This study shows that LPS is not mitogenic in cultures containing B cells, or B cells and accessory adherent cells or ME, unless T cells are present. This observation rules out models of induction of antibody synthesis in which it is assumed that the delivery of a mitogenic signal by the interaction of LPS with the membrane of the B cell is in itself sufficient for B-cell induction (19). Further, it makes unlikely the proposed extrapolation of such a model to other so-called thymus-independent antigens, e.g., PVP, levan, dextran, and SIII (19). The mitogenic action of LPS appears to be due to its ability to complete an inductive stimulus to B cells (13). We interpret the observed thymus dependence of the B-cell response to LPS in light of a model in which two signals are obligatory for B-cell induction (14). The first signal in the inductive pathway is delivered to the antigen-sensitive cell via a conformational change in the receptor upon interaction with antigen. The second signal is delivered via the thymus-derived cooperating system. Since LPS can induce immune responses to both immunogenic and nonimmunogenic ligands (9-13) we envision that one signal is delivered to the B cell via specific binding of the ligand to the B-cell antigen receptor, while a second signal is delivered as a result of T-cell cooperation via membrane-bound LPS. This has been termed abnormal induction (20). In this example LPS is the foreign membrane-bound determinant in question although histocompatibility antigens (21, 22), viral determinants, or surface bound lectins could act similarly. In light of the above model, one observation should be pointed out. LPS inhibits the induction of a SRBC response in normal Peyer's patch cells to which adherent cells or ME is added. This inhibition appears to be a T-cell-mediated effect because it is abolished by partial depletion of the T-cell population by antitheta treatment. Since the induction of IgM producing PFC is being measured, the T-cell-dependent LPS inhibition could act either (a) by induction of T-cell "suppression" (23, 24) of the normal cooperating system required for a SRBC response, or (b) by the induction of such high levels of cooperating function (13) as to be inhibitory to a SRBC IgM response. Our observations contrast sharply with prior reports which describe LPS as a thymus-independent antigen (2-4) and a B-cell mitogen (5-8) capable of stimulating immune responses in the absence of T-cell cooperation (2-12). This demonstration of the thymus dependence of LPS stimulation has been possible because Peyer's patches from congenitally athymic (nude) mice are functionally a highly purified B-cell population devoid of T cells and accessory adherent cells. In this respect, earlier studies relied on nude spleen cultures and spleen cultures from thymectomized, lethally irradiated, and bone marrow-reconstituted mice (3, 4, 6-13). These spleen cultures which contain B cells and accessory adherent cells are recognized to be deficient but not devoid of the thymus-derived contribution to the inductive stimulus (12, 13). It could be argued that the presence of T cells and adherent cells is in fact required for the antigen-specific effect and not for the LPS effect. However, this is unlikely since our experiments show that LPS is not directly mitogenic for B cells and does not stimulate background anti-SRBC PFC. It seems unlikely that Peyer's patch antigen-sensitive cells differ from antigen-sensitive cells in the spleen in their mechanism of induction. We have shown that Peyer's patch B cells can be specifically induced by antigen, and Peyer's patch T cells mediate cooperating and killer functions. Alternately, the possibility that Peyer's patch B cells were not stimulated by LPS as a result of prior cryptic exposure to LPS (13) in the intestinal tract was excluded since cultures containing B cells, T cells, and adherent cells or ME were stimulated to DNA synthesis by LPS. The reason that certain antigens appear to be thymus independent may be that their repeating polymeric nature permits inductive interactions at very low levels of thymus-derived cooperation (see reference 20 for quantitative considerations). It has been stated that the inductive properties of all thymus-independent antigens are directly related to their ability to act as B-cell mitogens (19). The observation that LPS is thymus dependent for its B-cell mitogenic activity makes us question the thymus independence of any antigen.