Morphological characters should be assessed in an ecological and evolutionary framework before their use in phylogenetic analyses. We have attempted such an assessment for the tarsal complex of the Tenrecoidea. The 30+/− extant species of these small mammals live in a diverse range of microhabitats. They exhibit markedly different positional behaviors, encompassing four basic locomotor repertoires: terrestrial running and walking, scansorial climbing, digging, and swimming. Articular surfaces from the upper, middle, and lower ankle joints in 10 Malagasy tenrecid, 1 potamogalid, 1 solenodontid, and 1 macroscelidid species were compared. The results were tested against published data on the correlation between function and morphology in other therian locomotor specialists. Descriptive accounts, supported by quantitative analyses, demonstrate significant differences in many aspects of tarsal morphology that may be explained by function-based hypotheses within the context of the tenrecoid heritage. Three closely related tenrecines (Hemicentetes, Setifer, and Echinops) show divergences in form that are consistent with their respective semifossorial, terrestrial, and scansorial/arboreal modes of locomotion. In addition to functional–adaptive considerations, we propose several synapomorphies for the tenrecine and oryzorictine clades that appear to corroborate their monophyly. There are apparent convergences between the habitual diggers from different subfamilies (Hemicentetes and Oryzorictes), and there are adaptive differences within subfamilies (e.g. arboreal Echinops vs. terrestrial Setifer). The few likenesses between Potamogale and Limnogale cannot be supported as homologies, and the proposal of a recent phylogenetic affiliation between them is therefore rejected here. Manifest differences in tarsal form between Geogale and the oryzorictines support recognition of the subfamily Geogalinae. As expected, the tarsal complex of Solenodon is fundamentally unlike that of the tenrecoids. Finally, the similarities in detail between Macroscelididae and Potamogalidae reflect the stabilization of UAJ during plantarflexion, and therefore such attributes are rejected as synapomorphies. Traits with clear species-specific adaptations are a potential interference in cladistic analyses and cannot be meaningfully used without ecology-based character assessment. While this practice may ultimately reduce the size of a database, it will nonetheless result in taxonomic properties with lasting value against which phylogenetic hypotheses may be tested with confidence.