Concern about world wide local recurrence rates for rectal cancer of 20-45%, together with anxiety at the recent proliferation of adjuvant therapies, led us to review the efficacy of total mesorectal excision (TME) with which no adjuvant therapy had been combined. Precise, sharp dissection is undertaken around the integral mesentery of the hind gut, which envelopes the entire mid rectum. This procedure adds to operative time and complications but has been claimed to eliminate virtually all locally recurrent disease after "curative" surgery. Independent analysis (J. K. M.) of prospective follow-up data extended over a 13-year interval (1978-91; mean 7.5 years). The actuarial local recurrence rate after curative anterior resection at 5 years is 4% (95% Cl 0-7.5%) and the overall recurrence rate is 18% (10-25%). 10-year figures are 4% (0-11%) and 19% (7-32%). In view of the high-risk classification used for the North Central Cancer Treatment Group (NCCTG), which has led to a trend to chemoradiotherapy, a similar group of high-risk Basingstoke cases was constructed for comparison purposes. This group included 135 consecutive Dukes' B (B2) and Dukes' C cancer operations, both anterior resection and abdominal-perineal excision, for tumours below 12 cm from the anal verge. Results from TME alone are substantially superior to the best reported (NCCTG) from conventional surgery plus radiotherapy or combination chemoradiotherapy: 5% local recurrence at 5 years compared with 25% and 13.5%, respectively; and 22% overall recurrence compared with 62.7% and 41.5%, respectively (Dukes' B cases [B2], 15%; Dukes' C cases, 32%). Meticulous TME, which encompasses the whole field of tumour spread, can improve cure rates and reduce the variability of outcomes between surgeons. Far more genuine "cures" of rectal cancer are possible by surgery alone than have generally been believed or are currently accepted. Better surgical results are an essential background for the more selective use of adjuvant therapy in the future.