The frequent practice of X-raying all injured ankles is reviewed. A simple test using basic equipment to measure the injuries was assessed. A ratio comparing the swelling, expressed as the bimalleolar diameter of the patient's injured ankle to that of the uninjured ankle, gave a highly reliable method of predicting whether a fracture was seen on X-ray or not. All the fractured ankles had a bimalleolar ratio above 1.065. Patients whose ratios were less than 1,065 formed 62 per cent of the series, and as these contained no fractures, the use of X-rays in these cases is questioned.