Isolated habitats generally have fewer species at local spatial scales than more connected habitats. However, over larger spatial scales, the response of species richness to variation in the degree of isolation is variable. Here, we hypothesized that the effects of habitat isolation on patterns of regional level species richness may depend at least in part on the level of disturbances those habitats receive. We tested this hypothesis in a microcosm experiment using an aquatic community consisting of container dwelling protists and rotifers by manipulating disturbance and dispersal to experimental regions factorially. In disturbed regions, regional species richness was lower in regions with isolated patches compared to regions where patches were experimentally connected by dispersal. A likely mechanism for this result is that dispersal from adjacent undisturbed local patches allowed disturbance-intolerant species a temporary refugia, thereby allowing regional coexistence of disturbance-tolerant and intolerant species. In contrast, without disturbances (and thus no temporal heterogeneity) it is likely that dispersal homogenized communities, leading to overall lower richness with higher dispersal. Our results emphasize the importance of simultaneously considering multiple limiting factors, disturbance and dispersal in this case, as well as the spatial scale of the response, in order to fully understand factors that control biodiversity.