FISHERY BULLETIN: VOL. 71. NO. 3 



tion of the run. The 43's amount to a little over 

 1% and the 73's less than 0.1% of the total run. 

 Rounsefell relies on correlations between fish 

 returning at different times of the season to 

 prove one population. It would seem entirely 

 possible that the total returns from two groups 

 of discrete populations that run prior to and 

 after a given date could be negatively correlated. 



Rounsefell (1958:142) used his Figure 54 to 

 show that "one can not compare pack during 

 one period with escapement in another period 

 as escapement was not uniform in different 

 l)arts of the season." This figure actually shows 

 that the percentages of the annual salmon catch 

 that was packed during different j^ortions of the 

 season for 1895 to 1919 and for 1921 to 1950 

 were almost identical and in reality demonstrat- 

 ed that the catch has been taken mostly from 

 the midseason runs as was shown by Thompson 

 (1950) and by Thompson and Bevan (1954) in 

 the early 1950's. Rounsefell (1958:143) how- 

 ever concluded that the Karluk runs had not 

 been overfished: ". . . . It is not the rate of ex- 

 ploitation that has changed, but the ability of 

 the populations to withstand even a greatly 

 diminished rate." This in effect was the same 

 as Thompson's conclusion except that Rounse- 

 fell offered no explanation for this decline in 

 productivity. He recommended in part: 1) con- 

 trol of predatory species of fish in Karluk Lake, 

 after demonstrating that partial control of 

 predators can result in increased predation, and 

 2) regulation to protect more midseason fish, etc. 



Owen, Conkle, and Raleigh (1962),^ after 

 studying the distribution of spawners of differ- 

 ent ages at different times each season, conclud- 

 ed that: 



( 14) Based on information and analyses now on hand: 

 (a) A change in the lacustrine environment has not 

 been demonstrated, (b) Long-term physical changes in 

 the spawning habitat have not been demonstrated, (c) 

 Changes in the proportion of the total escapement 

 utilizing the different specific spawning areas are indi- 

 cated, inasmuch as the fishery over the years has been 

 concentrated on the midseason portion of the run, and 

 the specific spawning areas are occupied in season se- 

 quence, (d) Changes in the biological characteristics of 



the spawners have been suggested, principally in the 

 prolonged lake residence indicated by a higher propor- 

 tion of 4-freshwater sockeye salmon. Other associated 

 changes are the disappearance of the formerly predomi- 

 nant 5-year cycle and the appearance of larger relative 

 numbers of 4-freshwater grilse. 



{ 15) Changes (c) and (d) above can both be explained 

 as possible effects of the concentrations of the commer- 

 cial fishery on the midseason Karluk sockeye salmon 

 run, leaving spawning escapement to occur in the fall 

 and in the spring. It has recently been shown that the 

 age composition of returning sockeye salmon is chiefly 

 determined by the age composition of their parents. 



They recommended that research be directed 

 at evaluating the productive quality of the season 

 escapements themselves. 



Ricker (1972) discussed Rounsefell's conclu- 

 sions concerning the existence of separate races 

 of sockeye salmon in Karluk Lake on the basis 

 of his interpretation of variations in abundance 

 of the different combinations of freshwater and 

 saltwater ages in different segments of the Kar- 

 luk run. Ricker concluded that: 



.... As 1 see it. however, none of the information 

 presented precludes the possibility of considerable dis- 

 creteness of stocks arriving at different seasons, provid- 

 ed the stocks are distinguished by having different 

 proportions of the different life-history types, as is actual- 

 ly the case (personal communication from Dr. J. B. 

 Owen). [Ricker, 1972:41.] 



.... Thus the balance of local evidence, as well as 

 the analogy with other areas, favours the existence of a 

 number of separate sockeye stocks at Karluk Lake. 

 [Ricker, 1972:42.] 



Gard and Drucker (1972)'' demonstrated the 

 existence of these races by significant differ- 

 ences between early sockeye salmon spawners 

 in the lateral tributaries of Karluk Lake, in the 

 upper Thumb. O'Malley River, and Canyon 

 Creek and those which spawn later in these ter- 

 minal streams, and on the Thumb Beach in 

 Karluk Lake near the outlet of the lower Thumb 

 River. The late spawners showed greater mid- 

 eye to fork length and greater fecundity at com- 

 parable lengths. Gard and Drucker did not in- 

 clude comparison with fish spawning in the 

 Karluk River but concluded that ". . . . Differ- 

 ence in length, age and fecundity among spatial 

 and temporal segments of red salmon have been 



4 Owen, J.B., C. Y. Conkle, and R. F. Raleigh. 1962. 

 Factors possibly attectmg proauction 01 socKeye saimon 

 in Karluk River, Alaska. U.S. Bur. Commer. Fish., Auke 

 Bay Biol. Lab., Manuscr. Rep. 62-8, 57 p. 



•^Gard, R., and B. Drucker. 1972. Differentiation and 

 cause of decline of sockeye salmon of the Karluk River 

 system, Alaska. Unpubl. manuscr. Auke Bay Fish. Lab., 

 Natl. Mar. Fish. Serv., NOAA, Auke Bay, AK 99821. 



630 



