Loading...
HomeMy WebLinkAboutAPA18271Nl&rn3~& t:l ~liD&®©@ Susitna Joint Venture Document Number 1Bd.7 Please Return To r:>OCUMENT CONTROL Sampling Red Salmon Fry by Lake Trap in the Wood River lakes, Alaska BY ROBERT L. BURGNER Contribution No. Ill, College of Fisheries, University of Wash- ington, Seattle, Washington. This is a revised version of the second of two installments of the author's thesis accepted in 1958 by the University of Washington in partial fulfillment of the requirements for the degree of Ph.D. CONTENTS Abstract 319 Introduction 319 Methods 322 Results of Littoral Area Sampling 327 Species Captured 327 Seasonal Trend of Lake-Trap Catches, 1956 329 Fry Distribution and Behavior 331 Relation of Stage of Fry Development to Magnitude of Trap Catches 333 Relation of Red Fry and Threespine Stickleback Catches 336 Evaluati.on of Results 341 Evaluation of Trap Catch Pattern 341 Problems in Use of Inshore Gear to Measure Abundance of Red Fry 311 Pelagic Area Sampling 342 Summary 345 Acknowledgment 347 Literature Cited 347 317 7. Sampling Red Salmon Fry by Lake Trap in the Wood River Lakes, Alaska Robert L. Burgner ABSTRACT A sampling system utilizing floating lake traps to measure abundance of young red salmon (Oncorhynchus nerka) dur.ing lake residence was established in the Wood River lakes, Bristol Bay, Alaska, in 1953 and tested over a six- year period. Catches of red salmon fry were heaviest during the first month after breakup of the lake ice in spring, and fell off sharply after the middle of July. The pattern t.>!: catch was shown to be associated in some degree to timing of lake-ice breakup, thermal development, and growth and behavior of the red salmon fry. The magnitude and pattern of catch varied considerably from year to year. Similarity in seasonal changes in catch pattern of red salmon fry and three- spine sticldebacks (Gasterosteus aculeatus) showed that environmental changes which influenced both species alike were creating variability in availability of the fish to the gear. Varying availability of the red fry presents a serious problem in indexing their abundance in littoral areas. In 1957 and 1958, pelagic areas were fished by setting floating lake traps to determine their feasibility as a test of, or substitute for, littoral area sampling. The pelagic area trap used did not provide a reliable measure of fry abundance. The correspondence in movement patterns of red salmon fry and three- spine sticklebacks suggests that they tend to inhabit feeding areas simul- taneously, thus increasing the opportunity for interspecific competition. INTRODUCTION THREE AGE groups of young red salmon, comprising fish in their first, second, and third years, are present each year in the Wood River lakes. During the first year the young reds emerge from the gravel and move into the lake areas to feed. A few migrate to sea. Normally the fry re- main to winter into theh second year in the lakes, migrating seaward in their second summer~ A portion remains behind for another year, migrating in the third summer. Age at seaward migration and growth and mortality rates of Wood 319 320 Alaska Red Salmon River red salmon have been shown to be influenced by climate and by lake population density of young salmon and their competitors (Burgner, 1962). These conclusions were drawn on the basis of samples of seaward- migrating smolts and returning adults. In order to understand more fully the interplay of mortality factors, periodic measurements of fluctuatiuns in abundance and growth of young salmon during lake residence were needed. These meas1ll'ements would allow the investigator to follow the success or failure of each year class in relation to the prevailing cop.di- tions in the ecosystem. In addition, information as to behavior and dis- tribution of the young salmon would help to determine the adequacy of measures of abundance thus far developed. It is the purpose of this paper to discuss problems connected with sampling free-swimming young salmon during their first year of resi- dence in the Wood River lakes, and to present results of a trial sampling system which utilized floating lake traps. In the Wood River lakes, as in many other river systems of Bristol Bay, a major portion of red salmon spawning occurs on lake beaches and in large rivers between lakes, the remainder in the creeks tributary to the lakes. Although sampling of creeks for downstream migrant fry following their emergence from the gravel has been used, this method provides an index of abundance only to a select portion of the total popu- lation. For this reason, attention was turned to methods of sampling free-swimming young salmon in the lakes after their entry from creek, river, and lake-beach spawning beds. Until recent years, efforts to sample the free-swimming fry of red, or sockeye, salmon in lakes had been generally unsuccessful, and their whereabouts was a matter of conjecture. Babcock {1904) reported that in 1903 fry could not be seen in Seton Lake after May; he commented that "heretofore no systematic study of the movements of young sockeye (0. nerka)· has been recorded." Chamberlain (1907) observed red salmon fry feeding over the entire surface of Yes Lake, southeast Alaska, on quiet evenings in August and September. He concluded that they remained at depths during daytime and rose to the surface during evening to feed on plankton and insects. In late August and September of 1905 he obtained samples after dark by surface hauls with a 130-foot seine in Yes Lake, and in the Naha lakes later in the same season. In most ether lakes under study young red salmon were not as readily observed during their residence. At Lake Aleknagik in Bristol Bay, red fry were reported seen in mshore areas as late as July and early August in 1908 (Marsh and Cobb, 1909) and 1909 (Bower, 1924). At Crawford Lake, British Columbia, Foerster {1925) observer1 young sockeye through- out the summer in the evening feeding on inseds at the surface of the water. However, of the fish at Cultus Lake, B. C., he stated: "They did not possess the habit of coming to the surface of the lake in quest of insects during the evening as did those in Crawford Lake . . . and none Sampling Red Salmon Fry 321 could be distinguished in the upper water during the night, attracted by the phenomenon of the diurnal movement of the plankton crustacea." Ricker {1937) indicated that, after emergence from the gravel, young sockeye salmon fry at Cultus Lake disappeared into the lake and were probably to be found feeding at depths in the vicinity of the Jake thermo- cline. He commented: "Extensive feeding in the upper epilimnion seems barred by the scarcity of food there . . . unless diurnal migrations of the crustacea are sufficient to add greatly to the number inhabiting the region in early morning or late evening." Scientists sampling young salmon residi!lg in Karluk Lake, Kodiak Island, met with V'd.rying degrees of success. Higgms (U.S. Bureau of Fisheries, 1929) remarked: "A good collection of young salmon was made in Karluk Lake with a special net. It never has been possible before to secure young salmon during their sojourn in the lake. n However, in commenting on the general success of the sampling ventures, he stated: ''It has been impossible to determine the mortality rate between the fry stage and the seaward migrant stage owing to the difficulties involved in collecting adequate samples of fingerlings" (Higgins, 1939). During fresh-water studies conducted at Karluk Lake from 1950 through 1954, Fisheries Research Institute personnel were successful i1'1 sampling all year classes of young by means of gear used along lake beaches (C. E. Walker, personal