Most new cryptic species are described using conventional tree- and distance-based species delimitation methods (SDMs), which rely on phylogenetic arrangements and measures of genetic divergence. However, although numerous factors such as spatial population structure and gene flow are known to confound phylogenetic and species delimitation inferences, the influence of these processes on species estimation is rarely evaluated. Using large amounts of exons, introns, and ultraconserved elements obtained using the FrogCap sequence-capture protocol, we compared conventional SDMs with more robust genomic analyses that assesses spatial population structure and gene flow to characterize species boundaries in a Southeast Asian frog complex (Pulchrana picturata). Our results showed that gene flow and introgression can produce phylogenetic patterns and levels of divergence that resemble distinct species (up to 10% divergent in mitochondrial DNA). Hybrid populations were inferred as independent (singleton) clades that were highly divergent from adjacent populations (7–10%) and unusually similar (<3%) to allopatric populations. Such anomalous patterns are not uncommon in Southeast Asian amphibians, which brings into question whether the high cryptic diversity observed in other amphibian groups reflect distinct cryptic species—or, instead, highly structured and admixed metapopulation lineages. Our results also provide an alternative explanation to the conundrum of divergent (sometimes non-sister) sympatric lineages―a pattern that has been celebrated as indicative of true cryptic speciation. Based on these findings, we recommend that species delimitation of continuously distributed “cryptic” groups should not rely solely on conventional SDMs but should necessarily examine spatial population structure and gene flow to avoid taxonomic inflation.

The relative roles of rivers and refugia in shaping the high levels of species diversity in tropical rainforests has been widely debated for decades. Only recently has it become possible to take an integrative approach to answer these questions with genomic sequencing and paleo-species distribution modeling. Here, we tested the predictions of the classic river, refuge, and river-refuge hypotheses on diversification in the arboreal West and Central African snake genus Toxicodryas. We used dated phylogeographic inferences, population clustering analyses, machine learning-based demographic model selection, species paleo-distribution range estimates, and climate stability modeling to conduct a comprehensive phylogenomic and historical demographic analysis of this genus. Our results revealed significant population genetic structure within both Toxicodryas species, corresponding geographically to river barriers, and divergence times ranging from the mid to late Miocene. Our demographic and migration analyses supported our interpretation that rivers have represented strong barriers to gene flow among populations since their divergence. Additionally, we found no support for a major contraction of suitable habitat during the last glacial maximum, allowing us to reject both the refuge and river-refuge hypotheses in favor of the river barrier hypothesis. This study highlights the complexity of diversification dynamics in the African tropics and the advantage of integrative approaches to studying speciation in tropical regions.