Roman Yukilevich1, Fumio Aoki2, Scott Egan3 and Linyi Zhang4 1Department of Biology, Union College, Integrated Science and Engineering Complex, Schenectady, New York 12308, USA 2Department of Ecology and Evolution, Stony Brook University, Stony Brook, New York 11794, USA 3Department of Biosciences, Rice University, Houston, Texas 77005, USA 4Department of Biological Sciences, George Washington University, Washington, D.C. 20052, USA Correspondence: yukilevrunion.edu

Speciation often involves the evolution of multiple genetic-based barriers to gene flow (i.e., “coupling”). However, barriers may exhibit a diversity of evolutionary interactions during speciation. These dynamics are important in reinforcement, where selection may favor different prezygotic isolating barriers to avoid maladaptive hybridization. Here we study the interaction between evolution of sexual and habitat isolation. We first review the empirical literature where both barriers were explicitly considered, and then develop a population genetic model of reinforcement. Most studies of both sexual and habitat isolation were found in phytophagous insect systems. In 76% of these studies, both barriers coevolved; the remaining cases either showed only habitat isolation (21%) or only sexual isolation (3%). Our two-allele genetic mechanism model of each barrier also found that these often coevolved, but habitat isolation was generally more effective during reinforcement. Depending on the fitness of hybrids (e.g., Dobzhansky–Muller incompatibilities) and initial migration rate, these barriers could either facilitate, curtail, or have no effect on each other. This indicates that basic parameters will alter the underlying evolutionary dynamics, and thus the nature of “speciation coupling” will be highly variable in natural systems. Finally, we studied initially asymmetrical migration rates and found that populations with higher initial emigration evolved stronger habitat isolation, while populations that initially received more immigrants exhibited stronger sexual isolation. These results are in line with observations in some empirical studies, but more data is needed to test their generality.