Results obtained during the first 5 years of a randomized study of postoperative radiotherapy (50 Gy) are presented. Criteria for randomization were fulfilled in 494 of 861 patients with Dukes' B and C tumors, when the trial was closed. Severe complications from radiotherapy approximated 10%. Probability of survival without local failure within 24 months was significantly higher after radiotherapy in patients with Dukes' C tumors, and the time of local failure was delayed 1 year. Patients with Dukes' B tumors had no benefit from radiotherapy. Risks of distant metastases and death were not influenced by radiotherapy in the main groups. Plasma-CEA measurements were evaluated blindly, and radiotherapy changed the critical levels of CEA for detection of recurrent cancer. It was concluded that patients with Dukes' C tumors may benefit from radiotherapy and plasma-CEA levels are influenced by radiotherapy, which may be important, when these are used in screening for recurrent cancer.