One hundred and three sequential Stanmore knee replacements were reviewed retrospectively on two occasions with a maximal follow-up period of nine years three months. This knee prosthesis, which is hinged, was successful in alleviating pain, stabilising an unstable knee and modestly increasing the arc of flexion. Walking capacity was increased and flexion contractures were reduced. There were seven cases of infection and four of fracture around the prosthesis. All these proved difficult to treat and two knees with both fracture and infection needed amputation. Eight knees were revised for aseptic loosening and a further 14 were found to have radiological signs of loosening. The results have been analysed by the methods advocated by Tew and Waugh and give a cumulative success rate of 80 per cent at seven years, provided success is judged solely by whether the prosthesis is still in situ. The role of the Stanmore knee as a primary arthroplasty is discussed.