The sera of 131 patients with anti-liver-kidney microsome antibodies (anti-LKM) detected between 1973 and 1979 in two different laboratories were re-examined. (1) Eighty-six anti-LKM corresponded to the description given by Rizzetto, Swana & Doniach (1973), with a pattern of fluorescence predominating on the 3rd portion of the proximal tubules (P3). This group comprised 45 cases of idiopathic chronic hepatitis or idiopathic cirrhosis and one case of halothane-induced hepatitis. (2) Forty-five anti-LKM gave a different pattern on male mouse liver and male rat kidney: (a) fluorescence was greater on centrolobular than on periportal hepatocytes; (b) the first and second portions of proximal tubules (P1 and P2) predominated over P3; (c) P1 fluorescence was equally intense as P2 and (d) P3 cells were heterogeneous with one cell out of 20 more positive than the rest. Absorption tests confirmed that the corresponding antigen was also present in the liver microsomal fraction. A retrospective clinical study discovered tienilic acid-induced hepatitis in all cases. We suggest naming this new antibody 'anti-LKM2'.