The tension of a repaired rotator cuff was evaluated in nineteen patients who had a repair of a full thickness rotator cuff tear. The tension of the repaired cuff was measured at the operation using a simple spring scale. The tension was evaluated regarding the size of the tear, the duration of the symptom, the presence of trauma, and the post-operative results using a UCLA score. The average of the tension at the arm in 0, 30, and 60 degree elevations were 39.2±18.4N, 23.5±17.2N, and 14.2±13.4N respectively. The average tension of the patient who suffered from a trauma was 20.3±15.8N, whereas the one in the patients who had no history of trauma was 35.0 ±18.0N. The slight positive relation, not statistically significant, was found between the tension and the size of the tear. We could not find a significant relation between the tension and the range of motion or the muscle power in this study. The UCLA score was significantly higher in those patients who had less tension of a repaired rotator cuff. We have to be careful not to put too much tension on the rotator cuff when we repair it. Too much tension might damage the muscles and musclotendious units of the rotator cuff or fail to unite the cuff to the bone, resulting in dysfunction of the rotator cuff postoperatively. Then, how much is "too much"? Only a few papers have described the details of the tension of a repaired cuff. Our results show that the lower the UCLA score in patients with a higher tension of the repaired cuff. These results suggest that the tension of the repaired cuff, indeed, changes the results of a rotator cuff repair.