VAN CLEVE and BEVAN: DECLINE OF KARLUK SALMON RUNS 



demonstrated. These findings support the theory 

 that distinct subpopulations of red salmon sep- 

 arated by space and time exist in Karluk Lake 

 and that these subpojuilations possess widely 

 varying reproductive cai)acities. . . ." These re- 

 sults confirmed the hypothesis of separate 

 races put forward by Thompson (1950), Thomp- 

 son and Bevan (1954). and Owen et al. (see 

 footnote 4). 



In Bulletin No. 10 of the International North 

 Pacific Fisheries Commission (1962:91), the 

 annual escapement and return of sockeye salm- 

 on to Karluk Lake were shown in Table 8. These 

 figures were taken from Rounsefell's Table 12 

 (Rounsefell, 1958:112) which were "corrected 

 for climate, for odd- and even-numbered years, 

 and for density of young in the lake, 1890-1948" 

 although that density had never been measured. 

 The princij)al result of such "corrections" based 

 upon hypothetical relationships was to reduce 

 the variability of the observed figures. 



Rounsefell (1958, Append. D) calculated the 

 total run into the Karluk for each year before 

 1921 by multiplying the catch by 1.538, on the 

 following basis: Rich (in Gilbert and Rich, 1927) 

 says, "The spawning escapement [of 1926] was 

 the best in many years, and in all probability 

 was the best that has ever been observed by 

 the few white men who have visited the lake."*^ 

 Rounsefell then assumed that if Rich's state- 

 ment were correct, he could conclude that earli- 

 er escapement must have been proportionately 

 smaller than in more recent years and there- 

 fore he (Rounsefell) used the number of sockeye 

 salmon (15,000) which Shuman (1950) said 

 spawned in Moraine Creek in 1947 and the 

 number of fish Rutter (quoted by Chamberlain, 

 1907) estimated to be in Moraine Creek in 1903, 

 to obtain a ratio of 22/15 of the escapement in 

 1947. Since the escapement in 1947 was 485,000 

 fish, the escapement in 1903 was estimated as 

 22/15 of 485,000 or 711,000 fish. The catch in 

 1903 was 1,320,000 fish, which would make the 

 escapement only 35% of the total run. Rounse- 

 fell admitted that the calculation was probably 

 not valid, but used it nevertheless to compute 

 the number of fish in the total run each vear 



" This was certainly a rhetorical statement with little 

 foundation other than Rich's impression of a "big"" run. 



before 1921 as the catch divided by 0.65 oi- 

 multiplied by 1.538. 



Since Shuman's count for Moraine Creek 

 escapement was obtained by a weir, while Rut- 

 ter's estimate was from a stream survey. Rounse- 

 fell's factor was too small. In 1952. Bevan and 

 Walker (1955) counted the spawners in Moraine 

 Creek on weekly stream surveys and obtained a 

 peak count of 2,730. The total count through a 

 weir maintained in Moraine Creek in 1952 gave 

 an escapement of 7,921. From this it can be in- 

 ferred that Rutter's estimate of escapement in 

 1903 may have been as little as one-third of the 

 actual numbers. In 1948, Shuman counted 

 62,000 fish into Moraine Creek. Calculations 

 similar to Rounsefell's for 1948 and 1952 pro- 

 duce an escapement to Moraine Creek of 17% 

 and 54% of the total run. While Moraine Creek 

 does not appear to be representative of the entire 

 run as Rounsefell himself suggested, Rounse- 

 fell's conversion factor is about the average of 

 these two extremes. 



A careful reading of Gilbert and Rich (1927) 

 also shows that they felt that Rutter's estimate 

 of the total number that spawned in Moraine 

 Creek in 1903 was about one-half as large as 

 the true figure. 



The assumptions involved in the above "cor- 

 rections" and conversion of catch to total run 

 undoubtedly make them too conservative so 

 that the sizes of total runs before 1921 must 

 have been larger. However these figures were 

 used in Figure 10 of International North Pacific 

 Fisheries Commission (1962:92) to calculate 

 the average relationship between spawners and 

 returns, and thus give a conservative picture of 

 the decline in the Karluk sockeye salmon runs. 

 The lower rate of return per spawner shown 

 for 1929-1948 compared with 1870-1928 would 

 indicate that the actual reduction in the Karluk 

 run must have been much greater than was 

 shown by Rounsefell. No basis has been found 

 for the division of the data into two parts in 

 1928. This division was apparently an arbitrary 

 decision by Rounsefell. There is no question 

 that the return per spawner after 1928 was less 

 than before 1928, and the decline in productiv- 

 ity is obvious, but there was no reason to sepa- 

 rate the earlier from later periods at 1928. 



The Karluk system was compared with the 



631 



