BDF1 mice were made tolerant by a single i.v. injection of 1 mg of DNAP-gamma1 or by weekly i.v. injections of 0.2 mg of DNP-gamma1 given for a month. In both instances, spleen cells of tolerant animals were fractionated to obtain pure populations of T cells (nonimmunoglobulin-bearing cells), referred to as tolerant T cells, and B cells (immunoglobulin-bearing cells) referred to as tolerant B cells (immunoglobulin-bearing cells) referred to as tolerant B cells. The control cells were similarly fractionated to obtain normal T and B cells. Mixtures of tolerant T cells and normal B cells, or conversely, normal T cells and tolerant B cells were used to repopulate lethally irradiated recipients. These recipients were then immunized with dinitrophenyl-keyhole limpet haemocyanin and in certain instances with other antigen horse red blood cells. The immune response to both antigens was measured using the direct hemolytic plaque assay. It was found that both T and B cells were tolerant and that tolerance was hapten specific at both T- and B-cell levels. While B-cell tolerance was demonstrated at a 1/1 T/B ratio, a 4/1 T/B ratio was necessary to show T-cell tolerance. Thus, the hapten-specific carrier-determined tolerance involves not only B cells but also T cells. The implication of this finding for the cellular mechanism of tolerance in an experimental model closely related to self tolerance is discussed.