We adapted our mouse model of allergic contact hypersensitivity to nickel for the study of tolerance. Sensitization in this model is achieved by the administration of nickel ions with H2O2; nickel ions alone are unable to prime naive T cells, but can restimulate primed ones. A 4-wk course of oral or i.p. administration of 10 mM NiCl2 to naive mice induced tolerance, preventing the induction of hypersensitivity for at least 20 wk; long term desensitization of nickel-sensitized mice, however, required continuous NiCl2 administration. When splenic T cells of orally tolerized donors, even after a treatment-free interval of 20 wk, were transferred to naive recipients, as with lymph node cells (LNC), they specifically prevented sensitization of the recipients. The LNC of such donors were anergic, because upon in vivo sensitization with NiCl2 in H2O2 and in vitro restimulation with NiCl2, they failed to show the enhanced proliferation and IL-2 production as seen with LNC of mice not tolerized before sensitization. As few as 102 bulk T cells, consisting of both CD4+ and CD8+ cells, were able to specifically transfer tolerance to nickel. A hypothesis is provided to account for this extraordinarily high frequency of nickel-reactive, suppressive T cells; it takes into account that nickel ions fail to act as classical haptens, but form versatile, unstable metal-protein and metal-peptide complexes. Furthermore, a powerful amplification mechanism, such as infectious tolerance, must operate which allows but a few donor T cells to tolerize the recipient.