We have examined the association of alcohol intake during various life periods, by beverage category, with breast cancer risk in Greece. A hospital-based case-control study was performed in Athens, involving 820 women with breast cancer as well as 795 orthopedic patients and 753 healthy visitor controls. Relative risk patterns were very similar with either control series, which were therefore combined to increase precision of the estimates. Drinkers of beer were at significantly elevated risk for breast cancer [odds ratio (OR), 1.34 (95% confidence interval, 1.05-1.71)]. However, drinkers of other alcoholic beverages were not at increased risk. Among beer drinkers there was no dose-response, and drinkers of both beer and other beverages had a lower OR compared to drinkers of beer only. By contrast, drinkers of 3 or more glasses of alcohol per day, mostly of spirits, were at elevated risk for breast cancer [OR for 3 - < 4 glasses per day, 3.01 (1.14-7.95); OR for 4 or more glasses per day, 3.79 (1.05-13.71)]. Reported frequency of consumption was a stronger predictor of breast cancer risk than either duration-weighted total consumption or consumption before the age of 30 years. There were no coherent patterns for interaction with menopausal status, obesity or use of menopausal estrogens. The association of very low levels of alcohol intake with breast cancer risk may be due to confounding, whereas drinking 3 or more glasses of alcoholic beverages daily appears to genuinely increase breast cancer risk, perhaps by acting as a late-stage growth enhancing factor. However, the data also are compatible with a linear relationship that has no threshold.