Cryptic diversity impacts model selection and macroevolutionary inferences in diversification analyses

Species persist in landscapes through ecological dynamics but proliferate at wider spatial scales through evolutionary mechanisms. Disentangling the contribution of each dynamic is challenging, but the increasing use of dated molecular phylogenies opened new perspectives. First, the increasing use of DNA sequences in biodiversity inventory shed light on a substantial amount of cryptic diversity in species-rich ecosystems. Second, explicit diversification models accounting for various eco-evolutionary models are now available. Integrating both advances, we explored diversification trajectories among 10 lineages of freshwater fishes in Sundaland, for which time-calibrated and taxonomically rich phylogenies are available. By fitting diversification models to dated phylogenies and incorporating DNA-based species delimitation methods, the impact of cryptic diversity on diversification model selection and related inferences is explored. Eight clades display constant speciation rate model as the most likely if cryptic diversity is accounted, but nine display a signature of diversification slowdowns when cryptic diversity is ignored. Cryptic diversification occurs during the last 5 Myr for most groups, and palaeoecological models received little support. Most cryptic lineages display restricted range distribution, supporting geographical isolation across homogeneous landscapes as the main driver of diversification. These patterns question the persistence of cryptic diversity and its role during species proliferation.