LAUSR

The 71 currently known species of dwarf geckos of the genus Lygodactylus are a clade of biogeographic interest due to their occurrence in continental Africa, Madagascar, and South America. Furthermore, because many species are morphologically cryptic, our knowledge of species-level diversity within this genus is incomplete, as indicated by numerous unnamed genetic lineages revealed in previous molecular studies. Here we provide an extensive multigene phylogeny covering 56 of the named Lygodactylus species, four named subspecies, and 34 candidate species of which 19 are newly identified in this study. Phylogenetic analyses, based on ∼ 10.1 kbp concatenated sequences of eight nuclear-encoded and five mitochondrial gene fragments, confirm the monophyly of 14 Lygodactylus species groups, arranged in four major clades. We recover two clades splitting from basal nodes, one comprising exclusively Malagasy species groups, and the other containing three clades. In the latter, there is a clade with only Madagascar species, which are followed by a clade containing three African and one South American species groups, and its sister clade containing six African and two Malagasy species groups. Relationships among species groups within these latter clades remain weakly supported. We reconstruct a Lygodactylus timetree based on a novel fossil-dated phylotranscriptomic tree of squamates, in which we included data from two newly sequenced Lygodactylus transcriptomes. We estimate the crown diversification of Lygodactylus started at 46 mya, and the dispersal of Lygodactylus among the main landmasses in the Oligocene and Miocene, 35-22 mya, but emphasize the wide confidence intervals of these estimates. The phylogeny suggests an initial out-of-Madagascar dispersal as most parsimonious, but accounting for poorly resolved nodes, an out-of-Africa scenario may only require one extra dispersal step. More accurate inferences into the biogeographic history of these geckos will likely require broader sampling of related genera and phylogenomic approaches to provide better topological support. A survey of morphological characters revealed that most of the major clades and species groups within Lygodactylus cannot be unambiguously characterized, either by unique character states or by a diagnostic combination of character states. Thus, any future taxonomic work will likely benefit from integrative, phylogenomic approaches.