Family history of breast cancer was examined among 1159 women who themselves had breast cancer and among 11,590 control subjects in a retrospective case-control study conducted in 1976 among a sample of female nurses in the United States. Having a mother with breast cancer increased a woman's risk of developing the disease by 80%, while having a sister with such a history increased the risk by 150%. Adjustment for potential confounding variables by logistic regression analysis did not alter these estimates appreciably. Both these associations were stronger at younger ages of onset of cancer, and in a positive maternal history was correlated with a larger increase in risk in women who menstruated for more than 35 years.