A prospective study was made of 92 patients who underwent emergency colorectal surgery during a 1-year period. A dedicated emergency theatre allowed half of the patients to be operated on between 09.00 and 17.00 hours with greater seniority of operating surgeons. The overall mortality rate was 14 per cent and the primary resection rate was 79 per cent. The mortality rate was 12 per cent for right-sided resection with anastomosis and 24 per cent for left-sided resection without anastomosis, including those undergoing Hartmann's procedure which had a 35 per cent mortality rate. Immediate left-sided anastomosis was performed safely in all 14 patients in whom it was attempted. Firms headed by consultants with and without a special interest in colorectal disease made differing use of primary resection and immediate anastomosis (67 versus 41 per cent, P < 0.05) and Hartmann's (7 versus 25 per cent, P < 0.05) and non-resectional (15 versus 29 per cent, P not significant) procedures. Retrospective surveys of perioperative deaths examine inadequate management and subsequent mortality rates, but do not reveal the significant morbidity rate that occurs in survivors. This audit revealed significant diferences in the management of patients with colorectal emergencies between, firms headed by specialist and non-specialist consultants. In order to minimize these diflerences we believe that recent advances in colorectal practice should be included in the training of all surgeons.