In the last few years the improvement of our knowledge of the pathogenesis of type I diabetes mellitus has been possible mainly because of the development of studies of the role played by genetic factors and the better definition of lesion mechanisms. The availability of experimental models such as "non obese diabetic (NOD)" mice and "Bio-Breeding (BB)" rats, which develop a form of type I diabetes similar to that of humans, has provided many data regarding the relevance of T lymphocytes in the pathogenesis of the disease, and made it possible to anticipate the immunopathological steps leading to pre-insulitis, insulitis and, finally, to pancreatic beta-cell destruction. The analysis of sera of patients with type I diabetes has demonstrated the presence, also in the preclinical states, of several autoantibodies directed against specific autoantigenic structures of beta cells, which have come to be useful for early diagnosis and in the monitoring of the disease. However, hard evidence for a relevant role of these autoantibodies in the pathogenesis of the disease does not exist. At present, we can affirm that the expansion of autoreactive T lymphocytes specific for membrane antigens of beta cells is the most relevant immunological event for the induction and sustenance of insular lesions. Autoreactive T lymphocytes may be able to activate several effector systems of lesions through the production of multiple combinations of cytokines. The aetiology of the disease is certainly multifactorial and involves both genetic and environmental factors. With respect to the former, the most accurate studies have been performed on the HLA system. It has been clearly shown that type I diabetic patients more frequently display HLA DR3 and DR4 specificities. The results obtained by the application of molecular techniques have suggested considering as risk factors the occurrence of DQB1 0302 DQB1 0201 alleles, the presence of a neutral amino acid residue in position 57 of the DQ beta chain instead of aspartic acid, as well as an arginine residue in position 52 of the DQ alpha chain. With regard to acquired aetiological factors, the hypothesis that primary lesions of pancreatic islets could be due to some viral infections capable of triggering, through several undefined mechanisms, persistent and self-sustaining autoimmune reactions, has gained some credit.