Molecular and morphological analyses of variation often conflict with historical species descriptions based on a few characters and small samples sizes. Here we present a molecular phylogeny together with a quantitative morphologica analysis of the species in Sturnira in Ecuador. The 438 terminal taxa or organisms included in the anlaysis occur within a total of 10 ingroup lineages, which contain considerable substructure. Some species, as recognized by their morphologica traits, form paraphyletic arrangements with other taxa. We could not distinguish the close species pairs S. erythromos/S. bogotensis and S. ludovici/S. oporophilum in morphospace and therefore when distinct lineages were recovered genetically, they initially contained mixed membership of specimens identified using morphological criteria. Similarly the qualitative character states that diagnose S. luisi in its original description are not recovered in a quantitative analysis of morphological variation and thus S. luisi cannot be mapped to a single lineage in a molecular phylogeny. We presen additional evidence to corroborate the existence of S. perla as a species. We found a remarkable geographic structure within some species containing sister pairings, with lineages having a clear eastern or western distribution in relation to the Andes. Our analysis demonstrates the potential for conflict between character-based diagnoses, analysis of morphological variation and molecular phylogenetics in the identification of species and supports a combined approach to this problem.