Comments on the systematic status of specimens belonging to the genus Viridovipera (Serpentes: Viperidae: Crotalinae) from Sichuan and Yunnan provinces of southwestern China, with a redescription of V. yunnanensis

Although the genus Viridovipera has been relatively well studied recently, the status and diagnosis of V. yunnanensis are still unclear. In this study, two samples collected from southwestern Sichuan (Huili), which are putative V. yunnanensis, were analysed in a molecular phylogeny for the first time. These two samples formed a well supported clade that was distinct from its congeners. Multivariate morphometric analysis (principal components analysis and canonical variate analysis) including these and 14 other new Chinese specimens showed that male specimens from Tengchong, Nujiang (Yunnan), and northern Burma are very similar to both the holotype and paratypes of V. yunnanensis in external morphology, and are morphologically distinct from their congeners. A similar trend was found in two new female specimens from Huili, Sichuan. Other new Chinese specimens can usually be allocated to either V. stejnegeri or V. gumprechti, but specimens from northeastern India remain ambiguous in their affinities. Several specimens were inconsistently allocated to species in the phylogenetic and the morphometric analysis; this may be a signal of introgression of mtDNA across species boundaries following hybridization between species. We conclude that although V. yunnanensis represents a valid taxon, which in parts of its range is sympatric with both V. gumprechti and V. stejnegeri, the exact morphological and geographical limits of this species are still not clear. Hemipenial characteristics show only subtle and inconsistent differences between species within Viridovipera and may not be a useful diagnostic character at the species level. We provide a redescription of the holotype of V. yunnanensis and a comparison with similar congeneric species, as a starting point for further investigations of hybridization among them, which ideally should be based on new collections in Sichuan and Yunnan provinces of China, the northeastern region of India and northern Burma, where at least two Viridovipera species seem to co-exist.