTH CAROLINA Robert M. Lewis/ E. Peter H. Wilkens/ and Herbert R. Gordy^ ABSTRACT Atlantic menhaden exhibit three different stages — larva, prejuvenile, and juvenile — during their stay in the estuary. For specimens collected from the White Oak River estuary, N.C., the length-weight rela- tion was logg Y = -8^1104 + 3.6050 (log^ X) for larvae, log^ Y = -16.9638 + 6.3083 (log^ X) for prejuveniles, and logg Y = —5.2298 + 3.1452 (logp X) for juveniles, where Y =: weight in mg and A' = length in mm. Larvae and prejuveniles concentrated in the low salinity-freshwater zone upstream. Juveniles tended to move downstream toward the higher salinity water. Condition factors of larvae and prejuveniles increased toward the higher salinity zone. Atlantic menhaden spawn and hatch in coastal oceanic waters from Maine to Florida. The larvae enter estuaries where they transform in- to juveniles near the freshwater zone. The re- lation of young menhaden length and weight to time, salinity, and location within an estuary provides insight on the environmental require- ments of menhaden during a critical phase in their life cycle. We collected young menhaden from a small estuary in North Carolina, from March to September 1969 with a tidal net (Lewis et al., 1970) to study changes in the length- weight relation. The lower portion of the White Oak River estuary (28 sq km) is shallow with depths from to 3.0 m and distances from opposing shores from 1 to 3 km. The intracoastal waterway crosses the lower estuary and is maintained at a depth of approximately 4 m. The upstream portion narrows into a river up to 4.6 m deep. During the study period we generally found that the change from brackish to fresh water oc- curred between 18 and 24 km upstream from Bogue Inlet. The exact location of this low sa- linity zone was influenced by tide, rainfall, and direction and speed of the wind. The mean tidal ^ National Marine Fisheries Service, Atlantic Coastal Fisheries Center, Beaufort, NC 28516. " National Marine Fisheries Service, Southeast Fish- eries Research Center, 75 Virginia Beach Drive. Miami, FL 33149. Manuscript accepted August 1971. FISHERY BULLETIN: VOL. 70, NO. 1. 1972. range at Bogue Inlet is 2.2 ft (0.67 m) (U.S. Coast and Geodetic Survey, 1970). At 21 km upstream the average river flow is 14.7 cfs (0.42 mVsec) (North Carolina State Board of Health, 1954) . A map showing the location of the White Oak River estuary is shown in Lewis and Mann (1971). We selected 14 stations from Bogue Inlet to 34 km upstream (Wilkens and Lewis, 1971). Stations ranged from 2 to 5 km apart and were selected to be representative of the various sa- linities encountered. We also sampled in fresh water to determine how far menhaden move up- stream. Salinity measurements were taken within 1 m of the surface. During collections, spot checks of salinity between the surface and bottom indicated that in this shallow estuary thorough mixing generally occurred. Diff"er- ences between measurements at one location were due to flooding and ebbing tides. Our menhaden collections ranged from 15 to several thousand individuals. We measured and weighed all fish to the nearest 0.5 mm total length and 0.1 mg in collections containing less than 26 and subsampled the larger collections. Since both length and weight variances in the sub- samples were small, we considered our estimates of length and weight to be reliable. Our mea- surements of total length were based on the greatest dimension between the most anteriorly projecting part of the head and the farthest tip 115 FISHERY BULLETIN: VOL. 70. NO. I of the caudal fin when the caudal rays are squeezed together (Hubbs and Lagler, 1949). We separated young menhaden into three stages on the basis of body form and the length- weight relation of individuals within each stage. Length and weight ranges of all the fish used in the study are given in Table 1, An illustra- tion of each stage (larva, prejuvenile, and ju- venile) that occurs during the first year in the estuary is shown in Figure 1. Allometric growth, with stanzas for larvae, prejuveniles, and juveniles is shown in Figure 2. The inflection points, indicating change in slope, are 30 and 38 mm for length, and 70 and 469 mg for weight. We considered specimens less than 30 mm and 70 mg as larvae; they are long and slender, and even at 30 mm total length the body depth is only 4 mm or less. In the next group, Table 1. — Lengths and weights of Atlantic menhaden from the White Oak River estuary, N.C., arranged in order of increasing weight classes. Weight Length range Number of menhaden Weight Length range Number of menhaden mg mm mg mm 0.0- 4.9 8-16 23 100.0- 199.9 29-34 38 5.0- 9.9 14-20 46 200.0- 299.9 33-37 26 10.0-14.9 17-21 43 300.0- 399.9 35-38 6 15.0-19.9 20-23 48 400.0- 499.9 37-43 5 20.0-24.9 20-24 45 500.0- 599.9 39-41 10 25.0-29.9 22-26 32 600.0- 699.9 40-44 14 30.0-34.9 23-26 30 700.0- 799.9 41-44 11 35.0-39.9 24-27 31 800.0- 899.9 44-45 6 40.0-44.9 25-28 29 900.0- 999.9 45-49 4 45.0-49.9 25-29 10 1,000.0-1,499.9 47-54 41 50.0-54.9 26-29 20 1,500.0-1,999.9 53-60 27 55.0-59.9 27-29 26 2,000.0-2,499.9 58-62 16 60.0-64.9 27-31 26 2,500.0-2,999.9 61-66 11 65.0-69.9 27-31 19 3,000.0-3,499.9 68-71 5 70.0-74.9 28-31 22 3,500.0-3,999.9 71-74 8 75.0-79.9 29-32 15 4,000.0-4,499.9 75-77 4 80.0-84.9 29-32 12 4,500.0-4,999.9 76-82 5 85.0-89.9 29-31 7 5,000.0-5,499.9 81 3 90.0-94.9 28-32 U 5,500.0-5,999.9 81-83 2 95.0-99.9 29-32 6 B Figure 1.— Atlantic menhaden (a) larva 27.0 mm total length (TL) ; (b) prejuvenile 32.0 mm TL; and (c) ju- venile, 64.0 mm TL. The alimentary tracts are shown as they were visible in the preserved specimens used in drawings. 116 LEWIS, WILKENS, and GORDY: YOUNG ATLANTIC MENHADEN o o o X o 7.0 10 9.0 B.5 8.0 7.5 7.0 6.5 6.0 6.5 6.0 4,5' 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 LENGTH (MM) 20.0 30 40.0 60.0 80. 100 .0 10,000.0 I I I I Y = -5. 2298 + 3.1452 X Juveniles Y = -16 9638 + 6. 3083X Prejuveniles Y=-8.1104+3.6050X Larvae I I 5.000.0 260 O X 100.0 O . 10.0 I no 2.2 2,4 2.6 2.8 3.0 3.2 3.4 3.6 3.8 4.0 4.2 4.4 4.6 LENGTH (LOGe MM) Larval menhaden were most abundant in March, prejuveniles in late March and April, and juve- niles by the beginning of May (Wilkens and Lewis, 1971). Large catches of larval and pre- juvenile menhaden within the freshwater-low salinity zone (Table 2) suggest that favorable conditions for growth are present. Condition factors {W/V, where W = weight in mg, L = length in mm, and 5 = value for the slope of weight on length for each growth stanza) of larvae and prejuveniles increased with time as the result of growth and develop- ment. The apparent lack of growth of larvae and prejuveniles in the low salinity-freshwater zone during April is probably due to large num- bers entering this zone, putting on fast growth, moving out of the zone, and being replaced by new groups (Tables 2 and 3). Juveniles, which have the same body form as adults and which are scattered in schools throughout the estuary, showed no change in condition factor with time or salinity. Figure 2. — Regression of weight on length for larval, prejuvenile, and juvenile Atlantic menhaden collected in White Oak River estuary, N.C., in 1969. (We separated the lengths and weights into three groups after visual observation of the data and fish. Lines were then fitted by least squares regression based on data in each group.) prejuveniles, we included specimens from 30 to 38 mm and 70 to 469 mg. In this stage there is a rapid increase in body depth, but little increase in length. Fish above 38 mm and 469 mg we classed as juveniles. Huntsman' found that the relation between length and weight is similar for juveniles and adults. Both stages have a similar body form, only their color and size being dif- ferent. We did not find any adults in our estu- arine study. Larvae enter the lower estuary and move up- stream to the freshwater-low salinity zone where they go through a prejuvenile stage before com- pleting their transformation into juveniles. ' Huntsman, Gene R. 1971. Growth by year class of Atlantic menhaden. (Unpublished manuscript.) NMFS Center for Estuarine and Menhaden Research, Beaufort, NC 28516. LITERATURE CITED HUBBS, C. L., AND K. F. Lagler. 1949. Fishes of the Great Lakes Region. Cran- brook Inst. Sci., Bull. 26, 186 p. Lewis, R. M., W. F. Hettler, Jr., E. P. Wilkens, and G. N. Johnson. 1970. A channel net for catching larval fishes. Chesapeake Sci. 11: 196-197. Lewis, R. M., and W. C. Mann. 1971. Occurrence and abundance of larval Atlantic menhaden, Brevoortia tyrannus, at two North Carolina inlets with notes on associated species. Trans. Am. Fish. Soc. 100: 296-301. North Carolina State Board of Health. 1954. The White Oak River Basin. N.C. State Board Health, PoUut. Surv. Rep. 2, 122 p. U.S. Coast and Geodetic Survey. 1970. Tide tables. East Coast of North and South America including Greenland, 1971. U.S. Coast Geod. Surv., 290 p. Wilkens, E. P. H., and R. M. Lewis. 1971. Abundance and distribution of young At- lantic menhaden, Brevoortia tyrannus, in the White Oak River estuary. North Carolina. Fish. Bull., U.S. 69: 783-789. 117 FISHERY BULLETIN: VOL. 70, NO. 1 Table 2. — The distribution and mean total length (mm) of menhaden by date collected, kilometers upstream from Bogue Inlet, and salinity i%o) in the White Oak River estuary, March-August 1969. Date 1969 Mar. 17 Mar. 27 Apr. 1 Apr. 9 Apr. 16 May 1 May 14 May 27 June 26 July 16 Aug. 27 16 kilometers: Salinity __ 1.5 4.1 8.0 5.2 13.7 15.4 1.8 10.2 15.1 Abundance index* 4.3 298.4 48.8 111.0 279.1 O.I 5.2 4.2 4.6 0.3 Mean total length — 28.5 28.9 30.0 29.2 32.1 — 42.4 53.4 58.8 — 18 kilometers: Salinity 1.5 0.2 0.1 3.8 1.8 7.4 11.7 0.2 3.0 8.9 Abundance index 9.6 0.4 1,053.4 497.7 526.5 155.2 0.4 1.1 16.3 6.6 Mean total length 29.9 — 28.5 27.5 30.8 30.5 — — 43.5 51.0 — 21 kilometers: Salinity 0.2 1.2 4.2 — __ Abundance index 48.7 54.2 65.3 __ 0.5 — , _- . Mean total length 30.8 28.3 27.2 — — — — — — — 24 kilometers: Salinity 1.5 4.4 0.5 Abundance index 13.3 1.8 6.0 1.5 1,533.3 392.6 0.7 13.9 0.4 1.4 I.O Mean total length 30.1 29.6 29.2 26.6 28.5 29.4 — 42.6 — — ~ 28 kilometers: Salinity 0.6 Abundance index 0.5 1.4 0.2 0.6 O.I 12.1 0.3 0.2 Mean total length — — 29.5 — — — — 41.8 — — — 31 kilometers: Salinity __ Abundance index 0.3 10.3 1.0 Mean total length — — — 29.1 — — — — — — — ^ Abundance index is the number of young menhaden for 100^ of water. Table 3.— Mean condition factors of young Atlantic menhaden collected in the White Oak River estuary, N.C., in 1969. Salinity {%„) 0.1-0.9 1. 0-1.9 2.0-2.9 3.0-3.9 4.0-4.9 5.0-5.9 >6.0 Larvae Prejuveniles Juveniles 1969 Mar. 17 Mar. 27 Apr. I Apr. 9 Apr. 16 May 1 Mar. 17 Mar. 27 Apr. 1 Apr. 9 Apr. 16 May 1 May 27 May 27 June 26 July 16^ 0.292 0.293 0.302 0.291 0.332 0.310 0.331 0.317 0.355 0.321 0.358 0.299 0.394 0.376 0.342 0,364 0.381 0.396 0.402 0.497 0.303 0.309 0.326 0.253 0.331 0.326 0.357 0.444 0.305 0.387 0.435 .555 .551 .526 0.314 0.390 0.599 0.540 0.476 0.420 0.517 0.547 0.542 0.516 0.498 0.507 0.535 ' Sample sizes after July 16 were too small to show trends 118 GROWTH OF PREMIGRATORY CHINOOK SALMON IN SEAWATER Bernard M. Kepshire, Jr., and William J. McNeil^ ABSTRACT A potential demand exists in sea farming for premigratory juvenile Pacific salmon that have been acclimated to seawater. This paper reports experiments on growth of premigratory chinook salmon (Oncorhynchus tshawytscha) acclimated to water of 33^r salinity and lower and describes a simple mathematical model to evaluate rate of growth. Although chinook salmon raised in these experiments experienced low mortality in water of high salinity, their growth slowed. Reasons for slow growth at high salinity are discussed. Pacific salmon reproduce in fresh water, but only two species — pink (Onchorhynchus gorbuscha) and chum (0. keta) salmon — survive direct transfer as fry from fresh water to full-strength seawater (Weisbart, 1968). The ocean serves as the early nursery ground for these two spe- cies. The other species — including sockeye (0. nerka) , coho (O. klsutch) , and chinook (0. tshawytscha) salmon — require freshwater nur- sery areas. Juvenile salmon undergo a period of adjust- ment when they first enter the sea in order to regulate water and salts in body fluids and tis- sues. This adjustive phase for chum salmon fry lasts about 30 hr and is characterized by an immediate depression of activity, increased con- centration of salts in body fluids, and dehydra- tion of body tissues (Houston, 1959) . A slightly longer adjustive phase of 36 to 40 hr has been reported for yearling coho salmon (Conte et al., 1966; Miles and Smith, 1968). Early exposure to water of low salinity can "trigger" the physiological adaptation to sea- water of salmon species which typically remain in fresh water for several months as juveniles. Acclimation of premigratory young chinook salmon to water of 30^( salinity by exposing them to gradual increments in salinity has been described by Wagner et al. (1969). Black (1951), Coche (1967), and Otto (1971) found ^ Department of Fisheries and Wildlife, Oregon State University, Marine Science Center, Newport, OR 97365. also that coho salmon fry were better able to tolerate water of high salinity after having first been exposed to water of low salinity. Other evidence suggests that the growth of juvenile coho and chinook salmon is influenced by salinity. Coho salmon fry were observed by Canagaratnam (1959) to grow faster in water of 12 to 18^;, than in fresh water. Otto (1971) reported faster growth of juvenile coho salmon at 5 and 10'/,r salinity than at higher salinities or in fresh water. Bullivant (1961) found no significant diflPerence in growth of juvenile chi- nook salmon in water of and W/U salinity. However, Bullivant's fish grew more slowly at 35/^f salinity than at the two lower salinities. This paper reports comparisons of the growth of juvenile chinook salmon raised in water rang- ing in salinity from to 33/i:f . The experiments were conducted at the Oregon State University Port Orford Marine Research Laboratory, Curry County, Oreg. GENERAL PROCEDURES Two groups of chinook salmon used in these experiments were obtained as eyed eggs from the Fish Commission of Oregon, Elk River Hatchery, in winter 1969. Group I fish were divided into five subgroups of 200 each on Feb- ruary 24 (47 days after hatching). Group II fish were divided into six subgroups of 300 each on March 5 (18 days after hatching). Indi- vidual subgroups were introduced to water of Manuscript accepted August 1971. FISHERY BULLETIN: VOL. 70, NO. 1, 1972. 