Fifteen patients underwent a shoulder arthrodesis utilizing standard dynamic compression plate fixation, but with limited postoperative immobilization with only an abduction pillow. In each case, the position of the extremity relative to the scapula and trunk was recorded immediately postoperatively, at regular intervals until fusion, and at follow-up evaluations. Thirteen of 15 shoulders fused without change of intraoperative position after an average postoperative period of 4 months. One patient lost position in the early postoperative period secondary to inadequate fixation, but subsequently fused. Another who demonstrated a persistent non-union at 2 1/2 years was subsequently explored and underwent a bone graft. Four patients complained of residual symptomatic hardware, with two requiring surgical removal of the plate and screws. All but one patient were satisfied with the clinical result at follow up. Only two patients were within 5 degrees of the preoperatively determined position of 30 degrees abduction, 30 degrees forward flexion, and 30 degrees internal rotation. However, almost all were able to function satisfactorily. The authors concluded that shoulder arthrodesis utilizing rigid internal fixation without postoperative cast or brace immobilization maximizes patient comfort without compromising the success of arthrodesis. However, control of arm position remains inexact and additional modifications are needed to ensure fusion position and to minimize disability.