Groups of rabbits were injected with either bovine serum albumin, sheep red cell stroma, or keyhole limpet hemocyanin to which 2,4-dinitrophenyl and/or p-azophenyl arsonate groups had been coupled. Groups of animals received either doubly coupled antigen or an equivalent mixture of singly coupled antigens. Materials were injected intravenously as a solution or subcutaneously and intramuscularly in complete Freund's adjuvant. The presence of dinitrophenyl groups on the immunizing antigen could suppress, partially or completely, the antibody response to p-azophenyl arsonate when this hapten was located on the same molecule. Suppression was dependent on the ratio of haptenic groups on the molecule, appeared to be greatly affected by the method of immunization, and could be demonstrated in all three antigen systems. Partial suppression was manifested in decreased frequency and delayed appearance of the response as well as decreased maximal antibody titers. These findings appear irreconcilable with the possibility of direct clonal selection of antibody-producing cells by unprocessed antigen.