Thirty-seven patients with low rectal carcinoma were treated by transanal resection between January 1979 and December 1988. Adjuvant radiotherapy was used in eight patients before operation, in 13 after operation and six patients had both preoperative and postoperative radiotherapy. The selection criteria for transanal resection were low, superficial tumours treated conservatively with curative intent (group 1, 18 patients) or patients medically unfit for (12) or refusing (seven) abdominoperineal resection (APR) (group 2, 19 patients). In group 1, the local recurrence rate was 11 per cent with an overall 5-year survival rate of 100 per cent. In group 2, the local recurrence rate was 53 per cent with an overall 5-year survival rate of 35 per cent. There was no postoperative mortality and minimal morbidity. Salvage of local failure was by APR in six patients. The survival rate after local recurrence was 70 per cent at 1 year and 30 per cent at 5 years. Transanal resection can be proposed as a curative procedure for selected low, small, minimally infiltrating and well differentiated adenocarcinomas. Local control might be improved by postoperative radiotherapy for Astler-Coller stage B1 tumours or those above 3 cm in diameter. Further trials should examine the results of high-dose (60 Gy) preoperative radiotherapy so that transanal resection might be employed for favourable post-irradiation stage (0 or A) lesions. This new strategy will benefit from a better evaluation of tumour response after primary radiotherapy and an improved assessment of histological excision margins.