The purpose of this study was to analyze age-related changes in the coracoacromial arch and correlate these degenerative changes with rotator cuff tears. We obtained 80 shoulders from 40 cadavers. The mean age at death was 58.4 years. We performed a gross examination of the rotator cuff and the acromion and histological examination of the coracoacromial ligament. The statistical significance of any difference for each group considered was determined by Student's t-test. The rotator cuff was normal in 66 specimens; there was an articular-side partial tear in 4 cases, a bursal-side pal tial tear in 6 cases, and a full-thickness tear in 4 cases. Age was correlated with increasing incidence and severity of cuff tears. We noted age-related degenerative changes in the coracoacromial ligament, degeneration of the acromial bone-ligament junction, and acromial spur formation. Anterior acromial spur was not related to the morphology of the acromion. We observed an increased incidence of bursal-side and complete cuff tears when the acromion was curved or beaked. Degenerative changes in the undersurface of the acromion were also present when the rotator cuff was normal. Bursal-side and complete cuff tears were associated with severe degenerative changes in the acromion in 100% of cases. Articular-side cuff tears were not related either to acromial morphology or degenerative changes in the coracoacromial arch. The association between cuff tears and acromial spur was more evident in the presence of a type III acromion. Our results would suggest that the incidence and severity of rotator cuff tears are correlated with aging and with the morphology of the acromion. Rotator cuff tears that involve the bursal side are often associated with changes in the coracoacromial ligament and the undersurface of the acromion. However, degenerative changes in the coracoacromial arch are always related to aging, also in the presence of a normal rotator cuff. Articular-side partial tears do not cause damage to the undersurface of the acromion.