Men and women in Turkey and Brazil completed the Ambivalent Sexism Inventory (Glick & Fiske, 1996) and measures of attitudes about wife abuse. In both nations hostile sexism (HS) and benevolent sexism (BS) positively correlated with attitudes that legitimize abuse. Regression analyses revealed that HS accounted for unique variance, but BS (once HS was controlled) was unrelated to wife abuse attitudes. These results: (a) add to the evidence for the cross-cultural validity of ambivalent sexism, (b) suggest that HS supports the justification of violence against wives, and (c) imply that the ostensible protectiveness of BS is contingent, failing to shield women from abuse if they are deemed to have challenged a husband's authority or violated conventional gender roles. (Contains 3 tables.)