Modern demographic models in conservation biology incorporate demographic and environmental stochasticity. Variability in vital rates, relationships among vital rates, and relationships between vital rates and the environment can affect projections from demographic models. We used a 30-year data set to estimate monthly mass-specific survival in hispid cotton rats (Sigmodon hispidus). Survival probabilities were estimated for both sexes and 3 mass classes using a multistate model and program MARK. We tested for differences in monthly survival among 3 mass classes, relationships between survival and environmental variables, and autocorrelation in survival. We found lower survival in summer for the smallest and medium mass classes. Survival in the largest mass class was positively autocorrelated during February and March, indicating that winter survival in large cotton rats is characteristic of a season, but that other monthly survival rates could be treated as unrelated to each other or to environmental variables.