The ability of environmental variation to affect species coexistence is much studied, yet environmental variation is not always important. I present an approximate calculation for the long-run growth rate of a species in the presence of spatially and temporally correlated environmental variation. I then perform a factorial numerical experiment, varying the mean seed dispersal distances, competition radii, and overwinter seed survival probabilities for two competing species for an array of variational regimes, noting the effects on their long-run growth rates. I find, first, that purely spatial variation has a greater capacity for influence than variation with a temporal component. Second, spatiotemporal variation can promote coexistence as strongly as purely temporal variation or more so, given the right species traits. Third, if the environmental variation has a spatial component, traits which enable species to become spatially segregated promote coexistence most strongly. That is, it is the possibility of spatial segregation which gives spatial variation its large potential to promote coexistence.