Thirty-six patients were studied prospectively to assess the benefit of 'springing' the pelvis in traumatized patients, to confirm or refute a fracture of the pelvis. None of the patients was multiply injured and half of those with fractures were elderly, sustaining their injuries in simple falls. Springing the pelvis was a poor predictor of the presence or absence of a pelvic fracture, at best it yielded a specificity of 71% and sensitivity of 59%. Its routine use in clinical examination should be abandoned.