A policy of selective colonoscopy after resection for colorectal cancer was evaluated to assess the benefit of detecting missed synchronous and metachronous tumours in patients under 70 years of age with a good prognosis, and the possible risk of missing such lesions in those not screened. With preoperative barium enema and palpation at laparotomy, synchronous cancers and adenomas were found in 2 and 28 per cent, respectively, of 275 consecutive patients undergoing resection. Eighty-five patients were selected for postoperative colonoscopy, which was repeated at least once in 32, there being a total of 60 repeat examinations. Adenomas, but no cancers, were found in 19 patients (22 per cent) on first screening, but on repeat colonoscopy early metachronous cancers were detected in three patients (9 per cent) and adenomas on 21 examinations (35 per cent). Only one of the patients followed by clinical examination developed a symptomatic tumour. This suggests that postoperative colonoscopy is unlikely to benefit the majority but may be extremely important for young fit patients with continuing polyp formation who are at high risk of developing metachronous cancer.