communication). Inshore availability of the young was found to fluctuate considerably during months of the year when the lake was ice free. Walker considered that fry abundance might be measured by inshore gear, but that inshore availability of yearlings was too brief and irregular to make indexing of their abundance possible. Yearlings were believed to be in deep and/ or offshore waters for all but approxi- mately two weeks of the year. Efforts to capture the young in offshore areas were generally unsuccessful, although small collections were ob- tained by purse seine. Higgins mentioned (1932-34) that Holmes was successful in collecting young red salmon in Chignik Lake, Alaska Peninsula, but results of this work have not been published. Methods and areas of collection were not given. More recently Krogius and Krokhin {1948, 1956) reported success in dP~.ermining the vertical distribution of young red salmon in Lake Dalnee, Kamchatka, by means of fine-mesh gill nets (mesh 10-12 mm). They pr\~sented graphs. to show the vertical distribution of young red salmon in percentage of all fish caught for each month in 1938 and 1939. Num- bers, siz.es, or ages of young in the catches were not indicated, and ef- feet of net selection was not discussed. They concluded that after emer- gence from the gravel the young live in the littoral part of the lake for one to two months, then move to the pelagic areas where they remain until seaward migration. During summer months the young were said to inhabit only the epilimnion but in the fall they moved deepe:.~ and by 322 Alaska Red Salmon winter inhabited all depths. Hydrological factors and food distribution were felt to be responsible for the vertical distribution of young reds in the lakes. Rees {1957) was successful in capturing sockeye young by means of gill nets at the forebay of Baker Dam, Washington. However, these fish were presumably on their seaward migration, and thus did not represent fish still resident in the lake. Johnson {1956, 1958) was successful in sampling "underyearling'' sockeye in British Columbia lakes in August and September by means of tow nets. A cone-shaped net 3 feet in diameter was found most satis- factory. It was towed by two outboard boats running parallel, with no lines or bridles directly preceding the mouth of the net during towing. The most successful means of sampling was the use of surface tows during evening hours. Tow nets of this type have been used to study abundance, distribution, and size of age 0 sockeye salmon in lakes of the Skeena River system and in Shuswap Lake in the Fraser system. METHODS Adequate methods of sampling lake stages of free-swimming young red salmon prior to seaward migration had not been developed at the time the Wood River investigation began. It was soon found that even though the young salmon might, in put, be in deeper layers of the lakes, they were also available in considerable numbers along the lake shores during summer months. During the summer of the first year of study, 1949, it was possible to make extensive collections of young salmon of both age 0 and age I throughout the la..lte system (see length-frequency compilations by the writer in Koo, 1962). From 1949 through 1951 various methods of measuring abundance were tested: beach seining, tow netting, purse seining, cast netting, abundance counts by visual observation, and sampling with lake traps. Disadvantages of beach seining were that success was dependent on visi- bility unless hauls were made "blind," and the catch was dependent on distribution of fish at the moment the set was made. Tow netting was unsuccessful using a single boat and several types of nets, and the-method was discarded. Purse-seine sampling was ruled out because success was dependent on weather and efficiency of set, boat and manpower re- quirements were heavy, and the fry would be missed until they were of sufficient size to move offshore. Cast-net sampling was discarded as too limited. The effectiveness of visual enumeration was dependent on the changes in light, turbidity 7 and water surface distortion from wind. Impressions of relative abundance in inshore areas could be obtained by the individual observer, but difficulties arose in assigning numerical values to abundance and in retaining impressions for comparisons with estimates of subsequent visits and subsequent years. Sampling Red Salman Fry 323 Sampling by lake trap appeared to have several basic advantages: (1) conditions of fishing could be duplicated with precision, (2) the trap could be fished over an extended period to give a measure of average abundance during the time intervals between tending, and (3) samples could be ob- tained independent of conditions for visual observation, eliminating per-- sonal error in estimates. Because of these advantages in sampling by lake traps, attention was directed toward determination of trap efficiency in capturing young reds of assorted sizes and ages found in the lakes. There were certain operational disadvantages to the use of pile traps. They were difficult to install, could not be shifted readily inshore or off- shore with changes in lake level, and were not versatile for use on dif- ferent types of lake bottom. After limited experiments with pile-trap de- signs in 1950, the writer turned to floating-trap designs for trial in 1951. Following experiments with design and operation of a floating lake trap in 1951, a modified type was designed in 1952 for use in later experiments. Saran plastic screen of eight meshes to the inch wa.s used for trap and lead instead of the conventional cotton mesh, the onl.i type of small web then available. Saran screen was found to have certain important advan- tages over cotton web in that it did not rot, required less skill in hanging, was more transparent in the water, and was more easily cleaned. 6' I I I I I I ~ H ,. •' ol II lo II II II ,, 1 !; . ,. I ol ,L--------t.:.::..r ,• ...... ,/ .... , .. ,- s•' ., ./ "t ' I I I I I I I I I I I I I 6 '10" I ' I ' I I I I I I I I I I I ,.J .. _ I / I I .1' I ,, ... 1951 model 1952 model T I I I I I I ~· I I I I I J. T I I I I I 6' I I I I I I ..... Fig. 1. Web outlines of the 1951-and 1952-model floating lake traps, showing the changes in design incorporated in the 1952 trap to increase the efficiency of capture and retention of red salmon fry. 324 Alaska Red Salmon The web outlines of the 1951 and 1952 traps shown in Figure 1 illustz "lte the primary changes in design suggested by study of fry behavior in and about the traps. The changes made were based on observations that red fry in the trap sought the widest and deepest parts of the heart and pot. The principle involved in the 1952 trap design was to lead fish from narrow to wide, sluill.ow to deep, inside the trap as far as possible. The pot depth was increased from 6 feet to 6 feet 10 inches in the 1952 design. The lead and the remainder of the trap were 6 feet in de,Ith. Dimensions and design of the 1952 trap are given in Figures 2 and 3. Four 5-gallon cans were used for floats, and 3-pound leads in the bottom corner angles and at the bottom of the inner "V" aided in holding the trar. web in position. An angle-iron h:)ld the bottom of the outer "V" in position a.tJ.d served as an attachment for the rigid inner edge of the trap lead. The traps were set out from sloping beaches in approximately 7 feet of water, with the 6-foot-deep lead extending out from shore line and touching bottom until the outer apron of the trap was reached. As the lake level changed, the traps were shifted inshore or offshore to ac- commodate the changes. A single offshore kedge ar ..:hor of 20-to 25- pounds was sufficient to hold the traps in position even in stormy weather. Load to thoro T : . I ! I : • ;:.. 