Beacham et al.: Geographic basis for population structure in Oncorhynchus tshawytscha 



233 



years within locations. In the Harrison River- Lillooet River 

 drainage, the Birkenhead River was treated as a single 

 population with four years of samples in this analysis. 



Results 



Variation within populations 



The number of alleles observed per locus ranged from 

 12 to 54, and fewer alleles were observed at the loci with 

 dinucleotide repeats (Ogo2, Ogo4, Oke4, Omy325, Ots2, 

 and Ots9) than at the loci with tetranucleotide repeats. 

 Heterozygosities were generally lower for the loci with 

 dinucleotide repeats ( mean //.=0. 7 1) than for the other loci 

 (mean //^=0.91) (Table 2). Genotypic frequencies at each 

 locus within sampling location and year generally con- 



formed to Hardy-Weinberg equilibrium expectations, with 

 the notable exception of Otsl02. At this locus, 57% of the 

 HWE tests were statistically significant, and in all cases 

 there was an excess of observed homozygotes. One or more 

 nonamplifying allele(s) was likely present at this locus in 

 many of the Fraser River populations surveyed. Observed 

 heterozygosities were less than expected heterozygosities 

 for all loci analyzed on the manual gels. Unequal amplifi- 

 cation of alleles (Wattier et al., 1998) may have resulted 

 in the failure to detect large alleles in some heterozygous 

 individuals with ethidium bromide staining. There was no 

 evidence of linkage between any of the microsatellite loci 

 used in this study, but four of the 52 samples surveyed in 

 the study exhibited significant linkage disequilibrium in 

 more than 109^ of the pairwise comparisons between loci 

 ( Table 1 ). These sample locations were Hairison River, Tete 

 Jaune, Fontoniko, and Bessette Creek. 



