Sky, sea, and forest islands: Diversification in the African leaf‐folding frog Afrixalus paradorsalis (Anura: Hyperoliidae) of the Lower Guineo‐Congolian rain forest

Abstract

Aim To investigate how putative barriers, forest refugia, and ecological gradients across the lower Guineo‐Congolian rain forest shape genetic and phenotypic divergence in the leaf‐folding frog Afrixalus paradorsalis, and examine the role of adjacent land bridge and sky‐islands in diversification. Location The Lower Guineo‐Congolian Forest, the Cameroonian Volcanic Line (CVL), and Bioko Island, Central Africa. Taxon Afrixalus paradorsalis (Family: Hyperoliidae), an African leaf‐folding frog. Methods We used molecular and phenotypic data to investigate diversity and divergence among the A. paradorsalis species complex distributed across lowland rain forests, a land bridge island, and mountains in Central Africa. We examined the coincidence of population boundaries, landscape features, divergence times, and spatial patterns of connectivity and diversity, and subsequently performed demographic modelling using genome‐wide SNP variation to distinguish among divergence mechanisms in mainland (riverine barriers, forest refugia, ecological gradients) and land bridge island populations (vicariance, overwater dispersal). Results We detected four genetically distinct allopatric populations corresponding to Bioko Island, the CVL, and two lowland rain forest populations split by the Sanaga River. Although lowland populations are phenotypically indistinguishable, pronounced body size evolution occurs at high elevation, and the timing of the formation of the high elevation population coincides with mountain uplift in the CVL. Spatial analyses and demographic modelling revealed population divergence across mainland Lower Guinea is best explained by forest refugia rather than riverine barriers or ecological gradients, and that the Bioko Island population divergence is best explained by vicariance (marine incursion) rather than overseas dispersal. Main conclusions We provide growing support for the important role of forest refugia in driving intraspecific divergences in the Guineo‐Congolian rain forest. In A. paradorsalis, sky‐islands in the CVL have resulted in greater genetic and phenotypic divergences than marine incursions of the land bridge Bioko Island, highlighting important differences in patterns of island‐driven diversification in Lower Guinea.

Comparative phylogeography of West African amphibians and reptiles

“…2017; Charles et al. 2018; Leaché et al. 2019a), although comparative studies that test for shared divergences among species across important biogeographic areas in this region have not been conducted.…”

Evolutionary history of the Cameroon radiation of puddle frogs (Phrynobatrachidae: Phrynobatrachus), with descriptions of two critically endangered new species from the northern Cameroon Volcanic Line

“…The identification of "P. hylaios" from Mt. Nlonako by Herrmann et al (2005) also remains questionable based on the biogeographic patterns (mountains north of the Sanaga River harbor usually different evolutionary lineages in amphibians; Portik et al, 2017;Charles et al, 2018;Leaché et al, 2019) until the specimen is thoroughly investigated. Last, "P. hylaios" on the photograph in Channing & Rödel (2019) also does not correspond to the morphology of P. hylaios (e.g., missing dark face mask).…”

Genome‐scale sampling suggests cryptic epigenetic structuring and insular divergence in Canada lynx

“…Historical demography was assessed using the program ∂a∂i (Gutenkunst, Hernandez, Williamson, & Bustamante, 2009), which can be used to assess the fit between various demographic scenarios and the expected allele frequency spectrum according to diffusion equations; moreover it is suitable for our neutral SNP data (Shafer, Gattepaille, Stewart, & Wolf, 2015). For model selection, we used the demographic models outlined in Charles et al, 2018 ( Figure S5, Table S4), within the model selection framework described in Portik et al (2017) and compared to simulations generated using the process within Barratt et al, 2018. Further details on the methods involving our demographic inferences can be found in Supporting Information.…”