The current paradigm of lymphocyte activation, the two-signal model, has developed from the premise that recognition of antigen alone is insufficient to stimulate naive B cells, as this could potentially induce autoreactive responses, and that cognate T-B interaction is necessary to induce a full B-cell response. Recent evidence suggests, however, that T-cell-independent B-cell activation is part of the humoral immune response to pathogens, and therefore that antigen alone, or antigen plus signals provided by cells other than T cells, can provide all the necessary signals to induce a B-cell response. Furthermore, the presence of secreted IgM produced either as natural antibodies by CD5+ B-1 cells or as antigen-induced IgM by conventional (B-2) cells was shown to affect the kinetics and magnitude of the IgG response significantly. These data and the observed rapid kinetics of in vivo humoral responses seem at odds with a model that predicts that full B-cell activation and expansion is delayed until sufficient T-cell help is generated. I will therefore argue here that, in response to an infection, initial clonal B-cell expansion and secretion of IgM occurs in a T-cell-independent fashion (phase I) driven by the presence of antigen, and that secreted IgM serves as an autocrine growth factor at this time. B-cell-T-cell interaction occurs only during phase II of the response, thereby initiating the germinal center reaction, isotype switching and memory B-cell development. Hence, this model provides an explanation of how B-cell responses are induced rapidly in vivo at a time when T-cell help is rare.