An earlier report documented that, in a community-dwelling sample of 317 older married couples, cognitive impairment in a wife,vas associated with depressive symptoms in her husband (Moritz, Kasl, and Berkman, 1989). No similar effects were found for wives. Here we examine the extent to which marital closeness moderates the impact of a spouse's cognitive impairment, the stability of influences over 3 years, and gender differences in the associations. Analyses of covariance, controlling for respondent risk factors and potential confounders in the spouse, showed that marital closeness moderated the impact of a wife's cognitive impairment, with husbands in close marriages affected more strongly than husbands in less close marriages. These effects held over 3 years. In addition, husbands became less depressed following the death of a severely impaired wife, whereas widowers whose wife had been unimpaired at baseline were more depressed. None of these effects were found for wives.