A computer-based model of the knee was used to study forces in the cruciate ligaments induced by co-contraction of the extensor and flexor muscles, in the absence of external loads. Ligament forces are required whenever the components of the muscle forces parallel to the tibial plateau do not balance. When the extending effect of quadriceps exactly balances the flexing effect of hamstrings, the horizontal components of the two muscle forces also balance only at the critical flexion angle of 22 degrees. The calculations show that co-contraction of the quadriceps and hamstring muscles loads the anterior cruciate ligament from full extension to 22 degrees of flexion and loads the posterior cruciate at higher flexion angles. In these two regions of flexion, the forward pull of the patellar tendon on the tibia is, respectively, greater than or less than the backward pull of hamstrings. Simultaneous quadriceps and gastrocnemius contraction loads the anterior cruciate over the entire flexion range. Simultaneous contraction of all three muscle groups can unload the cruciate ligaments entirely at flexion angles above 22 degrees. These results may help the design of rational regimes of rehabilitation after ligament injury or repair.

We recovered 23 meniscal bearings from 18 failed bicompartmental Oxford knee prostheses. They had been implanted for one to nine years. The minimum thickness of the retrieved bearings was measured and compared with the thickness of 25 unused bearings. The mean penetration rate, calculated by two methods, was either 0.043 or 0.026 mm per annum. This compares with 0.19 mm per annum reported for the Charnley hip. The use of a fully congruous meniscal bearing prosthesis can reduce wear in knee arthroplasty to a very low rate.

We measured the sagittal laxity in 70 knee replacements at least six months after surgery, using a KT 1000 arthrometer. With an unconstrained prosthesis (the Oxford meniscal knee) anteroposterior stability was normal in joints known to have intact cruciate ligaments. There was increased laxity in those which lacked an anterior cruciate ligament. In knees with an intact anterior cruciate ligament, sagittal laxity did not increase with time.

The Oxford Knee, a resurfacing prosthesis with a meniscal bearing, can be used for either bicompartmental or unicompartmental arthritis. The first 103 unicompartmental cases are presented at a mean time since operation of 36 months (range 21 to 56 months). In those cases with surviving arthroplasties, pain was relieved in 96%. The full range of pre-operative flexion was maintained and flexion deformity was improved from a mean of 6.7 to 5.4 degrees. Stability and alignment were restored to normal in nearly all the knees. Absence of the anterior cruciate ligament was associated with a significantly greater incidence of failure. Six failures occurred in 37 knees lacking a normal anterior cruciate ligament (16.2%); three occurred in 63 knees with a normal anterior cruciate ligament (4.8%) (p less than 0.02). Criteria for the future selection of patients have been deduced from our experience. The operation is recommended for knees with severe unicompartmental osteoarthritis in which all the ligaments are still intact.

We have studied the kinematics of the knee in the sagittal plane, using a four-bar linkage as model, and assuming that a "neutral fibre" in each ligament remains isometric throughout flexion. We devised a computer program to calculate the distance separating any pair of points, one on each bone, for various cruciate attachments at various angles of flexion. The parameters for the linkage in four cadaveric knees were obtained by marking the centre of attachment of the cruciate ligaments with tacks and taking lateral radiographs. The movements of the bones were then calculated, in the computer model, for various attachments of "replacement" ligament fibres, the distance between the attachment sites being plotted against the angle of flexion. It was then possible to define zones around the isometric attachment points within which changes in length would be predictable. Our results show that the position of the femoral sites of attachment of both anterior and posterior cruciate replacement was more critical than that of the tibial attachments.

After a unicompartmental Oxford knee replacement, a clear lateral radiographic view of the components may be obtained without the shadow of prosthetic components in the other compartment. Radiographs of 20 knees were studied; with the patient supine and the muscles relaxed, views with the knee at full extension and 90 degrees of flexion were obtained and the movement of the meniscal bearings over this range of flexion was measured. The bearings were found to move backwards on the tibia through an average distance of 4.4 mm (range 0.0 to 13.5 mm) in the medial compartment and 6.0 mm (range 1.6 to 13.0 mm) in the lateral compartment. These movements were in the same direction as that observed in cadaver specimens but smaller in magnitude. At 90 degrees of flexion, radiographs were obtained with the tibia twisted manually to the limits of medial and lateral rotation. The average movement of the bearings between these extremes was found to be 6.6 mm in the medial compartment and 5.1 mm in the lateral; their movements on the tibia were in opposite directions in the two compartments. Bearing movement was still present in knees examined five years after operation.

Arthrogryposis multiplex congenita is believed to be a specific clinical entity which is aetiologically unrelated to the "arthrogryposis-like" deformities of known neurological diseases such as myelomeningocele and myelodysplasia. The observation that the condition appeared to be three times as common in Melbourne, Australia, as in four centres in the United Kingdom (Wynne-Davies and Lloyd-Roberts 1976), prompted this survey of 132 patients: 73 from the United Kingdom, 34 from Australia and 25 from Wilmington, Delaware, USA. The survey aimed to established the same criteria for diagnosis in the three countries and to search for prenatal and genetic aetiological factors. It was shown that all centers treated more newborn children with this disorder during the 1960s than either before or after that period. All cases were sporadic and there was no family association with talipes equinovarus, congenital dislocation of the hip or hereditary neuromuscular disease. "Environmental" findings from all three centers were similar and it was concluded that arthrogryposis multiplex cogenita is a non-genetic disease of early pregnancy, associated with a variety of unfavourable intra-uterine factors. In addition, an unknown but possibly viral environmental agent may have been present to a significant extent only during recent decades and is now declining.

The mechanisms controlling and limiting movement and serving to transmit load between the femur and the tibia are discussed. Having accounted for the transmission of all components of force and couple across the joint and noted the load-bearing role of the menisci, some principles which might guide the design of knee prostheses are deduced. It is shown that current designs transgress some of these principles. An experimental prosthesis is then described, which incorporates analogues of the natural menisci. The possible practical application of this novel principle has been studied in cadaveric human joints and in living patients.

1. A predictable pattern of degeneration occurs on both the femoral head and the acetabulum and this pattern is age dependent.

2. The degenerative areas on the femoral head are related to habitual non-use.

3. The hip is shown to be anatomically incongruent, and the dome of the acetabulum, a predictable area of degeneration, is shown also to be an area of habitual non-use.

4. The possible relationships between age-dependent degenerative changes and senile degenerative joint disease is discussed and the importance of changing geometry stressed.