Sixty-six patients who had a total of eighty-six double-stem silicone implants in the first metatarsophalangeal joint were followed prospectively for an average of 5.8 years (range, two to fifteen years). There were two groups of patients: thirty-four patients (thirty-seven implants) who had degenerative joint disease (including those who had hallux rigidus or in whom a previous operation on a bunion had failed) and thirty-two patients (forty-nine implants) who had rheumatoid arthritis. The implants were used only if the patient was a candidate for an excisional arthroplasty or an arthrodesis; they were not used in patients who wished to maintain or adopt very active use of the foot (such as in running, jogging, and tennis) or to wear very high heels. Twenty-eight (82 per cent) of the thirty-four patients in the first group were completely satisfied and three (9 per cent) were somewhat satisfied. However, three patients (9 per cent), all of whom had had a failed bunionectomy, were dissatisfied; the ages of these three patients were less than the average age of all patients in the first group. Radiographs showed a fracture in three implants, but the patients had a good clinical result and an additional operation was not warranted. Twenty-seven (84 per cent) of the thirty-two patients in the second group were completely satisfied, four (13 per cent) were somewhat satisfied, and one (3 per cent) was dissatisfied. Radiographs showed a fracture in five implants. Four of the implants caused no symptoms, and the result was good; the fifth one was fragmented and was removed because of symptoms. Radiographs showed radiolucent areas around the implant and hypertrophic changes in many patients. There was no evidence of synovitis, such as that caused by silicone, either clinically or radiographically. We found that the double-stem silicone implant was effective in reconstructing the first metatarsophalangeal joint but emphasize our belief that it should be used only in carefully selected patients.