White sturgeonsAcipenser transmontanuswere sampled in three lower Columbia River reservoirs from 1987 to 1991 to describe population dynamics, the ability of these stocks to sustain harvest, and differences among reservoir and unimpounded populations. Significant differences were observed among reservoirs in white sturgeon abundance, biomass, size composition, sex ratio, size of females at maturity, growth rate, condition factor, and rate of exploitation. No differences among reservoirs were detected in fecundity, natural mortality rate, or longevity, in part because of sampling difficulties. Recruitment rates and densities in reservoirs were inversely correlated with growth rate, condition factor, and size of females at maturity. Differences in population dynamics resulted in substantial differences in sustainable yields. Maximum yields per recruit were predicted at annual exploitation rates between 5 and 15%. Most characteristics of reservoir populations were less than or equal to optima reported for the unimpounded lower river; as a result, yield per recruit, reproductive potential per recruit, and the number of recruits were less in reservoirs than in the unimpounded river. Comparisons with pristine standing stocks suggest that the unimpounded river may approximate preimpoundment conditions for white sturgeon. We conclude that potential yield from impounded populations has been reduced by dam construction, which restricts populations to river segments that may not include conditions optimal for all life stages. Alternatives for enhancement of reservoir populations might include improved passage at dams, increased spring flow to improve spawning success, transplants from productive populations, hatchery supplementation, and more intensive harvest management.