In a prospective randomized study, one- and two-layer anastomoses were compared following subtotal oesophagectomy and gastric substitution with cervical oesophagogastric anastomosis. After 54 one- and 53 two-layer procedures the rates of anastomotic leakage were the same (19 per cent). After a mean follow-up of 44 weeks, 13 of 51 patients (25 per cent) undergoing one-layer anastomosis and 28 of 50 (56 per cent) having the two-layer procedure complained of cervical dysphagia and required dilatation. The anastomotic strictures were fibrotic in 11 of 51 patients (22 per cent) undergoing one-layer anastomosis and in 24 of 50 (48 per cent) receiving the two-layer operation. Strictures were malignant in two and four patients (4 and 8 per cent) respectively. The lower incidence of fibrotic stricture following one-layer anastomosis was significant (P < 0.01), but not that of malignant stricture. With comparable leakage rates, one-layer anastomosis is superior to the two-layer procedure because of the lower incidence of fibrotic stricture.