The plain radiographic findings on 20 patients with traumatic aortic rupture were analyzed. A mediastinal-width to chest-width (M/C) ratio was calculated at three thoracic levels and compared to two matched groups of patients without aortic tears. Defining an M/C ratio of 0.25 or larger at the level of the aortic arch as abnormal would identify 95% of cases with ruptured aortas, and result in 25% false-positive studies in traumatized patients. A ratio greater than 0.28 retains 85% sensitivity for aortic rupture while increasing specificity to 100%. An analysis of the prevalence and location of rib fractures showed that there was little relation between the presence of fracture and existence of an aortic tear. Therefore, an acutely injured patient with an M/C ratio of greater than 0.25 should be considered highly likely to have an aortic rupture.