Previous studies have demonstrated morphological differences between hatchery-reared coho salmon (Oncorhynchus kisutch) of hatchery origin and wild-reared coho of wild origin. We tested for a genetic component to this divergence by comparing coho from hatchery and wild populations both reared in the same hatchery environment and for an environmental component by comparing hatchery- and wild-reared coho both of wild origin. As in the previous studies, wild-reared fish from wild populations had greater head dimensions, larger median fins, and deeper bodies than did hatchery-reared fish from hatchery populations. This difference, summarized by the first principal component (PC1) of the size-adjusted data, was related to rearing environment rather than to genetic differences between hatchery and wild populations. Genetic divergence (or maternal effects) did occur between hatchery and wild populations along PC2 and PC3, but this divergence was slight compared with the environmentally induced differences between the two types of fish along PC1.

We compared agonistic behavior of newly emerged coho salmon (Oncorhynchus kisutch) between hatchery and wild populations using mirror image stimulation tests. We used hatchery populations from two different regions of Vancouver Island B.C., each matched with a wild population from its region. In both comparisons, hatchery juveniles were more aggressive than wild juveniles. Rates of aggressive display increased with time since emergence for both hatchery and wild fish, as did the differences in behavior between the two types. By the sixth day of observation (13 d postemergence), the overall effect of fish type was highly significant for all aggressive behaviours. Since the individuals compared were reared from eggs under identical conditions, these differences are presumably genetic. Comparisons involved relatively few families from each population. However, because heritability was moderate to low within populations, and variance between population types exceeded variance among families within populations, these results indicate real differences at the population level. These results may have important implications for programs to rebuild wild populations using hatchery transplants and for selective breeding programs to develop domestic stocks of coho.

Differences in growth rate, proximate composition, body morphology, and time of downstream migration between two populations inhabiting tributaries of the Miramichi River, New Brunswick, that differed in distance from the head of tide, temperature, and flow velocity were studied. Rocky Brook, located 132.6 km above the head of tide, had lower temperatures and higher average flow velocities than Sabbies River, located 42.5 km above tide. Growth rate and proximate composition were similar between populations, but body morphology and time of downstream migration differed significantly between populations. Individuals from Rocky Brook had more fusiform bodies and larger paired fins than their counterparts in Sabbies River. Rocky Brook fish also left the tributary in the fall rather than in the spring as was the case in Sabbies River. The generality of the flow regime–body morphology relationship observed was tested and confirmed by predicting differences in morphology of juvenile salmon in other rivers based on a knowledge of their flow regimes. It is suggested that early migration by the Rocky Brook fish is related to the higher energetic costs of overwintering in that stream. We hypothesize that the phenotypic similarity in growth rate and proximate composition and geographic variation in body morphology and timing of migration have an adaptive basis.Key words: geographic variation, polygenic traits, population biology, Atlantic salmon, morphology, migration, proximate composition, growth, adaptation

Allozyme variation was examined in sockeye salmon (Oncorhynchus nerka) from 83 distinct spawning sites representing all major sockeye-producing river systems in Canada. Of 33 nonselected loci examined, only 14 were highly polymorphic (q > 0.05) and 10 were less polymorphic (0 < q ≤ 0.05). No two populations were fixed for different alleles at any locus, but allele frequencies ranged from 0.01 to 0.86 at PGM-1* and from 0.07 to 0.89 at ALAT*, the two most variable loci. Mean heterozygosity ranged from 2.3 to 5.6% (mean 4.1%) across all sites. Hierarchical analysis was used to partition relative gene diversity among river systems (6.3%), major drainages within a river system (2.9%), nursery lakes within drainages (7.0%), spawning sites within lakes (1.0%), and individuals within spawning sites (82.8%). Extensive differentiation among nursery lakes affords excellent opportunities for genetic stock identification within river systems, but the relatively weak regional structuring limits opportunities for coast-wide stock identification. Genetic variation at highly structured loci corroborates the view that modern populations in Canada originated from sockeye that survived the late Wisconsin Glaciation in the Bering and Columbia refuges, and also suggests the existence of coastal refuges in British Columbia.

We describe a conditional maximum likelihood procedure for estimating stock composition in a mixed-stock fishery, provided that samples can be collected from the contributing stocks in isolation from each other and that characters exist that differ significantly between stocks. The procedure presented can use discrete (electrophoretic, meristic) or continuous (morphometric) data or any combination of these data. The procedure is tested by simulations and is used to estimate stock compositions of chum salmon (Oncorhynchus keta) sampled in a test fishery near Vancouver Island, B.C., in 1981. The estimated composition in the test fishery agreed closely with the results of previous tagging studies in the same area.