A prospective epidemiological study was undertaken to determine the workload and patient characteristics for a putative trauma centre in a large defined area. One thousand and eighty-eight patients were included: 430 brought in dead, 309 hospital deaths and 349 survivors. Types of injury were: blunt 76 per cent, penetrating 3.6 per cent, burns 5.8 per cent, other 14 per cent. The incidence of blunt injury was 19/100,000 for patients arriving alive at hospital and accounted for 0.08 per cent of new A & E attendances. Eight per cent of blunt injury patients were children, 68 per cent were adults and 24 per cent elderly. Major causes of injury were: road accidents 67 per cent and falls 26 per cent. In patients arriving alive after blunt injuries, those who subsequently died were significantly older, more severely injured and more physiologically impaired. Hospital mortality was 45 per cent for blunt, 43 per cent for penetrating injuries, and 67 per cent for burns. TRISS methodology indicated 53 per cent of hospital deaths from blunt injuries were unexpected. Practically, it is questionable whether the incidence of major injuries is sufficient to provide the volume of patients necessary to sustain a Level I Trauma Centre. Nevertheless, concentration of injury service is essential, since no hospital receives sufficient patients to develop and maintain expertise.