In man, there are two major forms of diabetes: insulin-dependent diabetes (IDDM) and non-insulin-dependent diabetes (NIDDM). The pathogenesis and cause of IDDM and NIDDM are thought to differ. Insulin-dependent diabetes generally occurs in a young age group, the onset is more abrupt, and a seasonal incidence has been reported. In some patients with IDDM, inflammatroy cells have been observed in the islets of Langerhans, expecially in the course of the disease. Moreover, the risk of IDDM is greater in persons with certain HLA types, and antibody against islet cells has been found in the circulation of patients with newly diagnosed IDDM. The hypothesis that viruses are one cause of IDDM is supported by its abrupt onset, its seasonal incidence, the presence of inflammatory cells in the islet of Langerahans and the destruction of beta cells. In addition, numerous case reports have shown a temporal relationship between the onset of certain viral infections (e.g., mumps, rubella and Coxsackievirus B group) and the subsequent development of diabetes. The best evidence that a virus can produce diabetes in animals comes from studies with EMC virus. Infection of mice with the M-variant of EMC virus produces a diabetes-like syndrome that in many respects resembles IDDM. However, the development of diabetes after infection with the M-variant of EMC virus was not consistent; sometimes only about 10 to 30% of mice developed diabetes, while at other times up to 60% developed diabetes. Statistically significant differences were consistently found upon repetition of the same experiment and between cages within experimental groups. Moreover, when the virus was passaged in mouse embryo fibroblant cell cultures, the diabetogenic activity of the virus markedly diminished in contrast to when the virus passaged in mice. These finding suggested that the stock pool of EMC virus was made up of two populations of virus; one that had a tropism for insulin-producing beta cells and was diabetogenic and the other that did not have a tropism for beta cells and was nondiabetogenic. Growth of virus in fibroblast cell cultures may have Favored replication of the nondiabetogenic variant. In animals, both beta-tropic and non-beta-tropic variants would have had an opportunity to replicate. plaque purification of the M-variant of EMC virus was resulted in the isolation of two stable variants; one diabetogenic and designated D variant and the other nondiabetogenic and designated B variant. When the D variant was inoculated into SJL/J male mice, hypoinsulinemia and hyperglycemia developed 100? of the animals. In contrast, none of the mice inoculated with the B variant developed diabetes. Histologic examination of pancreata from mice infected with the D variant revealed insuitis and necrosis of beta cells, whereas islets from mice infected with the B variant showed little, if any, change. Moreover, 60% of the lslet cells from mice infected with the D variant contained viral antigens when stained with fluorescein-labelled anti-EMC virus antibody, whereas less than 5% of islet cells from animals infected with the B variant contained viral antigens. The bouyant density, size and shape of both purified variants were identical. There were no differences in the pattern of capsid polypetides between the two variants. Molecular hybridization studies with a (³H)-c-DNA probe of the D nariant failed to distinguish the D and B variants. The oligonucleotide fingerprints of the viral genomes show that one spot of oligonuucleotide of about 25 base length is missing from the RNA of B variant, but present in the RNA of D variant. Antigenically, the D variant and B variant could not be distinguished by a sensitive plaque neutralization assay. Antibody made against the D variant neutralized both the D and B variants. Conversely, antibody made against the B variant neuralized equally well the B and D variants. Competetion radioimmunoassays also failed to reveal any major differences bet