Concanavalin A (Con A) administered in vivo inhibited the development of insulin-dependent diabetes mellitus (IDDM) in nonobese diabetic mice. By experimental day 310 only one of 10 Con A-treated animals had developed IDDM compared to six of 11 cohort controls receiving buffered saline and 65% of colony controls. Treatment consisted of an initial 70-day course of Con A at a dose of 10-mu-g/g b. wt., i.p. in sterile saline every other day. After two deaths occurred in the Con A group, treatment was stopped at day 70. The remaining animals were given two 1-week booster doses beginning on experimental days 118 and 189. The single animal in the Con A group that eventually developed IDDM at day 293 had not received an injection for 107 days. Histological examination revealed peri-islet lymphocytosis but no islet infiltration in the Con A group. Surviving control animals showed both peri-islet and islet infiltrates. One-week Con A treatment significantly suppressed in vitro responses of spleen cells to Con A and allogeneic lymphocytes. The treatment increased the frequency of "blast-sized" cells in vivo and decreased the CD4+/CD8+ ratio among resting splenocytes. It is concluded that polyclonal T cell activation by Con A provides protection against autoimmune diabetes in nonobese diabetic mice concomitant with phenotypic and morphologic lymphocyte changes.