In the present study we used an adoptive transfer model with athymic nude mice to characterize the T cells involved in initiating and mediating skin allograft rejection. It was found that skin allograft rejection in nude mice required the transfer of immunocompetent T cells and that such reconstitution did not itself stimulate the appearance of T cells derived from the nude host. Reconstitution with isolated populations of Lyt-2+/L3T4- T cells resulted in the rapid rejection of MHC class I-disparate skin allografts, whereas reconstitution with isolated populations of L3T4+/Lyt-2- T cells resulted in the rapid rejection of MHC class II-disparate and minor H-disparate skin allografts. By correlating these rejection responses with the functional capabilities of antigen-specific T cells contained within the reconstituting Lyt-2+ and L3T4+ T cell populations, it was noted that skin allografts were only rejected by mice that, as shown by in vitro assessment, contained both lymphokine-secreting Th cells and lymphokine-responsive Tk cells specific for the alloantigens of the graft. The ability of two such functionally distinct T cell subsets to interact in vivo to reject skin allografts was directly demonstrated in H-Y-specific rejection responses by taking advantage of the fact that H-Y-specific Th cells are L3T4+ while H-Y specific Tk cells are Lyt-2+. Finally, the importance of in vivo interactions between functionally distinct Th/T-inducer cells and T killer (Tk)/T-effector cells in skin allograft rejection was demonstrated by the observation that normal B6 mice retain Qala and Kbm6 skin allografts because of a selective deficiency in antigen-specific Th cells, even though they contain T-effector cells that, when activated, are able to reject such allografts. Thus, the ability to reject skin allografts is neither unique to a specialized subset of T cells with a given Lyt phenotype, nor unique to a specialized subset of helper-independent effector T cells with so-called dual function capability. Rather, skin allograft rejection can be mediated by in vivo collaborations between T-inducer cells and T-effector cells, and the two interacting T cell subsets can express different Lyt phenotypes as well as different antigen specificities.