A nonapeptide epitope contained in the 65-kD species of heat shock protein of many mycobacterial species shares homology with a peptide on a mammalian cartilage link protein (1), with four of the amino acids identical. Structures containing this epitope, including proteoglycans extracted from human rheumatoid synovial effusions (2), induce proliferation ofan arthritogenic, rat T lymphocyte line (3) that has been subdivided into two clones, A2b and A2c, which, respectively, induce adjuvant arthritis or suppress the developing and established phases ofthis experimental model of autoimmunity (4, 5). Both clones proliferate in the presence of the epitope, as do T lymphocytes from some rheumatoid arthritis patients (6), but it is not established ifthe epitope plays a role in the pathogenesis of the arthritis. Administration of the recombinant (r) 65-kD shock protein antigen intraperitoneally did not induce arthritis in rats, but rendered such rats resistant to further attempts to induce arthritis with mycobacteria (1) or streptococcal cell walls (7). It is known, however, that adjuvant arthritis development requires the arthritogens to be administered in an oil vehicle by either the intradermal or subcutaneous routes (reviewed in reference 8), and that the peptidoglycan component of the mycobacterial cell wall is essential (8).Using high levels of the r65-kD protein in an oil vehicle, and even combining the protein with low levels of a synthetic, nonimmunogenic adjuvant, CP20961 (9), to mimic the peptidoglycan portion of mycobacteria, we have been unable to induce arthritis with the r65-kD shock protein, though the induction of resistance was confirmed. At high levels, the synthetic adjuvant CP20961, which is known to be arthritogenic (9), induced arthritis in rats rendered resistant to mycobacterial arthritis by the r65-kD shock protein. Type II collagen-induced arthritis also developed in r65-kD protein-pretreated rats. These results strongly suggest that the r65-kD shock protein produces an antigen-specific resistance to classical, mycobacterial adjuvant arthritis, since it is much less effective at inhibiting the arthritis induced by both CP20961 and type II collagen . Furthermore, the pathogenesis ofthe actual adjuvant arthritis does not appear to involve the 65-kD protein.