Lateral lymph node dissection has been advocated for advanced low rectal tumours. However, its benefit is debatable because of the possibility of postoperative bladder and sexual impairment. To assess the role of lateral lymph node dissection 95 patients who underwent the procedure between 1981 and 1991 were reviewed and compared with 83 who had resection of rectal cancer without lateral node dissection. Only ten (11 per cent) of the 95 patients had lateral lymph node involvement; all had Dukes' C tumours. Lymphovascular invasion was present in 50.6 per cent of all patients and neural invasion in 27 per cent of 65 examined specimens. Local recurrence, distant metastasis and overall 5-year survival rates were 7, 9 and 76 per cent respectively in patients undergoing extended lymphadenectomy and 16, 7 and 72 per cent respectively in those who had resection alone. There were no statistically significant differences in survival between the two groups for any Dukes' stage. Recurrence, metastasis and survival were related more to venous or neural invasion and tumour spread than to node dissection. These results demonstrate that patients with Dukes' A tumours do not benefit from lateral lymph node dissection, and that local recurrence rates in those with Dukes' B and C lesions are not significantly decreased. It is concluded that extended lymphadenectomy is unlikely to provide significant benefit to patients with low rectal cancer.