In extensive rotator cuff tears with rupture of the supra- and infraspinatus muscles, we have obtained good and very good results persisting for 3 years or longer only in 40% of cases treated by resection and debridement, and in 25% of our cases the head of the humerus was found to be no longer centrally located in the joint socket on X-ray examination 10 years after treatment. It was therefore proposed that an active muscle flap taken from the anterior part of the deltoid muscle (part III according to Fick) should be used to span the trophic defect in the rotator cuff, being sutured into healthy tissue following freshening up. This seems logical, since the transferred muscle flap remains innervated and vascularized and also works synergistically with the rest of the rotator cuff. The flap interposed in this way between the greater tubercle of the humerus and the acromion functions in the same way as a double-bellied muscle and prevents displacement of the head of the humerus. In this way an active rotator cuff is produced, which also takes over the function of stabilizing the humeral head. After follow-up of at least 1 year (average 19 months), the result is aesthetically satisfactory in all 50 shoulders treated in this way; 47 are free of pain or painful only some of the time; in 32 active elevation of over 120-degrees is possible; 21 have symmetrical force ratios in elevation and 15 for outward rotation in 90-degrees abduction. All reconstructed shoulders were characterized by fatigability in elevation. A clinically satisfactory result has been obtained in 39 patients (78%), and 17 humeral heads were recentered by the operation. An EMG 6 months after the operation showed that the muscular deltoid flap contracts in synchrony with the supra- and infraspinatus muscles. Arthrography confirmed the continuity of the rotator cuff. After at least 1 year, 41 of these shoulders were examined by ultrasound, which showed an intact rotator cuff in 35. There were 14 very good and 10 good results in this group. Anterior discontinuity with thinning of the cuff toward the back was found in 6 shoulders. These accounted for 5 poor results and confirms the connection between anatomy and clinical condition. This muscle transplant is only possible in the presence of an accessible cuff remnant and providing there is no substantial amyotrophy in the deltoid muscle resulting from inactivity or from paralysis of the axillary nerve. It would be possible, however, to carry out an arthroplasty with a no constrained prosthesis at the same time, even in the presence of stage II arthrosis of the shoulder.