H-2-deficient variants, selected from the murine lymphoma RBL-5, were recently shown to be less tumorigenic in syngeneic mice compared with the corresponding H-2-positive wild-type cell line. The present study focused on the detailed cellular interactions leading to the specific rejection of H-2-deficient cells in vivo. In vivo experiments with mixed tumor inocula, measuring either tumor outgrowth or rapid elimination, showed that the presence of H-2-positive lymphoma cells did not suppress the rejection of H-2-deficient cells. Conversely, H-2-deficient cells did not activate a rejection of H-2-positive cells. H-2-deficient cells were selectively eliminated even when they were present in a 10-fold excess compared with the H-2-positive lymphoma cells in the same inoculum. In vitro, H-2-deficient cells were more sensitive to killing by spleen cells from mice treated with interferon or natural killer-inducing agents. The presence of cold H-2-deficient cells did not activate the killing of H-2-positive cells, and H-2-positive cells did not inhibit killing of H-2-deficient cells. Such mixing experiments revealed a more efficient cold target inhibition by the H-2-deficient cells, although no difference was seen between these and the H-2-positive wild-type cells in target binding assays. When Thy-1.2+, nylon wool adherent or phagocytic populations were removed from normal spleen effectors, the difference in spontaneous cell-mediated cytotoxicity between the H-2-deficient and the H-2-positive cells persisted. This result was also obtained when spleen cell effectors from nude mice or asialo GM1+ effectors, positively enriched by fluorescence-activated cell sorter, were used. The results show that the selectivity in the rejection of H-2-deficient cells was not determined in the afferent arm alone. The present data fit with a previously proposed model where natural killer cells would bind equally well to both major histocompatibility complex class I negative and class I positive target cells, but only the latter would be able to present a postbinding inhibitory signal allowing them to escape killing which results in outgrowth in vivo and poor cold target competition in vitro.