A 7 1/2-year consecutive series is presented from a district hospital with a policy of referring all rectal carcinomas to one surgical firm. The performance of lower anterior resections has limited the rate of abdominoperineal excision with permanent colostomy to 11%. Of 115 patients in whom curative resection was attempted, 69 had anastomoses below 5 cm and 39 had mural resection margins of less than 2.5 cm. Surgical priority, however, was given to complete excision of the visceral rectal mesentery or mesorectum. At an average of 4.2 years postoperatively, three pelvic recurrences have developed but there have been no staple-line recurrences in patients who had "curative" surgery. The corrected cumulative probability of survival at 5 years is 87% and the tumour-free survival by Dukes stage is A 94%, B 87%, and C 58%. Patients with low tumours did no less well than those with high tumours, when treated by anterior resection. On this evidence, it is often safe to limit mural clearance and thus preserve the anal sphincters, provided that the mesorectum is excised intact with the cancer.