Beacham et al: Population structure and stock identification of Oncorhynchus keta 
253 
populations from the Utka River in west Kamchatka 
and the Anadyr River in northeast Russia. In our study, 
some similarities were observed between Magadan and 
the Sea of Okhotsk regional population groups, as well 
as between the groups from east and west Kamchatka. 
Winans et al. (1994) did not include populations from 
Primorye or Sakhalin Island in their survey, but Ginat- 
ulina ( 1992 ) had previously demonstrated clear differen- 
tiation of mitochondrial genotypes between the popula- 
tions from these two regions. The results of our study 
revealed regional separation of populations between 
these two areas. Our analysis supports the concept of 
regional groups of populations, and generally supports 
the concordance in patterns of population differentiation 
derived from analysis of allozymes and mitochondrial 
DNA variation. Allendorf and Seeb (2000) reported a 
concordance between results from allozyme and mic- 
rosatellite analyses of population structure for sockeye 
salmon (O. nerka). 
Genetic differentiation of Russian chum salmon gen- 
erally follows a regional structure because proximate 
populations are generally more similar to each other 
than to more distant populations. However, there were 
some cases of populations from one region clustering 
with populations from another region. One example 
was the Utka River population from west Kamchatka 
clustering with populations from northeast Russia. An 
association between the Utka River population and the 
Anadyr River population was also reported by Winans 
et al. (1994) in an analysis of allozyme variation. Be- 
cause Winans et al. (1994) and authors of the present 
study analyzed the same sample from the Utka River 
population, concurrence between the allozyme and mi- 
crosatellite analyses was not unexpected. However, the 
number of fish sampled for the Utka River population 
was the fewest for any of the west Kamchatka popula- 
tions (Table 1), and it may be that an increase in sample 
size for this population would result in estimated allele 
frequencies that would be more similar to those of other 
populations in west Kamchatka. Additionally, the Tugur 
River population was most closely associated with the 
population from the Amur River. In recent geologic his- 
tory, the Tugur River may have been once part of the 
Amur River drainage, but now flows into Tugur Bay on 
the Sea of Okhotsk. A common origin between the Amur 
River and Tugur River populations may account for the 
current association between the two populations. 
Distinctive groups of populations surveyed were those 
from the Primorye, Sahkalin Island, the northern Sea 
of Okhotsk, and northeast Russia, and a strong regional 
clustering of these populations was observed in the 
dendrogram analysis. Most of these population groups 
were characterized by slightly lower genetic variation 
compared with other populations surveyed in Russia. 
For other salmonids, populations from regions with re- 
duced genetic variation have formed distinctive clusters 
in dendrogram analysis (Beacham et al., 2006). 
Russian chum salmon populations displayed on aver- 
age less genetic differentiation than did populations 
from western Alaska and adjacent areas. Despite the 
Russian populations being surveyed from a larger geo- 
graphic area than were populations in western Alaska 
(Beacham et al., in press), and thus there was greater 
likelihood of differentiation due to isolation by distance, 
comparisons of locus-specific F st values between the two 
groups indicated that Russian populations were less 
differentiated (lower values in 11 of 14 loci, P= 0.057). 
This result indicates that there may be more straying 
among Russian populations than among those in west- 
ern Alaska, possibly as a result of adaptation to harsh 
enviromental conditions or hatchery development and 
broodstock transfer in Russia. Alternatively, less dif- 
ferentiation would also be observed if Russian chum 
salmon had colonized available habitats more recently 
than had chum salmon in western Alaska. 
Stock identification 
Accurate, economical, and practical methods of stock 
identification are required to determine the migra- 
tion pathways of juvenile and maturing salmon, and 
to manage fisheries that may intercept salmon during 
their migration to natal spawning grounds. Effective 
stock identification techniques are based on characters 
that display stable differentiation among groups to be 
discriminated, and these characters must be examined 
easily in a rapid and cost-effective manner. Allozymes 
provided the characters for initial genetically based 
population surveys and stock identification of Russian 
chum salmon (Winans et al., 1994, 1998). Later, single 
nucleotide polymorphisms (SNPs) were used to estimate 
the genetic structure of the population (Sato et al., 2001); 
therefore estimation of stock composition in mixed-stock 
samples can proceed (Sato et al., 2004). In an analysis of 
30 haplotypes from mtDNA, Sato et al. (2004) were able 
to indicate some level of regional structure in popula- 
tions in Russia, but the exact nature of the geographic 
structure was uncertain. In our analysis of microsatel- 
lite variation, clear differentiation was observed among 
regional groups of populations, and populations from 
Primorye were the most distinctive. 
Accuracy of estimated stock compositions is directly 
influenced by the baseline used in the estimation pro- 
cedure, and the level of genetic differentiation among 
regional groups of populations is a key component. How- 
ever, sample sizes of populations in the baseline are also 
an important part because the accuracy of estimation is 
related to the population sample size (Beacham et al., 
2006). Fewer than 60 fish were sampled for many of the 
populations sampled in our survey, and increasing sam- 
ple sizes to approximately 150 fish per population would 
likely lead to all regional estimates of stock composition 
being in excess of 90% accurate in all simulated single- 
region mixture samples. 
Surveys of genetic variation of salmon populations al- 
low stock identification in mixed-stock fisheries, where 
the origins of fish contributing to mixed-stock fisheries 
are determined by comparing the genetic characteristics 
of fish in the fishery samples to the genetic character- 
istics of fish from potentially contributing populations. 
