Clinically evident anastomotic dehiscence was studied in 1466 patients who had undergone resection of a large-bowel adenocarcinoma. The overall incidence of anastomotic leakage was 13%, but the incidence varied between surgeons (range 0.5% to over 30%). Morbidity and mortality were significantly higher in those patients in whom the anastomosis failed to heal primarily. If these results are extrapolated to the national level, it should be possible by achieving results closer to those in patients without leakage to reduce overall post-operative mortality after resection of large-bowel cancer by 2% and to achieve an appreciable reduction in morbidity. Both factors are clinically important and, taken together, could result in appreciable saving of revenue.