One hundred and eighty patients had a jejunoileal bypass performed during the years 1971-1982. By leaving only 14 in. (35 cm) of intestine in continuity a mean weight loss of 34.4 per cent (s.d. = 8.5) was achieved over 2 years and, unless the operation had to be reversed for complications, this weight loss was maintained. The improvement in quality of life for a majority of patients should not be undervalued. Two-thirds of patients required admission for complications and eight patients died (4 per cent). Many of these problems were provoked by an inability to control eating. There have been no hospital deaths since 1976 which we attribute to better management of complications and a policy of early reversal for patients with excessive weight loss and signs of metabolic failure. Despite performing jejunoileal bypass less often in recent years we are still frequently reversing patients with electrolyte disturbances, metabolic failure, urinary calculi or arthritis. Thirty patients (16.7 per cent) have been reversed, half more than 5 years after bypass. Metabolic failure may occur even after many years of stable weight reduction. Because this is not well known the insidious onset of new weight loss and malaise may not be recognized, or not associated with the bypass many years before. Indefinite outpatient surveillance is mandatory. Changes in the operation have not significantly affected results. There has been no serious liver dysfunction in the 7:7:CJ group but this may reflect better management of lesser metabolic disturbances. Jejunoileal bypass remains the most effective operation for gross obesity and, with experience, can be performed safely. However, the complication rate and difficulty maintaining satisfactory follow-up on large numbers of young patients makes it an unacceptable procedure on any major scale.