The discovery of HLA antigen associations with juvenile-type insulin-dependent diabetes mellitus (IDDM) provided strong evidence separating this disorder, or group of disorders, from maturity-type noninsulin-dependent diabetes, as well as adding to the evidence for an immunologic pathogenesis. In addition, it was hoped that the use of these disease-marker associations in appropriate studies might clarify the genetics of IDDM. While these associations have provided a useful tool to further investigate the genetics and pathogenesis of IDDM, the mode or modes of inheritance of this group of disorders remain an area of great controversy. Susceptibility to IDDM is currently being proposed as being inherited as a single autosomal dominant, as a single autosomal recessive, as recessive and some dominant forms, in an intermediate gene dosage model, in a heterogeneous three-allele or two HLA loci model, and as a two-locus disorder. The arguments for each of these proposals is presented, as well as the problems of each. We surmise that the weight of evidence supports the heterogeneity hypothesis but that the modes of inheritance of IDDM will be fully resolved only when we can more reliably identify the diabetogenic genotype, rather than being limited in our investigations to the study of only full-blown clinical disease.