Between 1978 and 1986 colorectal adenomas without invasion were removed in 552 patients under 76 years of age. Patients were randomly allocated to different follow-up groups with intervals of colorectal examination varying from 6 to 48 months after the initial polypectomy. A large bowel carcinoma developed in three patients and new adenomas in 87 patients. Based on the morphology of the initial adenomas possible relationships between prognostic factors and the risk of new adenoma formation were assessed using life-table analysis. The advantage of a possible decrease in the risk of developing carcinoma was offset by the disadvantage of eight severe complications occurring in 1818 colonoscopies, one of which proved fatal. The occurrence of one of the three carcinomas was considered a failure of the follow-up programme.