We evaluated the results in eighty-three patients (ninety-five knees) who had had a high tibial osteotomy for either unicompartmental osteoarthritis or osteonecrosis. The operations were performed between 1965 and 1976. The mean length of follow-up was 8.9 years (range, five to fifteen years). The early results were promising: at two years 97 per cent and at five years 85 per cent of the knees had either an excellent or a good result. At subsequent follow-up, however, only sixty knees (63 per cent) had an excellent or good result, and in the remainder recurrent pain had developed. Twenty-two knees (23 per cent) had been revised to a total knee arthroplasty because of pain. The alignment obtained by the osteotomy was not as important in determining the long-term result as we had previously believed. Although recurrent varus deformity was observed in more than one-quarter of the knees, it was not necessarily associated with an unsatisfactory result. The passage of time was the most important factor in determining the result, as only fifteen (37 per cent) of the knees that had been followed for more than nine years were pain-free. We now believe that total knee arthroplasty is a more suitable operation for patients who are more than sixty years old and that high tibial osteotomy should be reserved for patients who have a strenuous occupation or who wish to continue to participate in sports activities.