Many clinicians still associate oesophagectomy for oesophageal carcinoma with low cure rates, poor palliation and prohibitive peri-operative mortality. Surgical advances have rendered such perceptions inaccurate. A prospective study of all patients undergoing surgery for oesophageal cancer in an Australian teaching hospital between 1979 and 1993 has been undertaken. Selection, staging, pre-operative preparation, surgical technique and postoperative care were all carefully controlled. One hundred and thirty-seven patients were explored. Twenty-one were inoperable. One hundred and sixteen underwent resection with intent to cure. Hospital mortality for oesophagectomy was 1.7%. There were no cases of clinical anastomotic leakage. Eighty-nine per cent achieved excellent to good swallowing. The median survival for all cases was 14 months and the 5 year survival was 18%. Median survival for resected cases was 18 months and the 5 year survival was 26%. The long-term survival was related to postoperative stage of the disease but not to tumour type. Oesophagectomy for oesophageal cancer will restore good swallowing in 90% of cases. Operative mortality should be less than 5% and the overall 5 year survival 20-30%. Early tumours can often be cured (ca in situ 100%, stages I and II 50-60%), indicating the benefits of early detection. Poor survival in advanced disease (stage III 15%, stage IV 0%) on a background of low surgical mortality indicate the need for better staging and more effective adjuvant therapies.