We used data for mountain sheep (Ovis canadensis) in Arizona to examine the theory that populations with <50 individuals will become extinct. We found no significant difference in extinction rate between populations with <50 individuals (n= 12; four extinctions) and populations with>50 individuals (n = 8; two extinctions). We question three assumptions of the theory: 1) that different races of mountain sheep respond in a similar manner regardless of habitats, 2) the criteria used to define a population, and 3) the accuracy of population estimates.