Genomic analyses in Drosophila do not support the classic allopatric model of speciation

The allopatric model of speciation has dominated our understanding of speciation biology and biogeography since the Modern Synthesis. It is uncontroversial because reproductive isolation may readily emerge as a by-product of evolutionary divergence during allopatry unopposed by gene flow. Recent genomic studies have found that gene flow between species is common, but whether allopatric speciation is common has rarely been systematically tested across a continuum of closely related species. Here, we fit a range of demographic models of evolutionary divergence to whole-genome sequence data from 93 pairs of Drosophila species to infer speciation histories and levels of post-divergence gene flow. We find that speciation with gene flow is common, even between currently allopatric pairs of species. Estimates of historical gene flow are not predicted by current range overlap. Whilst evidence for secondary contact is generally limited, a few sympatric pairs showed strong support for a secondary contact model. Our analyses suggest that most speciation processes involve some long-term gene flow, perhaps due to repeated cycles of allopatry and contact, without requiring an extensive allopatric phase.