Nineteen measurements were made on 136 skulls belonging to seven mustelid species: Meles meles (Eurasian badger), Mustela nivalis, (weasel), Mustela erminea (stoat), Mustela putorius (polecat), Lutra lutra (otter), Mustela furo (ferret), and Mustela vison (American mink), and polecat-ferret hybrids. To investigate shape, size-related effects were eliminated by dividing all measurements by their geometric means. Canonical variate analysis was used to reveal major interspecies distinctions. Excluding the ferrets and polecat-ferrets from the analysis, only 3.2% of the skulls misclassified (one mink, one weasel, and two stoats). Three groups separated on the first canonical axis: 1) badgers, 2) polecats, mink, and otters, and 3) stoats and weasels. The important variables were width of zygomatic arch and height of sagittal crest opposed to the postorbital distance, condylobasal length, and basilar length. Otters separated out on the second canonical axis; the most important variables were postorbital breadth and width of the postorbital constriction opposed to the basioccipital width. There was reasonable separation of polecats from mink on a combination of the second and third canonical axes. On the latter the most important variables were postorbital breadth opposed to postorbital distance. Addition of the ferret data showed that they lay closest to, and overlapped with, the polecats. The stoat and weasel data alone gave complete separation, with height of sagittal crest and width of zygomatic arch opposed to basioccipital width. However, using size-in data the best separation was the relationship between postorbital breadth and either basioccipital width or postorbital distance. Sexual dimorphism was demonstrated in the skulls of badgers but was shown to be relatively insignificant when compared to the interspecific differences. J. Morphol. 260:57–64, 2004. © 2004 Wiley-Liss, Inc.