A definite cytotoxic activity was developed in a BALB/c (H-2d) anti-DBA/2 primary mixed leukocyte culture (MLC), which received interleukin 2 (IL-2) on day 3 of culture. This cytotoxic activity was minor histocompatibility antigens (MIHA)-specific at the stimulator level, and was not developed in a syngeneic (BALB/c anti-BALB/c) MLC. The addition of IL-2 on day 3 of culture was crucial; no or very weak cytotoxic activity was developed in MLC receiving IL-2 on day 0 or on both day 0 and day 3. Only appropriate MIHA-allogeneic tumor cells were lysed as the target of the cytotoxic activity. The cytotoxic activity seemed MIHA-specific also at the target level; it lysed tumor cells of DBA/2 mouse origin but not those of BALB/c (syngeneic) origin. Phenotypes of the cytotoxic effector cell were Thy-1+ Lyt-2+. We concluded from these results that MIHA-specific cytotoxic T lymphocytes (CTL) were generated in the MIHA-allogeneic primary MLC. In this newly developed system, we studied genetic and antigenic requirements for primary anti-MIHA CTL responses in vitro. We demonstrated; among spleen cells (SC) of seven B10 H-2-congenic strains only SC of B10.D2 strain whose major histocompatibility complex (MHC) (H-2d) was compatible with the responder MHC effectively stimulated responder BALB/c (H-2d) SC for an anti-MIHA (DBA-C57BL-common) CTL response. Similarly, only SC of two out of seven C x B recombinant inbred strains (C x B.H and C x B.D), which were compatible at the MHC with responder SC, activated responder BALB/c SC for the response. The possibility that cells responding to H-2 alloantigens suppressed the anti-MIHA response was ruled out. Additional experiments showed that compatibility at the H-2K-end or the H-2D-end of the MHC was sufficient for a definite anti-MIHA response. These provided formal evidence that primary anti-MIHA CTL responses in vitro were MHC-restricted at the stimulator level. We then showed that sonication-disrupted SC or Sephadex G-10 column-passed nonadherent SC failed to stimulate responder SC for a primary anti-MIHA CTL response, whereas G-10-passed nonadherent SC responded well to adherent stimulator cells. Further study demonstrated that Ia+ adherent cells were the most active cell type as stimulator. Finally, we confirmed that the primary anti-MIHA CTL responses to adherent stimulator cells was MHC-restricted.