A case-control study is being conducted in which a questionnaire is used to collect information on past history of drug usage (including menopausal estrogens) and the normal weekly dietary intake. Breast cancer cases from 3 selected populations (ages 45--74) are identified: 1) the Caucasians in Hawaii, at high risk for breast cancer; 2) the Japanese in Hawaii, at intermediate risk; and 3) the Japanese in Fukuoka, Japan, representing a low-risk population. For each case, a neighborhood and hospital control matched by race and age is interviewed. Attempts are made to verify the drug history with the subject's personal physician, and a subsample of the study population received a repeated dietary interview to assess its reliability. Pathologic slides of the breast cancer cases are reviewed by three consulting pathologists to confirm cancer diagnosis and to identify possible histologic differences in the 3 populations. In all, 200 cases and 400 controls will be interviewed from each study population.