1 : . : . . j ' ! . : . . I . • 1----------------------15'8"-------------------------------l To anchor+ Seal• II"" • 11 Fig. 2. Plan view of 1952-model floating lake trap used at Lake Nerka. The pot was emptied from a skiff alongside the trap. The cork-line lead was released at point of attachment to the outer "V'' opening and the pot web was pulled up in such a manner that the catch was concentrated at the surface in the rear corner of the pot (See Fig. 4, p. 326). The fish were then brailed by means of a dip net. All species of fish in the traps were enumerated if the catch was small. Larger catches were sampled in the following manner: (i) Large fish such as adult chars, -rainbow trout, salmon, grayling·, and pike were dipped out by a large-mesh dip net and counted. (2) The remaining fish confined in the corner of the pot were mixed as thoroughly as possible by stirring and a sample then dipped out quickly by means of a wire-mesh dipper. The volume of the sample ·was obtained Fig. 3A. The 1952-model floating lake trap. View from lead. Fig. 3B. View from shore. 326 Alaska Red Salmon Fig. 4A. Fish in the pot prior to brailing operations. Fig. 4B. The pot being lifted in preparation for removing the catch. by displacement and the fish of each species were counted. Then the volume of all remaining fish was determined and the number of each species in the total catch estimated from the composition and volume of the sample. Lake Nerka was chosen for trial of the lake-trap system of sampling because it is the most important red salmon producer in the Wood River lakes, is centrally located, and is, presumably, most representative. In 1952 several littoral lake-trap sites were fished on Lake Nerka. Of Sampling Red Salmon Fry 327 these, four were selected to be fished from breakup of lake ice to mid- summer each year. Three requirements sought in choice of trap site were: (1) evenly sloping beach of suitable gradient, (2) relatively smooth bottom, and (3) protection from prevailing winds. It ·was found in earlier experiments that both catch and ease of tending the trap were affected if the traps were subjected to heavy wave action. Beginning in 1953 the tr~ps were installed and fished as consistently as time and funds permitted. The annual fishing periods for each trap are listed in Table 1. Location of the four littoral sites 011 Lake Nerka are shown in Figure 5. (Seep. 328) The traps were tended from the Insti- tute field station at Cabin Bay. In the round trip by boat a distance of 25 miles was traversed to tend the four traps. TABLE 1. DATES OF OPERATION FOR THE FOUR MAIN LAKE TRAPS, LAKE NERKA, 1953-58 Year Cabin Bay Catherine Cove Pike Bay Anvil Bay 1953 June 5-Aug. 27 June 8-Aug. 12 J1me 6-Aug. 13 June 11-Aug. 11 1954 June 7-July 2 June 14-July 2 1955 June 15-Aug. 6 June 18-Aug", 1 July 3-Aug. 1 June 29-Aug .. 1 1956 June 16-Sept. 12 June 16-Aug. 10 June 13-Sept. 12 June 22-Aug. 12 1957 June 6-Sept. 1 June 6-Aug. 9 June 8-Aug. 9 June 7-Aug. 8 1958 June 2-Aug. 22 June 1-Aug. 22 E-A-periments conducted in 1951 showed that efficiency of retention of fish by the lake trap varied with species, and that the daily catch ob- tained by brailing a trap once a day could not be expected to equal that obtained by brailing several times a day. The time of brailing each day would thus be important if the rate of capture varied with time of day. No consistent pattern of catch of the major species in relati0n to times of daylight and darkness was found in limited experiments in 1951. How- ever, as a precaution, each trap was brailed on a daily schedule at as near the same time of day each year as work schedules permitted. RESULTS OF LITTORAL AREA SAMPLING Species Captured Species of fish caught in the lake traps, listed in approximate order of numerical abundance, were: threespine sticklebacks (Gasterosteus aculeatus) red salmon .(Oncorhynchus nerka): fry, yearlings, age n, and adults ninespine sticklebacks (Rmgitius pungitius) arctic char (Salvelmus alpinus) slimy sculpin (Cottus cognatus) coho salmon (OncorhynchuB kisutch) Alaska blackfish (D:J.llia pectoralis) rainbow trout (Salrno gairdneri) • Imhore lake-trap sit~s A OHshore lake-trap si.tes Fig. 5. Location of lake-trap sites at Lake Nerka. .. Sampling Red Salmon Fry pond smelt (HyjxJmesus olidus) arctic grayling (Thymallus rrrcticus) northern pike (Esox lucius) round whitefish. (PYosopUmz cylirulraceum) pink salmon (Oncorhynchus gorbuscha) king salmon (Oncorhynchus tshawytscka) Dolly Varden (Salvelinus malma) 329 The list includes all species known to reside in Lake Nerka. The burbot (Lota lota) may be present in Lake Nerka but has not been caught to date in various gear used. Seasonal Trend of IAke-Trap Catches, 1956 If it is assumed that the la.'l\:e traps were fished with the same efficiency throughout the season, the size of the catch at a trap site such as Cabin Bay was determined by the combination of several variables. The most obvious were: 1. Actual over-all abundance of the species in the area, which we wished to measuTe. 2. Changes in vertical and horizontal distribution of young salmon resulting from changes in weather or other factors that affected availability to the gear. · 3. Changes in behavior associated with growth and feeding habits of the species that affected catchability. 4. Random movement within the area, not necessarily connected with (2) and (3) above. 5. Reduction in numbers by mortality. As a result of these factors, catches of the several species in the lake traps were extremely variable from day to day and between different periods of the season. The problem resolved into one of determining how closely the trap catches actually reflected {1) the actual over-all abundanee of individual species in the area, and (2) how, if at all, cor- rections could be made for the other factors enumerated above. Catch pattE!rns of the two most abundant groups, threespine sticklebacks and red salmon fry (age D), were studied from this standpoint. The 1956 catches of red fry in the four floating lake traps fished at Lake Nerka will be presented first to show seasonal pattern of catches. Each trap was installed at its site as soon as the lake ice disappeared in June. Except during a two-day absence of personnel, the traps were brailed daily. Catherine Cove and Anvil Bay traps were removed on August 10 and 12, re:·Jpectively, Pike Bay and Cabin Bay traps on Sep- tember 12. Daily numbers of red fry caught in the four traps during June, July, and August are presented in Figure 6. Two features of the catch data are particularly apparent for all four trap locations. First, the day- 330 Alaska Red Salmon to-day fluctuations in numbers of fry caught were large; second, the catches were much heavier in June and early July than jn late July and August. These points require further discussion. 10 5 4 ................. Lake ice-covered at trap-site area. Trap not tendea. ~ .............................. L&&A&&&A ........ &Ulll.a.u.IL ......... __ ..II.Ih • .___ANV_I_L BA-Y .... _ ..................... .. CA 'I'HERINE COVE I I •••••••••••!••+•<t••••1'••••••••••• ..::: !! 10 "' II u I ~ 5 '"' 4 ... I >. 3 rl 2 .... e1 Cl PIKE BAY 1· 0 15 I 10 5 4 3 2 CAIHN DAY ··r· r , .. , .... , .... , •• •• , ._,..... 