In a prospectively randomized study, the effect of adjuvant chemotherapy with 5-FU on survival and recurrence rates was analyzed in 299 evaluable patients with colorectal carcinoma who either underwent a curative or a palliative resection. In the treatment group, chemotherapy consisted of the intravenous administration of 12 mg/kg daily of 5-FU for 4 consecutive days, then 6 mg/kg on alternate days, to the point of toxicity, or to a maximum of five doses, followed by 12mg/kg weekly for 1 year. Some degree of drug toxicity was seen in the majority of patients, was rarely severe, and there have been no drug-related deaths. Analysis of the survival curves and disease-free interval curves reveal definite evidence of drug benefit in two unfavorable subgroups, namely patients with Dukes C tumors and in patients whose tumor was located in the rectum. In the chemotherapy groups, patients who were treated to toxicity (WBC less than 4000 mm3), the disease-free interval was significantly longer than the nonleukopenic patients. We conclude that the addition of 5-FU to the surgical treatment of colorectal carcinoma provides a small, but significant benefit in patients with colorectal cancer in certain unfavorable subgroups, namely patients with Dukes C lesions and patients with rectal carcinoma.