Résumé
Delimitation of closely related species is often hindered by the lack of discrete diagnostic morphological characters. This is exemplified in bumblebees (genus Bombus). There have been many attempts to clarify bumblebee taxonomy by using alternative features to discrete morphological characters such as wing shape, DNA, or eco-chemical traits. Nevertheless each approach has its own limitations. Recent studies have used a multisource approach to gather different lines of speciation evidence in order to draw a strongly supported taxonomic hypothesis in bumblebees. Yet, the resulting taxonomic status is not independent of selected evidence and of consensus methodology (i.e. unanimous procedure, majority, different weighting of evidence). In this article, we compare taxonomic conclusions for a group of taxonomically doubtful species (the Bombus lapidarius-group) obtained from the four commonly used lines of evidence for species delimitation in bumblebees (geometric morphometric of wing shape, genetic differentiation assessment, sequence-based species delimitation methods and differentiation of cephalic labial gland secretions). We ultimately aim to assess the usefulness of these lines of evidence as components of an integrative decision framework to delimit bumblebee species. Our results show that analyses based on wing shape do not delineate any obvious cluster. In contrast, nuclear/mitochondrial, sequence-based species delimitation methods, and analyses based on cephalic labial gland secretions are congruent with each other. This allows setting up an integrative decision framework to establish strongly supported species and subspecies status within bumblebees.