Guinea pigs which were injected with either whole bacilli or purified soluble glycerol teichoic acid (GTA) usually exhibited a rise in hemolysin titer to GTAcoated erythrocytes. The only exceptions were those animals having high baseline titers of natural anti-GTA antibodies. Rats yielded better responses than guinea pigs and produced significantly higher responses to the soluble antigen than to the cellular GTA. Rats reared on a GTA-free diet were predominantly free of natural antibodies to GTA and furnished a more clear-cut model for assaying immune responses. Using this model, it was shown that adsorption of GTA to homologous erythrocytes before injection resulted in poor responses, suggesting that such spontaneous adsorption does not account for the good responses to soluble antigen. In GTA-deprived rats, positive skin tests were induced only with bacilli, whereas migration inhibitory factor was induced with both bacilli and soluble antigen. Hemolytic plaques in immunized rats were increased over controls with both kinds of immunogen, but the GTA-deprived rats responded better than conventional ones, and hemolytic plaque responses to bacilli were better than those to soluble antigen. This reversal of the serum hemolysin results may be due to delayed suppression by soluble GTA or to antibody cycling. The guinea pig data, combined with results from GTA-deprived rats, suggest that high antibody levels resulted in depressed antibody synthesis, perhaps because antibody cycling was initiated. No evidence was found to explain the superior responses to soluble antigen, but it did not seem related to formation of immune complexes or adsorption to erythrocyte membranes.