119 FISHERY BULLETIN: VOL. 70, xNO. 1 increasing salinity according to the schedules outlined in Tables 1 and 2. Both groups of fish received the Oregon Moist Pellet diet. The young salmon were fed five times daily beginning 30 days after hatching. After the fish had attained an average weight of 1 g, the frequency of feeding was reduced to three times daily. They were provided more food than they would consume at each feeding. Fish were raised in 100-gal pl>^vood tanks which were lined with fiber glass. Water was introduced to each tank at the rate of one-half gallon per min. Incoming fresh and salt water were premixed in head tanks to obtain desired salinities. Salinities were calculated from the proportions of premixed seawater and fresh water, and density of water in fish tanks was measured periodically with hydrometers to in- sure that salinities remained at their calculated levels. The first experiment (Group I fish) began on January 8 with newly hatched alevins. Group I fish first received food on February 7, and se- lected subgroups were exposed to water of 9 or lT/(( salinity beginning on February 24. The five subgroups were first weighed on February 27. The experiment ended on May 7. The second experiment (Group II fish) began on February 16 with newly hatched alevins. All six subgroups of fish were first exposed to water of 5 or 9'/, salinity on March 6 while still in the alevin stage, and they remained at these salinities for 18 days. The fish were first fed on March 18 and first weighed on April 7. The experiment ended on May 6. Mortality of the 11 subgroups of fish during the test periods ranged from to 6 /r of the orig- inal number of fish placed in the tanks. Even the maximum mortality {6^r ) was considered to have no appreciable effect on the comparisons of growth. The average wet weight of fish in each sub- group was determined at 14-day intervals from random samples of 30 fish. Excess water was blotted from anesthetized fish before weighing. Fish were weighed separately in a flask contain- ing a known weight of water and were returned to their respective tanks after each weighing. Table 1. — Exposure of Group I chinook salmon to water of increasing salinity. Date of hatch — January 7, 1969. Subgroup Age (days after hatchi water' !ng) at which of given sal fish inlty were placed 47 66 80 la lb 9 %. 17 17 Ic 17 17 33 Id 17 24 33 le 9 24 33 ' Temperature of incoming water averaged 10.7° C for fresh water and 10.8° C for seawater. Table 2. — Exposure of Group II chinook salmon to water of increasing salinity. Date of hatch — February 15, 1969. Aga (days after hatchi ng) at which fi sh were placed Subgroup in wafer' ot given sal ini ty 18 36 54 66 %, %, ^o 0/ AW ila 5 18 18 18 Mb 9 18 18 18 lie 5 18 25 33 lid 5 18 25 33 lie 9 18 25 33 llf 9 18 25 33 ' Temperature of incoming water averaged 11.9° C for fresh water and 12.0° C for seawater. OBSERVATIONS ON GROWTH Growth rate was calculated for each subgroup from the periodic measurements of wet weight. Growth was assumed to be exponential over each period considered, and a value for the daily in- crement in body weight, which can be expressed as a iiercent of body weight per day, was ob- tained from the expression Wt (1 + h) (1) where Wt Wo h t weight at the end of the period, weight at the beginning of the period, the compounded daily increment of body weight, and days. It is convenient to solve equation (1) for (1 + h) by converting the terms to common logarithms and taking the antilog, i.e. log (1 + h) = log Wt log Wo (2) 120 KEPSHIRE and McNEIL: GROWTH OF PREMIGRATORY CHINOOK To clarify the concept of daily increment of body weight, fish that can maintain an increase in body weight of 2.0 and 3.0 "^r per day, for example, will double their weight in approxi- mately 35 and 23 days, respectively. Fish held in water of 0, 17, and IS'/cc salinity grew at a faster rate and were heavier at the end of the experiments than fish of the same age transferred to water of 24, 25, and SS'/cc salinity. The observed mean weight of fish in individual subgroups is plotted against age of fish in Fig- ure 1. Equation (2) can be rewritten in linear form to calculate statistics which are useful for making comparisons of rate of growth among test groups of fish. The linear model is: (log Wt — log Wo) = log (l + h) t (3) Slope of the regression line is given as log (1 + h). This model requires the regression line to pass through the origin since (log Wt — log T^o) = ait = 0. Group I fish were weighed on six occasions over a period of 70 days. We have estimated values of log (1 + h) and h for each of the five subgroups of Group I by calculating the five regressions of (log Wt — log Wo) on t. Because the weight of Group II fish was measured on only three occasions, we have not applied a sim- ilar analysis to the second experiment. Application of regression methods to obser- vations on Group I fish indicates that fish in fresh water and water of 17^f salinity (Sub- groups la and lb) grew at a significantly faster rate than fish exposed to water of 33%r salinity (subgroups Ic, Id, and le). Equations for the five subgroups are given in Table 3 along with the 95 '^r confidence interval estimates of log {1 + h) and the approximate confidence interval estimates of h. Figure 2 shows growth curves for the fastest (Subgroup la) and slowest (Sub- group le) growing fish. The periodic measure- ments of weight are plotted in Figure 2 to show their correspondence with the growth curves calculated by use of equation (1). 30 2.5 - 2.0 1.5 5 1,0 0.5 Salinity O 0% O 17-18%. A 33 7c _1_ _1_ _1_ 50 64 78 92 106 Age (days after hotching) 120 Figure 1. — Growth in weight of experimental subgroups of juvenile chinook salmon. DISCUSSION Chinook salmon used in these experiments were exposed to salt water much earlier in life than they normally would experience in nature. Group I fish were acclimated to high salinity (24%r) 66 days after hatching and 36 days after commencement of feeding. Group II fish were acclimated to high salinity (25'/, o) 54 days after hatching and 24 days after commencement of feeding. There were only 66 deaths (3.7%) among the 1,800 fish which had been exposed to salinities of 24, 25, and 33/{o for periods of 25 and 54 days. The average rate of growth in water of high salinity (24'/cc and above) varied between 2.1 and 2.3% increment in body weight per day. These fish doubled their weight in 30 to 33 days. The average rate of growth in water of low sa- linity (17%f. and 0%o) was 2.6 and 2.7% per day. These fish doubled their weight in 26 to 27 days. Although these experiments demonstrate that juvenile chinook salmon can be acclimated to full- strength sea water in an early age, it is apparent that water of high salinity causes a reduced rate 121 FISHERY BULLETIN: VOL. 70, NO. 1 Table 3. — Regression of (log W^ — log Wq) on time for Group I fish. The approximate 95 "Jf confidence interval estimates of h are taken from the confidence limits of log (1 + h). Sub- group Regression equation 95% confidence limits of log (I + A) Approximate 95% confidence limits of A la (log If^ - log W^) = 0.01 168< lb (log IV ^- log (Cg) = 0.011 18« Ic (log »'j — log ffp) = 0.01006< Id (log ;fj - log W^) = 0.00945< le (log Jfj - log W^) = 0.00880i 0.01168 ± 0.00099 0.01118 ± 0.00084 0.01006 ± 0.00087 0.00945 It 0.00077 0.00880 ± 0.00042 2.7 ± 0.2%/day 2.6 ± 0.2%/day 2.3 ± 0.2%/day 2.2 ± 0.2%/day 2.1 ± 0.1%/day of growth. Reduced growth may come about in part because the young salmon expend more en- ergy to maintain an osmotic homeostasis in water of high salinity than in water of low sa- linity. Chinook salmon blood is isotonic with water of salinity between 10 and 13'/r (Coche, 1967). Houston (1959) thought that the increased ener- gy demands for osmoregulation combined with possible inhibition of electrolyte-sensitive com- ponents of the neuromuscular system might con- tribute to reduced growth of young salmon in water of high salinity. There is the further possibility that endocrine systems which are as- sociated with osmoregulation and growth in water of high salinity are not fully functional in premigratory juvenile salmon (Saunders and Henderson, 1970). O Subgroup la (h = 2.7% /day) Subgroup le (h=2 l7o/day) ~ 3 E 2 2 28 42 56 I 70 Time (days after first weighing) Figure 2. — Calculated growth curves for chinook salmon from subgroups la and le as calculated from equation (1). The observed growth is plotted to illustrate cor- respondence with calculated curves, "h" is the com- pounded daily increment of body weight. The acclimation of premigratory chinook, coho, and sockeye salmon to seawater may find future applications in aquaculture. Possibilities include the early release of young salmon from hatcheries into open ocean pastures to reduce costs of feeding and handling and to increase hatchery production. Other possibilities are to pen young salmon in saltwater bays or estuaries (Garrison, 1965; Mahnken et al., 1970) or to place them in raceways receiving waste salt water from coastal thermal-electric stations (McNeil, 1970). Large-scale aquaculture systems, similar to one under development in the Canadian Maritime Provinces (Gunstrom, 1970), would most likely benefit from early acclimation of juvenile salmon to seawater. The release of premigratory ju- venile chinook salmon acclimated to seawater should also be tested at hatcheries equipped with seawater pumping systems. The eflFects of early acclimation on ocean survival is unknown, but the greater availability of food and space in the ocean than in freshwater conceivably would pro- vide potential advantages to juvenile salmon which had been acclimated to seawater. ACKNOWLEDGMENTS Research on acclimation of juvenile salmon to seawater is administered by the Oregon State University Agricultural Experiment Station. Funds are provided by the National Oceanic and Atmospheric Administration's Sea Grant Pro- gram (Contract No. GH97) and National Ma- rine Fisheries Service (Project No. AFC-55). We wish to express our appreciation to Robert Courtright, Director of the Oregon State Uni- versity Port Orford Marine Research Labora- 122 KEPSHIRE and McNEIL: GROWTH OF PREMIGRATORY CHINOOK tory, for assistance and guidance with this project. We also wish to acknowledge helpful comments on the content and organization of this paper by Dr. Lauren R. Donaldson, Uni- versity of Washington, Anthony J. Novotny, Na- tional Marine Fisheries Service, and Harry H. Wagner, Oregon State Game Commission. LITERATURE CITED Black, V. S. 1951. Changes in body chloride, density, and water content of chum {Oncorhynchus keta) and coho (0. kisutch) salmon fry when transferred from fresh water to sea water. J. Fish. Res. Board Can. 8: 164-177. BULLIVANT, J. S. 1961. The influence of salinity on the rate of oxy- gen consumption of young Quinnat salmon (On- corhynchus ishawytscha) . N.Z. J. Sci. 4: 381-391. Canagaratnam, p. 1959. Growth of fishes in different salinities. J. Fish. Res. Board Can., 16: 121-130. COCHE, A. G. 1967. Osmotic regulation in juvenile Oncorhynchus kisutch (Walbaum). I. The salinity tolerance of 50-day-old fry. Hydrobiologia, 29: 426-440. CoNTE, F. P., H. H. Wagner, J. Fessler, and C. Gnose. 1966. Development of osmotic and ionic regulation in juvenile coho salmon {Oncorhynchus kisutch). Comp. Biochem. Physiol. 18: 1-5. Garrison, R. L. 1965. Coho salmon smolts in 90 days. Prog. Fish- Cult. 27: 219-230. Gjjnstrom, G. K. 1970. Canadian mariculture facility begins oper- ation. Am. Fish Farmer 2(1) : 8-11. Houston, A. H. 1959. Locomotor performance of chum salmon fry {Oncorhynchus keta) during osmoregulatory adaption to sea water. Can. J. Zool. 37: 591-605. AHNKEN, C. V. W., A. J. Novotny, and T. Joyner. 1970. Salmon mariculture potential assessed. Am. Fish Farmer 2(1): 12-15, 27. McNeil, W. J. 1970. Heated water from generators presents fish- culture possibilities. Am. Fish Farmer 1(11): 18-20. Miles, H. M., and L. S. Smith. 1968. Ionic regulation in migrating juvenile salmon, Oncorhynchus kisutch. Comp. Biochem. Physiol. 26: 381-398. Otto, R. G. 1971. Effects of salinity on the survival and growth of pre-smolt coho salmon {Oncorhynchus kisutch). J. Fish. Res. Board Can., 28: 343-349. Saunders, R. L., and E. B. Henderson. 1970. Influence of photoperiod on smolt develop- ment and growth of Atlantic salmon {Salmo salar). J. Fish. Res. Board Can., 27: 1295-1311. Wagner, H. H., F. P. Conte, and J. L. Fessler. 1969. Development of osmotic and ionic regula- tion in two races of Chinook salmon Oncorhynchus tshawytscha. Comp. Biochem. and Physiol. 29: 325-341. Weisbart, M. 1968. Osmotic and ionic regulation in embryos, alevins, and fry of the five species of Pacific salmon. Can. J. Zool. 46: 385-397. 