We evaluated the function of the supraspinatus tendon with a dynamic shoulder model. Active glenohumeral joint motion was simulated in 10 cadaveric shoulder specimens with hydrodynamic cylinder forces at the deltoid muscle and at the rotator cuff. Computerized regulation initiated standardized cycles of glenohumeral joint motion, where the isolated effect of the supraspinatus muscle could be studied. The efficacy of the supraspinatus muscle on elevation of the glenohumeral joint was measured with an ultrasonic sensor system. Pressures underneath the coracoacromial vault were recorded with capacitive sensors, as an indicator of the impingement at the shoulder. Elimination of force of the supraspinatus muscle led to a 6 percent decrease in elevation of the glenohumeral joint. The deltoid muscle was able to reverse this loss of elevation by a force increase of one third of the lost supraspinatus force. If no force was applied to the supraspinatus muscle, average pressures underneath the coracoacromial vault decreased 8 percent. It was concluded that the supraspinatus produces less torque and more glenohumeral joint compression than the deltoid. However, the supraspinatus has no effect on depression of the humeral head during elevation. The clinical consequence of our observations is that operative closure of supraspinatus tendon defects is not mandatory.