Abstract: Methods for the evaluation of the four antimitochondrial antibody subtypes in primary biliary cirrhosis - anti-M2, -M4, -M8, -M9 - are described. The importance of the application of different preparations for the demonstration of complement fixing antibodies and the detection of antibodies by ELISA or Western blotting is emphasized. Complement fixing antigens can be prepared by discontinuous isopynic sucrose density gradient centrifugation using mitochondrial subfractions derived either from beef heart (M2), rat liver (M4), or pig kidney (M8). Anti-M9 antibodies do not fix complement. For ELISA, the pyruvate dehydrogenase or the AT-Pase-associated antigen fraction (M2), the sulfite oxidase fraction (M4), and the chromatographically purified M8-fraction should be used. The same antigen fractions are suitable for Western blotting, but anti-M4 and anti-M8 cannot be visualized by this method. For the demonstration of anti-M9 by ELISA and Western blotting a purified fraction prepared from rat liver has to be applied. Correlating antimitochondrial antibody-subtypes with clinical condition and the natural course, there is convincing evidence that especially the presence of complement fixing antibodies against the subtypes M2, M4, and M8 is a reliable indicator for a more active course. Patients expressing only anti-M9 (without anti-M2) have biochemically all the typical features also found in classical anti-M2 positive primary biliary cirrhosis patients, but seem not to advance to late stages. Since these antimitochondrial antibody-subtypes are present even in very early stages without changing their pattern during the course, antimitochondrial antibody-profiles can also be taken as early prognostic parameters. The evaluation of the immunological activity by antimitochondrial antibody-subtype testing may further facilitate the decision whether therapy with ursodeoxycholic acid should be combined with steroids and/or immunosuppressive agents. The role of mitochondrial autoantigens in the induction of this chronic destructive bile duct process is also discussed. The concept is put forward that not bile ducts but naive(?) B-cells expose the different mitochondrial antigens, thereby stimulating autoreactive T-cells to provide a second signal for antibody production. The degree of breakage of tolerance to the different mitochondrial epitopes may be one crucial factor which determines the diversity of antimitochondrial antibody-subtypes in patients with primary biliary cirrhosis.