The bacterial carbohydrate dextran B512 is a thymus-independent (TI) antigen and a poor immunogen. Humoral responses consist primarily of IgM and no memory response is observed; rather, secondary responses to native dextran are similar to or suppressed compared with primary responses. However, immune responses to dextran can be enhanced. In this study we have used a protein-dextran conjugate that elicits a thymus-dependent (TD) immune response against dextran. Furthermore, we used the potent adjuvant cholera toxin (CT) for the dextran immunizations. This enables us to re-evaluate the phenomenon of poor secondary response to dextran and whether it can be abrogated. We show that native dextran-primed mice were not able to mount IgG anti-dextran antibody responses after repeated immunizations with the TD, protein-dextran conjugate. This was also apparent in the spleen, where almost no dextran-specific germinal centres were detected. However, the anti-protein antibody response was normal in these mice, demonstrating that it is only the anti-dextran-responding cells that are affected. The effect of CT adjuvant on these events was also evaluated. CT enhanced the humoral IgM anti-dextran responses as well as the splenic responses to dextran. But, the isotype profile was not altered, still no IgG was produced. In contrast, mice primed with the TD conjugate and repeatedly re-immunized with native, TI, dextran generated IgG anti-dextran responses. Our results indicate that it is probable that the lack of proper costimulation in the initiation of the response to dextran causes the suppressed secondary dextran responses. Furthermore, these results suggest that TI and TD forms of dextran activate the same type of B cells, since TI dextran-priming abrogated TD dextran IgG responses. The importance of the priming event for the induction of a classical memory response to carbohydrate antigens and the implications for vaccination strategies, are discussed.