FISHERY BULLETIN: VOL. 71, NO. 3 



however, no sockeye were seen beyond seven miles 

 upstream, and nearly the entire spawning population 

 confined itself to the lowest mile of the stream. . . . 

 [Schaefer 1951:26.] 



The Birkenhead sockeye supported a hatchery from 

 1905 to 1935, when all the British Columbia hatcheries 

 were closed. Here eggs were taken and the fry held 

 until absorption of the yolk sac. They were not reared 

 beyond the yolk-sac stage. ... It is apparent that al- 

 though the egg-takes fluctuated considerably they are 

 of about the same size in the later years as in the earlier 

 years of operation. This record, however, is somewhat 

 misleading as an inde.x to the relative sizes of the 

 populations from year to year, since a variable fraction 

 was taken by the hatchery men. From examination of 

 the remains of structures on the river banks, and from 

 conversations with local inhabitants and employees of 

 the Dominion Fisheries Department, it has been possible 

 to ascertain that the lake of eggs was maintained by 

 the expedient of including an ever increasing share of 

 the population in the operation. The first fence for 

 capturing sockeye was placed some distance above Owl 

 Creek. As the eggs became harder to get, fences were 

 erected at successively lower points until the last fence 

 was located at the lowest practicable place for estab- 

 lishing the conventional type of fence for capturing 

 upstream migrants (see Figure 9). In the last few years 

 of operation the captures at this fence fell so low that 

 they were supplemented by gaffing ripe fish out of the 

 stream below. At the lime of our study the preponderate 

 majority of sockeye were found spawning well below 

 the former site of the lowest fence, and the number of 

 sockeye ascending above that point could by no means 

 supply as many eggs as were taken by the hatchery in 

 its last years of operation. [Schaefer, 1951:32-33.] 



Schaefer's statement that the number of eggs 

 from the Birkenhead sockeye salmon was only 

 maintained by taking a larger and larger per- 

 centage of the spawning stock is not quite true. 

 In fact, the weir seemed to kill off most sockeye 

 salmon that spawned above it, except for those 

 that must have escaped upstream, and after 

 about 4 yr in each location an insufficient num- 

 ber came up to the weir to provide the numbers 

 of eggs required to fill the hatchery. When this 

 happened, the weir was moved downstream to 

 a new location where the process of taking eggs 

 was continued until the spawning stock had 

 been killed off and it was necessary to move the 

 weir again. When sockeye salmon hatcheries 

 were closed in British Columbia in 1935, Schae- 

 fer found that the weir had been moved down- 

 stream as far as it could be conveniently built 

 on the Birkenhead. But, even though no weir 

 had been installed on that stream after 1935, 

 most sockeye salmon spawning was still con- 



fined to the lower end of the Birkenhead River 

 until after 1947. Between 1947 and 1950 the 

 Canadian Prairie Farm Rehabilitation Admin- 

 istration attempted to divert the lower Birken- 

 head River into a newly dug channel. In addi- 

 tion Lilloet Lake was lowered 6 feet in 1950. 

 One result of the changes made by the Prairie 

 Farm Rehabilitation Administration was to de- 

 stroy a part of the original spawning grounds 

 of the lower Birkenhead River. (Memorandum 

 7 November 1959, M. C. Bell to L. A. Royal, 

 International Pacific Salmon Fisheries Com- 

 mission.) 



Bell concluded from his study of the Birken- 

 head that the upper part of that river is inhabit- 

 ed by an early run of fish ". . . that are from 8 

 to 10 cm less in length than the late fish . . . ." 

 which occupied the lower river. He stated that 

 some way would have to be found to stabilize a 

 spawning area for this late run to. correct the 

 harm resulting from the work of the Prairie 

 Farm Rehabilitation Administration. 



The implications of the effect of the Birken- 

 head weir are many. Of course the death of the 

 runs above the weir was due partly to the lack 

 of eggs, but then, if this killed off the run to a 

 particular section of a straight stretch of river, 

 this must mean that an independent "race" of 

 salmon inhabited, i.e., spawned, in that partic- 

 ular section of the stream. In other words, it 

 indicates that different parts of a continuous 

 section of a stream can be inhabited by distinct 

 races of sockeye salmon which are closely 

 adapted to the different conditions they en- 

 counter during the spawning and subsequent 

 rearing of eggs and young. This was substanti- 

 ated by Bell's report of 1959. 



While the Karluk weir has always been pro- 

 vided with gates through which the fish are 

 counted upstream, no study has ever been made 

 of the length of time sockeye salmon have been 

 delayed in finding their way through. Again 

 work at Hell's Gate by the International Pa- 

 cific Salmon Fisheries Commission has indicat- 

 ed that such delays can be fatal. Talbot (1950) 

 concluded that 14 days' delay at Hell's Gate 

 was sufficient to prevent successful comi)letion 

 of migration of sockeye salmon to any of the 

 streams above. Construction of the fish ladders 

 at Hell's Gate eliminated this cause of mortality 



638 



