Understanding population mechanisms that set range limits is key to predicting shifts in species distributions in response to climate change. In the simplest view, range limits occur at the point along an environmental gradient at which annual deaths first exceed annual births on average. However, in a stochastic environment, inter-annual variability in survival, birth, or growth could lead populations to decline even if average rates suggest a stable or growing population. In a six-year demographic study, we estimated the means and temporal variances of all vital rates for two species of alpine plants, Silene acaulis and Polygonum viviparum, for multiple populations arrayed along a thirty degree latitudinal transect ending near the southern range limits of both species. We found that, not only did mean survival, reproduction, or growth decrease as the southern range limit was approached, but also that the year-to-year variability of one or more of these vital rates increased for both species. Interestingly, the vital rates that contribute most to population growth (as assessed using population projection matrix models) varied least in space and time. These results highlight the potential role that gradients in environmental stochasticity (as distinct from gradients in average environmental conditions) may play in setting range limits, and suggest that to predict the consequences of climate change for range shifts, it may be important to consider changes in year-to-year climatic variability as well as changes in mean climate.