This study examined the correlates of hospital closure for short-stay, nonfederal hospitals in New York City. Twenty-four of the 94 hospitals analyzed closed between December 1975 and May 1981. The mean proportion of nonwhite inpatients was higher for those hospitals that closed (0.45) than for those that remained open (0.26) (P less than 0.01). Approximately 48.8% of the 3,308 inpatients in all closed hospitals combined were nonwhite, while 33.3% of the 29,130 inpatients in all open hospitals combined were nonwhite (P less than 0.0001). Multiple regression analysis showed the proportion of inpatients who were nonwhite to be a strong predictor of hospital closure. The racial composition of the surrounding neighborhood was also a positive correlate of closure. Negative predictors included mean occupancy rate, the presence of social work services, and a variable representing the interaction of the racial composition of inpatient and neighborhood populations. In separate analyses, the proportion of inpatients with Medicaid coverage also emerged as a strong, independent correlate of closure. Hospital closures reflect the interplay of complex demographic and economic forces. In the context of proposed changes in health care financing, these findings suggest that equity considerations are important in the preservation of medical services.