We carried out four experiments to assess the extent to which familiarity with certain objects in everyday life is related to gender and can account, at least partially, for the semantic category dissociation observed in a few brain-damaged patients. In the first experiment, 210 normal subjects, half males and half females, were given the names of 60 stimuli from the Snodgrass and Vanderwart's set, 30 belonging to living categories and 30 to non-living categories. The task was to rate their familiarity, based on the frequency with which one (i) thinks or speaks of a given item, (ii) sees it represented in the media, and (iii) is confronted with real exemplars. The three indices were highly correlated and their average value was, therefore, used. Females gave higher familiarity ratings to fruit, vegetables and furniture and males to tools. The second experiment was aimed to verify whether the gender difference was responsible for the category dissociation found following brain damage. A male patient with greater impairment for living categories and a female patient with greater impairment for non-living categories were requested to name the same 60 stimuli and their scores were analysed, partialling out the familiarity effect, measured both with the non-gender specific index of Snodgrass and Vanderwart and with the new gender-specific index. In either case, the category dissociation remained significant. To determine if the mean general population familiarity index was valid for the single subject, we studied whether a cohabitant first degree relative was able to predict a normal subject's familiarity better than the population index. Contrary to expectations, the better predictor was the population index. The test-retest reliability of each subject's familiarity ratings was satisfactory, but not higher than the correlation between the personal judgement of each subject and the population index.