Abstract
Simulated mixtures of juvenile sockeye salmon (Oncorhynchus nerka) were constructed using parasite data to represent proportionally the major component stocks of Fraser River and Lake Washington sockeye migrating within the Strait of Georgia, British Columbia, in 1982–84. Samples of migrating juveniles were also collected from Bedwell Harbour, South Pender Island, British Columbia, each year and analyzed for parasites and stock composition. The compositions of simulated and sample mixtures were estimated using a maximum likelihood stock composition model. Simulated mixture compositions were accurately estimated for most stocks for all year-classes. When significant misassignment occurred between stocks, the stocks were analyzed as a complex using the allocate-sum procedure. Sample mixture estimates correctly identified the dominant stock for each year-class, although for 1984 the dominant group was determined as a complex of three stocks because the individual stocks were not distinguishable. The results indicate that it is feasible to use parasites as natural tags to estimate stock compositions of migrating juvenile sockeye salmon in the Strait of Georgia.