The authors' explanation for the decline in intimate partner homicide is based on a theory of exposure reduction that helps to account for the especially pronounced decline in the rate at which married women kill their husbands. Women are much more likely than men to kill out of a perceived need or desire to protect themselves or their children from a violent or abusive partner. Factors that reduce exposure to violent relationships and provide protective alternatives to violence, therefore, should be more closely associated with variation in the rate at which women kill their partners than the rate at which men kill their partners. The exposure reduction theory is tested with data from a panel of 29 large U.S. cities for the years 1976 to 1992. The homicide data are from the Supplementary Homicide Reports. The dependent variables are the male and female intimate partner homicide victimization rates per 100,000 population, calculated separately for married and unmarried intimate partners. The domestic violence service information was compiled from six editions of national domestic violence service directories. The impact of domesticity on change in intimate partner homicide rates is estimated with measures of the rate of marriage and divorce for each city and time period. Two measures of the economic status of women are incorporated in the analysis: relative educational attainment and gender differences in earnings. Also included in the model are dummy variables for each city and period in the panel. The results of the analysis are generally supportive of the exposure reduction theory. The authors consider the importance of the results for subsequent research on intimate partner homicide and call for further evaluation of the efficacy of legal responses to domestic violence. 1 figure, 4 tables, 11 notes, and 65 references