Genetic, endocrine and immunological factors are probably involved in adjuvant polyarthritis. The nature of the vehicle and of the mycobacterial components administered also has a major influence. It was originally assumed that arthritogenicity and adjuvanticity of mycobacterial fractions such as wax D were intimately related. Our previous findings showed that the water soluble adjuvant (WSA) of M.smegmatis which could substitute for mycobacterial cells in Freund's complete adjuvant and induce delayed hypersensitivity was not arthritogenic in the Wistar rat. We have since observed that auto-immune diseases could be elicited by WSA. Therefore experiments were repeated using the very susceptible Lewis strain. The activity of cord factor and of various mycobacterial preparations suspended in mineral or in peanut oil was also evaluated in mice and in normal or hypophysectomized rats. Our present findings confirm the absence of arthritogenicity of WSA in the Lewis strain. They also indicate that cord factor with WSA does not suffice to induce a generalized adjuvant disease, but that a mycobacterial component which could be susceptible to lysozyme treatment is required also. However, the local inflammation of the injected limb was produced by a preparation of cord factor administered in mineral or even in peanut oil. This was observed in normal or hypophysectomized rats and in Swiss mice which were not susceptible to the generalized disease.