BACKGROUND The goal of this study was to compare immediate with delayed range of motion (ROM) following total shoulder arthroplasty (TSA). The hypothesis was that ROM gains would occur earlier with immediate motion but that there would be no difference in ultimate ROM or functional outcome. METHODS Sixty patients were randomized to immediate motion (IM) or delayed motion (DM) following TSA. A lesser tuberosity osteotomy was performed in all cases. ROM and functional outcome were compared at 4 weeks, 8 weeks, 3 months, 6 months, and 1 year postoperatively. RESULTS Compared with preoperative values, in the IM group, forward flexion improved from 106° to 141° at 1 year postoperatively, external rotation improved from 21° to 65°, and internal rotation improved by 2 spinal levels (P < .05). In the DM group, forward flexion improved from 104° to 144°, external rotation improved from 20° to 53°, and internal rotation improved by 4 spinal levels (P < .05). The 2 groups regained motion differently, but there were no significant differences in final ROM or functional outcome scores between the 2 groups. The IM group had higher functional outcome scores initially, but by 3 months postoperatively, there was no difference. The rate of osteotomy healing was 81% in the IM group compared with 96% in the DM group (P = .101). CONCLUSIONS Immediate ROM provides a more rapid return of function compared with a delayed ROM protocol following TSA. However, there are no differences in ultimate ROM or functional outcome between the 2 groups. Moreover, immediate ROM may lower the healing rate of a lesser tuberosity osteotomy.