The objective of this research was to determine the effect of residential racial segregation on all-cause and cardiovascular disease mortality in New York City. A cross-sectional study of residents in New York City was conducted linking mortality records from 1988 through 1994, to the 1990 United States Census data stratified by zipcode. All-cause and cardiovascular disease mortality rates for non-Hispanic blacks and whites were estimated by zipcode. Zipcodes were aggregated according to the degree of residential segregation (predominantly (> or = 75%) white and black areas) and mortality rates were compared. Multiple regression analysis was used to associate population characteristics with mortality. In New York City, although overall mortality rates of blacks exceed whites, these rates varied substantially by locality according to the pattern of racial segregation. Whites living in the higher (mainly white) socioeconomic areas had lower mortality rates than whites living in predominantly black areas (1473.7 vs 1934.1 for males, and 909.9 vs 1414.7 for females for all-cause mortality). This was true for all age groups. By contrast, elderly blacks living in black areas, despite their less favorable socioeconomic status, had lower mortality rates for all-cause, total cardiovascular disease, and coronary heart disease, than did those living in white areas, even after adjusting for available socioeconomic variables. Racial segregation in residence is independently associated with mortality. Within racially segregated areas, members of the dominant group, for all age, among whites and elderly blacks, enjoy outcomes superior both to members of the minority racial group of their community, and to members of the same race residing in other areas, where they are in the minority, independent of socio-economic status.