A CRITICAL review of the work of Barnett, Davies, Charnley and MacConaill1–5 indicates that there exists a tendency to explain the lubricating conditions in synovial joints by either the lubricating properties of synovial fluid, or by the lubricating properties of the articular surfaces. The proponents of these hypotheses applied the theories of hydrodynamic lubrication or boundary lubrication, respectively. Neither of these theories can explain, however, the lubrication in joints; both are much too crude for this purpose. The nearest approach to reality could be given, perhaps, by the recent elasto-hydrodynamic theories6,7, but even in these, when applied to synovial joints, a discrepancy exists between the factual evidence and the theoretical results. In fact, a discrepancy between the experimental and the calculated values of rates of shear, or between the calculated and the experimental values of viscosity of synovial fluid, is of an order of 100,000. That the synovial fluid is indeed required for proper lubrication and functioning of joints may be judged from the effect of a fall in temperature on the increased resistance of joints to movement, and also on a decrease in the exertible external force8. Both phenomena would be due to an increase in the viscosity of synovial fluid and, to some degree, also due to the changes in the elasticity of the cartilage. A decrease of the viscosity of synovial fluid and a change of the rheological characteristics of this fluid from the thixotropic type into a Newtonian type was observed in cases of rheumatoid arthritis9–11; in such cases the joint temperature was higher than anticipated from relative clinical activity8.