An epidemiological study and the simultaneous evaluation of anti-Hav antibody using radio immunoassay (RIA) and immunoadherence hemagglutination assay (IAHA) was performed during three hepatitis A epidemics in the Tours area (France). Fifty-seven sera from 35 subjects with viral hepatitis type A and 16 sera from nine children who did not develop any clinical signs of hepatitis were studied. The more explosive epidemic occurred in an institution for mentally retarded children (attack rate 68%). The two major outbreaks observed were due to the introduction in the institutions of individuals infected with hepatitis A virus. Two out of three of the index cases had a seafood dinner three to four weeks before onset of jaundice. Sera taken one week after jaundice were always found to be anti-HAV positive by both RIA and IAHA, and sera taken more than three days before the appearance of jaundice were negative by both methods. Sera taken at the peak of the transaminase elevation were anti-HAV positive by RIA but only one out of two were positive by Iaha. The anti-HAV titre by RIA increased from the time of the appearance of jaundice and the highest titres, over 1/20000 were seen only after several months. Observations of subjects in close contact with patients who seroconverted without any manifestation of hepatitis, confirmed the existence of clinically mute infections.