123 EFFECT OF ENCROACHMENT OF WANAPUM DAM RESERVOIR ON FISH PASSAGE OVER ROCK ISLAND DAM, COLUMBIA RIVER Richard L. Majors and Gerald J. Paulik° ABSTRACT The filling of Wanapum Reservoir in 1964 flooded the lower sections of the three fish ladders at Rock Island Dam, 61 km upstream from Wanapum Dam on the Columbia River. To maintain fish passage under the new hydraulic conditions, the lower portions of the center and left-bank fish ladders of Rock Island Dam were rebuilt and a new sequence of spill patterns inaugurated. The effectiveness of these modifications was evaluated by comparing results from a series of tagging experiments conducted in 1964-65 on spring chinook salmon (Oncorhynchus tshawytscha) and sockeye salmon (0. nerka) with the results of similar experiments in 1954-55 before Wanapum Dam was built. These comparisons indicated fish passage over Rock Island Dam had improved substantially between 1954-55 and 1964-65; tagged fish traveled over the dam in a shorter time, and higher percentages of the tagged groups were sighted passing over the dam under postencroachment conditions. Successful reproduction of Pacific salmon {On- corhynchus spp.) and steelhead trout (Salmo gairdneri) requires that sufficient numbers of adults in suitable physical condition reach the spawning grounds. Serious consequences can result from delays en route. Thompson (1945) , for example, showed that very few of the sockeye salmon (0. nerka) that were delayed more than 12 days by the Hell's Gate rock slide (Fraser River, British Columbia) reached their spawn- ing grounds. Thompson also suggested that shorter delays reduced the reproductive capa- bility of the survivors. Similarly, man-made fa- cilities such as hydroelectric dams, even though equipped with fish-passage facilities, can act as barriers and thus delay or otherwise interfere with the migratory behavior of salmonids on their way to the spawning grounds. One of the primary goals of the agencies re- sponsible for conserving the fish resources of the Columbia River is to seek ways of minimizing the eflfects of dams on the migration and spawn- ing success of the river's populations of salmo- nids. Although a variety of solutions to the ^ National Marine Fisheries Service Fisheries North- west Center, 2725 Montlake Boulevard East, Seattle, WA 98102. " Center for Quantitative Science, University of Wash- ington, Seattle, WA 98195. problem of dams impeding the passage of mi- grating spawners have been proposed and a number of these have been tried in the field, the pool type of ladder has proven to be the only practical means of passing large numbers of adult salmonids over the dams on the Columbia River. Many research studies aimed at improv- ing fish passage have been conducted over the past several decades. One result of this research has been the introduction of a number of im- provements in design and operation of the pool- and-weir ladder. In some cases fish passage over ladders can be substantially improved by modi- fication of spill patterns (Leman and Paulik, 1966). The present study was designed to evaluate the eflfectiveness of modifications in the fish lad- ders at Rock Island Dam and changes in the spill pattern which were made after the lower portions of the ladders were flooded by the res- ervoir of Wanapum Dam. This type of problem is apt to become more common as all existing sites for hydroelectric dams are utilized and the reservoir of one dam begins to encroach on the tailrace of the dam immediately above. Rock Island Dam, completed in 1934, was the first dam built on the Columbia River. It is in central Washington and about 725 km above the river's mouth (Figure 1). Originally, the dam Manuscript accepted July 1971. FISHERY BULLETIN: VOL. 70, NO. 1, 1972. 125 FISHERY BULLETIN: VOL. 70. NO. I Figure 1. — The Columbia River and locations important to the present study. was provided with two pool- and weir-type fish ladders, one adjacent to each bank. A third was added near the middle of the dam in 1936. Rock Island Dam was modified during 1951- 53. Six new generating units were added to the powerhouse (on the left side of the dam looking downstream), the reservoir was raised about 3.7 m, and regulating lift gates were installed in spillway bays 16-37 on the right side of the dam (Figure 2). Turbine discharge was in- creased from about 793 to 2,265 m^ per sec, the fish ladders were altered to enable them to function at the new reservoir level, and the at- traction flow at the entrance to the left ladder was increased to counteract the increased dis- charge from the turbines. Although the fishery agencies requested changes at the lower end of the right ladder to provide better entrance con- ditions and additional attraction flow, no imme- diate action was taken to implement these re- quests. The Federal Power Commission, in granting permission for the modification of the dam, re- FiGURE 2. — Rock Island Dam showing locations of the fish ladders, powerhouse, and spillway bays. served the right to require alteration of the low- er end of the right-bank ladder if substantial evidence were presented to show that alteration was necessary to protect runs of anadromous fish. Any such alteration was to begin before December 1, 1960. To determine whether the dam caused loss or delay to the runs and whether loss or delay was associated with the right-bank fishway, tagging studies were conducted in 1954-56. French and Wahle (1966) summarized the results as fol- lows: Point estimates of sockeye salmon losses ranged from to 42 percent. Tagging results (one season only) on spring chinook salmon indicated a loss of fish re- leased below the right bank ladder, but no loss when total tag returns from below and above dam releases were compared; data failed to show the dam caused losses of simimer chinook. Tagged salmon released below the dam were delayed 2 to 4 days. Altering the right bank fishway may cause more fish to use it, but there is no clear evidence that such alteration will reduce overall loss or delay.^ Although the evidence did not indicate that a major overhaul of the right-bank ladder was justified, one relatively minor change was made. In 1956, a concrete wall was built at the entrance to the ladder. This wall replaced a cyclone fence " Seasonal races of chinook salmon (0. tahawytcha) in the Columbia River system are classified as spring, summer, or fall chinook depending on the time of year that the adults enter the river to spawn. 126 MAJOR and PAULIK: ENCROACHMENT OF WANAPUM DAM RESERVOIR which, at high tailwater elevations, appeared to impede the entry of fish into the ladder. The wall also eliminated the surging action across the ladder that occurred under certain combi- nations of spill pattern and tailwater elevation. Even after these changes, fishery biologists continued to voice concern over the effectiveness of the right-bank ladder. Some held that hy- draulic conditions at the entrance to the right- bank ladder impeded fish passage over the ladder. In 1958, in response to the continuing concern about the fish-passage conditions at Rock Island Dam, the owners of the dam — Puget Sound Pow- er and Light Co. and Public Utility District No. 1 of Chelan County — financed a study to determine if fish passage could be improved by manipulat- ing the spill pattern. The study showed that fish could be guided to either the right or center ladders by spilling adjacent to the respective ladder. The experi- ment also indicated, but not conclusively, that when low tailwater elevations prevailed, spilling from gates 35, 36, and 37 (immediately adjacent to the right-bank ladder) confused the fish and interfered with their entry into the ladder (Leman and Paulik, 1966). The construction in 1960-64 of Wanapum Dam, 61 km downstream, brought about a fur- ther change in fish-passage conditions at Rock Island Dam. The lower portions of the fish lad- ders at Rock Island Dam were flooded by Wan- apum Reservoir. This condition, in the judg- ment of the Federal Power Commission, required certain modifications of the left and center fish ladders. The modifications were completed by 1963 — 1 year before Wanapum Reservoir was filled. The Commission also directed the owners of Rock Island Dam and of Wanapum Dam (Pub- lic Utility District of Grant County) to develop, in cooperation with representatives of the U.S. Department of the Interior and the Washington State Departments of Fisheries and Game, a program for the study and evaluation of the further effects of encroachment by Wanapum Reservoir on fish passage at Rock Island Dam. It is important to note that the question at hand was the effect of the encroachment of Wan- apum Reservoir on fish passage at Rock Island Dam and not the effects of Wanapum Dam on fish passage in the broader sense, i.e., passage over Wanapum Dam itself and passage through the newly formed forebay. Representatives of the participating agencies formed the "Rock Island Study Group" and en- gaged the junior author as a consultant to serve as chairman of the group. A major segment of the research program, initiated and supervised by the study group, consisted of a series of tag- ging experiments conducted under postencroach- ment conditions. These experiments were so designed that the results would be comparable to results available from preencroachment tag- ging in 1954-55 (French and Wahle, 1966).' The field work was conducted by experienced personnel of the National Marine Fisheries Ser- vice (formerly the Bureau of Commercial Fish- eries) under the supervision of the senior author. In this paper we describe tagging experiments at Rock Island Dam in 1964-65 and compare the results to those obtained in the earlier (1954-55) study. The primary purpose is to estimate the differences between the times required for sock- eye and spring chinook salmon (0. tshawytscha) to move from tagging sites below Rock Island Dam to the counting stations in the three (left, center, and right) fish ladders before encroach- ment and after encroachment. EXPERIMENT RATIONALE The basic experimental measures obtained from this type of tagging are (1) elapsed time from the release of tagged fish below Rock Island Dam to the sighting of tagged fish as they passed through a counting station near the top exit of the fish ladders and (2) the percentage of each release group passing over the dam. The elapsed times include (a) the time, if any, required for tagged fish to recover from possible effects of tagging, (b) the time required to locate and enter the fish ladders, and (c) the time required to ascend the ladders. A statistical analysis of the preencroachment tagging was employed to de- termine adequate sample sizes and release * French and Wahle also tagged in 1956, but because the tagged fish were released in a different manner and at different locations than in any other year, the 1956 experiments were excluded from our comparisons. 127 FISHERY BULLETIN: VOL. 70. NO, 1 frequencies needed in postencroachment tagging to be do^f certain of detecting a change of one- half day and 99 -^r certain of detecting a change of a full day in elapsed times, if such changes occurred between 1954-55 and 1964-65. If we assume that the basic condition of the tagged fish and the time required for tagged fish to re- cover from possible effects of tagging did not diflfer significantly between the 1954-55 and 1964-65 experiments, it follows that changes in elapsed times could be attributed to the ability of tagged fish to find and ascend the fishways. The efficiency of the fish-passage system at Rock Is- land Dam could thus be compared under pre- and postencroachment conditions. Although travel times were expected to be the most sensitive measure of encroachment effects, it is obvious that any significant drop in the percentage of tagged fish passing over the dam would indicate severe stress under postencroachment conditions. It might seem unrealistic at first to assume that tagged fish recovered from the possible effects of tagging equally well in the pre- and postencroachment phases of the study. One might expect, for example, that tagged fish were released into faster moving water in 1954-55 and into slower moving water in 1964-65 and that, accordingly, the tagged fish required longer to recover from the effects of tagging in the earlier phase of the study than in the latter. If this were true, we might have ended up measuring diflferences in recovery time of tagged fish rather than differences in the efficiency of the Rock Island Dam fish ladders. Although water velocities were not measured at the release sites, velocities measured in a model of Rock Island Dam (Ward, 1965)' were not uniformly diflferent under postencroachment conditions than under preencroachment condi- tions. In fact, velocities at the measuring point nearest the right-bank release site on the simu- lated model were generally higher after en- croachment than before. On the other hand, at the station closest to the left-bank release site, Ward, David A. 1965. Hydraulic model studies of the Rock Island fish attraction facilities. Wash. State Univ., Pullman, Div. Ind. Res., Inst. Technol Res. Rep. 65/9-4.3. Vol. 1—20 p., 29 fig., Append. I-II; Vol. 11—23 fig. (Processed.) velocities were higher at lower river flows and about the same at higher flows after encroach- ment. Observations made during the 1964-65 exper- iments revealed that large numbers of tagged fish tended to remain close to shore in protective eddies. According to French, tagged fish be- haved similarly during the 1954-55 experiments." These observations tend to support the assump- tion that tagged fish recovered from tagging equally well in 1954-55 and 1964-65. The flooding of the lower portions of the fish ladders at Rock Island Dam by Wanapum Reser- voir was not the only factor aflfecting fish pas- sage that changed between 1954-55 and 1964-65. Riprap was added to the left bank of the river below the dam, and the left and center fish lad- ders were modified extensively. Figures 3 and 4 show Rock Island Dam before and after Wana- pum Reservoir had been filled. New spill pat- terns designed to enhance fish passage were in eflfect throughout most of the 1964 and all of the 1965 tagging. The basic pattern was developed from findings of the 1958 study (Leman and Paulik, 1966) and modified slightly after ex- periments with the model of Rock Island Dam (Ward, 1965, see footnote 5). METHODS AND MATERIALS The basic experimental procedure was as fol- lows: salmon were trapped as they passed over the left ladder, transported to the release sites approximately 300 m below Rock Island Dam on either side of the river, then tagged and re- leased. Fish counters at the dam recorded the tags as the tagged fish passed the counting boards after reascending the ladders. TAGGING Two diflferent traps were used to capture the salmon. Sockeye were captured as they entered a trap placed at the upstream edge of the count- ing board in the left-bank fish ladder. Chinook salmon, which would not enter this trap, had to " Personal communication, Robert R. French, Fi-shery Biologist, Natl. Mar. Fish. Serv., Northwest Fish. Cent., Seattle, Wash. 128 MAJOR and PAULIK: ENCROACHMENT OF WANAPLM DAM RESERVOIR Figure 3. — Rock Island Dam before Wanapum Dam had been built. Figure 4. — Rock Island Dam after Wanapum Dam had been l)ui!t. Note how the rocks and the lower portion of the right-bank fish ladder, visible below the dam in Figure 3, have been inundated in Figure 4. 