A breast cancer case-control study in Atlanta and 5 counties of central New Jersey involving interviews with 960 white and 281 black cases younger than 54 years of age enabled assessment of reasons for the varying incidence rates among these 2 ethnic groups. Of interest was why rates of breast cancer are higher among older white women, a trend that is reversed among very young women (<40 years). Calculation of the prevalence of exposure to classic and speculative risk factors and associated relative risks enabled derivation of population attributable risks (PARs) for the various combinations of age and ethnic groups. A higher PAR was derived for older (40-54 years) white (62%) than black (54%) women, which appeared to account for the observed difference in incidence between the 2 ethnic groups. Most of the difference in PARs between older whites and blacks was accounted for by whites having fewer births, later ages at first birth and slightly higher risks associated with reproductive and menstrual factors. Consideration of only well-established breast cancer risk factors showed a PAR among older whites of 57%, an estimate comparable to those previously published. Slightly higher overall PARs were derived when analyses considered several speculative but modifiable risk factors, including years of use of oral contraceptives, body size and alcohol consumption. Many of the analyses among younger women (20-39 years) were limited by available numbers, but it appeared that very little disease occurrence in young black women was associated with the factors studied.