The terms we use to describe the geography of speciation are deceptively simple. Mention allopatry, parapatry, or sympatry, and most biologists will have a clear picture of the underlying conceptual model of range limits (and probably some strong opinions about their relative frequency). Yet from a genetic perspective, these definitions can often obscure more than they clarify. For example, it’s hard to know what exactly allopatry means in highly vagile species like seabirds. Sure, coastal populations in western and eastern Pacific might appear discrete on a map, but if it’s a relatively easy trip between the two areas for volant migrants, are they actually diverging in isolation? A population genetic approach to the same terms therefore rely on relative rates of gene flow, from none (allopatry) to a lot (sympatry). But as speciation is an inherently geographic process, this is also an imperfect solution. Whether these alternate definitions can cross predict has largely remained unclear.
Penalba et al’s new preprint addresses this question by looking at whether the geographic arrangement of contemporary bird species ranges predicts relative gene flow between species pairs during divergence. Using “suture zones” where eight Melaphagoid species pairs meet at biogeographic barriers across Northern Australia and New Guinea, the study collected SNP data and sequenced the mitochondrial gene ND2 to assess population structure, relative divergence, probability of migration throughout the speciation process, and the fit of various models of parapatric speciation, e.g. whether measures of genetic distance shows a linear or nonlinear increase. The authors also modeled changing species distributions through time with maximum entropy and environmental layers from the present, mid-Holocene, and last glacial maximum.
If you’re familiar with the comparative phylogeography literature, their basic results are unsurprising: taxa varied widely in their divergence history, and their contemporary distributions did little to predict rates of gene flow during divergence. Among the many possible explanations for this discordance, an important one is the inherent lability of species ranges through time, even though they can appear as fixed, static entities on human timescales. (Losos and Glor’s 2003 paper provides a nice discussion of range dynamism over geologic time in a phylogenetic context.) Penalba et al.’s species distribution modeling provides support for this hypothesis by identifying distance and total range connectivity throughout all time periods as a better predictor of gene flow during speciation than current geography alone.
But while individual taxon pairs showed idiosyncratic history, they shared a pattern of divergence featuring a rapid transition between high and low gene flow, likely indicating the point of speciation. Consistent with the “snowball” model of accumulating incompatibility loci, this finding ties nicely into our broader understanding of speciation genomics. As the authors note, however, distinguishing between this resistance to gene flow and patterns of linked selection can be difficult. It seems likely that this question and other mechanistic questions represent the next frontier in understanding the relationship between genomes and geography as one species becomes two.
Losos, J.B., Glor, R.E. 2003. Phylogenetic comparative methods and the geography of speciation. Trends in Ecology and Evolution.
Penalba, J.V., Joseph, L., Moritz, C. 2017. Current geography masks dynamic history of gene flow during speciation in northern Australian birds. bioRxiv.