129 FISHERY BULLETIN: VOL. 70, NO. 1 be taken in a larger floating trap positioned at the upstream end of the ladder. A conventional 1,000-gal (3.79-m3) tank truck transported sockeye salmon in 1964 and 1965 and Chinook salmon in 1965 but was not available for the 1964 chinook salmon experiments. In- stead, we used 1.2 m by 1.2 m by 1.1 m plywood boxes, equipped with aeration systems and mounted on -Vfton (680-kg) trucks. These units were suitable for transporting fish to the adja- cent left-bank release site but inadequate for moving more than seven or eight fish per unit to the opposite bank via Wenatchee, Wash., — a 48-km trip that took about 1/2 hr. Because of this limitation, only one-half as many chinook salmon were released on the right bank as on the left bank in 1964. Each batch of fish liberated was distinctively marked. Several types and colors of tags were used. Petersen plastic disks were used either alone or in combination with plastic bars and vinyl streamers. Nickel pins, inserted through the body just below the dorsal fin, provided the attachment. Tags were applied in pairs, so that the same color and type of tag showed on both sides of the fish. Tagging time seldom exceeded 30 sec per fish. ARTIFICIAL MANIPULATION OF SPILL PATTERN, AUGUST 3-5, 1964 The spill pattern throughout most of the 1964 and 1965 tagging was developed from results of experiments at Rock Island Dam in 1958 (Leman and Paulik, 1966) with subsequent refinements from a model study in 1964 and 1965 (Ward, 1965, see footnote 5). On August 3-5, 1964, however, gates 16 to 18 (adjacent to the center ladder) were closed and gates 34, 36, and 37 (adjacent to the right-bank ladder) were opened. This departure from the recommended spill pat- tern was undertaken to measure its eflfect on the passage of tagged fish over the dam. TAG OBSERVATION AND DATA RECORDING Four steps were taken to insure the accuracy of the tag observations: First, hydroscopes (floating "windows") were installed over the counting boards on the right- and left-bank lad- ders to suppress glare and surface disturbance. Second, all fish counters were tested for color blindness. Third, samples of tags were mounted on the tally boards to facilitate instant recogni- tion and recording of the tags. Fourth, fish counters were systematically rotated between ladders to distribute any bias by the counters be- tween the ladders. The gates in the fish ladders were open and the passing fish were counted 16 hr a day — 5:00 AM to 9:00 PM — during these experiments. The half-day units used to measure travel times were adapted to the counting schedule. Fish observed during the same 8-hr period in which they were released were assigned a travel time of one-quarter day or 0.5 half-day. Fish re- leased just before noon (as most were) were given a travel time of 0.5 half-day, if observed the same day, or a time of 1.0 half-day if ob- served the next morning. Tag observations were grouped by tag combination, ladder, and travel time in half days. Data were punched on IBM cards — one card containing the release and recovery data for each fish. The numbers of tagged salmon released below Rock Island Dam and later observed passing the dam in 1954, 1955, 1964, and 1965 are summarized in Table 1. Table 1. — Numbers of salmon that were tagged and re- leased below Rock Island Dam in 1954, 1955, 1964, and 1965, and the numbers and percentages of tagged fish that were later observed passing over the dam's fish ladders.^ Species of salmon Year and ^released l^'elow Tagged fish observed dam passing dam no. no. % Spring chinook 1954 155 60 38.7 1955 157 94 59.9 1964 103 93 90.3 1965 311 285 91.6 Sockeye 1954 1,485 1,176 79.2 1955 1,176 793 67.4 1964 951 895 94.1 1965 679 623 91.8 ' The numbers of spring chinook salmon released in 1954 and 1955 differ from those reported by French and Wahle (1966). They used the July 13 date suggested by Fish and Hanavan (1948) as the termination of the spring run and the beginning of the summer run. We used the scale method described by Koo and Isarankura (1967) to determine that the dotes of least overlap between the two races were July 30, 1954, and July 8, 1955. We believe that our separations, based on the more recent study, are the more accurate. 130 MAJOR and PAULIK: ENCROACHMENT OF WANAPUM DAM RESERVOIR TAG OBSERVATIONS AT ROCK ISLAND DAM, 1964 AND 1965 The 1964 and 1965 data are treated by spe- cies — spring Chinook salmon first, followed by sockeye salmon. In a later section these data will be compared to the 1954 and 1955 results to determine the effects of encroachment of Wanapum Reservoir on fish passage at Rock Island Dam. SPRING CHINOOK SALMON The tag release and tag observation data for 1964 and 1965 are presented in Table 2. In- cluded are the date and location of release, the number of fish tagged, the release area, and the number and mean travel time of tagged fish sub- sequently observed passing Rock Island Dam. Logarithmically transformed data are used throughout this paper for analysis of travel time; means are geometric means. Three times as many chinook (311) were tagged in 1965 as in 1964 (103). The small number released and the short duration of the tagging period in 1964 (May 19-27) can be at- tributed to difficulties in completing the access roads, difficulties with the trapping and trans- portation systems, and a shortage of fish in the left-bank ladder — the ladder where the trap was located. Tagging was from May 16 to June 9, 1965. The 1965 data are grouped, somewhat arbitrarily, into five time periods. Percentage Observed The overall percentages of tagged chinook salmon subsequently observed passing over Rock Island Dam were 90.3 in 1964 and 91.6 in 1965. Variability among the release groups was high, ranging from 67 to 129^; in 1964 and 60 to 115% in 1965. It is noteworthy that the number ob- served exceeded the number released for 6 of the 24 releases over the two tagging seasons. Table 2. — Numbers of chinook salmon that were tagged, released below Rock Island Dam and the numbers and mean travel times of tagged fish that were later observed passing over the dam's fish ladders, 1064 and 1965. Nurr iber and mean travel t ime in half-days of tagged fish observed possi ng dam Number of f ish tagged Period and 1 released be. low dam Left ladder Center ladder Right ladder Total' Year Dote Lefl bank Ri ght bank No. Travel tii me No. Travel time No. Travel time No. Travel time 1964 May 19 17 10 14.8 5 16.8 7 8.5 22 12.8 May 19 ._ 15 6 6.0 6 8.7 1 4.0 13 6.9 May 20 13 __ 7 13.2 6 13.4 2 14.8 15 13.5 May 20 __ 17 6 18.1 3 5.3 4 6.6 13 10.0 May 21 27 __ 9 10.5 7 4.2 4 8.1 20 7.2 May 21 5 1 7.0 3 12.6 __ 4 10.9 May 27 9 — 4 9.9 -- 2 8.7 6 9.5 1965 May 16 5 1 2.5 3 3.4 1 2.5 5 3.0 May 18 5 __ 2 11.5 2 4.0 1 2.0 5 5.3 1 May 21 19 12 6.8 4 2.7 3 4.0 19 5.1 May 22 41 — 33 5.4 9 15.8 5 5.7 47 6.6 May 23 11 8 7.0 3 5.2 - — 11 6.4 May 24 9 __ 4 4.7 __ 2 2.5 6 3.8 II May 24 __ 25 17 7.7 6 4.7 3 2.5 26 6.0 May 25 8 _. 5 3.1 2 22.5 2 12.5 9 6.6 May 26 15 — 8 4.8 — 1 0.5 9 3.7 May 27 30 23 8.2 4 5.9 5 3.9 32 7.0 III May 28 28 — 16 9.1 4 15.5 5 3.7 25 8.3 IV May 31 21 10 6.7 2 5.2 2 6.5 14 6.5 June 1 __ 27 14 6.4 4 4.1 4 2.4 22 4.9 June 4 10 5 2.7 3 8.0 ._ 8 4.1 V June 4 22 11 6.9 2 6.6 5 2.3 18 5.0 June 8 22 12 3.9 3 5.1 5 3.2 20 3.9 VI June 9 13 — 8 6.5 1 20.5 — 9 7.3 1 The total number observed ma) ,' exceed th le J lumber tagc ,ed. See text (p. 132) for expl anation. 131 FISHERY BULLETIN: VOL. 70, NO. 1 Possible explanations include the misidentifica- tion of tagfs by the counters and multiple obser- vations of the same tagged fish that passed over the dam. fell back, and survived to pass over the dam ajrain. The "falling: back" of salmon over dams is a frequent occurrence on the Columbia River (Johnson, 1965),' although recent studies^ have shown the magnitude of such fallback is not large (usually less than 5'^r ). The i^erceiitages of tagged fish recovered by release location (releases pooled by location within years) were not consistent for the 2 years. In 1964 the percentage of tagged fish released on the left bank and subsequently observed ex- ceeded that of the right bank— 95.5 to 81.1. In 1965 the comparable percentages were 88.4 and 94.9, respectively. Distribution by Ladder Of the tagged fish sighted in 1964, nearly one- half (46.2 "^r ) chose the left ladder; the center and right ladders lured 32.3 and 21.5^/r , respec- tively. In 1965, 66.3 ^r chose the left ladder, 18.2 vr the center ladder, and 15.4% the right ladder. Distribution between ladders was basic- ally the same for each release site within, but not between, years. The percentages observed in the left ladder were 47.6 and 43.3 for the left- and right-bank releases in 1964 but con- siderably higher (67.2 and 65.5) for fish re- leased from the left and right banks, respective- ly, in 1965. However, in both years similar per- centages of the fish not using the left ladder chose the center ladder (60.0 in 1964 and 54.1 in 1965) . Between-period comparisons are possible for 1965 only when the percentage of tagged fish using the left-bank ladder remained very con- sistent from i)eriod to period, varying only from 61.5 to 69.0. ' Johnson, James H. 1965. Fal'.back of adult chinook salmon at Ice Harbor Dam spillway, May 1964. Final Report to U.S. Army Corps of Engineers for Research Contract No. DA-45-164-CIVENG-63-286. Bur. Commer. Fish., Fi.sh-Passage Research Program, Seattle, Wash., 16 p. (Processed.) * Personal communication with Charles Junge of the Oregon Fish Commission with regard to experiments with tagged chinook salmon at Bonneville Dam during 1970. Travel Time from Release to Observation in Fish Ladders Travel times — by date of release, release lo- cation, and ladder in which the tagged spring chinook salmon were sighted — are presented in Table 2. Because of the small numbers of fish involved in the 1964 tests, their value is limited. The 1965 experiments provided the most sen- sitive analysis of the time required for tagged fish to pass over Rock Island Dam under en- croachment conditions. Results of analysis of variance tests of the hypothesis of no difference in mean travel time between fish released on the right and left banks in 1965 are summarized in Table 3. Regardless of how the data were grouped — whether travel times of the right-bank and the left-bank re- leases were compared period by period, whether adjacent periods were combined, or whether all periods were pooled — no statistically significant diff"erences were found. It is noteworthy, how- ever, that the mean passage time for fish released from the right-bank site was less than for fish released at the left-bank site. Thus, fish released from the right bank were finding and passing over the ladders at least as fast as, if not faster than, those released from the left bank. Table 3. — Analysis of variance tests of the hypothesis that spring chinook salmon, tagged and released on the left bank below Rock Island Dam in 1965, traveled over the Rock Island Dam fish ladders equally as fast as fish released on the right bank below the dam. Periods Mean travel tinne In half-days, all ladders combined (1 f-statistics Left-bank Right-bank and 268 df) releases releases 1 6.081 5.130 0.32 N.S.i II 4.630 6.159 0.93 N.S. III 8.249 7.009 0.29 N.S. IV 6.454 4.940 0.48 N.S. V 4.054 5.048 0.21 N.S. VI II 7.342 3.855 2.03 N.S. {1 /"-statistics and 279 df) 1 and 5.609 5.789 0.03 N.S. III and IV 7.553 6.078 0.84 N.S. V and VI 5.553 4.380 0.50 N.S. (1 /•'-statistics and 283 df) l-VI 6.097 5.486 0.63 N.S. ' N.S. = Not significant. 132 MAJOR and PAULIK: ENCROACHMENT OF WANAPUM DAM RESERVOIR Table 4. — Analysis of variance tests of the hypothesis that spring chinook salmon, tagged and released below Rock Island Dam in 1965, traveled over the right fish ladder at Rock Island Dam equally as fast as those traveling over the left and center ladders. Ladders Mean trave release 1 time in ho areas combi ilf days, ined /■-statistics' compared Left ladder Center ladder Right ladder (1 and 268 df) Right vs. left Right vs. center 6.210 6.929 3.396 3.396 10.60** 9.86** 1 = Highly significant at the 0.01 level, reject hypothesis of equal travel times, and conclude that travel time was significantly less through the right ladder. We also analyzed the 1965 data on a ladder by ladder basis. Although the data, even when compared on a period by period basis, are too limited to provide a sensitive comparison of the mean passage times between fish using the dif- ferent ladders, they did reveal that in every peri- od fish moved over the right ladder faster than over either the center or left ladder. Mean pas- sage times for right ladder versus the left ladder and right ladder versus the center ladder (all periods and both release sites pooled) are com- pared in Table 4." The data have been adjusted for simultaneous tests according to the method described by Dunn (1961). The diff"erences shown in Table 4 are highly significant. Thus the hypothesis that spring chinook salmon trav- eled from the tagging sites over the right ladder equally as fast as over the left ladder is strongly rejected as is the same hypothesis for the right versus the center ladder. In both cases the mean passage times are significantly less for fish using the right ladder. This means that ladder choice was the only observed factor clearly affecting passage time. Similar trends were noted in 1964. Mean passage time over the right ladder was less than for either the left or center ladders. The overall travel time in 1965 (5.8 half-days) was shown by an analysis of variance test to be significantly less than that of 1964 (9.8 half- days) . The F-value for this test was 17.72 with 1 and 165 degrees of freedom. * Because the passage of fish over the right-bank fish ladder had been a source of controversy among fishery biologists, we directed special attention to the right-bank ladder in the present study. SOCKEYE SALMON Fewer sockeye salmon were tagged in 1965 than in 1964. Analysis of the 1964 data revealed that the precision desired could still be achieved if sample sizes were reduced from 100 to 75 fish per release in each year. The tagging season was divided into five periods. With one excep- tion, each of these periods contained releases from the left and right banks. During period IV in 1965, both releases were from the left bank. Tagging was from July 15 to August 5 in 1964 and from July 14 to August 4 in 1965. Tag release and tag observation data are presented in Table 5. Percentage Observed The percentages of tagged sockeye from indi- vidual releases observed passing Rock Island Dam were similar for 1964 and 1965. The per- centage ranged from 84.3 to 120.8 in 1964 and from 81.2 to 98.8 in 1965. Overall, 94.1% of the tagged fish were observed in 1964 and 91.8% in 1965. Percentages observed from left-bank releases were not significantly diflferent from those released on the right bank in either year — 94.8 (left bank) versus 93.3 (right bank) in 1964 and 91.5 (left bank) versus 92.0 (right bank) in 1965. Distribution by Ladder Distribution by ladder was similar in both years. Of the tagged fish sighted, 55.6% used the left ladder in 1964 and 53.1% in 1965. The center ladder took 20.7 and 23.6% and the right ladder 23.7 and 23.3% in