I review species concepts, the taxonomy of phlebotomine sand flies, and some transmission cycles of leishmaniasis in order to illustrate the difficulties of classifying these vectors in a way that will be ideal both for medical parasitologists and sand fly specialists. Choices will have to be made between different classifications, either maintaining a practical one containing few vectorial genera (mostly Phlebotomus for the Old World and Lutzomyia for the Neotropics) or changing the generic names of many vectors so that the classification represents an evolutionary hypothesis. However, sand flies also transmit arboviruses and members of other sand fly genera bite humans, and so vectorial status alone might not provide the criteria for recognizing only a few genera. Vectorial roles are often determined by species-level co-evolution of susceptibility to Leishmania species, with selection being initiated and maintained by ecological contacts. There is only imperfect cocladogenesis of genus-level groups or subgeneric complexes of sand flies and Leishmania species. Natural hybridization between sand fly species has been recorded in several species complexes, and this highlights the need to focus on gene flow and the distribution of phenotypes of biomedical importance, not on taxa.