Two experiments were conducted to verify whether the gender stereotypes associated with Italian words are automatically activated, as claimed by Banaji and Hardin (1996) for English words. Subjects were presented with couples formed by a prime word followed by a target pronoun ("Lui" [He] or "Lei" [She]) on which a gender decision task was performed. The results of Experiment 1 only partially confirmed Banaji and Hardin's results (1996) for English: we obtained a general congruency effect, but a facilitatory effect with respect to control words only for words characterized by a natural gender ("babbo - lui" [father (masculine) - he]). No significant effect emerged for words with a stereotypical gender or a grammatical gender. In Experiment 2, the experimental procedure was modified so that the prime presentation and the ISI were both 100 ms longer than in Experiment 1. The experimental list was not modified. The results of Experiment 2 replicated the congruency effect found for words with a natural gender ("babbo - lui") and extended such a facilitatory effect to gender stereotyped words as well ("ingegnere - lui" [engineer (masculine stereotype) - he]). Again no significant effect was obtained for grammatical gender. Overall, Banaji and Hardin hypothesis that words characterized by a stereotypical gender would automatically prime the gender of pronominal targets was not confirmed.