Teel et al : Genetic analysis of juvenile Oncorhynchus kisutch 



649 



Nonetheless, the usefulness of the allozyme baseline that 

 we compiled for coho salmon is limited by two factors. First, 

 few samples in the baseline are from California and British 

 Columbia populations. Although the baseline appears to be 

 adequate to analyze stock mixtures of juvenile coho salmon 

 off Oregon and Washington, mixed stock analyses of sam- 

 ples from other marine areas, particularly to the north, re- 

 quires the sampling of additional populations. Second, our 

 study demonstrated that estimates of stock compositions 

 are not sufficiently accurate to effectively identify stock 

 groups that are absent from mixtures or present in small 

 proportions (Tables 4 and 6). Estimation accuracy can be 

 improved by using additional gene markers. These markers 



will likely be based on DNA variability because coho salmon 

 minisatellite (Miller et al., 1996; Beacham et al., 1996) and 

 microsatellite (Small et al., 1998a; 1998b; Beacham et al., 

 2001) loci show much higher levels of polymorphism than 

 do allozyme loci. Recently, variation at eight microsatellite 

 DNA loci and one Mhc locus in coho salmon populations 

 in British Columbia and Washington was used to estimate 

 the stock compositions of fisheries off the west coast of 

 Vancouver Island (Shaklee et al., 1999; Beacham et al., 

 2001). However, the use of highly polymorphic microsatel- 

 lite loci may not provide increased discrimination among 

 populations on large geographical scales because of allelic 

 convergence from multiple mutations (Nauta and Weiss- 



