VAN CLEVE and BEVAN: DECLINE OF KARLUK SALMON RUNS 



river, and with sufficient effort, the seines could 

 block the upstream movement of spawners just 

 as effectively by operating in the lagoon as they 

 had in the river itself. Gilbert and Rich (1927) 

 thought that the downward trend could be re- 

 versed if an adequate number of fish were al- 

 lowed to escape the fishery each year to spawn. 

 Their objective was to make sure that at least 

 1 million sockeye salmon sjiawned each year. 



Barnaby (1944) found that the relative num- 

 bers of fish which had migrated to sea as smolts 

 in their third and fourth years had changed. 

 From this, and from the small amount of phos- 

 phorous and nitrogen he found in Karluk Lake 

 compared with that in the spawning streams. 

 as well as from an increase in the proportion 

 of 4-yr smolts, he concluded that the length of 

 freshwater residence had increased, and thus 

 that the productivity of Karluk Lake had de- 

 creased. 



Thompson (1950), using the number of cases 

 of Karluk red salmon packed each day by one 

 packer, between 1900 and 1919. showed that 

 during the period of 1900 to 1904 the Karluk 

 run had a single peak which occurred early in 

 August (Thompson, 1950, Figure 10). During 

 the next 5 yr (1905-09), this midseason peak 

 decreased and while the run seemed to slack 

 off in early July, the daily pack remained al- 

 most constant at an average of about 7,000 

 cases (4,000 to 11,000) up to mid-September 

 (Thompson. 1950, Figure 11). Thompson's Fig- 

 ure 12 for 1910 to 1914 showed an almost uni- 

 form daily pack for the entire season from 

 early June to mid-September with a slight indi- 

 cation of peaks in June and September. These 

 "spring" and "fall" peaks had become well 

 marked in his Figure 13, which showed the 

 total daily case pack for 1915 to 1919. The 

 catch during that period was depressed from 

 30 June to late August, and the configuration 

 of the run found in 1921 had been established. 

 Thompson (1950) concluded that the midseason 

 fish, which had supported most of the catch in 

 the early fishery and hence were the most pro- 

 ductive segments of the run, had been fished 

 out by the fishery which concentrated most of 

 its effort on the middle part of the run. Further- 

 more, he stated that the productivity of the 

 entire Karluk run of sockeye salmon had been 



reduced as the run was cut back to the less 

 productive early and late fish. On the basis of 

 this analysis, Thompson and Bevan (1954) ad- 

 vocated managing the Karluk sockeye salmon 

 fishery so as to fish only the early (June) and 

 late (September) segments of the run, allowing 

 all the midseason fish to spawn. 



Discussing the similarity in the history of 

 the decline of Karluk sockeye salmon to that 

 of the Chinook salmon runs of the Columbia 

 River, and the sockeye salmon inins of the 

 Fraser. Thompson (1950) stated: 



In each case the change took place relatively early 

 in the history of the rivers. A large degree of indepen- 

 dence must exist between the different sections of the 

 runs in both cases. And the most productive part of 

 the season has been least protected, so that the best 

 part of the run has vanished. The independence of 

 different sections of the runs is strongly corroborated 

 by what we know of the red salmon runs in a third 

 river, the Fraser in Canada. Here tagging operations 

 have clearly shown that the early runs pass far into the 

 upper Fraser. The late runs spawn mostly in streams 

 near the mouth. Radical reduction of the early runs 

 occurred due to railroad construction along the main 

 river, but other and later runs continued. Even where 

 the early and late runs used the same streams, as in 

 the Adams River, they were depleted and rebuilt inde- 

 pendently of each other. It is not only logical but 

 inescapable that a run can be over-fished in one part 

 and under-fished in another, and I can see no other 

 explanation of the peculiar present character of the runs 

 in the Columbia, the Karluk, and the Fraser. 



Rounsefell (1958) concluded that the decline 

 since 1900 was not associated with fluctuations 

 in rainfall or temperature, but found reason to 

 believe that predatory fishes such as Dolly Var- 

 den. SalvcliiiKs nialnia, and Arctic char, S. 

 alpiniis, might have been a factor. He disagreed 

 with Thompson that different races of sockeye 

 salmon could exist in the Karluk watershed and 

 concluded that only one population of fish exists 

 in the Karluk by a study of the occurrence of 

 different ages offish in different years. 



Rounsefell has suggested that "the Karluk 

 sockeye salmon comprise one population, since 

 the number of fish of the same brood running 

 at different seasons, and even in different years, 

 are significantly correlated." Unfortunately, he 

 does not point out that the correlations present- 

 ed concern only a few of the i)ossible age combi- 

 nations involved and those that he does show 

 as being correlated make up only a minute frac- 



629 



