A prospective randomised multicentre trial compared pre-operative radiotherapy followed by surgery with surgery alone for rectal cancer < or = 12 cm from the anal verge. Of 468 patients (mean age 67 years, range 31-94, 273 males) who met the entry criteria, 228 were randomised to radiotherapy (3 x 5 Gy over 5 days within 2 days of operation) followed by surgery, and 239 to surgery alone. Randomisation was unknown in 1 patient. Follow-up to death or 5 years was achieved in 454 (97%) patients. 31 (7%) of the 468 patients died within 30 days of operation (radiotherapy and surgery 21 [9%], surgery alone 10 [4%]; P < 0.05). Cardiovascular and thromboembolic complications were more common after radiotherapy and surgery (30, 13%) than after surgery alone (8, 3%; P < 0.005). Of the 280 patients who had curative surgery, 52% of those who had radiotherapy and surgery and 56% of those who had surgery alone survived 5 years (P = 0.88). 395 patients attended outpatients clinics at least once. Local treatment failure was identified during follow-up in 82 patients [31/185 (17%) radiotherapy and surgery; 51/210 (24%) surgery alone; P < 0.05]. It occurred in 33 of the 258 patients who had a curative resection and attended outpatients [radiotherapy and surgery, 11/120 (9%), surgery alone, 22/138 (16%); P = 0.08]. Long-term survival was unaffected, but long-term local recurrence was reduced by the addition of low-dose radiotherapy to surgery. Peri-operative mortality was, however, increased.