Wildes et al. : Genetic variation in outer-coastal and fjord populations of Clupea pallasii in the eastern Gulf of Alaska 
387 
Table 3 
F ST values for Pacific herring estimated between sample pairs from geographic regions in the eastern Gulf of Alaska (above 
diagonal). Below the diagonal are P-values from pseudo-exact homogeneity tests of allele frequencies between collections. Signifi- 
cance of tests, based on 95% confidence intervals from permutation tests before (bold) and after (asterisk) corrections for multiple 
tests are indicated. /i=the number of individuals in the collection. 
Berners Bay- 
Lynn Canal 
n = 619 
Hobart Bay 
n=128 
Hoonah Sound 
n = 100 
Sitka Sound 
n = 206 
Nichols Bay 
n=97 
Prince William 
Sound n=191 
Berners Bay-Lynn Canal 

0.0004 
0 . 0011 * 
0 . 0008 * 
0.0013 
0 . 0004 * 
Hobart Bay 
0.1000 
— 
0.0009 
0.0001 
0 
0.0007 
Hoonah Sound 
< 0 . 0001 * 
0.0070 
— 
0.0005 
0.0025 
0 . 0009 * 
Sitka Sound 
< 0 . 0001 * 
0.1600 
0.0100 
— 
0.0011 
0 . 0011 * 
Nichols Bay 
0 . 0020 * 
0.0890 
0 . 0001 * 
0.2750 
— 
0.0017 
Prince William Sound 
0 . 0010 * 
0.0500 
0 . 0001 * 
0 . 0010 * 
0 . 0020 * 
— 
ka. McHugh (1954) points out that the expected broad 
latitudinal dines of morphological characters, such as 
vertebral counts and growth rates, differ sharply in 
some geographically adjacent areas within Southeast 
Alaska, despite similar environmental conditions affect- 
ing these phenotypes, and suggests that a degree of iso- 
lation may be responsible. Tagging studies corroborate 
these findings. Herring tagged in the Juneau area were 
not recovered in the Cape Ommaney reduction fishery 
on South Baranof Island (Rounsefell and Dahlgren, 
1935); however, fish tagged in Sitka Sound and Craig 
(southern Prince of Wales Island near Nichols Bay) were 
both detected in this fishery and were interpreted as 
evidence of an extensive movement and intermingling 
between these two regions (Skud, 1963). Previous ge- 
netic studies indicate that outer-coastal herring from 
Southeast Alaska were not significantly different from 
the majority of spawning herring in British Columbia. 
That study indicated that herring spawning at heads 
of inlets or ends of fjords migrate relatively short dis- 
tances to feed in the summer, whereas fish that spawn 
in exposed, coastal locations may migrate to the conti- 
nental shelf to feed (Beacham et al., 2008). 
Recent genetic studies of herring in the Northeast 
Pacific Ocean, have indicated that discrete populations 
exist in British Columbia (Beacham et al., 2008), and 
Puget Sound, Washington (Small et al., 2005), where 
some degree of geographic isolation or differences in 
spawning timing exist. The environments of outer- 
coastal areas in Southeast Alaska differ from those in 
interior waterways. This contrast may induce selective 
pressures on fjord/inland populations owing to selec- 
tion effects of salinity and may effectively isolate these 
groups and lead to differentiation at neutral microsatel- 
lite loci through drift. A salinity of 21 to 22 ppt (parts 
per thousand) was reported for the interior waters of 
Berners Bay (Harris et al. 2 ), compared to salinities of 
outer-coastal seawater. Outer-coastal water salinity is 
higher and directly affects vertebral counts (Schmidt, 
1917). Isolating mechanisms have been associated 
with specific salinity conditions on spawning locations 
(Bekkevold et al., 2005), and one microsatellite locus, 
Cpall2, is known to be influenced by divergent selection 
associated with salinity in Atlantic herring (Andre et 
al., 2010). Cpall2 does not appear to be under selection 
in Pacific herring in the Gulf of Alaska because our 
study showed this locus to be relatively homogeneous, 
significantly differentiating only Prince William from 
Hoonah Sound. 
Spawning time for herring varies annually in re- 
sponse to temperature. In 2009, initial spawning oc- 
curred 15-16 April for Nichols Bay, directly followed by 
Sitka Sound, 18-20 April, and three weeks later, 11-12 
May for Berners Bay (Pritchett and Hebert 3 ). Spawn- 
ing usually occurs in Berners Bay approximately three 
weeks later than in Sitka Sound because interior waters 
remain colder later into the season. 
Spawn timing and age of fish can be important con- 
siderations for collection of samples for genetic analy- 
ses. Spawning waves, where older fish spawn first, fol- 
lowed by younger year classes in the ensuing weeks, 
have been identified in Atlantic herring (McPherson 
et al., 2003). In age-structured populations with over- 
lapping generations, allele frequencies are predicted 
to differ among age classes because of “sweepstakes” 
recruitment, where a small number of spawners are 
disproportionately successful in reproducing offspring, 
compared with the massive number of spawners who 
fail to leave offspring (Jorde and Ryman, 1996). Age 
analysis conducted by Pritchett and Hebert 3 revealed 
that the largest age class of herring in Lynn Canal in 
2008 was 6 years, and 8+ years in Sitka Sound. Fish in 
2 Harris, P. M., S. W. Johnson, L. G. Holland, A. D. Neff, J. F. 
Thedinga, and S. D. Rice. 2005. Hydrocarbons and fish- 
eries habitat in Berners Bay, Alaska: Baseline monitoring 
associated with the Kensington Gold Mine. Alaska Fisheries 
Science Center Processed Report 2005-06, 44 p. Alaska 
Fisheries Science Center, NMFS, NOAA, Juneau, AK. 
3 Pritchett, M., and K. Hebert. 2008. 2009 report to the 
Alaska Board of Fisheries: Southeast Alaska-Yakutat her- 
ring fisheries. Fishery Management Report 08-65, 25 p. 
Alaska Department of Fish and Game, Anchorage, AK. 
