Specific antibody responses to influenza virus were obtained in vitro from human blood mononuclear cells (PBM). The addition of allogeneic T cells to responding PBM profoundly suppressed antigen-induced antibody responses, but had no effect on pokeweed mitogen (PWM)-induced polyclonal Ig formation. This raised the possibility that suppression by allogeneic T cells may result from the activation of antigen-specific T suppressor (Ts) cells rather than nonspecific allogeneic effects. The frequency of allogeneic Ts in PBM from a number of different donors, estimated in a series of limiting dilution analyses, was found to range from 0.8 X 10(-5) to 4.5 X 10(-5), which is more typical of antigen-specific than alloreactive T cells. By adding limiting numbers of allogeneic T cells to antibody-forming cultures stimulated simultaneously with two non-cross-reacting antigens (influenza A and B strain viruses A/X31 and B/HK), it was possible to demonstrate suppression of the response to one antigen, but not the other, in the same culture well. Moreover, the frequency of wells in which the response to both antigens was suppressed was not significantly different from that predicted from the calculated frequency of specific allogeneic Ts. These results show that allogeneic suppression was antigen specific, and was not due to non-specific allogeneic effects. By separating T cell preparations into Leu-3a+ (helper) and Leu-2a+ (suppressor/cytotoxic) T cell subsets, suppression was shown to be mediated by a radiosensitive Leu-2a+ T cell. The removal of Leu-2a+ cells from T cell preparations abrogated the suppressor effect and permitted T cell collaboration with B cells, across an HLA-A, -B, and -DR barrier. This result shows that in at least some combinations, suppression rather than major histocompatibility complex restriction is the reason for the failure of allogeneic T and B cells to collaborate in T cell-dependent antibody responses.