20 25 aol 5 ' 30 5 10 15 10 15 20 25 30 June July August Fig. 6. Daily catches of red fry at the four primary lake- trap sites 1 Lake Nerka1 1956. The day-to-day fluctuations in magnitude of fry catches in each trap appeared to occur without relation to fluctuations occurring in other traps. Rankings of the catches over short-term periods in various orders by date of catch failed to disclose that high catch in one trap was as- Sampling Red Salmon Fry 331 sociated with high catches in the others, or low with low. With no patte.L·n emerging to relate daily change in magnitude of catch between the dif- ferent trap sites, it was clear that daily fluctuations could not be related directly to immediate eifects of changes in general weather conditions at Lake Nerka. If we assume that the efficiency of the trap did not change from day to day, the changes in catch must then be assigned to local changes in distribution or behavior of the red fry. As for the seasonal trend of catches at the four traps, there was an unmistakable similarity in pattern. Catches were heaviest during the early season, then dropped sharply to very low level at all traps. This seasonal trend and differences seen between traps in details of catch pattern will be discussed in the light of observed behavior and distribu- tion of red fry w~ing late spring and early summer. These observations are described ln the following section. Fry DistributiOJ?, and Behavior Red fry that inhabit the lakes during the season originate in three major types of spawning areas-lake beach, rivers between lakes, and creeks. Time of emergence of fry from spawning-area gravel varies considerably among and within these areas. In 1956, a year of generally late emergence, the first free-swimming fry were observed in one of the creeks under study in early April, but larvae were still abundant in the gravel of this creek on May 16 (y¥. A. Church, personal communication). On one of the lake beaches, free-swimming fry were observed in late April at a time when :redds dug in other beach areas still contained un- hatched eggs. In spite of this difference in time, emergence was com- pleted in most areas by the time the first lake trap was installed in mid- June. However, in Anvil Bay, fry with yolk material still unutilized were caught in the lake trap as late as July 4. This indicated that in some areas, presumably portions of beach areas in which spawning is late and perhaps where tempering ground flow in the gravel was low, all fry had not emerged in time to be available for capture during the first days of lake-trap fishing. Fry f.mergence in the creeks and the rivers between lakes was es- sentially completed by the time lake traps were installed in the spring. Howe,·er, downstream migration does not necessarily occur simul- taneously with emergence. While fry do leave the rivers between the lakes soon after emergence, downstream migration of fry in most tribu- tary creeks is not completed for some time after breakup of lake ice. In many of the creeks a portion of the fry population remains to feed, and sometimes the fry acquire considerable growth before entering the lake. For enmple, in spring-fed Pick Creek on Lake Nerka, schools of fry showing good growth were very abundant along the creek •:.: late June and early July of 1956. Few fry are to be found in Pick Creek by 332 Alaska Red Salmon August, although small schools have been found in the upper cree~ as late as September 25. In general, however, the decline in numbers of fry remaining in the creeks is quite rapid, and i:>y late June these con- stitute a very minor portion of the total year class. Other minor sources of late-entry fry are small lakes such as Lynx lake and Hidden lake, tributaries to Lake Nerka. The young migrate out of these lakes into the main lake either as large fry in midsummer or winter over in the small lakes a year or two before leaving. Summer- migrant fry from these small lakes contribute in a very minor way to the main lake population. Red fry in the Wood River lakes are observed in abundance along the lake shores for at least a month after breakup of lake ice. When the lake level is high early in the season, they are to be found in droves in flooded grass along protected areas of the lake shores, and, except on stormy days, the fry also seek the few inches of water over the fine gravel of upper-beach areas along more exposed shores of the lake. In 1956, these conditions were found in the Wood River lakes as late a3 July 6-9, when the writer conducted visual surveys of fry along Little Togiak, Beverley, Mikchalk, and Kullk lake shorelines. The fry in Little Togiak Lake were found distributed around nearly the entire shoreline of the 7-mile-long lake. In lakes Beverley and Kulik where in general the fry emerge later, they were found still rather closely associated with the known spawning areas. July 13 was the last date fry were seen in abundance in shallow littoral areas at Lake Nerka. Distribution of fry in the lake is not confined to the shallow beach areas early in the season, for not only do they tend to move deeper on stormy days, but they can be found in numbers around island shorelines where no spawning is known to take place and which are separated from the main lake shore by deep water. It is also possible that the red fry popu- lations which inhabit the shallow litto1·;.:l areas are a minor portion of the total in the lakes. Although to date early-season fry have not been taken or observed elsewhere than in littoral areas, it may well be that this is because of the techniques employed. The fry are reasonably consistent in their behavior and range further offshore as they increase in size. During the early season, fry feeding in and over the shore grass ~u.bmerged by high lake level dart into the grass when frightened. As they develop, they are seen not only in the grassy areas but a.lso in open water a few feet deep, and tend to seek deeper water when disturbed. By mid-July in 1956 the lake level had dropped and the fry had largely disappeared from the inshore observation area at Cabin Bay, lake Nerka. Only occasional small schools were observed passing over lake bottom observation panels. These panels were placed from shore to a distance approximately 40 feet offshore and to a depth of 5 feet. From the end of July to mid-August fry were observed surfacing over the entire Cabin Bay area (maximum depth, 40 feet; width, ~mile}, Sampling Rea· Salmon Fry 333 and in similar localities in other areas of the lake. Later in the summer numerous large schools of fry were observed at the surface in mid-lake ;areas several hundred feet deep. Similar offshore movements during the season were observed in other years. Relation of stage of Pry Development to Magnitude of Trap Catches Changes in inshore distribution of the fry were reflected in the lake- trap catches. The early-season fry, although very numerous, were somewhat reluctant to lead out to the trap set at a depth of approxi- mately 6! feet, and sometimes accumulated in considerable numbers along 1 ~·.a inshore portion of the lead. However, by mid-July, schools we:..""e o·a:;served going out around the end of the trap, and still later in the season the major portion of the observed fry populations were off- shore from the traps. In conjunction with changes in behavior and movement of the fry there occurred changes in distribution. B«?cause of change in feeding habits, movements, distribution, and behavior of fry with increased size and age, it could hardly be anticipated that likelihood of capture by stationary gear would remain constant. Efficiency of the lake trap in capture of fry was therefore related to size and age of the fry in the trap area. During the season changes in level of lake-trap catches did correspond in a general way to the changes in abundance of red fry observed in lit- toral areas. However, because of the tendency for fry to move offshore as they increase in size, differences in level of trap catch at different periods during the season could not be