UTTER ET AL.: GENETIC POPULATION STRUCTURE OF CHINOOK SALMON 



ecological requirements of sockeye salmon during 

 their freshwater life history. This irregular distribu- 

 tion is accompanied by greater geographic hetero- 

 geneity of allelic distributions, perhaps reflecting 

 severe founder events and restricted gene flow 

 (Utter et al. 1984). One population of sockeye salmon 

 on the Quinault River (Washington coast) deviated 

 strongly from all other groups sampled, but the 

 possibility of a coastal unit of sockeye salmon, anal- 

 ogous to that of Chinook salmon (i.e., unit IV), 

 appears unlikely. Allele frequencies from Lake 

 Ozette on the Washington coast (W. K. Hershber- 

 ger*) were typical of noncoastal populations. Popula- 

 tions north of the Skeena River (approximately the 

 position of "A" in Figure 1) are distinguished by the 

 presence of Ldh-4 variation which is virtually ab- 

 sent from more southern groups (Utter et al. 1980; 

 Withler 1985), presumably reflecting postglacial 

 repopulation from a more northern refuge. 



Studies of population groups of chum salmon and 

 coastal cutthroat trout within Puget Sound and 

 Georgia Stait suggest similar genetic structures to 

 that observed in chinook salmon. Populations of 

 chum salmon from south Puget Sound were distin- 

 guishable from those of north Puget Sound and 

 Georgia Strait (Okazaki 1981). Populations of 

 Georgia Strait and the lower Fraser River were 

 likewise distinguishable from populations immedi- 

 ately north of Georgia Strait (Beacham et al. 1985). 

 Intensive subsampling of cutthroat trout within 

 Hood Canal and north Puget Sound indicated strong 

 and consistent differences between these regions 

 (Campton and Utter 1987). 



More comprehensive comparisons will be possible 

 as data accumulate on these and other species of 

 anadromous salmonids. Both the similarities and the 

 differences observed are of considerable interest in 

 gaining further insights into the determinants of 

 allele frequency variation, zoogeography, behavior, 

 and management of these species. 



Effects of Hatchery Operations 



Further consideration of the effects of hatchery 

 operations is also warranted. Hatchery operations 

 and transplanted hatchery fish do not appear to have 

 drastically altered the geographic distributions of 

 protein coding alleles among the major population 

 units. There is presently little question that hatchery 

 operations have homogenized allele frequencies 

 among many fall chinook hatcheries of the lower 



Columbia River (Simon 1972). However, the tem- 

 porally isolated spring and fall populations of this 

 region retain a greater similarity to one another 

 than to populations of other regions. Thus it seems 

 probable that the allele frequencies of unit V approx- 

 imate those existing prior to the present century in 

 spite of this region's large predominance of hatch- 

 ery fish. Hatchery populations established from (and 

 still reflecting) exotic origins (e.g., Carson and 

 Leavenworth Hatcheries) have not noticeably per- 

 turbed the allelic distributions of adjacent popula- 

 tions having indigenous origins (Utter et al. 1987^). 

 Where they exist (e.g., unit IV), indigenous wild and 

 hatchery populations within a unit are generally 

 separated by small genetic distances, reflected by 

 close aggregations in the dendrogram and principal 

 component clusters. 



Infrequent alleles do not strongly affect genetic 

 distance or heterozygosity, but their loss in hatch- 

 ery stocks relative to comparable wild populations 

 is a good indication of an inadequate number of 

 spawning individuals used to establish or maintain 

 a hatchery stock (Allendorf and Ryman 1987). A 

 comparison was therefore made of the average 

 number of alleles per locus and heterozygosity 

 between seven hatchery and six wild samples from 

 the Oregon coast, the most extensive collection of 

 hatchery and wild samples within a restricted geo- 

 graphic range made in this study (two statistically 

 indistinguishable combined populations each involv- 

 ing a hatchery and a wild sample were excluded). 

 The mean values were very similar (heterozygos- 

 ity— hatchery 0.137, wild 0.132; alleles per locus- 

 hatchery 1.74, wild 1.68) and were not significant- 

 ly different. Presumably, sufficient numbers of 

 breeders have been used in Oregon coastal hatch- 

 eries to prevent losses of heterozygosity or alleles. 

 However, the data provide no information concern- 

 ing possible losses of genetically distinct geographic 

 or temporal segments as a result of hatchery prac- 

 tices along the Oregon coast. 



The present data set also pertains to additional 

 aspects of hatchery management. Evidence con- 

 tinues to accumulate from numerous sources that 

 individual populations of anadromous salmonids 

 represent gene pools that are uniquely adapted to 

 a particular location and spawning time (see Ricker 

 1972). Stocks transferred to areas beyond those to 

 which they are locally adapted perform poorly 



»W. K. Hershberger, Univ. of Washington, Seattle, WA 98195. 

 pers. commun. December 1985. 



'Utter, F., P. Aebersold, M. Griswold, G. Milner, N. Putas, J. 

 Szeles. D. Teel, and G. Winans. 1987. Biochemical genetic vari- 

 ation of chinook salmon stocks of the mid-Columbia River. Pro- 

 cessed Report 87-19, 22 p. Northwest and Alaska Fisheries 

 Center, Seattle, WA 98112. 



251 



