This study assessed the usefulness of passing euthanized Atlantic salmon Salmo salar smolts through an Archimedean screw turbine to test for external damage, as compared with live, actively swimming smolts. Scale loss was the only observed effect. Severe scale loss was 5·9 times more prevalent in euthanized turbine‐passed fish (45%) than the live fish (7·6%). Additionally, distinctive patterns of scale loss, consistent with grinding between the turbine helices and housing trough, were observed in 35% of euthanized turbine‐passed smolts. This distinctive pattern of scale loss was not seen in live turbine‐passed smolts, nor in control groups (live and euthanized smolts released downstream of the turbine), which suggests that the altered behaviour of dead fish in turbine flows generates biased injury outcomes.