Records of Wood River sculpin (Cottus leiopomus) from 1893 to 2003 were examined and new data were gathered in 2004 and 2006 to estimate the species' historical range and to search for evidence of changes in its occurrence. Detections of Wood River sculpin were reported in 49 subwatersheds, primarily in the Idaho Batholith ecoregion. The remainder of the Wood River basin was classified as either potential historic range or unknown. Mark-recapture sampling was conducted to estimate the efficiency of electrofishing as a method to detect this sculpin. Sampling efficiency was calculated for 1 upstream pass and 1, 2, and 3 bidirectional passes. A mean sampling efficiency of 45.9% (range 15.4%–63.0%) was achieved using 3 bidirectional passes, block nets, and 0.942 minutes of electrofishing per m² of stream surface. Efficiency declined an average of 38% between electrofishing passes. The rates and reduced efficiencies observed between passes suggest that multiple passes and visits are needed to reliably estimate sculpin presence or absence.