We can now predict the development of type 1A diabetes in humans and prevent the disorder in animal models, but we cannot at present safely prevent type 1A diabetes in humans, although a series of clinical trials are under way and planned. A major lack in our current trial design is the inability to measure T lymphocytes directly responsible for beta cell destruction. Given the immunogenetics of type 1A diabetes and increasing knowledge of pathogenesis in the NOD mouse, we believe the disorder results from immune reactivity to a limited set of islet peptides, with reactivity to insulin a major determinant of disease. Insulin autoantibodies precede the development of diabetes in both humans and the NOD mouse. T lymphocytes isolated from the islets of the NOD mouse that recognize insulin peptide B:9-23 can transfer diabetes. Insulin expression within the thymus is correlated with genetic susceptibility, and insulin peptides can be used to induce diabetes and as an immunologic vaccine to prevent the disorder. Nevertheless, at present, routine measurement of anti-insulin T lymphocytes is not standardized. Better assays to monitor such autoreactivity are likely to be essential for the development and evaluation of preventive therapies.