We used radiotelemetry to monitor the migration behavior of juvenile hatchery and wild steelhead Oncorhynchus mykiss as they migrated through Lower Granite Reservoir and Dam on the lower Snake River, Washington. From 1996 to 2001, we surgically implanted radio transmitters in 1,540 hatchery steelhead and 1,346 wild steelhead. For analysis, we used the inverse Gaussian distribution to describe travel time distributions for cohorts (>50 fish) of juvenile steelhead as they migrated downriver. Mean travel rates were significantly related to reach- and discharge-specific water velocities. Also, mean travel rates near the dam were slower for a given range of water velocities than were mean travel rates through the reservoir, indicating that the presence of the dam caused delay to juvenile steelhead over and above the effect of water velocity. Hatchery steelhead took about twice as long as wild steelhead to pass the dam as a result of the higher proportions of hatchery steelhead traveling upriver from the dam. Because upriver travel and the resulting migration delay might decrease survival, it is possible that hatchery steelhead survive at lower rates than wild steelhead. Our analysis identified a discharge threshold (???2,400 m3/s) below which travel time and the percentage of fish traveling upriver from the dam increased rapidly, providing support for the use of minimum flow targets to mitigate for fish delay and possibly enhance juvenile steelhead survival.