Prognostic factors and treatment were analyzed for 2,578 patients to assess the impact of various forms of adjuvant chemotherapy on the natural history of operable stage II (node-positive) breast cancer. The outcome after surgery alone (or with radiotherapy) was determined in 1,014 patients in the natural history data base (NHDB). Adjuvant chemotherapy consisted of L-phenylalanine mustard (L-PAM; 130 patients); cyclophosphamide, methotrexate, and 5-fluoro-uracil (CMF; 645 patients); doxorubicin and cyclophosphamide (AC; 241 patients); CMF plus vincristine and prednisone (CMFVP; 263 patients); and 5-fluorouracil plus AC (FAC; 285 patients). L-PAM had minimal effect on relapse-free survival (RFS) compared to the NHDB, but all combination chemotherapy programs significantly improved RFS and survival compared to the NHDB. In women with 1–3 positive nodes, all combination chemotherapy programs produced similar results. In women with 4–9 positive nodes, the FAC regimen appeared to be associated with superior RFS compared to other programs, but all were superior to the NHDB. In women with 10 or more positive nodes, FAC was the only regimen associated with improved RFS. The use of a NHDB and known pretreatment characteristics, such as nodal status and tumor size, permits comparison of patients at similar risk of recurrence of breast cancer who have received adjuvant chemotherapy and provides leads for evaluation in future prospective clinical trials.