We reviewed, nine to twelve years postoperatively, the records on an original cohort of 289 arthroplasties (218 patients) in which a posterior stabilized knee prosthesis with an all-polyethylene tibial component had been inserted at The Hospital for Special Surgery. One hundred and eighty intact prostheses in 139 patients were available for this analysis. Fourteen knees in fourteen patients had had a revision procedure. Five of these fourteen patients had had a bilateral arthroplasty, but only one knee of each of the five patients had been revised. Forty-eight of the patients (sixty-six knees) had died less than nine years after the operation. Twenty-nine other knees (twenty-two patients) had been lost to follow-up before a nine-year evaluation could be performed. Considering all 194 knees (including the fourteen that had been revised), the result with the system of The Hospital for Special Surgery was excellent for 117 knees (61 per cent), good for fifty-one (26 per cent), fair for twelve (6 per cent), and poor for the fourteen knees (7 per cent) that had been revised. The 180 knees in which the prosthesis was intact were also rated with the new scoring system of The Knee Society: the average postoperative knee score was 92 points (range, 35 to 100 points), and the average score for function was 66 points (range, 0 to 100 points). Survivorship analysis showed that the average annual rate of failure was 0.4 per cent and that the over-all rate of success at thirteen years was 94 per cent.(ABSTRACT TRUNCATED AT 250 WORDS)