FISHERY BULLETIN: VOL. 80. NO. 1 



coho salmon but probably have an environmen- 

 tal component as well. A genetic basis, as shown 

 in rainbow trout, Salmo gairdneri, has also been 

 established for numbers of vertebrae (Winter et 

 al. 1980a), scales in the lateral series (Winter et 

 al. 1980a), scale rows (Neave 1944), gill rakers 

 (Smith 1969), branchiostegals (MacGregor and 

 MacCrimmon 1977), and anal fin rays (Mac- 

 Gregor and MacCrimmon 1977). Ricker (1970) 

 hypothesized that the meristic characters of 

 salmonids probably have both genetic and en- 

 vironmental components. The difficulty of deter- 

 mining the importance of these phenotypic 

 characters to the fitness of the stock does not pre- 

 clude the possibility that they could, through 

 selection or pleiotrophic effects, have a bearing 

 on fitness as suggested by Barlow (1961). 



The objective of this study was to characterize 

 stocks of coho salmon by using enzyme gene fre- 

 quencies, life history characters, and morpho- 

 logical characters. Secondarily, we hoped this in- 

 formation would help provide a basis for select- 

 ing donor stocks in Oregon hatchery programs. 

 The stocks were selected so that comparisons 

 could be made among geographical areas and 

 stream types and between hatchery and wild 

 stocks. We calculated a measure of a phenotypic 

 similarity and used cluster analysis to display 

 the relationships among stocks. Because cluster 

 analyses are arbitrary (Blackith and Reyment 

 1971), we used two clustering strategies. Factors 

 affecting genetic similarity were hypothesized 

 by determining environmental characteristics 

 common to the similar stocks. 



Although our analysis is primarily systematic, 

 we correlated the phenotypic characters with 

 variables characteristic of the stream systems. 

 Although correlations do not prove a functional 

 significance, they are included here because in- 

 ferences and hypotheses can be developed from 

 the correlations for future studies. 



METHODS 



Sampling 



We evaluated 10 characters for 15 hatchery 

 stocks (based on samples of 75-100 juvenile coho 

 salmon of the 1976 brood from 14 hatcheries in 

 Washington, Oregon, and California and 9 

 hatcheries from Oregon for the 1977 brood year) 

 and 12 wild stocks (based on samples of 30-100 

 juvenile coho salmon of the 1976 and 1977 broods, 

 collected by electrofishing from 12 Oregon 



stream systems). (See Figure 1 for locations of 

 hatcheries and stream systems.) Because some of 

 the hatcheries have used nonnative egg sources, 

 and stream systems have been stocked with 

 juvenile and adult coho salmon, few pure native 

 stocks remain. We did not use hatchery stocks or 



• HATCHERY 



• WILD 



1 



N 



TEN 



MADR 



CALIFORNIA 



Figure 1.— Map indicating sample site locations of wild and 

 hatchery coho salmon stocks. Location codes are as follows with 

 the hatcheries in parentheses: ALSE. Alsea River (Fall Creek 

 Hatchery); BEAV, Beaver Creek; BIGC, BigCreek(BigCreek 

 Hatchery); COLM. Columbia River (Cascade Hatchery in 1976 

 and Bonneville Hatchery in 1977); COQL, Coquille River; 

 COWL, Cowlitz River (hatchery stock reared at Cascade 

 Hatchery in 1976 and Big Creek Hatchery in 1977); KLAM, 

 Klamath River (Irongate Hatchery); MADR, Mad River (Mad 

 River Hatchery); NEHA, Nehalem River; NEST, Nestucca 

 River; NONE, North Nehalem River (North Nehalem River 

 Hatchery); QUIL, Quilcene River (Quilcene River Hatchery); 

 QUIN, Quinault River Hatchery); ROGU, Rogue River (Cole 

 Rivers Hatchery); SALM, Salmon River Hatchery); SAND, 

 Sandy River (Sandy River Hatchery); SILZ, Siletz River; 

 TENM, Tenmile Lakes; TRAS, Trask River (Trask River 

 Hatchery); TRIN, Trinity River (Trinity River Hatchery); 

 UMPQ, Umpqua River (hatchery stock collected from Smith 

 River and reared at Cole Rivers Hatchery). 



106 



