REISENBICHLER and PHELPS: GENETIC VARIATION IN CHINOOK AND COHO SALMON 



Figure 6. — Coho salmon — common-allele frequencies 

 {q ) fo r several pro tein-coding loci. Each horizontal bar 

 is 4V'g(l - q )l2n in length and approximates the 95% 

 confidence interval; n = number of fish scored. Fre- 

 quencies for fewer than 25 fish are not shown and were 

 not used in analysis. 



Bogachiel R. 

 Calawah R. 

 Clearwater R. 

 Dickey R. 

 Hoh R. 

 Oueets R. 

 Quinault R. 

 Soleduck R. 

 Snohomish R. 

 Snow Creek 

 Quinault NFH 

 Soleduck H. 



^ bGala-3 



0.0 



0.2 



0.4 



0.6 



0.8 



Bogachiel R. 

 Calawah R. 

 Clearwater R. 

 Dickey R. 

 Hoh R. 

 Oueeis R. 

 Quinault R. 

 Soleduck R. 

 Snohomish R. 

 Snow Creek 

 Quinault NFH 

 Soleduck H. 



ldh-3,4 



1.0 



0.0 



0.2 



0.4 



0.6 



0.8 



1.0 



Bogachiel R. 

 Calawah R. 

 Clearwater R. 

 Dickey R. 

 Hoh R. 

 Oueets R. 

 Quinault R. 

 Soleduck R. 

 Snohomish R. 

 Snow Creek 

 Quinault NFH 

 Soleduck H. 



Np-1 



\ \ I I \ \ \ r- 



0.0 0.2 0.4 0.6 0.8 



Frequency of 100 allele 



1.0 



Table 6. — Coho salmon — hierarchical analysis of electrophoretl- 

 cally detectable gene diversity for coho salmon from the Quillayute, 

 Hoh, Queets, Quinault and Wynoochee Rivers. Analysis was 

 based on 58 loci, including 36 that were monomorphic. The hierar- 

 chical design is shown in Figure 3. 



Relative gene diversity (%) 



Locus 



Total gene 



diversity 



(Ht) 



Within 

 samples 



Among 

 broods 



Average 



Average 

 excluding 

 Pnp-1 



0.021 



0.016 



95.09 0.09 



97.64 



0.12 



Within 

 drain- 

 ages 



0.85 



0.93 



Among 

 drain- 

 ages 



3.97 



1.31 



analysis; Grant et al. 1980; Smouse and Ward 

 1978). 



Samples without data for bGala-1 or Idh-3,4, 

 the most variable loci, were omitted from the 

 UPGM cluster analysis (Fig. 7) and multidimen- 

 sional scaling (Fig. 8). Both analyses showed that 

 fish from Quinault NFH were distinct from wild 

 fish; much of this distinctiveness occurred at the 



bGala-2 locus (Fig. 6). Fish from Snow Creek and 

 the Snohomish River clustered among the wild 

 fish from the north coast. The results were simi- 

 lar when Pnp-1 was excluded from the analysis, 

 except that fish from the upper Queets River were 

 no longer distinct from the other wild fish. 



DISCUSSION 



Wild Populations 



Variation in allele frequencies among drain- 

 ages for chinook salmon was not statistically sig- 

 nificant. The inability to detect differences among 

 drainages could have resulted from 1) low statis- 

 tical power (probability of rejecting Hq if it is 

 false) because we had too few broods or because 

 variation in racial composition of juveniles in dif- 

 ferent years inflated the estimates of error vari- 

 ance, 2) our exclusive reliance on data for genes 

 that can be sampled by electrophoresis, or 3) a 

 lack of true genetic difference among groups. We 



691 



