We used snow-tracking surveys to determine the probability of detecting Canada lynx Lynx canadensis in known areas of lynx presence in the northern Rocky Mountains, Montana, USA during the winters of 2006 and 2007. We used this information to determine the minimum number of survey replicates necessary to infer the presence and absence of lynx in areas of similar lynx density (approximately 2.8 lynx/100 km2) with confidence. The probability of detecting lynx in mountainous habitats that support resident populations was 0.80-0.99 when surveys were conducted on an 838 km2 grid with 10 km of search effort per cell. Snow-track surveys were highly successful at detecting the presence of Canada lynx over large landscapes. Two survey replicates established absence of Canada lynx with 95% certainty. The high probability of detection associated with snow-track surveys makes thismethod useful for documenting populations of Canada lynx in areas where their status is uncertain.