A unique experimental model is described, where natural immunologic tolerance to a well-defined soluble native antigen (murine C5) is examined in congenic strains of mice that differ only by the presence or the absence of C5. A highly sensitive hemolytic assay was developed to detect nanogram amounts of C5 as well as an assay of anti-C5 inhibition of C5 hemolytic activity. The latter was more sensitive than immunodiffusion. Two reciprocal approaches were used to study the cellular basis of tolerance in irradiated hosts of either strain. In the first, lymphoid cells from either strain were transferred to irradiated B10.D2OSN hosts that were lacking C5 and so would not hinder detection of anti-C5 antibody upon challenge with murine C5. Second, lymphoid cells from either strain were transferred to irradiated B10.D2NSN hosts, whose native C5 provided the antigenic stimulus. The immune response of whole nonadherent spleen cell suspension as well as mixtures of T and B cells (separated on the basis of surface immunoglobulin) from either strain were studied. In addition, the duration of tolerance and the antigen requirement to maintain it in irradiated C5-deficient hosts repopulated with C5-sufficient spleen cells was examined. The positive control of irradiated C5-deficient hosts repopulated with syngeneic spleen cells showed a primary and secondary response to immunization. In contrast, C5-sufficient spleen cells failed to respond both in the primary and the secondary response. Because the unresponsiveness was not caused by antigen carryover and was not antigen specific, it represents central tolerance. In C5-sufficient irradiated hosts (where immunization was not required and antigen was present in natural form and physiological concentration), transfer of C5-deficient cells mediated a drop in C5 levels to 10-20% of that noted in unreconstituted controls. T and B cell mixing experiments from the two strains into deficient or sufficient hosts demonstrated that tolerance is T cell dependent and that C5-sufficient or -deficient B cells could cooperate with nontolerant C5-sufficient T cells to produce significant anti-C5 antibody or mediate a significant drop in C5 levels. In addition, the presence of antigen was necessary to maintain tolerance. In conclusion, these results show that (a) natural tolerance to C5 is an active process that is T cell dependent and requires the presence of antigen; (b) in this natural model, clonal abortion does not seem to occur; and (c) both tolerant and nontolerant B cells retain the capacity to produce autoantibody.