the 2 years, respec- tively. For the left- and right-bank releases, respectively, the percentages using the left lad- der were 54.1 and 57.4 in 1964. Comparable percentages were 55.4 and 50.7 in 1965. Fish not choosing the left ladder were fairly evenly distributed between the center and right ladders in both years. The between-period consistency of the per- centage of tagged sockeye salmon using the pre- ferred left ladder was less evident than for chinook salmon. For sockeye salmon, the 133 FISHERY BULLETIN: VOL. 70, NO. I Table 5. — Numbers of mean travel times of sockeye salmon that were tagged, released below Rock Island Dam and the numbers, and tagged fish that were later observed passing over the dam's fish ladders, 1964 and 1965. Period Dote Number of f and released ■ish bel tagged ow dam N umber and mean travel 1 fish observed time in passing half-days of dam tagged Left ladder Cent er ladder Right ladder Totali Year Left bank Right bank No. Travel time No. Travel time No. Travel time No. Travel time 1964 1 July 15 July 16 July 17 76 24 113 21 16 48 2.7 2.2 5.9 16 9 29 2.2 1.9 4.8 40 7 29 2.4 2.0 6.8 77 29 106 2.4 2.1 5.8 II July 21 July 22 July 23 104 70 89 35 59 48 4.9 3.4 5.7 8 16 20 4.9 2.8 3.3 16 15 16 3.0 4.1 4.8 59 90 84 4.3 3.4 4.8 III July 28 July 30 92 80 45 56 1.6 5.0 25 20 1.3 3.3 19 11 3.2 5.5 89 87 1.7 4.6 IV July 31 August 3 85 70 — 53 38 2.3 4.4 16 10 3.4 2.8 13 15 3.3 5.6 82 63 2.7 4.4 V August 4 August 5 52 96 53 26 4.8 4.1 13 6 3.3 2.8 16 15 5.6 3.1 82 47 4.6 3.6 1965 1 July 14 July 15 71 68 38 38 3.1 4.2 8 9 3.8 2.3 23 20 2.4 3.5 69 67 2.9 3.7 II July 19 July 20 81 76 51 43 4.6 4.1 7 6 4.2 2.9 22 23 4.7 2.9 80 72 4.6 3.5 III July 27 July 28 75 75 36 28 3.7 2.5 19 28 2.8 2.1 13 16 3.4 1.9 68 72 3.4 22 IV July 29 July 30 69 66 25 25 4.3 2.0 18 19 4.5 2.7 13 13 4.1 2.6 56 57 4.3 2.4 V August 3 August 4 60 38 29 18 2.0 3.1 19 14 3.0 3.5 1 1 2.5 2.0 49 33 2.4 3.3 1 The total number observed may exceed the number tagged. See text (p. 132) for explanation. percentage varied from 40.1 to 62.8 in 1964 and from 44.2 to 61.8 in 1965. For chinook salmon, the range was 61.5 to 69.0 ''r in 1965, the only year in which adequate data were obtained. Travel Time from Release to Observation in Fish Ladders Travel time by date of release, release location, and ladder for the 1964 and 1965 experiments are presented in Table 5. Following analysis of the 1964 data, we will examine the 1965 exper- iments (p. 135). Analysis of variance tests of the hypothesis of no difference in travel time between left- and right-bank releases (all ladders combined) are given in Table 6. Because there were no re- leases from the right bank in period IV, the test compares the two left-bank releases. Within- release group variances were pooled to form an overall pooled estimate of the variance with 883 degrees of freedom. Because the mean travel times did not differ significantly between the Table 6. — Analysis of variance tests of the hypothesis that sockeye salmon, tagged and released on the left bank below Rock Island Dam in 1964, traveled over the Rock Island Dam fish ladders equally as fast as fish released on the right bank; the testing period IV involves two left-bank releases. Period Release locations compared Mean travel time in half-days, all ladders combined f-stotisticsi (1 and 883 df) Left bank Right bonk 1 Left vs. right 2.331 5.787 72.95 •• II Left vs. right 3.394 4.617 8.71 ** III Left vs. right 1.746 4.619 69.38 ** IV Left (July 31) vs. 2.651 14.55" Left (Aug. 3) 4.350 -- V Left vs. right 4.628 3.604 3.11 N.S. 1 «♦ = Highly significant at the 0.01 level, reject hypothesis of equal travel times, and conclude that travel time was significantly less for fish released from one bank than for those released on the other bank. N.S. = Not significant at the 0.05 level, accept hypothesis of equal travel times. July 15 and 16 left-bank releases, these releases were combined. The July 21 and 23 right-bank releases were similarly tested and combined. Note that in periods I to III, travel times for the left-bank releases were significantly less than 134 MAJOR and PAULIK: ENCROACHMENT OF WANAPUM DAM RESERVOIR for the right-bank releases. In period IV when both releases were from the left bank, the travel time of fish released on August 3 was signifi- cantly higher than that of fish released on July 31. Differences were not significant in period V when fish were released from both left and right banks. There is little doubt then that in periods I to III (when the spill pattern recommended by the study group was in operation) , fish released from the left bank were finding the ladders and trav- eling over the dam faster than those released on the right bank. During the last two periods (V and VI), the flows were intentionally switched from the center to the right and back again. The effects of this change on fish passage are discussed in greater detail in a later section. Further examination of Table 5 reveals two additional characteristics about the movement of tagged sockeye salmon over Rock Island Dam in 1964. First is the consistency of the relative passage times between ladders for fish from a given tag release. This means that regardless of how rapidly or slowly fish from a particular release moved, they did so more or less uniformly at all three ladders. Spring chinook salmon (Table 2) varied much more than did the sock- eye in this respect. Second, in five of six com- parisons fish released on the left bank negotiated the right-bank fish ladder faster than did fish released on the right bank. Considered jointly, these two features suggest that sockeye were capable of rapid lateral movement in the area downstream from the dam and that the passage of fish released on the right bank was somehow delayed whether they chose the left, center, or even the adjacent right-bank ladder. Table 7 depicts the travel times by period and ladder, ignoring release sites. Corresponding statistical tests of the hypothesis of no difference in travel times between the right and center and between right and left fish ladders are included. For these tests, a pooled estimate of the error variance with 880 degrees of freedom was com- puted from within-ladder variances for the 15 groups. Because we tested left versus right and center versus right ladders simultaneously, using the same within-period data, we modified the ^-test to control the type I error according to the method suggested by Dunn (1961). Only one difference is significant. In period III, the mean passage time through the center ladder was less than through the right ladder. In general, however, passage time does not ap- pear to be influenced by the ladder chosen. Analysis of variance tests of the hypothesis of no difference in travel time between fish released on the right and left banks (all ladders com- bined) in 1965 are presented in Table 8. The conclusions from these tests are mixed. Fish traveled over the dam faster from the left-bank release site than from the right-bank release Table 7. — Analysis of variance tests of the hypothesis that sockeye salmon, tagged and relased below Rock Island Dam in 1964, traveled over the right fish ladder at Rock Island Dam equally as fast as those traveling over the left and center ladders. Mean trove 1 time in half-d( 3ys, Period Ladders _ compared release areas combined — <-statistlcsi Degrees of freedom Left ladder Center ladder Right ladder 1 Left vs. right 4.018 3.530 0.981 N.S. (1, 159) Center vs. right — 3.358 3.530 -0.329 N.S. (1, 125) II Left vs. right 4.453 3.887 0.967 N.S. (1, 187) Center vs. right — 3.324 3.887 -0.893 N.S. (1, 89) III Left vs. right 2.989 __ 3.874 -1.494 N.S. (1, 129) Center vs. right — 2.014 3.874 -3.321 * (1, 73) IV Left vs. right 3.042 4.399 -2.043 N.S. (1, 117) Center vs. right — 3.155 4.399 -1.461 N.S. (1, 52) V Left vs. right 4.547 4.211 0.434 N.S. (1, 108) Center vs. right — 3.132 4.211 -1.215 N.S. (I, 48) ' N.S. = Not significant at the 0.05 level, accept hypothesis of equal travel times. * = Significant at the 0.05 level, reject hypothesis of equal travel times, and conclude that travel time through the center ladder was significantly less than through the right ladder. 135 FISHERY BULLETIN: VOL. 70, NO. 1 Table 8. — Analysis of variance tests of the hypothesis that sockeye salmon, tagged and released on the left bank below Rock Island Dam in 1965, traveled over the Rock Island Dam fish ladders equally as fast as fish released on the right bank. Mean travel time in half-days. Period all ladc lers combined /■-statistics' (1 and 613 df) Left-bank Right-bank releases releases 1 2.943 3.674 3.239 N.S. II 4.589 3.543 4.915 • III 3.374 2.182 12.867 *♦ IV 2.363 4.319 25.511 *• V 2.368 3.251 3.841 • 1 N.S. = Not significant at the 0.05 level, accept hypothesis of equal travel times. , . , i i ^■ * = Significant at the 0.05 level, reject hypothesis of equal travel times, and conclude that travel time for fish released on one bank was signifi- cantly less than for fish released on the other bank. •* = Significant at the 0.01 level, reject hypothesis, and conclude as above. site in period I but not significantly so. In peri- ods II and III, fish released on the right bank moved over the dam significantly faster than their left-bank counterparts. In periods IV and V, statistically significant differences were found only in the other direction, e.g., left-bank re- leases were faster than right-bank releases. Thus, there is no clear superiority of one re- lease location over the other. The effect of re- lease location on relative and absolute travel times changed from period to period. In con- trast, tagged sockeye salmon released from the left-bank site in 1964, before the spill was in- tentionally modified, moved past the dam faster than their right-bank counterparts. The main difference between the 2 years seems to be the decreased jiassage time for the right-bank re- leases of 1965 which, for every comparable peri- od, moved over the dam faster than their 1964 counterparts. Overall travel times (3.6 half-days in 1964 and 3.2 half-days in 1965) did not differ significantly despite the better performance by the right-bank releases. Next, it is appropriate to examine the effect of ladder choice on mean travel time in 1965 by period and with release areas pooled. The basic data and the corresponding tests of sig- nificance are given in Table 9. No significant differences were found. Ladder choice did not appear to influence travel time. The same re- sult was noted in 1964. SPILL PATTERN MANIPULATION On August 3, 4, and 5, 1964, spill was shifted from gates adjacent to the center ladder to the gate on the far right side of the dam. During this 3-day period, two groups of tagged fish (Au- gust 3 and 5) were released from the left bank and one (August 4) from the right bank. We will consider the left-bank releases first. The re- lease of August 3 was subjected to 3 days of the modified spill condition, whereas the release of August 5 was subjected to 1 day of the same condition. The left-bank release of July 31 pro- vided a crude "control" (no eflfect of modified Table 9.— Analysis of variance tests of the hypothesis that sockeye salmon, tagged and released below Rock Island Dam in 1965, traveled over the right fish ladder at Rock Island Dam equally as fast as those using the left and center ladders. Period Ladders compared Mean travel time in half-days, release areas combined Lefl ladder Center ladder Right ladder /"-statistics^ Degrees of freedom 1 Right vs. left 3.642 Right vs. center — II Right vs. left 4.354 Right vs. center -- III Right vs. left 3.096 Right vs. center -- IV Right vs. left 2.946 Right vs. center -- V Right vs. left 2.400 Right vs. center __ 2.906 3.577 2.340 3.464 3.201 2.868 3.051 N.S. 117) 2.868 0.004 N.S. 58) 3.639 1.903 N.S. 137) 3.639 0.006 N.S. 56) 2.502 1.766 N.S. 91) 2.502 0.155 N.S. 74) 3.288 0.402 N.S. 74) 3.288 0.080 N.S. 71) 2.236 0.019 N.S. 47) 2,236 0.472 N.S. 33) N.S. = Not significant at the 0.05 level, accept hypothesis of equal travel times. 136 MAJOR and PAULIK: ENCROACHMENT OF WANAPUM DAM RESERVOIR spill) for the two "experimental" releases on August 3 and 5. The percentag-es of the tagged fish observed in the right ladder were 15.9, 23.8, and 31.9 for the releases of July 31 (0-day modified spill), August 3 (3-day modified spill), and August 5 (1-day modified spill), respectively. Compara- ble percentages were 19.5, 15.9, and 12.8 for the center ladder and 64.6, 60.3, and 55.3 for the left ladder. Travel times for the three releases averaged 2.7, 4.4, and 3.6 half-days respectively. Thus, it appears that spilling from the right side tended to attract fish released on the left bank to the right ladder but at the expense of in- creasing overall travel time. The change in the spill pattern did not have a significant effect on tagged fish released on the right bank. For the July 30 "control" re- lease, mean travel time was 4.6 half-days; for the August 4 "experimental" release it was 4.6 half-days. Slightly more fish from the August 4 release (19.5^0 were attracted to the right lad- der than from the July 30 release (12.69r). Most striking is the similarity between the left-bank release of August 3 and the right-bank release of August 4. The percentages of tagged fish using various ladders for the releases of August 3 and 4 were: (see Table 5) left, 60.3 and 64.6; center, 15.9 and 15.9; and right, 23.8 and 19.5. Overall travel times were 4.4 and 4.6 half-days. Travel times by ladder were similar —4.4 and 4.8 half-days for the left, 2.8 and 3.3 for the center, and 5.6 and 5.6 for the right ladder. In summary, the departure from the basic spill pattern tended to attract fish to the right-bank ladder, especially those released on the left bank and in so doing, increased the overall travel time. These experiments support the eflScacy of the basic spill pattern as compared to the other pat- tern tested. COMPARISON OF PREENCROACH- MENT AND POSTENCROACHMENT TAGGING STUDIES The effect of the encroachment of Wanapum Reservoir on fish passage at Rock Island Dam is best measured by comparing the results of the pre- and postencroachment tagging studies. We shall consider spring chinook salmon first, followed by sockeye salmon. Three measure- ments — percentage observed, distribution by ladder, and travel time — provide the basis of our analysis. SPRING CHINOOK SALMON The results of the 1954 and 1955 tagging studies with spring chinook salmon are presented in Tables 10 and 11. Comparable data for 1964 