Abstract
Accurate species delimitation is critical for biodiversity conservation. Integrative taxonomy has been advocated for a long time, yet tools allowing true integration of genetic and phenotypic data have been developed quite recently and applied to few models, especially in plants. In this study, we investigated species boundaries within a group of twelve Pseudophrys taxa from France by analyzing genetic, morphometric and chemical ( i.e ., floral scents) data in a Bayesian framework using the program integrated Bayesian Phylogenetics and Phylogeography (iBPP). We found that these twelve taxa were merged into four species when only genetic data were used, while most formally described species were recognized as such when only phenotypic (either morphometric or chemical) data were used. The result of the iBPP analysis performed on both genetic and phenotypic data supports the proposal to merge Ophrys bilunulata and O. marmorata on the one hand, and O. funerea and O. zonata on the other hand. Our results show that phenotypic data are particularly informative in the section Pseudophrys and that their integration in a model-based method significantly improves the accuracy of species delimitation. We are convinced that the integrative taxonomic approach proposed in this study holds great promise to conduct taxonomic revisions in other orchid groups.