From June 1974 to October 1976, 288 patients with small cell undifferentiated lung carcinoma were entered into a randomized, controlled study comparing the two noncycle-active induction regimens of cyclophosphamide vs. the combination of cyclophosphamide, doxorubicin and imidazole carboximide (DTIC). Patients were stratified by extent of disease, previous radiotherapy and performance status. Responding patients and those who did not progress were then randomized to receive their initial regimen alone, or their initial regimen with added cycle-active therapy (vincristine, hydroxyurea and methotrexate). While only 4/34 (12%) evaluable patients treated with cyclophosphamide achieved a response (greater than 50% regression), a final total of 119/217 (57%) evaluable patients on the three drugs have responded (p = 0.005). The survival curve for all the combination-treated patients was significantly better than for those treated with cyclophosphamide alone (p = 0.012). There was no demonstrable statistical superiority in length of remission or survival for patients on the combination who received in addition cycle-active consolidation therapy. In the combination chemotherapy group, survival duration was longer for patients with limited disease than extensive disease (p = 0.035). There was a strong correlation between quality of remission produced by the combination and survival.