and 1965 are in Table 2. Percentage Observed The overall percentages of tagged spring chi- nook salmon observed passing Rock Island Dam were 38.7 and 59.9 in 1954 and 1955; they were 90.3 and 91.6 in 1964 and 1965. Although some of the significant increase may represent better tag retention or increased survival brought about by improved conditions for fish passage during the postencroachment study, it is likely that the precautions we took to improve the tag observations also were important. It is interesting to note that the greatest in- crease occurred for fish released from the right- bank site. In 1954 and 1955, sightings from right-bank releases were only 37.5 and 51.9%, whereas in the postencroachment years — 1964 and 1965— they were 81.1 and 94.9%. Few fish were released on the left bank in 1954, but the increase in the percentage of tagged fish ob- served for the other years (from 77.6 in 1955 to 95.5 and 88.4 in 1964 and 1965) while sig- nificant, is not as dramatic as for the right-bank releases. Distribution by Ladder The percentages of tagged spring chinook salmon in the left-bank fish ladder were 61.7, 74.5, 46.2, and 66.3 in 1954, 1955, 1964, and 1965, respectively. For the right ladder the percent- ages were 25.0, 16.0, 21.5, and 15.4 for the 4 years, respectively. This means that 13.3% used the center ladder in 1954, 9.6% in 1955, 32.3% in 1964, and 18.2% in 1965. Thus, there was no 187 FISHERY BULLETIN: VOL. 70, NO. 1 Table 10. — Tag release and observation data for spring chinook salmon seen passing over fish ladders at Rock Island Dam, 1954 and 1955.' Dote Tagging location Number of fisli fagged and released Number of fagged fish observed Left ladder Center ladder Right ladder Total 1954 June 23-25 Juno 29-July 2 July 7-9 July 13 July 16-22 July 29-30 1955 June 7-9 June 14-15 June 17-21 June 28-29 June 30-July 1 July 5-6 July 7-8 Right bonk Right bonk Right bank Right bank Left bank Right bank Left bank Right bank Right bank Left bank Right bank Right bank Right bank Left bank Right bank 26 36 57 26 3 7 8 14 13 9 19 15 19 17 43 6 5 7 15 3 1 3 7 7 10 6 10 3 8 16 3 2 1 2 1 1 1 2 3 1 7 2 5 1 1 1 2 2 2 2 1 4 ily 7-8 Right bank 43 16 1 4 ' The total number observed may exceed the number tagged. See text (p. 132) for explanation. 16 9 13 17 3 2 4 9 10 12 9 10 7 12 21 Table 11. — Travel time of tagged spring chinook salmon from tagging areas below Rock Island Dam to the Rock Island fish ladders. 1954 and 1955. Date 1954 Tagging location Mean travel time in half-days, all ladders combined June 23-25 Right bank 12.3 June 29-July 2 Right bank 7.4 July 7-9 Right bank 14.9 July 13 Right bank 13.8 July 16-22 Left bank 15.8 July 29-30 Right bank 10.6 1955 June 7-9 Left bank 6.5 Right bonk 19.4 Juno 14-15 Right bank 12.4 June 17-21 Left bank 24.2 Right bank 21.8 June 28-29 Left bank 20.3 June 30-July I Right bank 9.2 July 5-6 Left bank 19.6 July 7-8 Right bank 12.8 marked and repeatable difference in distribution by ladder between the postencroachment years, 1964 and 1965, and the preencroachment years, 1954 and 1955. The center ladder took a dis- proportionate share of the fish in 1964 (at the expense of the left ladder), but this was less pronounced in 1965. Travel Time from Release to Observation in Fish Ladders Apparently no large-scale mortalities and no great losses of tags were caused by encroachment (these assumptions are supported by the high percentages of tagged fish subsequently seen passing Rock Island Dam). The overall eflfect of encroachment is then best measured by com- paring the travel times between the pre- and postencroachment tagging studies. Although we have already shown that the 1965 travel time (5.8 half-days) was significantly less than the 1964 travel time (9.0 half-days), this difference does not overshadow the fact that both values are well below the comparable figures (12.4 and 16.0 half-days) for 1954 and 1955, respectively. We can only conclude that travel time of spring chinook has decreased markedly since encroach- ment. SOCKEYE SALMON Results of the 1954 and 1955 tagging studies with sockeye salmon are given in Table 12. Com- parable data for 1964 and 1965 are presented in Table 5. 138 MAJOR and PAULIK: ENCROACHMENT OF WANAPUM DAM RESERVOIR Table 12. — Tag releases, tag observations, and travel times for sockeye salmon seen passing over fish ladders at Rock Island Dam, 1954 and 1955. Period Date Tagging location July 7-8 Right bank July 16, 20 Left bank July 21-22 Right bank July 22-23 Left bonk July 27-29 Right bank July 28-30 Left bank August 3 Right bank August 6-12 Left bank August 5-12 Right bank July 19-20 Right bank July 21 Left bank July 22 Right bank July 26-29 Left bonk July 26-28 Right bank August 2-3 Left bank August 2 Right bank August 5-10 Left bank August 4-1 1 Right bonk Number of fish togged and released Nunnber of tagged fish observed' Mean travel time in half-days 22 23.3 168 12.5 151 10.1 196 8.4 75 7.7 111 7.3 80 6.6 128 4.8 245 6.4 118 7.3 46 7.0 6 14.9 224 7.0 151 5.2 119 7.1 12 3.8 38 7.1 79 7.5 1954 IV 1955 III IV 22 119 155 272 146 246 89 174 262 123 59 46 298 227 129 24 93 177 The number observed may exceed the number tagged. See text (p. 132) for explanation. Percentage Observed The overall percentages of tagged sockeye salmon observed passing Rock Island Dam in 1964 and 1965 (94.1 and 91.8, respectively) were significantly higher than those recorded in 1954 and 1955 (79.2 and 67.4). A similar change was noted for chinook salmon. As we mentioned earlier in discussing the results with chinook salmon, two factors — increased tag retention and improved facilities for observing and reporting tagged fish — probably contributed to the in- creased percentages of tagged fish that were observed. Unlike chinook salmon, for which a greater part of the increased percentage of fish sighted could be attributed to fish released on the right bank, the improvement in sockeye salmon was of the same magnitude for releases made on both banks. Distribution by Ladder Percentages of tagged sockeye salmon in the left ladder were 55.3, 64.3, 55.6, and 53.1 in 1954, 1955, 1964, and 1965, respectively. For the right ladder, the percentages were 13.2, 12.4, 23.7, and 23.3; for the center ladder they were 31.5, 23.3, 20.7, and 23.6. As with chinook salm- on then, the distribution of tagged sockeye salmon by ladder in postencroachment years, 1964 and 1965, did not differ significantly or at least consistently so from that observed in pre- encroachment years, 1954 and 1955. Travel Time from Release to Observation in Fish Ladders Comparisons of the mean travel times by period and area of release for 1954 versus 1964 and 1965 and for 1955 versus 1964 and 1965 are presented in Figure 5. On only 1 of 34 oc- casions was the travel time in the preencroach- ment tagging year less than for the correspond- ing postencroachment tagging year. The dif- ference was not significant. As with chinook salmon then, we found the time required by tagged sockeye salmon to pass the fish ladders at Rock Island Dam was con- siderably less after the onset of encroachment than before. 139 MAJOR and PAILIK: ENCROACHMENT OK WANAPl M D.VM RESERVOIR 25r LEFT-BANK RELEASE AREA — • 1954 — O 1955 * » 1964 A A 1965 RIGHT- BANK RELEASE AREA I II III IV V RELEASE PERIOD M III IV RELEASE PERIOD Figure 5. — Time required for tagged sockeye salmon to move over the Rock Island Dam fish ladders from the tagging areas below the dam, 1954, 1955, 1964, and 1965. SUMMARY AND CONCLUSIONS The lower portions of the fish ladders at Rock Island Dam were flooded by the reservoir of Wanapum Dam. At the direction of the Federal Power Commission, the fish ladders were mod- ified to maintain or enhance fish passage and a study was developed to evaluate the adequacy of the modifications. In 1964 and 1965 over 2,000 spring- chinook and sockeye salmon were tagged and released below Rock Island Dam; their subsequent move- ment over the fish ladders w^as noted. Three features — travel time, the percentage of tagged fish observed, and the distribution of tagged fish by ladder — were compared with data obtained in a similar tagging study in 1954 and 1955. Results clearly indicate that the fish passage over Rock Island Dam was better in 1964 and 1965 than in 1954 and 1955. Travel times were significantly shorter and higher percentages of tagged fish were sighted passing over the ladders under postencroachment conditions. LITERATURE CITED Dunn, 0. J. 1961. Multiple comparison among means. J. Am. Stat. Assoc. 56(293) : 52-64. Fisn, F. F., AND M. G. Hanavan. 1948. A report upon the Grand Coulee fish-mainte- nance project 1939-1947. U.S. Fish Wildl. Serv., Spec. Sci. Rep. 55, 63 p. French, R. R., and R. J. Wahle. 1966. Study of loss and delay of salmon passing Rock Island Dam, Columbia River, 1954-56. U.S. Fish Wildl. Serv., Fish. Bull. 65: 339-368. Koo, T. S. Y., AND A. Isarankura. 1967. Objective studies of scales of Columbia River chinook sa.\rtwn, Oncorhynchus tshawytsclia (Wal- baum). U.S. Fish Wildl. Serv., Fish. Bull. 66: 165-180. LEMAN, B., AND G. J. Paulik. 1966. Spill pattern manipulation to guide migrant salmon upstream. Trans. Am. P^ish. Soc. 95 : 397-407. Thompson, W. F. 1945. Effect of the obstruction at Hell's Gate on the sockeye salmon of the Fraser River. Int. Pac. Salmon Fish. Comm., Bull. 1, 175 p. 140 SCALE FEATURES OF SOCKEYE SALMON FROM ASIAN AND NORTH AMERICAN COASTAL REGIONS Kenneth H. Mosher' ABSTRACT Photographic plates of sections of sockeye salmon scales, with descriptions, and frequency tables of the number of circuli in the freshwater and first ocean zones illustrate the variations in scale features of fish over the range of the species in coastal regions of Asia and North America. Suggestions are also given for using these data to determine the geographical origin of sockeye salmon taken in offshore areas of the North Pacific Ocean and adjacent waters. Sockeye salmon {Oyicorhynchus nerka) are val- uable food fish of the Bering Sea and the north- ern part of the North Pacific Ocean. They spawn in coastal streams of Asia and North America but spend a portion of their lives feed- ing in oceanic areas. Upon the onset of sexual maturity, they migrate from the ocean, enter their natal streams, spawn, and then die. Be- cause sockeye return to natal streams to spawn, the species is divided into hundreds of individual populations (each from its own geographical area), which are self-reproducing units or "stocks." A major goal in fisheries research and man- agement of the sockeye salmon resource is to obtain enough spawning fish within each stream to provide the maximum catch to the fishery and to insure the perpetuation of each stock. This goal is difficult to attain in fishing areas where management agencies are uncertain of the geo- graphic area of origin of the stocks of fish that are being caught. Consequently, methods for de- termining the area of origin of sockeye salmon taken beyond their natal streams are needed. A number of methods have been used to de- termine the area of origin of sockeye salmon taken in offshore and coastal areas. These in- clude morphological studies (Fukuhara et al, 1962; Landrum and Dark, 1968), parasitologi- cal studies (Margolis, 1963), serological studies ' National Marine Fisheries Service, Northwest Fish- eries Center, 2725 Montlake Boulevard East. Seattle, WA 98102. (Ridgway, Klontz, and Matsumoto, 1962), tag- ging studies (Hartt, 1962, 1966; Kondo et al., 1965), and scale studies (Krogius, 1958; Kubo, 1958'; Kubo and Kosaka, 1959'; Henry, 1961; Mosher, Anas, and Liscom, 1961; and Mosher, 1963, 1968). Scale studies have become one of the most popular and successful methods; scale features, for example, are routinely used by in- vestigators of the International Pacific Salmon Fisheries Commission as one element in a tech- nique to determine the natal streams of sockeye taken near the mouth of the Fraser River and are also routinely used by investigators of the National Marine Fisheries Service (NMFS, formerly the Bureau of Commercial Fisheries) to determine continent of origin of sockeye salm- on taken in the Bering Sea and the central North Pacific Ocean. No detailed information, however, has been published on the variations in scale features among fish from diff'erent spawning regions along the Asian and North American coasts. Krogius (1958) specifically mentioned the need for an atlas illustrating scales from difl["erent Manuscript accepted September 1971. FISHERY BULLETIN: VOL. 70, NO. I, 1972. " Kubo, T. 1958. Study of sockeye salmon stocks by means of the growth pattern of scales (preliminary re- port). Fac. Fish., Hokkaido Univ. (Hakodate). Part I - 15 p. of Japanese text; Part II - 2 pi., 16 fig. in English. (Transl. of Part I, Int. North Pac. Fish. Comm. Doc. 206), 9 p. (Processed.) ^ Kubo, T., and J. Kosaka. 1959. A study of 5,^ age group red salmon stocks by scale growth formula [in Japanese with English abst., headings, tables, and fig. legends.] Suisan cho (Fisheries Agency of Japan), (Int. North Pac. Fish. Comm. Doc. 826), 27 p. (Processed.) 