To resolve the controversy initiated by Connor and Simberloff's (1978) analysis of similarity among the biota of the Galapagos Islands, we analyze the theoretical relationship between interspecific interactions and similarity. A counterintuitive result from our analysis is that all interspecific interactions, including competition, tend to increase similarity (given their effects on unconditional probabilities of occurrence). To have this effect, however, they must combine with the effects of interisland colonization. The results of Power (1975) and those of Connor and Simberloff (1978) suggest that interisland colonization and possibly interspecific interactions have had important effects on similarity in the Galapagos, but several difficulties prevent us from accepting their results. We conclude, with Connor and Simberloff, that the similarity of bird species lists for the Galapagos Islands does not support the importance of interspecific interactions in determining species distributions. In general, we...