Beacham et a\ ; Microsatellite DNA variation and estimation of stock composition of Oncorhynchus neika 



17 



tional maximum likelihood method. 

 Baseline genotypic frequencies for 

 each of the three stocks were calcu- 

 lated from the obsei^ved allele fre- 

 quencies under the assumption of 

 Hardy -Weinberg equilibrium. Each 

 baseline stock was resampled with 

 replacement in order to simulate 

 random variation involved in the 

 collection of the baseline samples 

 during the estimation of stock com- 

 position of each mixture. Hypo- 

 thetical fishery samples of 100-300 

 fish with fixed stock composition 

 were generated by randomly resam- 

 pling with replacement the baseline 

 stocks, and adding the appropriate 

 number of fish from each stock to 

 the mixture. Estimated stock compo- 

 sition of the mixture was then deter- 

 mined, and the whole process was 

 repeated 100 times to estimate the mean and standard 

 deviation of the individual stock composition estimates. 



Fishery samples 



In 1997, samples were collected fi-om three commercial 

 gillnet fishery openings in Barkley Sound, a gillnet test 

 fishery, a purse-seine test fishery, the recreational fish- 

 ery, and an aboriginal fishery. The commercial gillnet 

 fishery was conducted primarily in Barkley Sound, with 

 gillnet mesh sizes ranging from 114 mm (4.5 inches) to 

 133 mm (5.25 inches). The gillnet test fishery was con- 

 ducted farther inland at the head of Barkley Somid and 

 at the mouth of Albemi Inlet with a gill net 110 m (60 

 fathoms) in length and 180 meshes deep, and having 

 a mesh size of 114 mm. Samples from the purse-seine 

 fishery, the recreational fishery, and the aboriginal fish- 

 eiy were derived entirely from Albemi Inlet. The recre- 

 ational fishery was conducted near the head of Albemi 

 Inlet and the aboriginal fishery, conducted at the head 

 of Albemi Inlet and in the Somass River, was the most 

 terminal fishery. Estimated stock contributions to each 

 sample were determined as a point estimate from all the 

 fish in the sample, and standard deviations of the esti- 

 mates were derived fi'om bootstrap resampling of both 

 the baseline stocks and the mixture. 



Results 



Precision of estimation of allele size 



Standard deviations of the estimated allele sizes for 

 the heterozygous standard fish analyzed at each locus 



ranged from 0.00 to 1.30 and tended to increase with 

 allele size (Table 2). For both alleles at Ots3, 100% 

 of the estimated sizes for each allele spanned a 2-bp 

 interval. For alleles <110 bp at Omy77, 93% (56/60) 

 of the estimated sizes of the allele were in a 2-bp 

 interval, as were 90% of the estimated sizes of alleles 

 between 110 and 120 bp. Estimated sizes of alleles of 

 the standard fish that were analyzed at the other loci 

 were all estimated within a 4-bp interval for alleles 

 <200 bp, with 85% of the estimated sizes of the larger 

 allele (213 bp) at Otsl03 within a 4-bp interval. 



Variation within stocks 



All six microsatellite loci examined were polymorphic 

 for all three stocks. Obsei-ved heterozygosity of the 

 loci examined over all stocks was as follows: Omy77 

 0.70 (stock range 0.61-0.80), Ots3 0.67 (0.64-0.70), 

 OtslOO 0.75 (0.69-0.79), Otsl03 0.83 (0.81-0.86), 

 Otsl07 0.28 (0.17-0.40), and OtslOS 0.85 (0.80-0.89). 

 Significant departures (correction for three tests per 

 locus, a=0.0167) from the expected Hardy-Weinberg 

 distribution of genotypic frequencies were observed 

 at the Omy77 locus in all stocks, owing, in the case 

 of Sproat and Henderson lakes, to a deficiency of het- 

 erozygotes. A similar significant heterozygote defic- 

 iency was also detected at Otsl08 in Sproat Lake 

 sockeye salmon. Significant annual variation (correc- 

 tion for six tests per stock, a=0.0083) in allele fre- 

 quencies was observed at Omy77 in Sproat Lake and 

 Henderson Lake sockeye salmon, and at Otsl08 in 

 Henderson Lake sockeye salmon. No significant link- 

 age disequilibrium between any pair of loci in any 

 stock was obsei"ved. 



