Twenty-seven shoulders (in 26 patients) that had surgical repair of a rupture of the long head of the biceps tendon between 1962 and 1981 were evaluated at three to 22 years after injury (mean, 13 years). Thirty similar patients had been treated nonsurgically and were evaluated two to 15 years (mean, 4.6 years) after tendon rupture. Biomechanical testing was performed on ten patients in the surgical group and 13 in the nonsurgical group. Residual arm pain was infrequent in both surgically and nonsurgically treated groups. Residual subjective weakness at the elbow was reported in four of the surgical group and in 20 of the nonsurgical group. Surgically treated patients returned to work later than nonsurgical patients; however, 11 in the nonsurgical group were not able to return to full work capacity, versus only two in the surgical group. On biomechanical testing, the nonsurgical group had lost a mean of 21% of supination strength and 8% of elbow flexion strength but had no weakness in grip, pronation, or elbow extension. The surgical group had lost no strength in any of these testing modes.