We used a cadaveric model to examine the mechanical effects of changes in the conformity of the articular surfaces and the size of the humeral head component in glenohumeral arthroplasty. The experimental system permitted a manual clinical examination of the glenohumeral joint while sensors monitored the humeroscapular position and orientation as well as the forces and torques applied by the examiner. Four preparations were compared: an anatomical humeroscapular preparation and three glenohumeral arthroplasty preparations (one with anatomically sized components and a radius of curvature of the glenoid that was four millimeters larger than that of the humeral head, one with anatomically sized components and a radius of curvature of the glenoid that was equal to that of the humeral head, and one with a nonanatomical, large humeral head component and a radius of curvature of the glenoid that was equal to that of the humeral head). All motions, including flexion, external and internal rotation, and maximum elevation, were diminished with use of the non-anatomical, large humeral head component. Laxity of the joint on drawer and sulcus tests was not affected by the conformity of the articular surfaces but was decreased significantly by implantation of the large humeral head component. The kinematics of the glenohumeral joint were not markedly altered by reduction of the uniformity between the articular surfaces of the prosthetic components. In all preparations, obligate displacement of the humeral head associated with a passive range of motion occurred at smaller angles with the large humeral head component. CLINICAL RELEVANCE: Our results indicate the importance of selecting appropriately sized prosthetic components for the space available and point to the variables of shoulder arthroplasty that a surgeon must master to compensate for loss or resection of bone and to complement soft-tissue releases currently used in clinical practice.