An integral part of efforts to recover stocks of Pacific salmon Oncorhynchus spp. and steelhead O. mykiss in Pacific Northwest rivers is to increase passage efficacy and survival of juveniles past hydroelectric dams. As part of this effort, we evaluated the efficacy of a prototype surface bypass structure, the removable spillway weir (RSW), installed in a spillbay at Lower Granite Dam, Washington, on the Snake River during 2002, 2003, 2005, and 2006. Radio-tagged juvenile steelhead were released upstream from the dam and their route of passage through the turbines, juvenile bypass, spillway, or RSW was recorded. The RSW was operated in an on-or-off condition and passed 3–13% of the total discharge at the dam when it was on. Poisson rate models were fit to the passage counts of hatchery- and natural-origin juvenile steelhead to predict the probability of fish passing the dam. Main-effect predictor variables were RSW operation, diel period, day of the year, proportion of flow passed by the spillway, and total discharge at the dam. The combined fish passage through the RSW and spillway was 55–85% during the day and 37–61% during the night. The proportion of steelhead passing through nonturbine routes was <88% when the RSW was off during the day and increased to >95% when the RSW was on during the day. The ratio of the proportion of steelhead passed to the proportion of water passing the RSW was from 6.3:1 to 10.0:1 during the day and from 2.7:1 to 5.2:1 during the night. Steelhead passing through the RSW exited the tailrace about 15 min faster than fish passing through the spillway. Mark–recapture single-release survival estimates for steelhead passing the RSW ranged from 0.95 to 1.00. The RSW appeared to be an effective bypass structure compared with other routes of fish passage at the dam.