Both density-dependent factors and environmental stochasticity can impact the dynamics of free-ranging populations. The pathways through which these factors influence population dynamics can be complex and may be immediate or lagged, and cumulative effects of environmental factors have been reported. We examined the effects of the severity (snow depth and persistence and winter rainfall) of the current and previous winters on the probability that female adult and yearling white-tailed deer (Odocoileus virginianus) would produce a fetus, and that adult females would produce a male fetus. We used logistic regression and Akaike's information criterion to select the best models from a set of 11 a priori candidate models. The severity of the winter 1 year before gestation negatively impacted the probability that both adults and yearlings would produce a fetus. There was no evidence that the probability of yearlings or adults producing a fetus was affected by winter conditions while gestating. Further, there was no evidence that the severity of the winter during which a yearling was gestated affected its probability of producing a fetus as a yearling. As the severity of the winter of gestation increased, the probability of producing a male decreased, consistent with both the Trivers–Willard sex ratio adjustment hypothesis and the extrinsic modification hypothesis. We suggest that both the decreased probability of reproduction after severe winters and the variation in fetal sex ratio may ultimately increase lifetime fitness if they lead to the production of the fittest offspring given the available maternal resources.