A matched comparison was made between total hip arthroplasties done for osteonecrosis and those done for osteoarthritis. Patients were matched for age, sex, surgical approach, prosthesis, and surgeon. All received cemented Charnley replacements and had minimal follow-up of 10 years. There were 118 women and 70 men with a mean age of 58 years. Thirty-five patients were younger than 50. The mean follow-up was 17.8 years (range, 10 to 25.4 years). Revision rates for osteonecrosis and osteoarthritis were 18% and 19% (not significant). The mechanical failure rate (revision for loosening plus radiographic loosening) was greater, but not significantly so, in the osteonecrosis group. In patients younger than 50 years, the revision rate in osteonecrosis (50%) was significantly higher than that in any other group. Similarly, the mechanical failure rate in osteonecrosis was significantly greater in the patients younger than 50. Radiographic femoral loosening was greater in osteonecrosis at all follow-up intervals, and dislocations occurred more frequently in the osteonecrosis group than in the osteoarthritis group. Results of arthroplasty in both groups were comparable in patients over 50 years of age, although patients with osteonecrosis had an increased rate of dislocation. Patients with osteonecrosis who are younger than 50 years have a significantly higher rate of mechanical failure than those with osteoarthritis who are younger than 50 years. Cemented total hip arthroplasty should be recommended in this group with caution if at all.