Genomic and phenomic analysis of island ant community assembly

The taxon cycle is a deterministic perspective on island biogeography that emphasizes regular pathways in the processes of colonization, adaptation, speciation, and extinction. As applied to Pacific ants, taxon cycles occur when species from source areas adapt to marginal habitats and enter a phase of range expansion after colonization. After reaching remote islands, species are expected to adapt to interior habitats and trend toward ecological specialization, lower abundance and eventually extinction, pushed along this progression by competition from new arrivals to the marginal habitats. The taxon cycle is an integrative hypothesis, but rigorous tests of the idea over the years have been hampered by data and methodological constraints. Here, we combine phylogenomics and population genomics to examine the evolutionary dynamics of a hyper-diverse genus, Pheidole, in the Fijian archipelago and test the prediction that species trend toward specialization and lower abundance with level of endemism. We reject a strict version of the taxon cycle, in which species only trend in one direction after colonization. Although some lineages trend toward specialization and rarity, we show that generalist dispersive species appear to have re-evolved within an old island radiation. However, on the whole, we found demographic evidence consistent with the idea that colonizations are pushing original radiations toward low abundance and extinction. In all, we interpret these results as mixed support for the taxon cycle and outline a revised model of community assembly including roles for priority effects and bidirectional shifts in and out of interior habitats.