OBJECTIVE To assess the demands made on a regional trauma centre by a district trauma unit. METHODS Two part study. (1) Prospective analysis of one month's workload. (2) Retrospective analysis of one year's workload by using a computer based records system. Comparison of two sets of results. METHODS Accident unit in Gwynedd Hospital, Bangor. METHODS (1) All patients who attended the accident unit in August 1988. (2) All patients who attended the accident unit in the calendar year April 1988-April 1989. METHODS Workload of a district trauma unit. RESULTS In August 1988 there were 2325 attendances; 2302 of these were analysed. In all, 1904 attendances were for trauma; 213 patients were admitted to the trauma ward and 103 required an operation that entailed incision. Patients who attended the unit had a mean (range) injury severity score of 2-13 (0-25). Only two patients had injuries that a district general hospital would not be expected to cope with (injury severity score greater than 20). In the year April 1988-April 1989, 21,007 patients attended the unit. In all, 17,958 attendances were for orthopaedic injuries or injuries caused by an accident; 1966 patients were admitted to the unit. CONCLUSIONS Most trauma is musculoskeletal and relatively minor according to the injury severity score. All but a few injuries can be managed in district general hospitals. In their recent report the Royal College of Surgeons has overestimated the requirements that a British district general hospital would have of a regional trauma centre.