Risk of breast cancer following radiation exposure was studied, based on surveys of tuberculosis patients who had multiple fluoroscopic examinations of the chest, mastitis patients given radiotherapy, and atomic bomb survivors. Analysis suggests that the risk is greatest for persons exposed as adolescents, although exposure at all ages carries some risk. The dose-response relationship was consistent with linearity in all studies. Direct evidence of radiation risk at doses under 0.5 Gy (50 rad) is apparent among A-bomb survivors. Fractionation does not appear to diminish risk, nor does time since exposure (even after 45 years of observation). The interval between exposure and the clinical appearance of radiogenic breast cancer may be mediated by hormonal or other age-related factors but is unrelated to dose. Age-specific absolute risk estimates for all studies are remarkably similar. The best estimate of risk among American women exposed after age 20 is 6.6 excess cancers/10(4) WY-G-Y (10(6) WY-rad).