Our understanding of the role of HLA genes associated with insulin-dependentdiabetes mellitus (IDDM) is in disarray, despite recent improvements in the definition ofspecific alleles and haplotypes. Some genes are highly associated with IDDM, other genes are associated with resistance to IDDM, and some highly associated susceptibility genes are markedly influenced by trans-associated synergistic effects (DR3/4 heterozygotes) or protective effects (DR2/4 heterozygotes). This plethora of genetic associations has spawned the notion that there are many contributing susceptibility genes, which, in turn, has led to the search for shared structural features among different genes on IDDM-associated haplotypes. From a more mechanistic point of view, however, the wide range of variable IDDM associations, with both cis- and trans-encoded protective and/or synergistic effects, suggests a different approach. This article proposes a hypothesis in which the different HLA associations with IDDM can be simply explained by a single unifying concept: a hierarchy of affinities determines the interaction between a diabetogenic peptide and different class II molecules, and an individual is susceptible to IDDM if the class II molecule in that individual with the highest affinity for such a peptide is a DQβ susceptibility gene. The explicit formulation of this proposal and its genetic implications provide an explanation for HLA-encoded dominant “protection” and for some of the more subtle genetic observations related to cis and trans influences in IDDM susceptibility.