The ability of wild and hatchery yearling Atlantic salmon (Salmo salar) to hold a position against water currents was tested in relation to water temperature and season. This species would not swim, but held positions on the bottom of the test apparatus. Their performance was, therefore, rated on the basis of critical holding velocity instead of critical swimming velocity, often used for other species. Qualitatively, the relationships of critical holding velocity with both temperature and season were similar for wild and hatchery fish. However, wild fish were consistently more able to hold positions against water currents so that the mean water velocities at which they were unable to stem water currents were 25–84% greater than those for hatchery fish. In the winter, the mean water velocities against which both wild and hatchery fish could no longer hold a position were 21.9–50.0% lower than in the summer. Seasonal changes in performance generally increased or decreased with water temperature except in winter when performance declined while water temperature was almost constant. Performance was high and almost unchanging for temperatures above 8 °C, but fell sharply below this temperature. We suggest that the sharp decrease in performance at 8 °C explains the previous observation that this species suddenly moves into the stream bed when temperature falls during the autumn.