regarded as reflecting true changes in over-all abundance in that section of the lake. For comparison of abundance between years, a first requirement would appear to be that catches must be compared for periods when the young were in the same stages of development. The stages of fry development found at given dates within the season differed between traps sampling different portions of the lake fry popula- tion. This is illustrated by the comparison of four sets of small samples taken at Cabin Bay and Anvil Bay traps during the 1956 season (Table 2). Date Ju.,e 24 July 4 July 19 Aug, 2 TABLE 2. COIIIPARISON BETWEEN ANVIL BAY AND CABIN BAY LAKE TRAPS IN SIZE OF RED FRY SAMPLES TAKEN AT APPROXIMATELY SAKE DATES DURING 1956 SEASON* Anvil Day Cabin Bay Sample Length in Killimeters Wean Weight Sample Length in Kil1lmeters Standard Date Standard Size Mean in Grams Size lleB!l Deviation Deviation 8 27.5 1.50 0.15 June 26 29 29.1 1.75 13 27,6 1.26 0,15 July 1 25 29,5 1,58 58 27,7 1.58 0,17 July 19 24 36,2 S.25 59 27,0 1.49 0,16 Aug. 2 58 38.9 5,58 •Lengths and ~eights takan after the aa~p)es were preserved, Mean Weight in Grams 0.22 0,23 0.47 0.64 Fry captured in Cabin Bay were larger early in the season and showed a much greater increase in size subsequently. The continued small mean 334 Alaska Red Salmon weight and mean length of the Anvil Bay samples indicated that either .. growth was very retarded in this area or, more likely, that fry which emerged later were replacing fry which emerged early in the sampling area. Late emergence of fry in the Anvil Bay area is tha probable ex- planation as to why 1956 trap catches at Anvil Bay continued at a high level later in the season than at the other traps (Fig. 6). The smaller size of fry caught in Anvil Bay was a consistent feature in all years of sampling. Were the seasonal time-abundance patterns of the trap catches the same from year to year at a trap site, it would probably indicate close similarity in timing of fry development and in behavior. The magnitude of the catches from one year to the next could then be compared directly IJl '0 1 •••-enuu Lake still ice-covered in 1:rap, area • ............... Txap not tended. 1953 a Q IIIIIIIIIIIH...,._INM ......... HIIMIM-IM ... ..I.. ·e so • 1 1955 h I ,dl., ... u.l ...................... . 1956 2 1 1957 0 ......... -..... \"~ -~~~~~ .............. _.p.l&q~llf.,.,...,__,___,,..-.....,.-....,-...... -~ 1 5 15 20 5 10 15 20 25 3o 5 10 15 June July August Fig. 7. Daily catches of red fry in Anvil Bay lake trap, 1953, 1955, 1956, and 1957 seasons. SarnPling Red Salmon Fry .c Ill -.4 ... '6 1 Ill "0 •••••••••••••.•. Trap not tended 1953 ----·············· a o •· -· -- 1, 1955 ,, •• 11 ....•••. 1 ............................................................ . "' ;:1 0 .c .... .5 10 0 10 Humber of days a{ter ~rap-site area became ice-free Fig. 8. Daily catches of red fry in Anvil Bay lake trap, 1953, 1955> 1956, and 1957 seasons, gr~phed from date trap-site area became ice free. 335 by calendar date to the abundance of fry in the locality. However, a number of factors militated against such a happy occurrence, and the actual pattern of catches varied considerably from year to year. The seasonal patterns at Anvil Bay trap for the years 1953, 1955, 1956, and 1957 are used to illustrate this variability (Fig. 7). Three facts are at once apparent in Figure 7. First, the date Anvil Bay became clear of ice varied considerably from year to year. Second 7 there were very obvious differences in ssasonal time-abundance sequence of catches. 1'hird, the general level of fry catch was markedly different from year to year. The problem lay in determining whether these dif- ferences could be considered mea'lingful in terms of actual differences in 336 Alaska Red Salmon magnitude of fry populations within the sampling area, and if so, how the differences in level of catch could be compared quantitatively. The dates that the lake areas become ice free depend upon thickness and composition of the ice and snow cover that develop during the winter as well as upon the spring weather conditions prior to and dm•ifig breakup of lake ice. 'rhe timing of breakup does not necessarily show any dose relation to the $tages of development of the red fry population at that time, for the latter is dependent upon timing of parent sj)aWI1ing; gravel temperatures during egg incubation, h..ttching, larval development, and emergence; and relative survival of young that deveiop under different temperature regimens. b'ummer~ fall, and early winter conditions have little bearing on tlle timing of breakup of lake ice the following spring, but can have a very direct influence on rates of development and survival of incubating eggs and larvae. On the othe:r hand, bre:;~}~p of lake ice in the spring does herald a very rapid sequence of changes in the lake environment, and in spite oi initial differences in stage of development, fry behavior ma:y be rather closely associated with and influenced by the sequence of environmental changes that follow lake-ice breakup. This is certainly indicated by the improved relation between years in timing of the Anvil Taay fry catches when they are plotted according to number of days after Anvil Bay became ice free rather than by calendar date (compare Fig. 7 with Fig. 8). Relation of Red Fry and Threesplne stickleback Catches As seen in Figure 8, there still remained rna.rked differences between years in seasonal pattern of catches even when date of breakup was used as the beginning date of the season. There were not sufficient samples measured each year to show whet~er the differences between years could be explained by differences in the developmental stages of fry present in the catches. Therefore, other evidence was sought that might bear on the problem. This was found in the catch record of the threespine stickle- backs, Gasterosteus aculeatus, the species that rivals the young red salmon in abundance in t.'le \Vooo River lakes system. Catches of red fry and threespine sticklebacks at Anvil Bay are compared for four years in Figure 9. (Catches of stickleback fry hatched during the season are omitted in these comparisons.) In each year abundance of sticklebacks in the catches did not drop as early during the season as did that of red fry. Except for this difference there existed general similarity in seasonal pattern of catches of the two species. In 1953 the initial low level, the large catches, and the fall-off occurred for both species dt1Xing the same periods. Again in 1955, a similarJty in pattern occurred between species, but it contrasted with 19!i3, In 1956 5 and particularly in 1957, catch patterns of the two species were not as similar, yet they contrasted in the same general way with 1953 and 1955. Not only did seasonal catch Sampling Red Salman Fry 337 patterns of the two species tend toward similarity, but changes between years in general magnitude of the catches were similar. 4 3 2 l -Lake s-.;ill ice·-co,verEld in trap area. -Trap ~'.Jt tended. ·~'hreespine stici!>lebs.Cke, .19{?;3 o~~--·----~~DUUU~~--~~Ld~.