The radiographic examination of the shoulder, one of the most common studies performed in the emergency department of examine trauma patients, very often yields poor and incorrect results. A radiographic study of the shoulder requires at least two radiographs, and different projections-namely, a frontal anteroposterior view and some additional projections, at right angles to one another. The shoulder radiographs of 126 patients examined for acute trauma were retrospectively reviewed to investigate the adequacy of lateral and oblique apical views in detecting shoulder injuries missed on anteroposterior view. The antero-posterior view was the most effective one, showing 88% of the injuries. Even though the oblique apical and the lateral views were less sensitive than the antero-posterior view, detecting 82% and 60% of the injuries, respectively, they were more sensitive for some specific injuries missed by the other views. Thus, glenohumeral dislocations with or without fracture and some clavicular fractures were best depicted by the oblique apical view, while the fractures of the body and coracoid process of the scapula were best detected by the lateral view. However, the two views had a complementary role and, in our series of patients, both of them showed four injuries missed by the anteroposterior view. Therefore, we conclude that the radiologist should collaborate with the orthopedist to choose which projections to use in each patient on the basis of physical and clinical examinations.