The genetic predisposition to autoimmune disease in man is largely specific for each disease, indicating that these diseases are based not on a generalized breakdown of a tolerance mechanism, but on highly specific abnormalities of immune responsiveness which are subject to genetic transmission. In the presence of particular antigens encoded in the major histocompatibility complex (MHC) the relative risk of certain autoimmune diseases is increased, or in some cases decreased, and the increased risk has been widely attributed to linkage disequilibrium with unidentified disease-causing genes. Our studies on the inheritance of spontaneous autoimmune diseases in New Zealand mice have revealed that small numbers of dominant genes, some associated with the MHC but the remainder elsewhere in the genome, determine susceptibility to these diseases. That MHC-linked genes are not of paramount importance casts doubt on the linkage disequilibrium hypothesis. An alternative possibility is that major and minor histocompatibility antigens themselves predispose to the development of autoimmune diseases by altering the immune response repertoire through the effect of their clonal deletions on the network of paratope-idiotope clonal interactions. Such alterations could influence the chances of emergence, by somatic mutation, of pathogenic forbidden clones.