We investigated survival, reproduction, and population growth (l) for a declining elk (Cervus canadensis nelsoni) population in South Dakota, USA, 2011-2015. We obtained survival data from 125 calves and 34 yearlings. We determined survival and pregnancy rates for 42 adults (2-8 years old) and 39 old adults (!8 years old). We combined population vital rates into a matrix model, which indicated a slightly growing population but with considerable uncertainty ( l ¼ 1.03, 95% CI ¼ 0.93-1.13). Our elasticity analysis suggested asymptotic growth rates were most sensitive to proportional changes in old adult and adult female survival, followed by proportional changes in calf and yearling survival. Our life-stage simulation analysis further supported asymptotic growth rates being most sensitive to variation in survival, and most of the variation in l we observed was a consequence of variation in annual calf survival (R 2 ¼ 0.58). Annual calf survival was low (0.26, 95% CI ¼ 0.05-0.52), and puma (Puma concolor) predation was the primary causespecific mortality factor of calves (0.63, 95% CI ¼ 0.51-0.76). Adult female survival was near its biological maximum (0.95, 95% CI ¼ 0.87-0.99); therefore, managing for increased calf survival may be the most practical strategy for promoting elk population growth in this system. Managing this puma population at the lower end of the population objective and reducing white-tailed deer (Odocoileous virginianus) numbers (primary prey source) may allow for elk population growth in this system. Ó 2017 The Wildlife Society.