141 FISHERY BULLETIN: VOL. 70, NO. 1 areas and included in her paper were many pictures of scales of sockeye salmon of Asian stocks. A photographic atlas of sockeye salmon scales (Mosher, 1968) was the first step in de- termining racial origins, and it should be avail- able for reference when the present paper is studied. The purpose of this paper is to show varia- tions in age and scale characteristics among adult fish from various coastal areas over the range of the species so that workers planning to col- lect and analyze scale data to determine origin of sockeye taken at sea and in coastal waters are informed about scale features that are linked to various geographic localities. This report consists of two principal parts. The first comprises (1) photographs of sections of scales of adult sockeye salmon, as plates, for each freshwater age group from various areas over the range of the species; (2) frequency tables of the number of circuli in the freshwater and first ocean zones for fish taken from the var- ious areas; and (3) descriptions of the scales of sockeye salmon from the various areas. The second part is concerned with the selection of scale features for the determination of the origin of fish taken in offshore waters. METHODS AND MATERIALS In the preparation of this paper I was con- cerned with (1) the selection of scale samples of fish from various geographical areas, (2) the selection of scale features that are linked to var- ious stocks or geographical areas, and (3) the method of preparation of plates from photo- graphs of selected sockeye scales. I have dis- cussed each of these items separately. SELECTION OF SCALE SAMPLES OF FISH FROM VARIOUS GEOGRAPHICAL AREAS An important consideration in deciding which stocks of sockeye salmon to include in this paper was the relative number of fish produced in the various localities over the range of the species shown in Figure 1. Study of catch data seems to be the best way to determine the most abun- I50°E 60°N — 60°N Figure 1. — Approximate range of sockeye salmon in and around the North Pacific Ocean and adjacent seas. The distribution in the northern Bering and Chukchi Seas was estimated to include the northernmost known spawning streams on both continents. Sockeye salmon may be found in many streams within the range shown, but in only a few streams in some areas. Atkinson et al. (1967) shows detailed maps of streams where sockeye salmon have been known in the United States. The distribution at sea varies within and between years, depending on many factors, Manzer et al. (1965), Hartt (1962, 1966), and Kondo et al. (1965). In addition to the above refer- ences, Hanamura (1966, 1967) and Aro and Shepard (1967) were also sources of data for this figure. dant stocks because the catch is roughly propor- tional to the production of fish in an area. Table 1 shows the average catch for the 3 years, 1966- 68, the statistics for which are complete for Asia Table 1.— Sockeye salmon catch, average of 1966-68. Area Thousands of fish Metric tons Total all areas 27,297.2 70,387.0 Asia 9,512.6 20,972.6 Japan' 8,527.0 17,988.6 USSR 985.6 2,984.0 North America 17,784.6 49,414.4 Canada (British Columbia) 5,676.4 United States 12,108.2 North of Bristol Bay 2.8 Bristol Bay- 5,715.0 Alaska Peninsula^ 1,625.5 Cook Inlet 1,465.7 Copper River area* 832.2 Southeastern Alaska^ 952.1 Washington and Oregon" 1,514.9 ' Japan has no stocks of sockeye salmon. - Includes north side of Alaska Peninsula. ■' Includes Aleutian Islands, south side of Alaska Peninsula, Chignik, and Kodiok Island, ' Includes Resurrection Bay, Prince William Sound, Copper and Bering Rivers, ^ Includes Yakutat. " Includes the Columbia River (32.2 thousand fish). Source of data: International North Pacific Fisheries Commission, 1966, 1967, and 1968; and supplemental catch statistics supplied to the INPFC by the USSR. 142 MOSHER: SCALE FEATURES OF SOCKEYE SALMON and North America from International North Pacific Fisheries Commission (INPFC) sources. The approximate ranking on each continent of the importance of the coastal areas is as follows: Asia (from Hanamiira, 1906 and Krogins, 1958) Ozernaya River 1 Kamchatka River 2 Bolshaya River 3 Paratunka River 4 Apuka River 5 Okhota and Kukhtuy Rivers 6 North America {from catch data of Table 1) Bristol Bay 1 British Columbia 2 Alaska Peninsula 3 Washington and Oregon 4 Cook Inlet 5 Southeastern Alaska 6 Copper River area 7 Columbia River 8 North of Bristol Bay 9 The abundance and catch of sockeye salmon in most areas can fluctuate widely between years, however. Some of these variations in catch are revealed in Appendix Table 1 of the catch of the 5 years, 1964-68, which series includes one high production year in Bristol Bay — 1965. The distribution of spawning streams in Asia extends from approximately lat 66° N near the Anadyr River southward to the tip of the Kam- chatka Peninsula and the Kurile Islands, and westward to the Okhota and Kukhtuy Rivers on the northern coast of the Okhotsk Sea (Hana- mura, 1966, 1967). Berg (1948) indicated that the species was very rare in northern Hokkaido Island. Krogius and Krokhin (1956) concluded that approximately 90 ';r of the total sockeye catch along the Far Eastern Coast of the USSR was produced in the Ozernaya and Kamchatka Rivers of the Kamchatka Peninsula. The distribution of North American spawning streams extends from the Noatak River of Kot- zebue Sound in Northern Alaska, southward to the Columbia River of Oregon and Washington (Aro and Shepard, 1967; Atkinson et al., 1967). The streams can be conveniently grouped into three major geographical areas for study: (1) the Columbia River to and including British Columbia: (2) Bristol Bay, Alaska, and areas north of Bristol Bay; and (3) the area between Bristol Bay and British Columbia. In many years each area contributes about one-third of the North American catch. Normally, the catch of sockeye salmon north of Bristol Bay is insig- nificant in relation to the number of fish taken in the Bay, but consumption and barter of salm- on is substantial, especially by residents along the Kuskokwim River. Thus, plates of representative scales of fish from southern Kamchatka, Bristol Bay, and the areas north of Bristol Bay, central and south- eastern Alaska, British Columbia, and the Co- lumbia River (the coastal areas listed in Table 1) are included in the first part of this paper. The scale samples used in a previous study (Mosher, 1968) with a few samples, which have recently become available from additional areas, were used for this study.' The areas from which these samples were collected are listed in Table 2. Figure 2 shows the approximate location of the areas mentioned in the text and on the plates, SELECTION OF SCALE FEATURES LINKED TO VARIOUS STOCKS OR AREAS My previous paper (Mosher, 1968) shows in detail the features of sockeye salmon scales and the range of variations in many characters. This paper continues the study of sockeye salmon scales to show the relation of many of the var- iations to locality and how these variations in scale characters can be used to identify the main- land origin of sockeye salmon taken in offshore waters. A number of age groups have been found in all populations of sockeye salmon that have been studied. These age groups are based on the number of years the fish lived in fresh water and in the ocean. Over the geographical range of the species, individuals with scales showing freshwater ages from 0. to 4., ocean ages from .1 to .4, and total ages of 0.1 to * Contributions of the following agencies to the salmon scale sampling program are gratefully acknowledged: The Alaska Department of Fish and Game, Juneau, Alaska; the Fisheries Research Board of Canada, Na- naimo, B.C.; the Fisheries Agency of Japan, Tokyo, Japan; and the Fish Commission of Oregon, Portland, Oreg. In addition, special thanks are given to Dr. I. Lagunov of the Pacific Institute of Fisheries Research and Oceanography (TINRO), Petropavlovsk. Kamchat- ka, USSR, who kindly supplied a series of samples from USSR streams. 143 FISHERY BULLETIN: VOL. 70, NO. 1 Table 2. — Geographical areas wliere scales were taken from sockeye salmon. (Scales are available at the Na- tional Marine Fisheries Service, Northwest Fisheries Center, Seattle, Wash.) Asia 11. Alaska Peninsula area!' 16. Ketchikan orea- 1. Kamchatka River A. King Cove A. Portland Canal 2. Paratunka River A. Dalnee Lake B. Chignik C. Kodiak Island B. Moira Sound (1) North Arm 3. B. Blizhnee Lake Bolshaya River (1) Karluk (2) Red River (2) Kegan Creek 4. Ozernaya River (3) Frazer River C. Karta Bay 5. Okhtosk Sea 12. Cook Inlet orea- D. DolomI Lake and Stream 6. North Pacific Ocean and Bering Sea from A. Cook Inlet Fishery E. Hugh Smith Stream areas near the Kamcha tkc 1 Peninsula B. Kenoi River F. Clarence Strait C. Susitno River G. Eek Boy North America (1) Fish Creek H. Hetta Bay Nicholas Bay 7. Nome area A. Salmon Lake (2) Judd Lake (3) Alexander Creek 1. B. Unalakleet River D. Kasilof River J. Nichols Bay 8. Kuskokwim River E. Fish Creek, Knik Arm, Kenai Peninsula K. Klowock Creek 9. Bristol Bay A. Togiok River F. Skilok Lake G. Tustamina Lake (Bear Creek) L. M. Klokas Lake Deweyville B. Nushagak-Wood River System' H. Upper Russian Lake 17. British Columbia C. Kvichak Riveri 13. Copper River" A. Nass River D. Naknek River' A. Haley Creek B. Skeeno River (1) Brooks Lake 14. Yakutat C. Rivers Inlet (2) Branch River A. Situk River D. Smith Inlet E. Egegik River' 15. Petersburg area- E. Alert Bay F. Ugashik River' A. Kasheets F. Nimpkish River G. Bear River B. Stikin© River G. Eraser River 10. Aleutian Islands C. Salmon Boy 18. Columbia River A. Attu D. Tohlton Lake A. Main stem B. Adak E. Red Bay B. Wenatchee River C. Unalaska F. Port Houghton C. Okanogan River ' Samples from the five major Bristol Bay rivers were taken each year since 1954. but there is only one sample for some of the less important areas. - Additional samples were taken, but numbers of scales were small. There ore samples for most other areas for a number of years. ""3.4 occur, but in most localities most fish are in affe groups 1.2, 1.3, 2.2, and 2.3. The number of years spent in fresh water varies within and between many spawning areas and influences many of the scale characters (Mosher, 1963, 1968). Only a few adult fish of age 0. or 4. are found, some fish of age 3. are found in a few areas, but fish of age 1. and 2. are present in substantial numbers in most areas. Table 3 shows the percentage freshwater age composition of sockeye salmon in stream samples (except as noted) from areas around the North Pacific Ocean available at the NMFS Northwest Fisheries Center. Many features of sockeye salmon scales can be used in racial studies. Scientists at the NMFS Northwest Fisheries Center in Se- attle have examined about 50 diff"erent features ■^ Age designation follows the European system, Koo (1962a) : the number of winters the fish spent in fresh water since hatching, a decimal point, and the number of winters the fish spent in the ocean (see Mosher, 19G8, p. 259 and 262). such as counts of circuli, measurement of zones and portions of zones, and various ratios based on these counts and measurements; but only scale characteristics with the greatest differ- ence between Kamchatkan and Bristol Bay stocks have been described in our publications (Mosh- er et al., 1961; Mosher, 1963, 1968; Anas, 1964; Anas and Murai, 1969). In all of our studies — published and unpublished — the best features for racial studies have been in the fresh- water and first ocean zones of the scale. Because of the large number of Asian and North American spawning areas and the large number of age groups in some areas, it is evi- dent that space is not available for examples of scales representative of each area and age group over the geographic range of the species. Scales from numerous areas are similar in many char- acters; consequently I will group together scales from fish of more than one stock or spawning area that have a relatively similar ap])earance. The number of circuli in the freshwater and first ocean zones (counted as indicated on p. 36 144 MOSHER: SCALE FEATURES OF SOCKEYE SALMON 60°N 60°N 50° Figure 2. — Approximate 1. Okhota River 2. Kukhtuy River 3. Okhotsk Sea 4. Bolshaya River 5. Ozernaya River 6. Paratunka River (Dalnee and Blizhnee Lakes) 7. Kamchatka River 8. Apuka River 9. Anadyr River 10. Attu Island 11. Adak Island 12. Unalaska Island location of areas mentioned in the text and on the plates 13. Bear and Sandy Rivers 25. Kodiak Island 14. Ugashik River 26. Cook Inlet 15. Egegik River 27. Copper River area 16. Naknek River 28. Yakutat Bay 17. Kvichak River 29. Petersburg area 18. Nushagak-Wood River 30. Ketchikan area system 31. Nass River 19. Togiak Bay 32. Skeena River 20. Kuskokwim River 33. Rivers Inlet 21. Norton Sound- 34. Smith Inlet Yukon River 35. Nimpkish River 22. Kotzebue Sound 36. Fraser River 23. Noatak River 37. Columbia River 24. Chignik Bay and 37 of Mosher, 1963) are shown for the var- ious areas as frequency tabulations in Tables 4 and 5 for the age 1. fish; in Tables 6 and 7 for the age 2. fish; and in Table 8 for the age 3. fish. Inspection of Tables 4 and 6 shows that the mean number of circuli in the freshwater zone varies among some geographical areas, but that in the first ocean zone (Tables 5 and 7) there is a cline in number of circuli from least in the Adak Island fish to most in the central British Colum- bia areas of Rivers and Smith Inlets and the Nimpkish River. A decrease in the mean num- ber of circuli among stocks from central British Columbia southward to the Columbia River is also found. The Asian fish and those north of Bristol Bay have slightly more circuli, on the av- erage, than those of Adak Island and Bristol Bay. The scales from sockeye taken from certain geographic areas have similar frequency distri- butions of circuli in the freshwater or first ocean zone. These similarities are the basis for di- viding the coast of North America into certain broad areas. When I discuss the various geo- graphic areas, reference will be made to the appropriate frequency table. PREPARATION OF PLATES The scale plates for the report were made as follows: (1) The scale images produced by a scale projector like the one I described (Mosher, 1950) , at 82 x magnification, were photographed with a 35-mm single lens reflex camera on me- dium speed, fine grained film and processed to accentuate the contrast by minimum exposure 145 FISHERY BULLETIN: VOL. 70. NO. 1 Table 3. — Percentage freshwater age composition of samples of sockeye salmon from geographical areas of the North Pacific Ocean from those listed in Table 2. Locality Yeari Freshwater age Number in 0. 1. 2. 3. 4. sompla ~ — _ Percent - Asia* Ozernaya R. 1959 0.8 53.3 45.1 0.8 124 Bolshaya R. 1964 1.2 94.1 3.5 1.2 — 84 Kamchatka R. 1964 2.2 90.2 6.5 1.1 — 92 Parotunka R.: Dalnee Lake 1964 __ 2.9 44.3 52.8 — 70 Blizhnee Lake 1958 0.6 78.5 20.3 0.6 172 Okhotsk Sea 1957 51.5 42.1 6.4 — 501 Off S.E. Kamchatka 1965 — 21.9 59.2 16.9 2.0 201 North America: Kuskokwim 1959 70.5 28.7 0.8 — 122 Bristol Boy: Ugashik R. 1966 39.6 60.4 — 318 Egegik R. 1966 5.9 86.9 7.2 __ 305 Naknek R. 1966 32.0 67.2 0.8 — 356 Kvichak R. 1966 4.0 95.7 0.3 351 Nushagck R. 1966 94.7 5.3 __ — 322 Togiak R. 1955 3.3 80.1 16.6 __ __ 307 Bear R. 1959 __ 1.6 95.0 3.4 119 Aleutian Islands: Attu Isl. 1956 1.3 38.8 52.9 6.6 0.4 227 Adak Isl. 1956 _. 30.6 60.5 8.0 0.9 213 Unalaska Isl. 1956 38.9 37.5 22.9 0.7 144 Alaska Peninsula: King Cove 1957 45.7 52.7 1.6 __ 182 Chignik R. 1957 __ 40.0 59.3 1.7 175 Karluk R. 1959 48.1 47.2 4.7 106 Red River 1959 _. __ 85.1 14.9 107 Cook Inlet 1959 _. 83.9 16.1 118 Cook Inlet Fish Cr. 1959 __ 98.3 1.7 120 Southeastern Alaska: Copper R. 1959 __ 87.7 12.3 __ 122 Yakufat R. 1958 __ 31.5 65.6 2.9 105 Petersburg 1964 4.6 81.8 13.0 0.6 323 Ketchikan 1964 0.3 84.1 15.6 __ 352 British Coiunnbio: Mass R. 1964 35.0 65.0 __ __ 320 Skeena R. 1964 ._ 98.7 1.3 297 Rivers Inlet 1961 __ 98.0 2.0 248 Smith Inlet 1961 _. 