760 



Fishery Bulletin 100(4) 



resolve population differentiation (e.g. Rousset and Ray- 

 mond, 1995; Goudet et al., 1996). Nevertheless, empirical 

 support for this relationship is mixed (e.g. Pogson et al., 

 1995; Estoup et al, 1998; Scribner et al., 1998 ). We believe 

 our results suggest that mutation, which varies among 

 these marker classes, has relatively little bearing on 

 detecting genetic variation among walleye pollock popu- 

 lations. Instead, spatial differentiation may be limited 

 because populations are large (i.e. genetic drift is weak) 

 and gene flow among populations is high. The fact that 

 only one private allele was found at the most variable 

 microsatellite locus (Middleton Island, locus Teh 10. allele 

 147) is an argument for high gene flow. Further, the index 



of population structure Rgj, (Slatkin, 1995) developed to 

 reflect the high mutation rate and mutation process of 

 microsatellites is no larger than 6. The similarity of these 

 values suggests that gene flow is relatively high and that 

 mutation has relatively little influence on population dif- 

 ferentiation (Slatkin, 1995). 



Our second conclusion is that balancing selection may 

 influence some but certainly not all allozymes in walleye 

 pollock, as posited for marine species exhibiting low varia- 

 tion in allozyme allele frequencies (Pogson et al., 1995). 

 Our conclusion is supported by two observations. First, 

 microsatellites and mtDNA provide no greater resolution 

 of overall population structure than allozymes (Tables 5 



