A recent estimate suggests that the Scottish wildcat may be critically endangered. Nevertheless, there is still no uncontroversial method for diagnosing the Scottish wildcat. We analysed morphological differences between wild-living cats in Scotland on the basis of 20pelage characters, scoring from 1 (domestic cat) to 3 (wildcat), in combination with 40skull parameters and intestinal length. A cluster analysis, based on Principal Components derived from the scores for pelage characters, showed that the wild-living cats fell into three main groups without any a priori classification. Each group corresponds well to the traditional characteristics of wildcats, hybrids and domestic cats, respectively, and the former two each show higher levels of morphological homogeneity compared with the third group. The three groups are most significantly differentiated by seven pelage characters: (1) extent of dorsal stripe, (2) shape of tail tip, (3) distinctness of tail bands, (4) presence/absence of broken stripes and (5) spots, on flanks and hindquarters, (6) shape and number of stripes on nape and (7) on the shoulders. Most Group-1 cats (75.6%, n=74), but none of the other two groups, score more than 2 for all seven characters. All Group-3 cats (n=35) and some Group-2 cats (19.2%, n=26), but no Group-1 cats, scored 1 for one or more of the seven characters. We propose that Group-1, which is the furthest from the domestic cat in all criteria, should be used to define the Scottish wildcat. However, in practice, if a wild-living cat does not score 1 for any of the seven characters it should be treated as a wildcat in the field. These definitions provide a simple way of diagnosing a Scottish wildcat scientifically, as well as practically, which will effectively facilitate conservation action and the enforcement of protective legislation.