A cadaver study was performed to determine the effect of arm position and capsular release on rotator cuff repair. Artificial defects were made in the rotator cuff to include only the supraspinatus (small) or both supraspinatus and infraspinatus (large). The defects were repaired in a standard manner with the shoulder abducted 30 degrees at the glenohumeral joint. Strain gauges were placed on the lateral cortex of the greater tuberosity and measurements were recorded in 36 different combinations of abduction, flexion/extension, and medial/lateral rotation. Readings were obtained before and after capsular release. With small tears, tension in the repair increased significantly with movement from 30 degrees to 15 degrees of abduction (p < 0.01) but was minimally affected by changes in flexion or rotation. Capsular release significantly reduced the force (p < 0.01) at 0 degree and 15 degrees abduction. For large tears, abduction of 30 degrees or more with lateral rotation and extension consistently produced the lowest values. Capsular release resulted in 30% less force at 0 degree abduction (p < 0.05).