248 
Fishery Bulletin 106(3) 
were size fractionated in an ABI 3730 capillary DNA 
sequencer, and genotypes were scored by GeneMapper 
software 3.0 (Applied Biosystems, Foster City, CA) by 
using an internal lane-sizing standard. Allele identifi- 
cation between the two sequencers was standardized 
by analyzing the same approximately 600 individuals 
on both platforms and converting the sizing in the gel- 
based data set to match that obtained from the capil- 
lary-based set. 
Data analysis 
Each population at each locus was tested for depar- 
ture from Hardy-Weinberg equilibrium (HWE) by using 
genetic data analysis (GDA). Critical significance levels 
for simultaneous tests (34 populations, Table 1) were 
evaluated using Bonferroni adjustment (0.05/34 = 0.0015) 
(Rice, 1989). All annual samples available for a loca- 
tion were combined to estimate population allele fre- 
quencies, as was recommended by Waples (1990). F st 
estimates for each locus were calculated with FSTAT 
(Goudet, 1995), individual locus values were determined 
by jackknifing over populations, and the overall F st esti- 
mate was determined by jackknifing over loci (Goudet, 
1995). Inter-regional comparisons of F st estimates were 
determined by calculation of all appropriate pairwise 
point estimates of F st values, and then determining 
the mean and standard deviation of these values. The 
Cavalli-Sforza and Edwards (CSE) (1967) chord distance 
was used to estimate distances among populations. An 
unrooted neighbor-joining tree based upon CSE was 
generated using NJPLOT (Perriere and Gouy, 1996). 
Bootstrap support (by sampling loci) for the major nodes 
in the dendrogram was evaluated with the CONSENSE 
program from PHYLIP (Univ. Washington, Seattle, WA) 
and based on 500 replicate trees. Computation of the 
number of alleles observed per locus, as well as allelic 
diversity standardized to a common sample size, was 
carried out with FSTAT. 
Estimation of stock composition 
Genotypic frequencies were determined for each locus 
in each population, and the Statistical Package for the 
Analysis of Mixtures software program (SPAM, vers. 
3.7, Debevec et al., 2000) was used to estimate stock 
composition of simulated mixtures. The Rannala and 
Mountain (1997) correction to baseline allele frequencies 
was used in the analysis in order to accommodate the 
occurrence of fish in the mixed sample that were from 
a specific population having an allele not observed in 
the baseline samples from that population. All loci were 
considered to be in Hardy-Weinberg equilibrium, and 
expected genotypic frequencies were determined from 
the observed allele frequencies. Reported stock composi- 
tions for simulated fishery samples were the bootstrap 
mean estimate of each mixture of 150 fish analyzed, 
and mean and variance estimates were derived from 
100 bootstrap simulations. Both the baseline popula- 
tion and the simulated single-population were sampled 
with replacement in order to simulate random variation 
involved in the collection of the baseline and fishery 
samples. 
Results 
Variation within and among populations 
The observed number of alleles observed at a locus 
ranged from 21 alleles at Oke3 and Oki2 to 138 alleles 
at Onelll (Table 2). Lower expected heterozygosities 
were generally observed at loci with fewer alleles. The 
genotypic frequencies observed at the 14 loci generally 
conformed to those expected under Hardy-Weinberg 
equilibrium (HWE) after Bonferroni correction. For the 
Oke3 and OtsG68 loci, a minor HWE nonconformance of 
genotypic frequencies was observed, and observed hetero- 
zygosities were 2-6% less than those expected (Table 2). 
Genetic diversity, with respect to the number of al- 
leles observed, was evident among regional groups of 
chum salmon. Chum salmon populations from Primo- 
rye, the northern Sea of Okhotsk, and northeast Rus- 
sia displayed fewer alleles (mean 320 alleles) than did 
populations in Magadan, west Kamchatka, and east 
Kamchatka (mean 370 alleles) (Table 3). Chum salmon 
from the latter regions displayed approximately 16% 
more alleles than did those from the former regions. 
The greatest differentiation in allelic diversity was 
observed at those loci with greater numbers of alleles, 
particularly at locus Onelll. 
Population structure 
Genetic differentiation was evident among chum salmon 
populations from the different geographic regions sur- 
veyed. The F st value over all 34 populations and 14 loci 
surveyed was 0.017, and individual locus values ranged 
from 0.003 ( Onel02 ) to 0.054 ( Ots3 ) (Table 2). Chum 
salmon populations from Primorye and the Amur River 
were well defined compared with other regional popula- 
tions (Table 4). Populations from the southwestern por- 
tion of Russia (Primorye, Amur River, Sakhalin Island) 
were most distinct from those in more northern and 
eastern regions (Magadan, Sea of Okhotsk, Kamchatka, 
northeast Russia). 
Regional clustering of population samples was ob- 
served in the analysis of population structure. Strong 
clustering of population samples from the Primorye 
region was observed; the three population samples in- 
cluded in the analysis clustered together in 100% of the 
trees examined (Fig. 2). Similarly, strong clustering 
was observed in most of the population samples from 
Sakhalin Island, as well as the two population samples 
from the northern coast of the Sea of Okhotsk. The two 
population samples from northeast Russia clustered to- 
gether in 100% of the trees examined, together with the 
Utka River population sample from west Kamchatka. 
Although there was a general clustering of population 
samples from east and west Kamchatka, these regional 