100.0 .. __ 255 Nimpkish R. 1967 __ 64.5 30.4 5.1 79 Fraser R. 1964 0.7 98.0 1.3 __ 153 Columbia River 1964 0.2 88.1 11.7 — — 416 ' The year selected is the most complete for that area, is in a series with adjacent areas, or is the only year a sample is available in our series (Table 2). ^ Historical data on age composition of samples from Ozernaya, Bolshaya, and Kamchatka Rivers, and Lake Dalnee for a number of years from 1931 to 1960 are available in Hanamuro (1966). and maximum development (Adams, 1952; Mortensen, 1947). (2) Positive prints of sec- tions that showed the important features of the anterior field of the scales were made from these negatives on high contrast enlarging pa- per. (3) The positive prints of the scale sec- tions were assemV)led by area groups on mount- ing sheets and photographed to provide the plates. The scales were all photographed at the same magnification. Consequently, the rel- ative size of the scale features on the plates reflects the relative size of the scale features themselves. As indicated in my previous paper (Mosher, 1968), the texture, contrast, and distinctness of circuli vary greatly both on individual scales and between the scales of the same and different fish. Some scales can be photographed to show the features clearly; other scales, especially from some localities with many closely spaced and broken circuli, do not provide clear photographs of all features. 146 MOSHER: SCALE FEATURES OF SOCKEYE SALMON o m u >> w 'u > o 'o o C3 CO -^ o o w ID bo c3 o u 0) c o ;-l O) .—I CO CO 3 «H O c _o 3 jQ 'C -1^ to •i-H >. u c Ol 3 cr a> »-i «H (D bo 03 -1^ C O) o ;-! O) PL. ■J 0-" 5^ E 1^ n) "1 iL V ,*_ u >- , 0) a 1) n U u O 0) 0> IE U to D C "D D u- O 2 DO CQ_c — t Oei O D > »0 CN <> O O 00 p "O "O lO p ^ ■^' CO CN O ^o o* CO to CO CO o ^ "vr o 00 CO ' o I lO o ' O CO »0 »0 iq »0 lO O --- Ci CN oi C^^ lO CN Qs O; CN ^ •— »0 >0 to ^ -o cs d CO v6 CN ^ 00 N; o* rx — ' CO "^ -^r <> UO lO o p d — ^ CS CN T t I I rs CO o p d CN r< CO CO^CO^cq—lCO*— CO 00'-;fOrxCMC^CO I dddd — cs'^'o'^ 'Oi^-^dwo-^d ' O"— oi'^o-^^Jcod dcoTir — roco'^Tro^'^o^corvco ■^■^uoo^o^ioiocs'o o.'o^'Oio'^co'^dco<> ^•^ojdddddd uo IT) O "O ^ "0 I I I [ CO lO S3 CO 00 — ' ' ' ' O *— ^ CN O lO'OOiOO'OOpio I * * lo-^wSco-^-^coKco^ d ' — (N CS — o^oo^ooooooio UOOOsQWOCMpCO^ csdco'O'oco — d I I I C'jcoco"^oq»-;pv> I I I I — ^C'JTroov^'O^d ' ' ' ' >OCNlO^C00'0;UOiO S^CM — CSP'— ;0^'OCM0v ' ' ' 'ocof\rx^o^6 "OTtioo^co — 'OUOTt-^ O-Tooo — — ^CMO'Ococoro * • sJM-'csrs.'rx'ococorocN S cS d o C t- 00 a c * ^o vo rv 00 cs o ^CNCO'^iOSSrN.OOO-O — CNCO^'OSSN.OOOO "— E o.: D-^oorxsco *- S) wo "O "O "O yo^ 0*00^0 H M « ^ to (0 147 FISHERY BULLETIN: VOL. 70, NO. I CO U3 0> s o *-l SQ u a >, S o •c > C o CI] u 0) >. u o 03 be C] o c o (u as -- 2 (fl « 3 3 -t-> CO -3 >. C 0) 3 cr O) »^ «t-i 0) bo rt +^ C (U u n < ■5-° 5 0,> ^1 0) (1) :3o >- a: — 0) u_ iZ. U O 0) 31 'c U o c T) O < — 2"^ O -a .•^ o z D O o o , On o o > E o i^ t I I o^pcocooocqcNoicN Tt o^ C^ I I I I I 1 I I I I I CO >q ■'T t I I I I I I I I I * * I I I I I t i*otr>p»oop ^pp'Oioiqp'Oioio * * + I lOtoO'OiOiO^P^'^ PPPP ' * * * III IIO*— CvlOOcOCMCS *Oltl|t — CM CNJ ^ — I ico oo-— "^csiococqo^p ^(Nrv^'-;co o-K loo^ChCM-^oocoCNsq co i pu") Ou^opv>iOp'<)'^ I ^ ^ ^ c^ — 0-— ON^cocN o OsN.(>p*orocM00'^, "^ cqGOco^N.cs(N^. ] I I ] ocNvooCNcocoOcdrx wScocN — 0000 ' ^Tj-— _<>ocorN.p '^"Ipch'^. I ' ' ' oo^f^cocvisD— cdoio — o ' ' ' ' — -- CM — ■" '^. CNCN^CN iO'-;CNfOpr^ I I I I lTfCN*00> CO— ;0"^'0of^'^ '.-^coc*^io cno-o^^OiocsO 'cor^f^o locoioCN — ocqr^p 'ooO'^ '^fCNo^ — coc^iooi — — r— cs •— — iCM— ;rN. ooco^o^>;C^^o'>'ocq i t i '-■coco ^6cNrxor^o"^^o ' ' ' — .— CN »— . — o "— cMco-^'O'^ihsCOO^o •-csco'^iONor^coCho — csco'^io^Ji^cocso E^.t2 00 O CN C t- X ffl o * _Q E o o • • ■^ 010 "o ^c^c^c^ o M c9 ^ la o 148 MOSHER: SCALE FEATURES OF SOCKEYE SALMON Table 6. — Percentage frequency distribution' of circuli in the total freshwater zone of age 2. sockeye salmon col- lected in various years from 1956 to 1967. Number Asia Alaska British Columbia of circuli Ozernaya River- Blizhnee Lake^ Dal nee Lake^ Far North* Bristol Bay^ Attu Island* Adak Island" Un- alaska Island" Chignik" Cook Inlet** Karluk River" Yakutat= Ketch- ikan" Nass River"^ Nimp- kish River^i . - PtTCftlt - 9 — 0.5 — ~ — 0.4 — — — — — — — — — . 10 2.9 2.4 __ ._ 1.2 ._ __ __ _. .. 1 __ 10.3 __ 5.5 8.5 2.4 __ — •• 2 0.3 18.6 __ 9.3 16.1* __ __ 6.0 __ _^ 3 0.7 20. 1» __ 1.2 __ 12.9 13.3 0.3 ^_ 11.9 __ __ _^ 4 0.7 16.2 __ 4.5 0.2 14.0* 7.7 3.0 20.2* 0.5 __ __ 5 2.4 11.3 8.2 1.9 12.1 __ 0.9 7.3 6.9 19.0 4.0 — -• __ 6 5.6 6.9 ._ 12.7 4.4 8.8 3.7 9.7 10.2 0.5 13.1 8.5 1.1 -« 7 9.4 4.4 16.4* 6.8 7.9 0.2 6.9 12.5 11.8* 4.0 11.9 11.5 5.7 8 14.6 3.9 16.4* 10.0 7.4 0.4 7.9 12.1 11.5 10.0 8.3 15.5 10.2 1.0 9 16.3* 2.9 13.9 11.9 4.8 0.9 7.9 7.3 10.9 12.5 3.6 16.5* 13.6 3.1 20 14.6 1.5 — 9.4 13.1 3.3 2.5 10.2* 2.4 11.5 12.0 2.4 13.5 19.3* 5.2 1 13.2 0.5 7.0 14.8* 2.6 5.8 10.2* 0.4 11.8* 15.0 1.2 n.5 18.2 5.2 2 9.7 6.6 14.1 2.4 9.7 6.9 _. 10.5 18.5* ._ 9.5 10.2 3.1 3 5.9 3.3 11.3 2.6 13.1 6.9 0.4 5.9 14.5 5.5 8.0 2.1 4 4.2 0.4 7.4 1.4 15.5* 9.2 0.8 2.6 7.0 __ 1.5 8.0 2.1 5 2.1 __ 3.0 0.7 15.3 9.7 0.4 1.3 3.5 4.5 2.1 6 0.3 __ __ 0.8 0.9 13,3 7.9 __ 0.3 2.0 0.5 1.1 3.1 7 0.7 0.2 0.4 10.2 5.1 0.3 0.5 1.0 ^_ 5.2 8 5.3 __ 6.5 2.8 __ 0.7 __ __ 0.5 ^_ 4.2 9 10.5 __ __ 3.2 1.9 __ 0.3 __ 2.1 30 — 11.2 — — — 1.3 1.4 — — — — — — 3.1 1 12.5 0.7 0.5 __ __ .. 4.2 2 16.4* ._ __ 0.5 _. __ 7.3 3 16.4* __ __ 0.4 12.5* 4 __ 10.5 _. 0.4 __ 11.5 5 5.9 __ __ 0.2 __ 6.2 6 __ __ 5.3 __ __ __ __ 4.2 7 __ __ 3.9 __ __ _^ __ __ — 5.2 8 __ __ 1.3 __ _.. __ 5.2 9 — — — — — — -- — — — — — — — 2.1 Number of 72 51 38 61 216 105 139 54 62 76 50 21 50 22 24 fish 1 Actual 2 1959. 3 1958. * 1957. s 1966. « 1956. frequencies ; smoothed 1 occordi ng to Her iry (1961). ^ 1961. 8 1963. 1965. 10 ,964. >i 1967. * Indicates modes. DESCRIPTION OF SCALES FROM FISH OF VARIOUS AREAS The scale photographs are shown in three major series: ages 1., 2., and 3. Representative scales from each broad coastal area with rela- tively similar scales (including specific scale types from some areas of small production, when necessary) are shown on one plate. Two ex- ceptions to this grouping are made: (1) dis- tinctive scales from age 1. fish from North Amer- ican areas are shown on two plates and those for age 2. fish on one plate; and (2) scales from fish from North American areas north of Bristol Bay of ages 1., 2., and 3. are shown on the same plate. Scales of age 0. fish from all areas are shown on one plate, and those of age 4. fish from all areas are shown on another plate. A reference to the appropriate frequency tables of the number of circuli in the freshwater and first ocean zones (Tables 4 to 8) is made for each area group. These tables should be re- ferred to as scales from each group are dis- cussed. 149 FISHERY BULLETIN: VOL. 70, NO. I Table 7. — Percentage frequency distribution' of circuli in the first ocean zone of age 2. sockeye salmon collected in various years from 1956 to 1967. Number Asia Alaska British Columbia of Ozernaya Blizhnee Dal nee For Bristol Attu Adak Un- alaska Chignik'' Cook Karluk Yakutat'' Ketch- Noss Nimp- kish circuli River* Lake* Loke^' North' Bay=* Island" Island" Island" Inlet" River" ikan'" River- Riverii Pgrcint - II ._ — — 0.9 — — — — — — — 2 __ __ — -- 2.7 -- — — — — — — _. 3 __ __ __ 0.1 5.9 0.2 — — — — __ 4 __ 0.3 0.4 10.0 1.2 — __ __ _. __ 5 _. 1.2 0.8 12.6 3.4 __ __ __ _. _. 6 __ 3.4 1.2 13.3* 5.4 __ __ __ __ __ _. 7 2.0 2.0 8.0 2.6 13.2 7.0 .,_ __ __ 8 5.3 7.0 13.1 5.1 10.8 9.6 __ __ ._ ._ .. 9 0.3 7.9 9.8 17.9 7.3 8.4 13.7 __ 0.7 0.5 __ 20 1.7 1.0 13.8 12.7 20.9* 12.6 7.4 14.8* 0.4 3.0 1.5 — — — — 1 3.1 5.9 21.1* 16.4* 17.8 16.9* 5.8 11.2 4.0 6.9 4.5 0.5 2 6.6 11.8 21.1* 15.6 10.5 14.9 3.8 9.5 8.1 9.5 7.5 2.0 _. _, 3 16.0 13.2 13.8 13.5 4.5 13.2 2.3 8.9 8.9 10.9 8.5 2.4 3.5 4 24.0* 14.7 7.9 10.7 1.5 11.5 1.5 6.4 11.3 15.5 12.5 14.3 5.0 1.1 1.0 5 22.2 18.6* 5.3 6.6 0.5 6.7 0.7 3.9 14.5 18.4* 15.0 25.0* 10.0 3.4 3.1 6 14.2 16.7 2.0 3.7 0.2 2.7 0.2 2.0 16.1 13.8 14.5 21.4 15.0* 8.0 5.2 7 7.3 9.3 __ 1.6 0.1 1.4 0.2 1.2 16.9* 9.2 16.5* 16.7 14.5 20.5 6.2 8 3.5 4.9 __ 0.4 _. 1.3 0.1 1.1 13.3 6.9 13.5 13.1 13.0 29.5* 8.3 9 1.0 2.5 __ __ 0.9 0.4 5.6 3.3 5.0 6.0 12.5 21.6 16.7 30 1.0 — — — 0.3 — — 0.8 1.3 0.5 1.2 10.0 9.1 22.9* 1 ... 0.5 — — — — — 0.7 — — 6.5 2.3 16.7 2 3 4 5 ~ — ~ — — — — — — — ~ — 4.5 2.5 0.5 1 1 8.3 5.2 2.1 1.0 — — ~ ~ — — — — — — — — 1 . 1 2.3 1.1 6 7 ~~ ~~ ~~ ~ — — :: — :: — — *""■ ~~ — 2.1 1.0 Number of 72 51 38 61 216 191 203 140 62 76 50 21 50 22 24 fish ' Actual frequencies smoothec 1 accord i ing to Henry (1961). " 1961. = 1959. » 1963. ■■> 1958. « 1956. * 1957. i» 1964. s 1966. 11 1967. « 1956, ages 1 and 2 combined. * Indicates modes. KEY TO THE PLATES To compensate for the reduction of the original scale photographs to fit the printed page, use a 3 to 5x reading or magnifying glass to study them. The area enclosed by the first or central circu- lus is the focus or central platelet of the scale. The long black pointers near the focus indi- cate winter marks in the freshwater growth zone. A black stub pointer, if present, indicates the end of plus or transitional growth. If plus growth is i)resent, the circuli between the outer- most long black pointer and the black stub point- er are plus growth circuli. If no plus growth is present, the outermost black pointer indicates the end of freshwater growth as well as the last winter in fresh water. The white pointers bordered by black indicate the first winter mark in the ocean growth. The circuli between the end of the freshwater growth (or plus growth, if present) and this pointer are ocean-growth circuli and record the first year's growth in the ocean (the first ocean growth zone). The more widely spread circuli of this zone were deposited from May or June to Sep- tember or October (the summer growth) , where- as the more closely spaced circuli near the pointer were deposited during the autumn, winter, and early spring months (the winter growth). If a small white pointer'is present, it indicates an adventitious check in the first ocean growth 150 MOSHER: SCALE FEATURES OF SOCKEYE SALMON Table 8. — Percentage frequency distribution' of circuli (A) in the total freshwater zone and (B) in the first ocean zone of age 3. sockeye salmon collected in various years from 1955 to 1964. (A) (B) Number of circuli Asia North America Asia North America Asia" Blizhnea Loke^ Dalnee Lake^ Bristol Bay=^ Korluk River" Asia2 Blizhnea Lake^ Dalnee Lake* Bristol BayS Karluk River" Percent . . _ )2 _. 0.6 __ __ ._ __ .. 3 _^ 4.1 — — — — __ __ 4 __ 9.9 __ .„_ 0.4 5 0.5 15.1 __ 1.2 .. 6 2.9 19.2* __ .__ _, 3.3 _. 7 8.8 17.4 — ■_.. __ __ „_ 7.1 __ 8 15.9 11.6 __ 2.7 12.1 _. 9 18.1* 9,9 1.0 1.0 8.1 17.9 0.3 20 16.9 8.1 "- 2.0 — 2.9 — lO.I 19.6* 1.9 1 14.5 3.5 4.0 1.0 5.9 11.5 15.4 4.8 2 7.8 0.6 __ 10.0 3.2 10.8 1.2 16.9 10.8 7.6 3 2.2 — . — 12.0 6.6 16.4 5.8 18.9* 6.7 13.0 4 1.2 .— __ 13.0 9.5 20.1* 12.2 15.5 2.9 20.2 5 1.5 „ 19.0* 11.4 19.1 15.1 10.8 1.7 20.9* 6 1.2 _. __ 17.0 13.3 13.5 16.3 4.7 0.8 14.8 7 1.7 __ 10.0 12.7 6.4 18.0* 0.7 -- 8.6 8 2.5 __ 0.7 6.0 12.1 2.2 16.3 __ __ 3.8 9 2.0 __ 4.1 3.0 14.2* 1.2 10.5 __ 1.9 30 1.2 8.1 2.0 10.2 0.5 4.1 — 1.6 1 0.7 8.1 I.O 4.1 0.6 _. 0.6 2 0.2 __ 6.1 .. 1.2 _^ __ __ __ __ 3 __ __ 8.1 0.3 _^ _ — __ __ 4 5 ~ — 11. 5» 8.1 ~ ~ — — ~ ~ ~ 6 7 ~ — 6.8 11. 5* ~ ~ ~ — — — — 8 9 40 1 2 3 4 5 6 7 8 ^ — — 11. 5» 8.1 4.1 0.7 ~ ~ — — ~ ~ — — — 0.7 1.4 0.7 — ~ ~ — ~ Number of fish 102 43 37 25 79 102 43 37 25 79 ^ Actual frequencies smoothed according to 2 1962, from an area off southeast Kamch 3 1958. ' 1964. « 1955 and 1957 combined. » 1961. • Indicates modes. Henry (1961), latka. zone. Adventitious checks in other zones are not noted. AGE 0., ALL AREAS (Plate I) As noted previously, age 0. sockeye salmon are not common anywhere since fish of this spe- cies normally live for one or more years in a lake before migrating to the sea. Consequently, it was not possible to assemble frequency dis- tributions of the number of circuli in the first ocean zone for age 0. fish. It appears, however, that usually there are a few more circuli in the first ocean zone of scales of age 0. than on scales from fish of the same geographical area that have lived one or more years in fresh water. A few individuals of this age have been found at some time in almost every locality. Gilbert 151 FISHERY BULLETIN: VOL. 70, NO. 1 KAMCHATKA R. COPPER R. NUSHAGAK R. PETERSBURG NASS R SMITH INLET Plate 1. — Age 0., all areas. 152 MOSHER: SCALE FEATURES OF SOCKEYE SALMON KUSKOKWIM R. :^'^*fe&, AGE ^^^ Plate 2. — Ages 1., 2., and 3., Alaskan areas north of Bristol Bay. 153 FISHERY BULLETIN: VOL. 70, NO. 1 KAMCHATKA R. OZERNAYA R. BOLSHAYA R. OKHOTSK SEA LAKE DALNEE Plate 3. — Age 1., Asia. 154 MOSHER: SCALE FEATURES OF SOCKEYE SALMON UGASHIK R. NAKNEK R. EGEGIK R. KVICHAK R. NUSHAGAK R. TOGIAK BAY Plate 4. — Age 1., Bristol Bay. 155 FISHERY BULLETIN: VOL. 70, NO. 1 ><" -T^r^-.-rv- ATTU ISLAND :~ --«.:*5'S CHIGNIK R. UNALASKA RED R. KARLUK R. COOK INLET Plate 5. — Age 1., Aleutian Islands to Cook Inlet. 156 MOSHER: SCALE FEATURES OF SOCKEYE SALMON COPPER R. YAKUTAT A PETERSBURG KETCHIKAN Plate 6. — Age 1., Copper River to southeastern Alaska. 157 FISHERY BULLETIN: VOL. 70, NO. 1 MASS R. SKEENA R. FRASER R. COLUMBIA R Plate 7. — Age 1., British Columbia and the Columbia River. 158 MOSHER: SCALE FEATURES OF SOCKEYE SALMON FISH CREEK UNALASKA '' NIMPKISH R B Plate 8. — Age 1., North American areas with distinctive scales, Fish Creek type. 159 FISHERY BULLETIN: VOL. 70. NO. 1 A RIVERS INLET A NIMPKISH R. B Plate 9. — Age 1., North American areas with distinctive scales, Rivers Inlet type. 160 MOSHER: SCALE FEATURES OF SOCKEYE SALMON KAMCHATKA R. BOLSHAYA R. LAKE BLIZHNEE LAKE DALNEE' Plate 10. — Age 2., Asia. 161 FISHERY BULLETIN: VOL. 70, NO. 1 :,^w. UGASHIK R. NUSHAGAK R BEAR R Plate 11. — Age 2,, Bristol Bay. 162 MOSHER: SCALE FEATURES OF SOCKEYE SALMON ADAK ISLAND V.Udl KARLUK R. COOK INLET Plate 12. — Age '2., Aleutian Islands to Cook Inlet. 163 FISHERY BULLETIN: VOL. 70, NO. 1 COPPER R. NIMPKISH R. Plate 13. — Age 2., Copper River to the Columbia River. 164 MOSHER: SCALE FEATURES OF SOCKEYE SALMON FRAZER LAKE FISH CREEK COLUMBIA R. RIVERS INLET Plate 14. — Age 2., North American areas with some distinctive scales. 165 FISHERY BULLETIN: VOL. 70, NO. 1 OZERNAYA R. BOLSHAYA R OKHOTSK LAKE BLIZHNEE LAKE DALNEE Plate lo. — Age