**Background/Question/Methods ** Predicting species’ range shifts in response to climate change will require a greater understanding of the ecological mechanisms that give rise to range boundaries. Because range boundaries arise where, on average, populations are more likely to go extinct than to persist, range limits are an inherently population-level problem for which a demographic framework is useful. Using a seven year dataset for eight populations of riparian perennial monkeyflowers (*Mimulus cardinalis* and *M. lewisii*), I compared vital rates and population dynamics between central and marginal populations to investigate how spatiotemporal environmental variation gives rise to range limits. I used annual estimates of survival, growth and fecundity to construct stage-classified transition matrices and project stochastic population growth. I hypothesized that vital rates and projected population growth rates would be lower on average and more variable over time in marginal compared to central populations.

**Results/Conclusions **

For *M. lewisii*, survival and fecundity were lower and more variable in marginal than in central populations. Surprisingly, for *M. cardinalis* the converse was true. Differences in stochastic population growth rates between central and marginal populations mirrored the patterns observed for mean vital rates. Substantial year-to-year variation in projected population growth was also evident at all sites, regardless of position within the range. Populations that exhibited greater temporal variation in some vital rates did not always show greater temporal variation in projected population growth rates. These results highlight the potentially complex interactions between temporal and spatial variation in determining population dynamics across species’ ranges.