To successfully perform a total shoulder arthroplasty, it is essential to understand the normal anatomy and biomechanics of the shoulder joint. Currently, nonconstrained prostheses offer the most consistent and durable long-term results, allowing for the restoration of normal anatomy and motion close to that of the normal shoulder joint. To determine the effects of the disease process, surgical technique, and postoperative rehabilitation on glenohumeral biomechanics after nonconstrained total shoulder arthroplasty, a prospective clinical trial was undertaken to evaluate preoperative and postoperative shoulder motion clinically and roentgenographically in nine patients with severe arthritis. Preoperatively, the ratio of glenohumeral to scapulothoracic motion was 1:2; for every degree of glenohumeral movement there were two degrees of scapulothoracic motion. This was associated with significant pain and decreased motion in the shoulder and represented the patient's attempt to immobilize the glenohumeral joint for pain relief and maximize shoulder movement with scapulothoracic motion. After total shoulder arthroplasty, significant improvements were observed in pain relief, motion, and function; however, the ratio of glenohumeral to scapulothoracic motion was not significantly different. The abnormal ratio indicates that less motion occurs between the prosthetic components compared with a normal joint, whereas scapulothoracic motion is unchanged. Abnormal shoulder biomechanics seem to be a function of the underlying disease process and were not restored after total shoulder arthroplasty. Despite the successes to date, improvements on the current state of the art are still needed, and will occur with better understanding of the complex biomechanics of the shoulder joint.