Although macrophages from Mycobacterium bovis, strain BCG-infected C3H/HeN mice were highly cytotoxic to tumor cells in vitro, cells from BCG-infected A/J mice were not tumoricidal. Varying the dose of BCG or the time of infection did not evoke cytotoxic activity with A/J macrophages. Inflammatory responses to BCG infection in A/J mice were relatively normal: the yield of macrophages was about ¼ that from C3H/HeN mice; however, no differences between strains were detected for inflammation-induced increases in macrophage phagocytic capacity or in percentage of peroxidase-positive cells. Similarly, production of macrophage activation factor activity by tuberculin-stimulated BCG-immune spleen cell cultures was also intact in A/J mice. Thus, despite adequate inflammatory responses and production of active lymphokines, macrophages from BCG-infected A/J mice were not activated for tumor cytotoxicity. Under certain conditions, however, these macrophages could develop cytotoxic activity. Macrophages from BCG-infected but not control A/J mice developed tumoricidal activity after further in vitro treatment with lymphokines, bacterial lipopolysaccharides (LPS), or certain plant lectins. Peritoneal exudate macrophages first treated with lymphokines also developed cytotoxic activity but only after further treatment with LPS or lectins. The phenotypic expression of the A/J genetic defect in macrophage cytotoxic activity was very similar to the defect of lipid A-unresponsive C3H/HeJ mice except for the concentration of LPS required for cytotoxic activity by lymphokine-treated cells. Other responses of A/J mice to LPS were also different from those of the unresponsive C3H/HeJ mice. Responses of A/J mice to the lethal toxicity of LPS or of A/J macrophages to direct toxicity of LPS in vitro, responses influenced by the Lps gene, were normal. On the other hand, macrophage cytotoxic activity and spleen cell proliferative responses to LPS in vitro, responses also controlled by the Lps gene, were abnormal. We conclude that the genetic basis for the macrophage tumoricidal defect of A/J mice represents either a previously undescribed allelic change at the Lps locus or, what is more likely, gene mutation(s) at entirely different loci.