----~-1 4 3 2 Red fry, 1953 1 or-===~~~~~~========~ 2 Threespine sticklebacks, 1955 1 Red fry, 1956 Threes pine s·ticl!.ie!!a!;)(s, 1S55 Red fry, ~9S7 10 J{lly Fig. 9. Comparisons of daily catches of red salmon fry and threespine sticklebacks, Anvil Bay la.lte trap, 1953, 1955, 1956, and 1957 seasons. The similarity in seasonal catch pattern of red fry anct sticklebacks in the Anvil Bay lake trap suggests that the same factors were operating on both species to affect their catchability by inshore fixed gear during 338 Alaska Red Salmon the season. To examine whether or not this was a phenomenon peculiar only to Anvil Bay, the red fry and threespine stickleba:ck catch records were graphed for another iake-trap site, Cabin Bay (Fig. 10). Here the similarities ffi seasonal catch trends of the two species within individual years were even more strikin,g, and differences between years equally noticeable. Although early season catches of sticklebacks at C~in Bay in l .,., __ Lake still ice-covered at trap site. -----Trap not tended. ~~---··••··M~~~~~~ug~~ .......... M .. A~~·-·-•••-·•--·~--· .. i ned fry, 1953 0~~~~~~~~====~~====~==~ .; 1 .... .... Threespine sticklebacks, 1955 ~ 0~::::::::~~~~~~==~~~~~~~~==~1 Ill "tl r:: 5 : 4 g 3 :S 2 a "" 1 '5 Threespine sticklebacks, 1956 1: o 1-------" l&lol~~&~o~ol@!!!.!!.o....u.JL.. .. "..,......._ __ _....-.. _____ 1 u f'ls +' ~10 .... a 5 4 3 2 Red fry, 1956 1957 June July Fig. 10. Comparisons of daily catches of red salmon fry and threespine sticklebacks, Cabin Bay lake trap, 1953, 1955, 1956, and 1957 seasons. Sampling Red Salmon Fry 339 1955 were not matched by fry catches of similar relative ~onitude, the pattern in July and early August was mu_ch the same for the two species. Aside from this exception in early 1955, the major seasonal character• istics of timing and magnitude of catches were borne out by both red fry and threespine sticklebacks. 4 2 6 4 2 ~· D, E 0 .: 2 2 D ]Lm~ 21-30 3 samples ". 1023 August 2:cmples n=3lll :ot--~~~-_._ _ _._ _ __._ _ __.~__. _ __,~i.----f 0 ~ 6 " u 2 2 0 4 2 0 Sept•mber 3 samples n = 224 40 so 60 70 90 Total length In millimeters Fig. 11. Length frequencies of threespine stickle- backs (Gasterosteus aculeatus) in samples taken at Lake Nerka, 1957, combined by month, showing progression of modal groups ;Juring the season. (!''ish measured after preser·,ation in formalin. Length frequencies smoothed by moving avern.ges of threes.) 340 Alaska Red Salmon The question arises as to the significance of these relationships be- tween stickleback and red fry catches. Since the red fry were of a single age, changes in catch level could have been associated with behavior change as the f:rv developed. However, similar changes in catch level of sticklebacks could not be related in the same way to developmental stage because more than one age group was involved. Length-frequency analysis of threespine sticklebacks taken during the 1957 season indicated that at least two or three age groups were present in the catches. This is illustrated in Figure 11, in which all samples preserved from catches made by lake-trap, tow-net, and beach-seine gear are combined by monthly periods. Stickleback fry of the year did not appear in the catches until August, and were readily separated at that time from older fish because of their small size. As seen in Figure 11, they first appeared in numbers in September samples. During June, July, and early August the strong mode of smaller fish titus represented hatch of the previous summer. The progression of this size group and the trimodality of the June samples strongly suggested the presence of three age groups at the beginning of the year, which would place the maximum age attained at three years or more. 1 As the sticklebacks near maturity, sexual dimorphism in size occurs, and modal progression is probably somewhat obscured. The stickleback catches, therefore, differed from the red fry catches in that the former were not at a uniform age or stage of development, whereas the fry catches were all of the same year cla.ss. The relatively faster decline in numbers of red fry caught in the lake traps during the season may have been due to more uniform change in stagf~ of develop- ment and more marked change in behavior with increase in size. How- ever, because of the indications of general similarity in rise and fall of the catches of the two species during a season and of similar contrasts between seasons, it was evident that stage of development of red fry c ::mld not alone explain the differences found between years in seasonal pattern of red fry catches. Lrrimodality in length-frequency samples of threespine sticklebacks taken at Karluk and Bare Lakes on Kodiak Island was also shown by Greenbank and Nelson (1959). Otolith study confirmed the presence of three year classes. They assigned the group of smallest fish in May and June samples, with mode at about 30 rom, as represen'ting fry of the current season. It appears most likely that these fry were hatched the previous summer but were not taken by collecting gear at that time. In the 1957 Lake Nerka samples, the modal group of new rec:ruits was 23 mm total length at the time of September sampling. The June mode, 29 mm, for the youngest age group of Lake Nerka stickleb:lcks corl'e- sponds approximately with that found at Bare Lake in :May and June. SamPling Red Salmon Fry 341 EVALUATION OF RESULTS Evaluation of Trap Catch Pattern Catches of red salmon fry made in floating traps set along the lake shore have been evaluated in relation to observed behavior, weather conditions, and catches of threespine sticklebacks. This evaluation is summarized as follows: 1. There were marked differences between years in magnitude and seasonal pattern of catch of red fry in the lake traps. 2. Part of this difference was due to the influence of differences be- tween years in time of lake-ice breakup. 3. The seasonal timing and magnitude of catch of fry and threespine sticklebacks each year tended toward similarity in pattern. 4. While Inagl!itude of catches made during a year must be dependent to some extent on actual numerical abundance, factors influencing both fry and sticklebacks create seasonal variability in availability of these species. 5. The fact that catches of both species showed similarity in seasonal pattern suggests that environmental changes occurred, affecting the distribution. and behavior of both species in a similar manner. 6. The pattern also suggests that external factors which affected both species were more influential in determining the catch pattern than were seasonal changes in catchability of red fry due to behavior changes with increase in size. 7. The fact that the lake traps did record similar changes in catches of fry and sticklebacks indicates that the trap catches provided a record of actual changes in inshore availability of these two species. This was substantiated by visual observations of inshore abundance. 8. It should also be noted that the close correspondence in move.ment patterns of red fry and threespine sticklebacks suggests that, at least during this period of the year, the two species were in com- petition with each other for food. This was borne out by visual ob- servations, for it was commonplace to find schools of red fry and threespine sticklebacks completely intermingled in littoral areas, sometimes in dense aggregation. Problems in Use of inshore Gear to Measure Abundartce of Red F1~y The lake trap studies led to certain g~neral conclusions about adapta- bility of gear set in littoral areas for indexing red fry and threespine sticklel::ack ab..mdance. Changes in inshore availability of the fish present the same problem regardless of indexing method, whe1~her it be by lake trap, visual observation, photography, or other means. Essentially 342 Alaska Red Salmon the same difficulty would arise in systematic, periodic sampling at a single location by beach seine or other inshore mobile gear. Tests of reliability of inshore gear in reflecting actual fry abundance need in some way to be independent of fluctuations in availability of red fry in littoral areas. Other methods of measuring inshore abundance might reflect the same fluctuations in availability as did the traps without providing additional