The stress responses of migrating smolts of chinook salmon Oncorhynchus tshawytscha and steelhead Oncorhynchus mykiss (formerly Salmo gairdneri) to passage through three flumes (small baffled, large baffled, and unbaffled with corrugations) were determined by assaying plasma cortisol concentrations before and after fish passed through each flume. The flumes were tested under three conditions: nighttime, partly darkened in the daytime (400-900 lx at surface of water), and completely darkened in the daytime (1-4 lx). Flume design significantly affected post-passage cortisol concentrations in steelhead but not in chinook salmon smolts; concentrations were lowest in steelhead smolts that passed through the corrugated flume. In daytime tests, cortisol concentrations were significantly lower in chinook salmon smolts that passed through completely darkened flumes than in those that passed through partly darkened flumes. Cortisol concentrations did not, however, differ significantly in steelhead smolts that passed through partly and completely darkened flumes. Plasma cortisol concentrations followed a diel cycle in chinook salmon smolts held in both darkened (1-4 lx) and undarkened (3,800 lx) tanks and raceways for 40-60 h, being higher by day than by night. The cortisol response (postpassage minus prepassage concentration) of both species to passage through darkened flumes was smaller in day tests than in night tests.