In 1976, 117,557 women in the United States aged 30-55 years and without a history of cancer provided detailed information on current smoking habits. By 1986, 1,788 cases of breast cancer had been documented during 1,133,682 person-years. There was no association between current smoking and risk of breast cancer (multivariate-adjusted relative risk for smokers of greater than or equal to 25 cigarettes/day compared to nonsmokers: 1.02, 95% confidence interval, 0.96-1.22). Past smoking also was unrelated to breast cancer risk (relative risk, 1.08; 95% confidence interval, 0.96-1.20). The results did not differ by menopausal status. Tumor size and the presence of nodal metastases were unrelated to smoking. Smoking was weakly associated with estrogen receptor-positive tumors (relative risk for smokers of greater than or equal to 25 cigarettes/day compared with never smokers, 1.38; 95% confidence interval, 1.04-1.84), but there was no dose-response relationship across categories of current smoking. These results suggest that smoking and breast cancer are not materially related.