information on actual over-all abundance of young. It might be proposed that reliability of the procedure could be tested by correlating the gear catches with abundance measures at earlier or later stages. For instance, correlations between numbers of eggs de- posited the previous year, corrected for winter-kill if feasible, and lake-trap catches of red fry might be tried. However, this would not provide a satiffiactory test of the lake-trap method as a measure of abun- dance, since our objective is to measure fluctuations between years in mortality rate occurring between egg and free-swimming fry stages. The same objection applies to attempted correlations between measured abundance of fry and of subsequent surviving smolts. We prefer not to make any assumptions as to constancy of mortality rates from one stage to the next. An independent method of indexing abundance should be ap- plied, if possible, at approximately the same early-life stage, if it is to satisfy our objective. Two additional methods of sampling were begun in 1957 to determine the feasibility of pelagic sampling as a test of, or substitute for, littoral area sampling. The first method made use of floating lake traps set in positions over deep wate1 some distance from shore. The second method involved offshore sampling by means of a large net towed between two outboard skiffs. Evaluation of the pelagic trapping method will be pre- sented in the following section. The tow-net method of sampling will be described in a later report. PELAGIC AREA SAMPLING In 1957 a lake trap of the same design as the 1952-model trap was fished at Lake Nerka in four locations. These sites varied from 1/8 to 1/2 mile offshore .and from 1/2 to 4 miles apart. The only trap modifi- cation was the extension of the lead to 140 feet in length. An anchored 55-gallon drum served to hold the trap in position. The tra_p was attached to the drum by a single line at the pot end. The lake currents set up by wind action were normally sufficient to keep the trap and lead extended out full length "downstream" from point of anchor (Fig. 12). Depths over which the trap was fished ranged f:rom approximately 20 feet to over 100 feet. The trap was set on June 9 and, with the exception of two days, was tended daily until August 10. The loc~tions fished are shown in Figure 5. -: ~ '0 "' E ~ c ?: c c Fig. 12. Floating trap anchored to buoy off Elbow Point, Lake Nerka, lead trailing on the right. 0~----~------~--~----------~----------~----------------Dif :Amahuk: Elbow Point 600 Cabin Boy 10 IS I C :~i n I Arm l(anchor .. •d at pot) 3840 I I l I I ....a.l..!- I I l I l I I I I I I I I I ......,..._;_ 20 25 ]u n ,.. IO Elbow Point (anchon .. •d at ~nd of h!'od Thr,•.,·splnt.' I 5 20 July Elbow Point ( And!Oh·Ll ot pot) stickl,•backs 25 AU '!]Util Fig. 13. Daily catches of red salmon fry and threespine sticklebacks in floating lake trap fished in offshore positions, 1957. (Trap anchored at pot end except as indicated. Note that the ordinate scale is not the same for the two species.) 344 Alaska Red Salmon The primary problem. encountered in the offshore trap trials was the low level of the red fry catches {Fig. 13). Several sites were fished in an effort to see whether change in trap location might bring about in- creased catch. No marked improvement in catch level of red fry was obtained by shifting to new locations. The question exists as to whether low catches were due to scarcity of red fry in the areas o:r to inefficiency of the trap in leading the fry where there were no natural shore contours so that the fry could sound under- neath the trap lead. The stickleback catches were much higher in pro- portion to red fry catches than in inshore traps. This may have been par- tially due to differences between the two species in vertical distribution. Results of tow-net sampling in pelagic areas at night have also indicated that the sticklebacks remain nearer the surface. Catches presumably depend on dh·ection of movement of red fry with respect to lake surface currents and the anchored trap. 2 At Elbow Point, the final location, anchoring of the trap was reversed for a trial period by attaching the lead end rather than the trap end to the anchor buoy. In this position the fish must lead "downstream" in the lake current to the trap opening rather than upstream. No marked change in level of red fry catches was achieved, although the level of stickleback catches was dis- tinctly lower during the 13-day trial period while the trap was down- stream from the lead (see Figure 13 and Table 3). It would appear that sticklebacks, and possibly red fry, tend to move or lead upstream, and are thus caught and retained more efficiently when the trap is anchored with the lead trailing. TABLE 3. COMPARISON OF LEVEL OF .CATCHES OF R.ED FRY AND THREESPINE STICKLEBACKS AT ELBOW POINT BEFORE AND AFTER ANCHORING OF LAKE TRAP WAS REVERSED Number Trap Average Daily Catch Dates of Days Anchored Red Fry Three spine by Sticklebacks June 28-July 10 13 pot 6 530 July 11-July 23 1:3 lead 3 44 July 24-Aug. 9 17 pot 6 218 No trials were made with the tra.p anchored in a fixed position with both pot and end of lead fastened. This position, when crosswise to the lake current, would be difficult to hold on stormy d<l.yS. The test at Elbow Point indicated that the efficiency of the trap would vary with change of 2 Temperature sections taken at Lake Nerka have revealed that, as a result of wind action, the water masses in Lake Nerka seldom ap- proach horizontal stability during summer. Even on calm days there is usually a perceptible movement of surface layers. SamPling Red Salmon Fry 345 its axis with respect to lake current. Uniform efficiency is, of course, desirable. The 1957 trials of offshore lake trapping suggested that fluctuations in magnitude of catch from day to day and during the season were less severe than in inshore trap locations, and that offshore catches might persist later in the season than at inshore sites. These, of course, would be highly desirable features. It also appeared desirable to fish the off- shore traps later in the season, since tow netting for fry proved more successful at that time. In 1958 a trap was located again at the site off Elbow Point for the period June 2 to August 20 to determine whether late season catches of red fry would increase witn a shift of red fry from littoral to pelagic areas. The inshore sites at Cabin Bay and Catherine Cove were also fished to detect presence of fry in littoral areas. By July 25 red fry had virtually disappeared in catches of the two inshore traps and fry began to show up regularly after August 1 in catches of a tow net.fished during the season near the surface in pelagic areas. However, an increase in red fry catch in the pelagic area trap failed to materialize, and fry catches totaled only 23 for the entire period of over two and one-half months. The failure to catch red fry was partially attributed to low abundance, as indicated by low parent-spawning density, low fry catches in inshore traps and in the tow nets, and a small seaward migration in 1959. How- ever the stickleback catches were also many times lower than in the previous year in the offshore trap but only slightly lower in the littoral area traps. This would indicate that either vertical or horizontal dis- tribution of the fry and sticklebacks in the pelagic area differed between the two years or the pelagic trap had somehow been modified in effi- ciency. Because there was considerable difference between 1957 and 1958 in development and extent of thermal stratification, some difference in vertical distribution of fry and sticklebacks is believed to have oc- curred. From results of the tests in 1957 and 1958, it was concluded that catches of traps fished in pelagic areas did not provide a reliable meas- ure of red fry abundance. Certainly, modifications that materially in- crease the magnitude of catch must be achieved before offshore traps can be considered a promising alternate methcx:l of enumerating fry abun- dance. SUMMARY The present study concerned primarily trials of floating lake traps as a means of measuring fluctuations in abundance of young salmon in the Wood River laltes, Alaska, during their lake residence. A sampling system that utilized lake traps was established at Lake Nerka in 1953 and tested over a six-year period. 