After curative surgical resection, 621 patients with modified Dukes' stage B2, C1, or C2 colon carcinoma were randomly assigned to one of four treatment programs. These included chemotherapy with fluorouracil and semustine, immunotherapy with methanol extraction residue of bacillus Calmette-Guérin (BCG), combination therapy with fluorouracil, semustine, and immunotherapy, or close follow-up without adjuvant treatment. Treatment continued for 70 weeks. After a median of 51/2 years of follow-up, no significant differences were noted in either recurrence or survival rates among the four treatment programs. Leukemia developed in seven patients, all of whom had received fluorouracil and semustine. The results of this study do not support the use of chemotherapy with fluorouracil and semustine, immunotherapy with methanol extraction residue of BCG, or their combination as an adjuvant treatment program for patients at high risk for recurrent colon carcinoma. The data do, however, demonstrate the necessity for an untreated control group in a trial of adjuvant therapy for colon cancer.