Ecologists have long investigated why communities are composed of a few common species and many rare species. Most studies relate rarity to either niche differentiation among species or spatial processes. There is a parallel between these processes and the processes proposed to explain the structure of metacommunities. Based on a metacommunity perspective and on data on stream macroinvertebrates from different regions of Brazil, we answer two questions. 1) Are sets of common and rare species affected by similar niche and spatial processes? 2) How does the community composition of common and of rare species differ? The main hypothesis we test is that common species are mainly affected by environmental factors, whereas rare species are mostly influenced by dispersal limitation. We used variation partitioning to determine the proportion of variation explained by the environment and space in common and rare species matrices. Contrary to our expectations, evidence supported the idea that both common and rare species are affected mainly by environmental factors, even after controlling for the differing information content between common and rare species matrices. Moreover, the abundance of some common species is also a good predictor of variation in rare species matrices. Niche differences are unlikely to be the sole cause of patterns of rarity in these metacommunities. We suggest that sets of common and rare species react to similar major environmental gradients and that rare species also respond to processes that operate at a more fine-grained spatial scale, particularly biotic interactions. We extend the view that species sorting is the dominant process structuring metacommunities and argue that future studies focusing on rarity would benefit from a metacommunity perspective.