Interactions between species are widely understood to have promoted the diversification of life on earth, but how interactions spur the the formation of new species remains unclear. Interacting species often become locally adapted to each other, but they may also be subject to shared dispersal limitations and environmental conditions. Moreover, theory predicts that different kinds of interactions have different effects on diversification. To better understand how species interactions promote diversification, we compiled published genetic data for host plants and intimately associated herbivores, parasites, and mutualists. We first tested whether host and associate population structures were correlated --- an indication of associates locally adapting to hosts --- and tested for confounding correlations with geographic distance or climate variation. We used Bayesian multiple regression to estimate the effect of host plant genetic differentiation on associate genetic differentiation over and above the confounding effects of geography and climate. We found that plant and associate genetic structures are indeed often congruent, but isolation by distance and by climate are also common. Multiple regressions established that the effect of host plants on associates is robust to accounting for geographic distance and climate. Finally, associate genetic structure was significantly explained by plant genetic structure more often in antagonistic interactions than in mutualistic ones. This supports a key prediction of coevolutionary theory, that antagonistic interactions promote diversity through local adaptation of antagonists to hosts, while mutualistic interactions promote diversity via the effect of hosts' geographic distribution on mutualists' dispersal.

Interactions between species are widely understood to have promoted the diversification of life on Earth, but how interactions spur the formation of new species remains unclear. Interacting species often become locally adapted to each other, but they may also be subject to shared dispersal limitations and environmental conditions. Moreover, theory predicts that different kinds of interactions have different effects on diversification. To better understand how species interactions promote diversification, we compiled population genetic studies of host plants and intimately associated herbivores, parasites, and mutualists. We used Bayesian multiple regressions and the BEDASSLE modeling framework to test whether host and associate population structures were correlated over and above the potentially confounding effects of geography and shared environmental variation. We found that associates' population structure often paralleled their hosts' population structure, and that this effect is robust to accounting for geographic distance and climate. Associate genetic structure was significantly explained by plant genetic structure somewhat more often in antagonistic interactions than in mutualistic ones. This aligns with a key prediction of coevolutionary theory that antagonistic interactions promote diversity through local adaptation of antagonists to hosts, while mutualistic interactions more often promote diversity via the effect of hosts' geographic distribution on mutualists' dispersal.