Navigating through the high species diversity of wild bees (Hymenoptera: Apoidea) requires more than identification keys

Abstract In the context of pollinator decline and species-focused conservation measures, the need for data on wild bees continues to grow, and identifying them at a species level has become a crucial issue. This challenge is linked to a lack of naturalist skills, to the difficulty of the group, but also to a lack of tools to train people in this expertise, such as accessible, comprehensive, and up-to-date identification keys. Online multi-access keys are designed to facilitate identification, better illustrate criteria, and enable more fluid updates. These keys require a higher initial investment than dichotomous keys, but they should enable correct identification if they contain at least the information found in the corresponding traditional keys. In this study, we tested whether multi-access keys could help identify wild bees in France. We created three multi-access keys for the three most diverse genera of bees and tested them against dichotomous keys. These tests were carried out by 101 participants each having 15 min per specimen. The resulting 1320 identifications show that the multi-access format returns significantly less correct identifications from the 10 specialists (−19.4%), for longer identification time (+ 156 s). There were no significant differences between the keys for the less-experienced participants. Our results also highlight a high misidentification rate, 23–80% for specialists and 74–100% for beginners, stressing that species identification of wild bees in 15 min using only an identification key is ill-advised. This study suggests that morphological identification keys alone are not enough for fast and accurate species identifications of wild bees.