The long-term results were reviewed for seventy-two patients (seventy-five knees) who had had a bone-patellar ligament-bone intra-articular reconstruction of the anterior cruciate ligament between August 1984 and May 1992. The mean age of the patients at the time of the operation was forty-five years (range, forty to sixty years). Three patients had a bilateral procedure. The primary mechanisms of injury were accidents that occurred during skiing (thirty-two knees), tennis (fourteen knees), and soccer (five knees).We analyzed the responses to subjective questionnaires, the functional results, and the objective clinical data. The clinical examination included assessment of the range of motion, performance of Lachman and pivot-shift tests, and measurements with use of a KT-1000 arthrometer. All knees were evaluated with use of three common rating scales: that of Lysholm and Gillquist; that of The Hospital for Special Surgery, as modified by Insall et al.; and the International Knee Ligament Standard Evaluation Form.At the latest follow-up evaluation, at a mean of fifty-five months (range, twenty-six to 117 months), three patients reported pain or swelling. No patient reported giving-way or symptoms related to the patellofemoral joint. The mean range of extension was -12 to 6 degrees, compared with -8 to 42 degrees preoperatively, and the mean range of flexion was 112 to 150 degrees, compared with 52 to 154 degrees preoperatively. Flexion was limited to 112 degrees in one patient, but this was 5 degrees greater than that of the uninvolved knee. Sixty knees (80 per cent) had a negative pivot-shift test, and ten knees (13 per cent) had a grade of 1+. On testing with the KT-1000 device at maximum manual pressure, the mean difference between the injured and uninjured knees was found to have improved by 5.1 millimeters, from 6.4 millimeters preoperatively to 1.4 millimeters postoperatively (p < 0.01).The grade on the International Knee Ligament Standard Evaluation Form improved markedly; seventy-two knees (96 per cent) had a grade of C or D preoperatively, whereas seventy knees (93 per cent) had a grade of A or B postoperatively. The Hospital for Special Surgery score improved from a mean of 69 points preoperatively to a mean of 92 points postoperatively (p < 0.01). The mean score according to the scale of Lysholm and Gillquist increased from a mean of 63 points preoperatively to a mean of 94 points postoperatively (p < 0.01).All patients indicated that they were pleased with the result of the procedure. Bicycling was resumed at a mean of four months; jogging, at a mean of nine months; skiing, at a mean of ten months; and tennis, at a mean of twelve months.