The present knowledge of the HLA system and its biological function is summarized as a basis for the subsequent discussion of the associations between this system and insulin-dependent diabetes (IDDM) and some mechanisms that may explain them. Although the serologically detectable DR determinants are still the most handy markers, there is now increasing evidence from studies of restriction enzyme fragment length polymorphism (RFLP) in IDDM that DQ determinants may play a primary role in causing susceptibility and/or resistance to this disease. Thus, it is now evident that about 90% of DR4-positive diabetics carry the DQw8 determinant present in only about 65% of DR4-positive controls. Most recently, it has been claimed that an aspartic acid in position 57 of the DQB1 (DQ-beta-1) chain confers resistance to IDDM. Although this may be true, it does not explain the disproportionate decrease of DR2 or the particularly high risk of DR3/4 heterozygotes, which is still good evidence that several HLA genes are involved. Because Class II antigens show the strongest associations, the most plausible hypothesis about the mechanism(s) involves specific presentation of as yet unknown antigenic peptides to T-helper lymphocytes, which may induced the formation of both anti-islet cell antibodies and T-cytotoxic lymphocytes capable of destroying beta cells. However, T-suppressor lymphocytes also may be involved. If this hypothesis is correct, the most urgent task is to define the antigenic peptides in question, whether they are environmental (e.g., viral) or autologous.