Exploring rain forest diversification using demographic model testing in the African foam‐nest treefrog <i>Chiromantis rufescens</i>

Abstract Aim Species with wide distributions spanning the African Guinean and Congolian rain forests are often composed of genetically distinct populations or cryptic species with geographic distributions that mirror the locations of the remaining forest habitats. We used phylogeographic inference and demographic model testing to evaluate diversification models in a widespread rain forest species, the African foam‐nest treefrog Chiromantis rufescens . Location Guinean and Congolian rain forests, West and Central Africa. Taxon Chiromantis rufescens . Methods We collected mitochondrial DNA (mtDNA) and single‐nucleotide polymorphism (SNP) data for 130 samples of C. rufescens . After estimating population structure and inferring species trees using coalescent methods, we tested demographic models to evaluate alternative population divergence histories that varied with respect to gene flow, population size change and periods of isolation and secondary contact. Species distribution models were used to identify the regions of climatic stability that could have served as forest refugia since the last interglacial. Results Population structure within C. rufescens resembles the major biogeographic regions of the Guinean and Congolian forests. Coalescent‐based phylogenetic analyses provide strong support for an early divergence between the western Upper Guinean forest and the remaining populations. Demographic inferences support diversification models with gene flow and population size changes even in cases where contemporary populations are currently allopatric, which provides support for forest refugia and barrier models. Species distribution models suggest that forest refugia were available for each of the populations throughout the Pleistocene. Main conclusions Considering historical demography is essential for understanding population diversification, especially in complex landscapes such as those found in the Guineo–Congolian forest. Population demographic inferences help connect the patterns of genetic variation to diversification model predictions. The diversification history of C. rufescens was shaped by a variety of processes, including vicariance from river barriers, forest fragmentation and adaptive evolution along environmental gradients.