Beacham et al : Microsatellite DNA variation and estimation of stocl< composition of Oncorhynchus nerka 



21 



son Lake composition, estimated 

 stock compositions were 5.39c 

 (SD=3.9'7f) Henderson Lake, 

 34.3% (SD=8.2%) Sproat Lake, 

 and 60A7c{SB=8.97c) Great Cen- 

 tral Lake. No significant bias 

 was observed when Henderson 

 Lake sockeye salmon composed 

 57c or less of the mixture. 



Application of estimates to 1997 

 fisheries 



Although estimated stock con- 

 tributions varied according to 

 sampling period, sockeye salmon 

 fi-om Great Central Lake tended 

 to predominate in all fisheries 

 at any week (Table 6). However, 

 differences in stock composition 

 estimates among fishing gears 

 were evident. In the commer- 

 cial gillnet fishery, Great Central 

 Lake sockeye salmon constituted 

 about 70% of the catch (Table 

 6). In the gillnet test fishery, the 



proportion of Great Central Lake sockeye generally 

 varied between 55 and 75% prior to July 25th. In 

 the purse-seine test fishery, they accounted for about 

 50-55% of the catch. Higher proportions of Great 

 Central Lake sockeye salmon were observed in the 

 selective gillnet gear than in the more nonselective 

 purse-seine gear. For example, for the week ending 4 

 July, Great Central Lake sockeye were estimated to 

 have represented 70-757f of the catch in the commer- 

 cial gillnet fishery and in the gillnet test fishery, but 

 only about 40% of the catch in the seine test fishery. 

 Although the samples analyzed from the purse-seine 

 fishery were derived fi-om more inland locations than 

 those from the commercial and test gillnet fisheries, 

 the differences in proportions of Great Central sockeye 

 salmon more likely resulted from differences in gear 

 selectivity than ft-om differences in stock distribution 

 because fish fi-om all three stocks are generally dis- 

 tributed throughout Barkley Sound and Albemi Inlet 

 when present. 



Sockeye salmon stock ft-om Henderson Lake are the 

 smallest salmon exploited in the fishery, and thus the 

 most vulnerable to overfishing in the mixed-stock har- 

 vest that takes place. Henderson Lake fish, which do 

 not have to travel through Albemi Inlet in their spawn- 

 ing migration, were apparently caught in fisheries 

 throughout Albemi Inlet, although there was a high 

 degree of uncertainty about whether they were caught 

 in the aboriginal fishery at the extreme head of Albemi 



Inlet (Table 6). Henderson Lake sockeye salmon gen- 

 erally represent 10% or less of the catch, except for 

 sockeye salmon sampled after 18 July in the gillnet 

 test fishery, when the relative abundance of Hender- 

 son Lake sockeye salmon substantially increased. By 

 late July, Henderson Lake sockeye salmon constituted 

 nearly 30% of the gillnet test fishery sample. 



Discussion 



DNA variation at microsatellite loci is becoming an 

 increasingly important tool in fisheries research and 

 management (see review by O'Connell and Wright 

 [1997]). In salmonids, microsatellite loci are gener- 

 ally characterized by high levels of variability and dif- 

 ferentiation among spawning populations (Angers et 

 al., 1995; McConnell et al., 1997; Seeb et al., 1998), 

 even in very localized areas (Beacham and Dempson, 

 1998). The feasibility of applying biological markers 

 to salmon stock identification is enhanced when they 

 display limited annual variation. With temporal sta- 

 bility of the discriminating characters, annual surveys 

 of contributing populations are unncecssary once they 

 have been adequately characterized. As for other neu- 

 tral genetic markers (Wood et al., 1994; Beacham et 

 al., 1996), temporal stability of allele frequencies at 

 microsatellite loci has generally been observed in sal- 

 monid populations (Small et al., 1998). For popula- 



