The articular surface of the normal humeral head has a variable posterior and medial offset with respect to the central axis of the humeral shaft. Recreation of the normal humeral head shaft offset is postulated to be an important consideration during shoulder arthroplasty. However, the effect of humeral head malposition is unknown. The purpose of this study was to determine the effect of articular malposition after total shoulder arthroplasty on glenohumeral translation, range of motion, and subacromial impingement. Twenty-one human cadavers were dissected and tested with the use of an active or passive shoulder model. Range of motion and translation were recorded by means of an electromagnetic tracking device. The experiment was performed in 2 phases. For kinematics study, 11 cadaver shoulders were positioned both passively and actively from maximum internal rotation to maximum external rotation at 90° of total elevation in the scapular plane. Three rotator cuff and 3 deltoid muscle lines of action were simulated for active joint positioning. Passive joint positioning was accomplished with the use of a torque wrench and a nominal centering force. The testing protocol was used for the natural joint as well as for 9 prosthetic head locations: centered and 2- and 4-mm offsets in the anterior, posterior, inferior, and superior directions. Repeated-measures analysis of variance was used to test for significant differences in the range of motion and translation between active and passive positioning of the natural joint as well as all prosthetic head positions. (2) For impingement study, 10 cadaver shoulders were used in a passive model, loading the tendons of the rotator cuff with a 30-N centering force. The humerus was passively rotated from maximum internal rotation (1500 Nmm) to maximum external rotation (1500 Nmm) by means of a continuous-recording digital torque wrench. Trials were performed with the use of centered, 4-, 6-, and 8-mm offset heads in the anterior, posterior, superior, and inferior positions before and after removal of the acromion and coracoacromial ligament. The relation between change in mean peak torque (with and without acromion), passive range of motion, and humeral head offset was analyzed by means of repeated-measures analysis of variance. In the kinematics study, total range of motion and all humeral translations were greater with passive joint positioning than with active positioning (P =.01) except for total superior-inferior translation and superior-inferior translation in external rotation. Anterior to posterior humeral head offset was associated with statistically significant changes in total range of motion (P =.02), range of internal rotation (P =.02), range of external rotation (P =.0001), and total anterior-posterior translation (P =.01). Superior to inferior humeral head offset resulted in statistically significant changes in total range of motion (P =.02), range of internal rotation (P =.0001), anterior-posterior translation during external rotation (P =.01), and total superior-inferior translation (P =.03). In the impingement study, there was a significant increase in torque from centered to 4-mm inferior offset (P =.006), 6-mm inferior offset (P <.001), and 8-mm inferior offset (P <.001). There was no significant increase in torque with superior, anterior, and posterior offsets. Glenohumeral motion significantly decreased from 129° for centered head to 119° for 8-mm superior (P =.002), 119° for 8-mm anterior (P =.014), 118° for 8-mm inferior (P <.001), and 114° for 8-mm posterior (P =.001). Humeral articular malposition of 4 mm or less during prosthetic arthroplasty of the glenohumeral joint may lead to small alterations in humeral translations and range of motion. Inferior malposition of greater than 4 mm can lead to increased subacromial contact; offset of 8 mm in any direction results in significant decreases in passive range of motion. Therefore if subacromial contact is to be minimized and glenohumeral motion maximized after shoulder replacement, anatomic reconstruction of the humeral head-humeral shaft offset to within 4 mm is desirable. (J Shoulder Elbow Surg 2001;10:399-409.)