The results of the first 333 Charnley total hip arthroplasties that were performed with cement at the Mayo Clinic were reviewed a minimum of fifteen years postoperatively. Data were available for 166 of 170 hips of patients who were still alive. One hundred and thirty patients died, and thirty-seven hips were revised. At the time of this study, 80 per cent of the living patients had no pain, and 152 of the 160 hips remained much better than before the operation. Kaplan-Meier analysis of probable loosening of one or both components, on the basis of roentgenographic evidence, demonstrated a probability of loosening of 3 per cent incidence at one year after operation, 13 per cent at five years, 19 per cent at ten years, and 32 per cent at fifteen years. The probability of failure (that is, revision or symptomatic loosening) was 0.9 per cent at one year, 4.1 per cent at five years, 8.9 per cent at ten years, and 12.7 per cent at fifteen years. We did not identify a dramatic increase in the incidence of loosening or failure at any of the follow-up periods (one, five, ten, or fifteen years). With the Mayo Clinic clinical and roentgenographic system for scoring the hips, we found that ninety-seven hips had a good or excellent result; fifteen, a fair result; and thirteen, a poor result. (The scoring could not be completed for forty-one hips). The functional results deteriorated slightly over time.(ABSTRACT TRUNCATED AT 250 WORDS)