The autoimmune nature of insulin-dependent diabetes mellitus (type 1 diabetes) has been definitively established during the past ten years only, owing essentially to the development of the NOD and the BB rat models. Three of the four criterias required for defining an autoimmune disease have been demonstrated in these animal models. IDDM is accompanied by immunological stigmatas including circulating autoantibodies and insulitis (lymphocytic infiltration of the islets of Langerhans), it is attenuated or prevented by immunosuppressors and it is transferable from diabetic to non-diabetic mice or rats, via T lymphocytes. Only the fourth criterium, namely the induction of the disease by immunization with an autoantigen, has so far not been met. As is the case for many, if not most, autoimmune diseases, the pathogenesis of IDDM is, however, far from being totally understood. Many aspects including the circumstances favoring escape from self-tolerance, the role of the genetic background, the nature of the pancreatic antigens involved in the initiation and perpetuation of the disease, the effector mechanisms responsible for the elimination of the insulin cells and, most importantly, the conditions for restoring tolerance are at the forefront of immunopathologists' concerns. We provide in this review an account of the present situation in these different areas of research. There is no doubt that a cure or a prevention of the disease will be available in the forseeable future. Experiments on animal models have already initiated several clinical trials and epidemiological studies, and this is probably only the beginning of a long list.