Olsen et al : An examination of spatial and temporal genetic vanation in Theragm chalcogiamma 



753 



Gulf of Alaska \ 



If 



,/^ 



X 



T" 



T 



T 



Figure 1 



Map of the Gulf of Alaska and Bering Sea showing the location of walleye pollock 

 spawning aggregations sampled for this study. 



single study of mitochondrial DNA (mtDNA) sequence 

 variation (Shields and Gust, 1995), but it is not evident 

 in an unpublished study of microsatellite variation at two 

 loci (reviewed in Bailey et al., 1999). Conversely, genetic 

 difTerentiation among eastern Bering Sea and Gulf of 

 Alaska populations is supported by the microsatellite 

 study but is not supported by a study of mtDNA-RFLP 

 (restriction fragment length polymorphism) variation 

 (Mulligan et al., 1992). These discordant results may 

 reflect attributes of walleye pollock population biology, 

 the unique evolutionary properties of each marker class, 

 or both. Nevertheless, an interpretation of the collective 

 data is not possible because the studies were conducted as 

 much as 22 years apart, and many samples were not from 

 spawning aggregations. 



Nonrepresentative sampling may bias attempts to 

 define genetic population structure in walleye pollock, 

 particularly on a fine spatial scale (within sea basins). For 

 example, some studies using allozymes (Grant and Utter, 

 1980) and mtDNA (Shields and Gust, 1995) have included 

 population samples taken during summer and fall, even 

 though walleye pollock aggregate for spawning in late 

 winter and early spring. These samples may obscure ge- 

 netic differentiation because walleye pollock populations 

 are believed to mingle in summer and fall during feeding 

 migrations (Bailey et al., 1999). Also, the sample size used 

 in some studies may be inadequate to detect population 

 structure. The sample size in walleye pollock studies thus 

 far has been typically less than 50, and the minimum 

 sample size has been 22 (e.g. Grant and Utter, 1980; Mul- 

 ligan et al., 1992). Sample sizes in this range (22-50) are 

 generally insufficient to detect statistically significant 

 differences in allele frequencies between populations that 

 exhibit weak population structure (e.g. Fgj <0.02; Goudet 

 et al., 1996; Ruzzante, 1998; Waples, 1998). These studies 

 suggest that a sample size of 50 should be considered an 

 absolute minimum for high gene flow species and that 



sample sizes of 100 or greater may be necessary when al- 

 lele frequencies differ by 5% or less. 



In our empirical study, we accounted for the sources of 

 sample bias described above and addressed two important 

 questions: 1) Do allozymes, mtDNA, and microsatellites 

 provide concordant estimates of inter- and intraregional 

 population structure in walleye pollock? 2) Is the genetic 

 variation in walleye pollock populations stable from year 

 to year? We tested for spatial patterns of genetic variation 

 among six population samples from three regions: Gulf of 

 Alaska, eastern Bering Sea, and eastern Kamchatka. We 

 also tested for interannual stability of the genetic signal 

 in replicate samples from three of the North American 

 populations. Our results showed that two of 24 polymor- 

 phic allozyme loci (SOD-2*. MPI*) reveal significant spa- 

 tial heterogeneity at the inter- and intraregional level. In 

 general, the mtDNA data were concordant with these two 

 allozymes, revealing significant genetic variation between 

 North American and Asian (eastern Kamchatka) popula- 

 tions, as well as evidence of intraregional genetic varia- 

 tion, particularly among the Gulf of Alaska populations. 

 The microsatellites, although highly polymorphic, showed 

 little spatial variation. The allozyme and mtDNA data 

 provided evidence of interannual genetic variation in two 

 North American populations. F^j values showed this in- 

 terannual variation is of similar magnitude to the spatial 

 variation in these populations. 



Materials and methods 



Sample collection 



Tissue samples for allozyme and DNA analysis were 

 obtained from six spawning aggregations of walleye pollock 

 representing three major geographic regions: Gulf of Alaska, 

 eastern Bering Sea, eastern Kamchatka (Fig. 1, Table 1). 



