We investigate the rate of increase of protists introduced at low density into experimental communities comprising different sets of coexisting species. The experiment separates (1) the main effect of six introduced species, (2) the main effect of eleven resident communities, and (3) the introduction×community interaction. Introduced species differ markedly in their capacity to invade, and only one, Paramecium, invades all of the communities. Most of the introduced species invade some communities, and there is a large introduction×community interaction that comes about from bringing together particular combinations of species. For example, the rate of increase of Amoeba is greater when introduced into communities containing Paramecium than when introduced into communities in which Paramecium is absent. The omnivore Blepharisma increases faster in the presence of one of its prey Tetrahymena, notwithstanding the fact that it is also a potential competitor with this prey for bacteria. There is little indication that the species richness of resident communities (over the limited range available) affects the success of invasion. The results suggest that understanding invasions depends as much on detailed knowledge of idiosyncratic biological interactions as on general properties of community structure.