Understanding the dynamics of communities in space and time requires reconciling ecological and evolutionary processes, including colonization, adaptation, speciation and extinction. In practice, this has been challenging because empirical data obtained by traditional methods and predictive models typically focus on particular processes driving local community assembly and biogeographical structure. In this issue of Molecular Ecology, by using phylogenomics, population genomics and phenomics approaches, Darwell et al. show that ant community assembly on islands is governed by predictable eco-evolutionary trends of geographical range expansion, adaptive radiation and local population decline. The authors provide one of the most robust lines of evidence that the evolutionary progression of island communities may often be directional and repeatable, as predicted by the concept of taxon cycles.