We evaluated the postoperative use of sequential methotrexate and fluorouracil followed by leucovorin in 679 patients with primary breast cancer, histologically negative axillary nodes, and estrogen-receptor-negative (less than 10 fmol) tumors. No survival advantage was observed with this therapy as compared with no postoperative therapy during four years of follow-up (87 percent vs. 86 percent; P = 0.8). However, there was a significant prolongation of disease-free survival among women who received this therapy as compared with those who did not (80 percent vs. 71 percent; P = 0.003). An advantage was observed in both the patients less than or equal to 49 years old and those greater than or equal to 50. At four years, treatment failure was reduced by 24 percent in the younger group and by 50 percent in the older group. The rates of both local and regional and distant metastases were decreased. These benefits, achieved without the use of an alkylating agent, were associated with tolerable side effects. Multivariate analysis testing for potential interactions failed to identify subgroups of patients who did not benefit from the therapy. These results, although promising, do not obviate the need for additional trials to evaluate potentially better regimens of therapy, but they do suggest that sequential methotrexate-fluorouracil should be used in the control arm in such studies. Their use is also justified for the treatment of patients who refuse to participate in clinical trials, provided the patients meet the eligibility criteria of the present study. Since women with tumors too small for conventional analysis of estrogen-receptor and progesterone-receptor concentrations were not included in this study, we do not recommend systemic treatment for them outside of a clinical trial.