124 
Fishery Bulletin 99(1) 
more extensive allozyme surveys show divergence within 
broodlines in several North American regions (Beacham et 
ah, 1985, 1988; Gharrett et ah, 1988; Shaklee et ah, 1991). 
Genetic divergence is usually attributed to random genet- 
ic drift, adaptation to local environmental conditions, or 
both, which can occur because salmon home to their natal 
streams to spawn (for an overview, see Allendorf et ah, 
1987). 
The genetic compositions of populations are also mold- 
ed by other events in their evolutionary histories. If popu- 
lations share multiple genetic characteristics, then they 
may share a common origin (colonization or gene flow) or 
may be derived from different ancestral lines that have ex- 
perienced homogenizing selection pressures (similar his- 
torical environments). The far North Pacific Ocean, Bering 
Sea, and Sea of Okhotsk are of special interest in the evo- 
lution of salmon population structure because these ba- 
sins encountered the most extreme environmental condi- 
tions during the late Pleistocene Epoch. During the last 
100,000 years, the environment ranged from extremely fa- 
vorable (modern conditions) to extremely harsh, which in 
some areas undoubtedly led to extirpation of many salm- 
on populations. Comparison of the nature and extent of 
genetic variation within regions to variation among re- 
gions is one means of exploring the recent history of a spe- 
cies. Reports of genetic variation among chum salmon (O. 
keta\ Wilmot et ah, 1994; Winans et al., 1994; Seeb and 
Crane, 1999), sockeye salmon (O. nerka\ Varnavskaya et 
ah, 1994a, 1994b; Wilmot et al, 1994), and pink salmon 
(Gharrett et ah, 1988; Varnavskaya and Beacham, 1992; 
Shaklee and Varnavskaya 1994) are mostly limited to a 
single region with few reference populations from other 
regions, or they skip across intervening regions. In addi- 
tion for sockeye salmon, there are relatively few informa- 
tional loci and the variation among populations within a 
region is often large. Consequently, it is difficult to formu- 
late a coherent picture for any Pacific salmon species. The 
data available for a broader range of Asian populations 
(Glubokovsky et ah, 1989) are based on too few loci to pro- 
vide a strong basis for broader comparisons. 
One advantage of studying pink salmon is that they are 
distributed almost continuously throughout the northern 
region. Tagging studies indicate that pink salmon from 
streams in large contiguous areas of the coast make simi- 
lar movements and may occupy the same areas within the 
high seas during portions of their ocean migrations (Tak- 
agi et ah, 1981). Although it is not clear to what extent 
this movement pattern reflects either a shared history or 
shares physical effects (such as the direction of prevailing 
currents), populations within a region presumably expe- 
rience similar marine environments. The few genetic da- 
ta that address the relatedness of even-year pink salmon 
from both Pacific coasts are limited by the number of ge- 
netic characters examined or by geographic area (Ghar- 
rett et ah, 1988; Zhivotovsky et ah, 1989; Shaklee and Var- 
navskaya, 1994). 
Our study analyzed numerous allozyme loci in popula- 
tions from the even-year broodline of pink salmon from 
Asian waters and compared those data with data from 
western Alaska populations (Gharrett et ah, 1988). The 
study included most of the geographic groups identified 
by Takagi et ah, (1981). We substantially expanded the 
number of allozyme loci sampled in Asian even-year popu- 
lations to determine the genetic structure of those popu- 
lations and to investigate the genetic relatedness within 
and among large areas of the pink salmon range. The 
questions we addressed are 1) Is there evidence of genetic 
structure for even-year pink salmon populations? 2) How 
does the genetic structure of Asian even-year pink salmon 
relate to the adjacent western Alaskan and Aleutian is- 
land populations? and 3) How does the genetic structure 
of Western Alaska and Asian pink salmon relate to geo- 
graphic and oceanographic features and to recent geologi- 
cal history? 
Materials and methods 
Tissue samples from returning adult spawners were col- 
lected 1) from four river systems on southern Sakhalin 
Island (Dolinka, Lutoga, Monetka, and Ochepukha rivers) 
between 6 August and 19 September 1990; 2) from two 
river systems in the Magadan region of Russia (Tauy River 
on 26 July 1990 and Uglekanka River on 3 September 
1990); 3) from seven river systems on the Kamchatka Pen- 
insula (Utka, Pymta, Kol, Bistraya, Vorovskaya, Karaga, 
and Ossora rivers) between 29 July and 3 September 1990 
(Fig. 1); 4) at hatcheries on three streams in Japan (Kush- 
iro, Tokushibetsu, and Yurappu Rivers) between 21 and 29 
September 1990; and 5) in five collections taken at differ- 
ent times during the return to a hatchery in Sawmill Bay 
in Prince William Sound, Alaska, between 27 August and 
8 September 1988. There are no natural spawning runs of 
pink salmon in Japan; consequently, the hatchery samples 
are all that were available to represent Japan. Also, the 
samples from Prince William Sound, in the center of the 
North American range are intended to provide an idea of 
the extent of difference between Asian and North Ameri- 
can pink salmon for the entire suite of allozyme loci exam- 
ined in the Asian fish. The Znamenka River, a tributary of 
the Ochepukha River, was sampled repeatedly and consid- 
ered separately in some analyses. 
Pieces of heart, eye, liver, and skeletal muscle were sam- 
pled in the field and frozen on wet ice, frozen gel-packs, or 
dry ice until they were transferred to temporary storage at 
-20°C or to liquid nitrogen dewars at -196°C. Long-term 
storage was maintained at -85°C. The samples were ana- 
lyzed by using horizontal starch-gel electrophoresis (Ut- 
ter et al., 1974). Eight different gel buffers were used (Ta- 
ble 1). Proteins were revealed by using standard staining 
recipes (Aebersold et al., 1987). All allozyme data were 
collected at the NMFS Auke Bay Laboratory. An Excel™ 
file of allele frequency data can be downloaded by anony- 
mous ftp from ftp: / / wwwabl.afsc.noaa.gov in file SIDAI 
pink_salmon / evenasia. 
Variability at isoloci (Allendorf and Thorgaard, 1984) 
was assigned to one of the loci, and the other was treated 
as monomorphic, which for low frequencies (P<about 0.15) 
has little influence on the analysis (Gharrett and Thom- 
ason, 1987). The genotypic frequencies observed at each 
