Clinical experience with the Copeland cementless surface replacement arthroplosty (CSRA) of the shoulder now spans 17 years. Between 1986 and 1997, 79 CSRAs (42 total shoulder replacements and 37 hemi-arthroplasties) were performed for primary osteoarthri-tis of the shoulder. Total shoulder replacement was done in 12 men and 30 women with a mean age of 71.5 years (range, 50-87 years). Hemiarthroplosty was used in 9 men and 28 women; 5 patients had bilateral hemiarthroplosty. The mean age was 73.4 years (range, 53-88 years). Thirty-nine total shoulder arthroplasties and thirty hemiarthroplasties with a follow-up of more than 2 years were available for review. The mean follow-up was 7.6 years (range, 48 months to 13 years) for total shoulder replacement and 4.4 years (range, 24 months to 6.5 years) for hemior-throplasty. The Constant scores improved from an age-adjusted Constant score of 33.8% (20.0 points) to 94% (61.9 points) for total shoulder replacement an from on age-adjusted Constant score 40.0% (25.3 points) to 91% (58.1 points) for hemiarthroplasty. Active elevation improved by a mean of 59.9degrees to a mean of 1280 for total shoulder replacement and to a mean of 124degrees for hemiarthroplasty. Of the patients, 89.9% considered the shoulder to be much better or better as a result of the operation. Radiographically, one humeral implant and three glenoid implants had evidence of loosening. Four revisions were performed in the total shoulder replacement group. No revision surgery was needed in the hemiarthroplosty group. The results of this series are at least comparable to those reported for stemmed prostheses with a comparable length of follow-up. The results of total shoulder replacement and hemiarthroplasty in osteoarthri-tis of the shoulder seem to be comparable. With use of the CSRA prosthesis, several severe complications mainly concerning the humeral shaft and periprosthetic fractures can be avoided. Should the need for revision surgery or arthrodesis arise, these procedures are eas-ily performed, as bone stock has been maintained and no loss of length has been encountered. It does seem that the humeral component does not need a stem or cement for fixation.