Two hundred hemiarthroplasties were studied to compare uncemented Austin Moore and cemented Thompsons prostheses. The following variables were compared: preoperative mobility, ASA grade, Abbreviated Mental test scores, surgeon grade, peri-operative mortality, infection rate, operation time, intra-operative fractures, dislocation rate, thigh pain and revision rate to total hip replacement. Both groups were comparable in all respects. The incidence of intra-operative fractures, infection, postoperative dislocation and conversion to Girdlestone arthroplasty was higher in the uncemented group. The mortality remained the same in both groups. The incidence of significant thigh pain was higher in the uncemented group (13%), of whom half (6%) had to undergo a conversion to a total hip replacement within the 24 months follow-up period. Our current study of changed practice concluded that cemented Thompsons prosthesis performs better than uncemented Austin Moore when the incidence of hip pain and subsequent conversion to a total hip replacement were considered. (Hip International 2004; 14: 189-92).