Anterior shoulder instability is increasing in incidence because the general population is becoming more and more active. Many factors influencing the occurrence of a chronic problem have been identified and many surgical solutions proposed. Nevertheless, the recurrence rate after treatment remains relatively high. We believe that experimental investigations leading to improved anatomic and biomechanical knowledge are an important pathway for developing better treatment modalities.In the first part of this thesis, we tried to develop and validate an experimental model for anterior shoulder instability. In cadaver shoulders with an intact muscular envelope, a series of sequential cutting experiments was done. Lesions of the capsuloligamentous complex were created at different locations of the joint and repeated in the presence of rotator cuff lesions. The consequences for stability were evaluated with five tests. An almost linear relationship between the extent of the created damage and the ensuing degree of instability was found. The model, however, was confronted with a few deviations from this covariability. Because we had the feeling that these could be explained by anatomic factors and because an exhaustive literature review failed to explain the aberrations, we initiated the second part of the thesis. Here, a series of anatomic and anatomo-clinical studies led to an expansion of the present anatomic knowledge. In the course of these dissections, a new ligament - the posterosuperior glenohumeral ligament - was discovered as well as an expanded role for the tendon of the latissimus dorsi.