We measured population abundance and density of Dall’s sheep (Ovis dalli) before and during gray wolf (Canis lupus) and coyote (Canis latrans) harvest for 3 years (1998–2000) on two similar and adjacent study sites with treatment (canid harvests) and reference (no canid harvest) in interior Alaska, USA. Between 1998 and 1999, density of Dall’s sheep in the treatment area increased by 0.746 ± 0.163 sheep/km2. Between 1999 and 2000, after a winter with above-normal, crusted snow, density of sheep decreased by 2.1 ± 0.14 sheep/km2. Because of this large decline, sheep density on the treatment area decreased by 1.3 ± 0.08 sheep/km2 over the entire monitoring period. Sheep in the reference area showed no significant change in density between 1998 and 2000. We hypothesize that Dall’s sheep in the treatment area initially benefited from the harvest of wolves and coyotes via reduced predation and predation risk and that resultant high sheep densities coupled with severe winter weather caused a dramatic decline in Dall’s sheep between 1999 and 2000. Thus, this study illustrates the potential consequences of interactions between density-dependent and density-independent factors for the conservation of large mammals, especially those residing at high latitudes and altitudes. These results should serve as a cautionary tale for those wishing to increase ungulate numbers via predator control without regard to other ecological factors, such as the proximity of the prey population to ecological carrying capacity (K).