Our study assessed the accuracy of shoulder ultrasonography in detecting complete and partial rotator cuff tears. We performed a prospective study in 24 patients scheduled for shoulder arthroscopy for impingement, partial-thickness or full-thickness (complete) rotator cuff tears, adhesive capsulitis, or arthritis. Before surgery, all patients underwent diagnostic shoulder ultrasonography followed by a shoulder arthroscopy with examination of the rotator cuff. The arthroscopic findings of rotator cuff integrity were compared with the diagnoses made from ultrasonography. The ultrasonographer was not blinded to the history and examination findings, and the surgeon was not blinded to the results of the ultrasonography. Ultrasound had sensitivity and specificity rates of 80% and 100%, respectively (positive and negative predictive rates of 100% and 88%), in the diagnosis of full-thickness rotator cuff tears and sensitivity and specificity rates of 71% and 100%, respectively (positive and negative predictive values of 100% and 88%), for partial-thickness tears. Although the potential for bias exists in this study, the results suggest that diagnostic ultrasound is effective in the diagnosis of rotator cuff tears.