For capital breeders, mass may affect reproductive potential. Reproductive expenditure may reduce future reproductive potential, particularly when resources are scarce. To test the hypothesis that reproductive success and the costs of reproduction vary according to mass and population density, we analyzed 25 yr of data on bighorn ewes (Ovis canadensis). The number of adult females was first limited by yearly removals, then allowed to triple. We found no survival costs of reproduction for ewes aged 4-7 yr. For ewes aged 8-14 yr, survival was density dependent for barren ewes but not for ewes that weaned lambs. Failure to lamb was rare and negatively correlated with fertility the following year. At low population density, lactation had a negative effect on mass gain but had a limited reproductive cost. At high density, heavy ewes had higher reproductive success than light ewes, and the reproductive cost and somatic costs of reproduction increased. The cost of reproduction was greater for light than for heavy ewes. Survival of weaned lambs to 1 yr was affected by population density but not by maternal mass or previous reproductive success. In large mammals, manipulations of reproductive effort are problematic, but long-term monitoring of individual mass and reproductive success under varying conditions of resource availability can provide insights into the evolution of life histories.