1. A series of experiments on the tensile strength of the anterior capsular mechanism have been performed. These show that in the young the weakest point is the glenoid labral attachment, whereas in the elderly calcification of tissues makes the capsule and subscapular tendon weaker.

2. It has been shown previously that glenoid labral detachment is the common injury in the young at the time of an acute dislocation, whereas capsular rupture and subscapularis tendon damage occur in the elderly.

3. These findings suggest that in an acute anterior dislocation of the shoulder the shoulder integuments give at their weakest point, and that it is the site of this weakest point and not the mechanism of injury which influences the liability to recurrence.

1. Arthrography demonstrates two types of injury to the capsule in acute anterior dislocations of the shoulder.

2. The first is a capsular rupture which does not appear to lead to recurrent dislocation of the shoulder unless there is concomitant humeral head damage. In this group healing is complete in ten days and it should be safe to start exercises early.

3. The second is associated with labral detachment from the glenoid and most heal with immobilisation for three weeks. Failure to heal leads to recurrent dislocation. It is not known whether immobilisation had any influence on healing of the lesion in these patients and this remains the subject of further investigation.

4. In recurrent anterior dislocation of the shoulder there is constant enlargement of the subscapular bursa, the outline of which becomes continuous with the inferior pouch. Axial arthrographs show either an absence of the glenoid labral outline or an enlarged entrance to the subscapular bursa.

5. Ruptures of the supraspinatus portion of the tendinous cuff were seen in five patients out of a total of twenty-seven acute dislocations, suggesting that this associated injury is more common than was previously believed.

POSTERIOR DISLOCATIONS

6. When the dislocation is voluntary there is marked elasticity of the capsule but the joint is only unstable in one direction when examined under anaesthesia. Both shoulders appear equally affected when examined radiologically under general anaesthesia even though the patient only has the ability to dislocate one.

7. All patients with voluntary dislocation had a curious voluntary muscle control and were able to contract the anterior and posterior parts of deltoid separately. Each dislocation was preceded by scapular movement.

8. No evidence of increased joint laxity was found in other joints in any of the patients.

9. In two patients with acute dislocations the defect of the humeral head was seen after the initial dislocation and in the third patient it occurred at the time of the second dislocation. In all three there was a spill of fluid beneath the subscapularis but no leakage into the axilla as occurred in anterior dislocation with capsular rupture. The capacity for healing appeared greater than in anterior dislocations with labral detachment; one patient treated in a sling had a better functional result than another treated with the shoulder in lateral rotation.