Beacham et al. : Population structure of Oncorhynchus keta across the Pacific Rim 
257 
whereas the southern lineage was observed in the Alas- 
ka Peninsula, Kodiak Island, and areas to the south 
and east. The two lineages were reported to overlap in 
the northern Alaska Peninsula. 
Development of DNA-level markers provided addition- 
al markers for genetic evaluation of population struc- 
ture of chum salmon, and surveys of mitochondrial DNA 
variation have been reported. Differentiation among 
Russian populations has been reported (Ginatulina, 
1992; Brykov, 2003; Polyakova et al., 2006), as well as 
in Japanese populations (Sato et al., 2001). In an analy- 
sis of mtDNA variation across the Pacific Rim, Sato et 
al. (2004) reported that there were three major lineages 
of chum salmon, with populations from Japan, Russia, 
and North America comprising three distinct regional 
groups. Chum salmon from Japan were observed to be 
the most distinct, with less divergence between popula- 
tions from Russia and western Alaska. 
Minisatellite variation was used by Taylor et al. 
(1994) and Beacham (1996) to survey variation in 42 
chum salmon populations across the Pacific Rim. Three 
regional groups of populations showed that those from 
Japan were the most distinct, followed by a second (less 
distinct) group comprising Russian and Yukon River 
populations, and a third group comprising southeast 
Alaska and British Columbia populations. 
Microsatellites have been used to evaluate chum 
salmon population differentiation and structure on a 
local and regional basis (Chen et al., 2005; Beacham et 
al., 2008a, 2008b, 2008c, in press). In those studies, as 
in the previous allozyme-based studies, regional groups 
of populations were observed, with the regional groups 
consisting of adjacent river populations or local popula- 
tions that were genetically similar within one region. 
The results from the current study were remarkably 
similar to the results of the allozyme-based study re- 
ported by Seeb and Crane (1999), with populations from 
Korea, Japan, Russia, Kotzebue Sound, Norton Sound, 
the Yukon River, and northern Bristol Bay determined 
to be in one major lineage. Populations from southern 
Bristol Bay and the northern Alaska Peninsula were 
intermediate, and populations on the south side of the 
Alaska Peninsula, Kodiak Island, and areas to the 
south and east to Washington State were determined 
to be in a second major lineage. 
Successful transplantation of salmon within the range 
of a species has the potential to alter genetic popula- 
tion structure. Population structure of chum salmon 
has been influenced to some degree by transplantations 
within its range. For example, due to frequent trans- 
plantations associated with hatcheries, most Japanese 
populations have received some level of transplantation 
of non-natal fish. Although initial studies indicated that 
the effect of transplantations were minimal in Japanese 
populations (Okazaki, 1982a), more recent work has 
shown that some current run-timing variation in popu- 
lations may be a result of transplantations. Beacham 
et al. (2008b) reported that allozyme monitoring indi- 
cated that successful introduction and establishment 
of broodstock from the Chitose River on the Sea of 
Japan coast of Hokkaido Island to the Gakko River on 
the Sea of Japan coast of Honshu Island accounts for 
observed temporal differentiation in the existing Gakko 
River population. Transplantations have also occurred 
in Russian and North American populations, but there 
is little evidence for a demonstrable change in popula- 
tion structure as a result of transplantations. 
Although most production of Japanese chum salmon 
is currently derived from hatcheries, there is little evi- 
dence that hatchery production has resulted in reduced 
genetic variation of the populations, in relation to chum 
salmon in other portions of the range. Initially, Kaeri- 
yama (1999) indicated that, on the basis of allozymes, 
Japanese populations were less variable than Russian 
wild populations. In our study, on the basis of 14 mic- 
rosatellites, we found no evidence that Japanese chum 
salmon populations were less genetically variable than 
Russian or North American chum salmon. In fact, the 
opposite result was observed, with higher levels of ge- 
netic variation observed in Japanese populations com- 
pared with chum salmon from other regions across the 
Pacific Rim. 
Population structure of chum salmon across the Pa- 
cific Rim was demonstrated to have a regional basis. 
A regionally based population structure is generally 
required for genetic stock identification estimation be- 
cause an important assumption is that the portion of 
the mixed-stock sample derived from unsampled popula- 
tions is allocated to sampled populations from the same 
region. This assumption reduces the cost and complexity 
of developing a baseline for stock composition analysis. 
Chum salmon population structure thus meets the im- 
portant condition that unsampled populations contribut- 
ing to mixed fishery samples will likely be allocated to 
sampled populations in the same region. 
Populations in the major river drainages surveyed all 
clustered together within a drainage, with the excep- 
tion of the Yukon River, where lower river summer-run 
populations clustered with populations from the Kus- 
kokwim River in western Alaska and the Nushagak 
River in northern Bristol Bay. Similar results were also 
reported in the allozyme survey conducted by Seeb and 
Crane (1999), who suggested that genetic exchange may 
have occurred between the Kuskokwim and Nushagak 
rivers during the last glaciation because both rivers 
were headwaters to a Bering Sea Land Bridge river that 
drained into the Bering Sea (Hopkins, 1967; Lindsay 
and McPhail, 1986). The ancient mouth of the Yukon 
River was farther south than at present (Hopkins, 1967; 
Knebel and Creager, 1973), increasing the probability 
of genetic exchange among ancestral populations of the 
Yukon, Kuskokwim, and Nushagak rivers. 
Chum salmon likely had a different pattern of dis- 
persal from refuges after the last glaciation ended in 
the Pleistocene Era some 10,000 years ago than did 
either sockeye salmon or Chinook salmon. For exam- 
ple, evaluation of genetic diversity in Asian and North 
American populations of sockeye salmon and Chinook 
salmon have indicated that there were similar levels 
of genetic diversity between populations from these 
