In 20% of patients exposed to hepatitis C virus, infection is transient but, after a few months, the patient remains susceptible to infection with the same strain. Protective immunity is short-lived. This suggests that recovery is related to the cellular immune response, which presumably lyses infected cells, and that the need during recovery for a virus-neutralizing anti-envelope response, is transient. In 80% of patients the infection is persistent, and it seems that antigenic variation of the envelope proteins allows the virus to escape neutralization by anti-envelope responses. The fact that this antigenic variation occurs at a much lower rate in agammaglobulinaemic subjects suggests that the major immune pressure producing this variation is humoral. How the virus-infected cells avoid lysis by cytotoxic T cells, which can be demonstrated in small numbers in the infected liver, remains unclear. The recent observation, that HCV infects CD8 lymphocytes, raises the possibility that virus infection of CD8 cells may impair their function and contribute to persistent infection. The mechanisms of production of cryoglobulin and of autoantibody formation are both unclear.