346 Alaska Red Salmon All fish species known to be present in the lake were captured in the traps. The two most abundant species, young red salmon and th.reespine sticklebacks, were captured in greatest numbers. Catches of red salmon fry were found to be heaviest during the first month after breakup of the lake ice in late May or early June, and to drop off by mid -July to a low level. The pattern of ca:f;<· .h was shown to be associated to some degree with timing of spring br(;cikup, thermal development, and growth and behavior of t.lJ.e red fry. In early summer red salmon fry were found to inhabit the littoral areas, but gradually moved to habitats in deeper water as they developed. Aside from the gross feature of decrease in catch by midsummer, the magnitude and pattern of red fry catch varied considerably from year to year. In order to determine whether these patterns of catch could be explained on the basis of _changes in abundance and developmental stage alone, red salmon fry catches were compared with those of threespine sticklebacks. Similarity in seasonal patterns of catch of the two species was found, and differences between years in pattern tended to be the same for both species. These results showed that availability to the gear was variable, which complicates the use of the lake traps for abundance measure- ments. A decrease in inshore catches of both red fry and sticklebacks by mid- summer suggests that the general offshore shift in distribution of both species was not related primarily to developmental stage, since the stick- leOO..cks in the catches were comprised of several age groups at different stages of development. The correspondence in movement patterns of red fry and threespine sticklebacks suggests that they tend to inhabit feeding areas simulta- neously, thus increasing the opportunity for interspecific competition. Varying availability of the fish in littoral areas presents a problem in indexing of abundance by means of stationary gear as well as by beach seine or other mobile gear used only in inshore areas. In 1957 and 1958 flc .. ting lake traps set in positions over deep water some distance from shore were fished to determine the feasibility of their use as a test of, or substitute for, littoral area sampling. Catches of red salmon fry in traps set in offshore locations were much lower than in inshore locations, and the proportion of sticklebacks was higher. Catches of red fry and sticklebacks failed to increase in the pelagic areas with the shift in their distribution away from littoral areas in midsummer. The shift was indicated by decrease in catch in inshore areas and increase in catch in tow nets fished .in pelagic areas. It was concluded that the pelagic traps used did not provide a reliable measure of red fry abundance. SamPling Red Salmon Fry ACKNOV/LEDGMENT 347 This study was part of the research program of the Fisheries Research Institute on the Nushagak red salmon runs. The work was financed by the Bristol Bay salmon canning industry. The author is particularly indebted toW. F. Thompson, Director of the Institute during these studies, for his guidance and inspiration. Thanks are also due in particular toW. A. Church for his extensive as- sistance in conduct of the sampling program at Lake Nerka, to Richard B. Thompson for collection of data at Nerka station as a member of the Institute staff and later as a member of the U.S. Fish and Wildlife Serv- ice, and to John R. Gilbert, Ole A. Mathisen, and Ted S. Y. Koo for cooperation and assistance while working on closely related problems. Finally, I wish to thank Miss Pearl Mooney for her assistance in prep- aration of my original thesis, and William F. Royce for review and criticism of the rgvised manuscript. LITERA'l'URE CITED Babcock, J. P. 1904. Fisheries Commissioner's Report for 1903. Brit- ish Columbia Dept. of Fisheries, Victoria. 15 p. Bower, W. T. 1924. Alaska fishery and fur-seal industries in 1922. Rept. U.S. Commissioner of Fisheries for 1923, Appendix No. IV, p. 1-79. Washington, D. C. Burgner, R. L. 1982. Studies of red salmon smolts from the Wood River lakes, Alaska. See elsewhere in this volume. Chamberlain, F. M. 1907. Some observations on salmon and trout in Alaska. Rept. U.S. Commissioner of Fisheries for 1906, Washington, D. C. 112p. Foerster, R. E. 1925. Studies in the ecology of the sockeye salmon (On- corhynchus nerka). Contrib. Canadian Biol., N. S., 2(16):335-422. Greenbank, J., and P. R. Nelson. 1959. Life history of the threespine stickleback Gasterosteus aculeatus Linnaeus in Karluk and Bare Lake, Kodia..~ Island, Alaska. U.S. Fish and Wildl. Serv., Fishery Bull., 59(153):537-59, Higgins, E. 1929. Progress in biological inquiries, 1927. Rept. U.S. Commissioner of Fisheries for 1928, Appendix No. VI, p. 199-247. Washington, D. C. ----. 1932. Progress in biological inquiries, 1930. Rept. U.S. Com- missioner of Fisheries f6'r 1931, .Appendix No. ill, p. 553-626. Wash- ington, D. C. ----. 1933. Progress in biological inquiries, 1931. Rept. U.S. Com- 348 Alaska Red Salmon missioner of Fisheries for 1932, Appendix No. m, p. 441-529. Wash- ington, D. C. --~--. 1934. Progress in biological inquiries, 1932. Rept. U.S. Com- missioner of Fisheries for 1933, Appendix No. n, p. 79-147. Wash- ington, D. C. ----. 1939. Progress in biological inquiries, 1936. Rept. U. S. Com- missioner of Fisheries for 1937, Appendix No. III, p. 1-61. Wash- ington, D. C. Johnson, W. E. 1956. On the distribution of young sockeye salmon (On- corhynchus nerk~) in Babine and Nilkitkwa Lakes, B. C. J. Fisheries Research Board Canada, 13(5):695-708. -----. 1958. Density and distribution of young sockeye salmon. (On- corhynchus nerka) throughout a multibasin lake system. J . Fisheries Research Board Canada, 1.5(5):961-82. Koo, T. S. Y. 1962. Age and growth studjes of red salmon scales by graphical means. See elsewhere in this volume. Krogius, F. V., and E. M. Krokhin. 1948. Ob urozhainosti molodi krasnoi (Oncorhynchus nerka Walbaum). Izvestiia Tikhookeanslcovo. Nauchno- issledovatelskovo Instituta Rybnovo Khoziaistva i Okeanog. 28:3-27. [On the production of young sockeye salmon Oncorhynchus nerka (Wal- ba.um). Fisheries Research Board of Canada, Trans!~ Ser., No. 109 .• 1958.] ---em-. 1956. Prichiny Ko!ebanii chislennosti krasnoi na Kamchatke. Trudy Problemnykh i Tematicheskikll Soveshchanii ZIN (6): 144-49. [Causes of the f!uctuations in abundance of sockeye salmon in Kam- cha.tka. Fisheries Research Board of Canada, Transl. Ser., No. 92, 1959.] Marsh, M. C.~ and J. N. Cobb. 1909. The fisheries oi Alaska. in 1908, In Rept. U.S. Commissioner of Fisheries for 1908. Washington, D. C. 78 p. Rees, W. H. 1957. The vertical and horizontal distribution of seaward migrant salmon in the forebay of Baker Dam. Fisheries Research Papers, 2{1):5-17. rwash. State Dept. of Fisheries] Ricker, W. E. 1937 . .ibe food and food supply of sockeye salmon(On- corhyncltus nerkaWalba.um) in Cultus Lake, British Columbia. J. Bioi. Board Canada, 3 (5):450-68.