W e have histocompatibility (HLA) genotyped 24 families with two or more juvenile, insulin-dependent, ketosis-prone diabetic siblings. This criterion for family selection was used to obtain a homogeneous form of diabetes within a sibship, because diabetes appears to be a genetically heterogeneous disease. 58 diabetic and 53 nondiabetic sibs and 40 parents were studied. 55% of the diabetic pairs were concordant for both HLA haplotypes (expected 25%), 40% were concordant for one haplotype (expected 50%), and 5% were discordant for both haplotypes (expected 25%). These values are significantly different from the expected values (P < 0.001). On the other hand, the inheritance of haplotypes among the nondiabetic sibs in these families was not significantly different from the expected mendelian segregation.When comparing 20 pairs of HLA identical (sharing two haplotypes) with 15 pairs of haploidential (sharing one haplotype) diabetic sibs for the intrapair difference in age of onset of disease, we found that the HLA identical sibs were significantly more concordant for age of onset (3.9 yr difference) than the haploidential (7.3 yr difference) (P < 0.05). The same type of analysis for the difference in seasonal incidence in months revealed that the HLA indentical sibs were more concordant (1.8 mo difference) than the haploidentical sibs (3.2 mo difference) (P < 0.025). Furthermore, the HLA identical diabetic sibs were more likely to develop diabetes in the winter months (78%) than the haploidentical diabetic sibs (21%).No particular HLA haplotype or antigen seemed to be associated with any particular clinical feature.These data are compatible with the theory of genetic heterogeneity of juvenile, insulin-dependent diabetes. It is suggested that there are one or more diabetes response genes in the HLA region playing an important role in the pathogenesis of juvenile, insulin-dependent diabetes in the families studied here. It is, however, possible that other genes, not associated with the HLA complex, may play an etiologic role in some cases of juvenile, insulin-dependent diabetes, resulting in lack of association between HLA and some forms of diabetes.