Rabbits of allotype a1a3 were injected on days 0, 2, and 4 with mixtures containing equal amounts of pigeon erythrocytes (Prbc) coupled to para-azobenzenearsonate (AA) and to para-azobenzene-N-trimethylammonium (TMA). On day 6, the allotypes of antibody from plaque-forming cells (PFC) of the blood were determined by observing the inhibition of plaque formation by anti-allotype sera. Anti-AA PFC appeared to consist for the most part of cells making antibody of allotype a1 since 65% of them were inhibited by anti-a1 serum and only 8% by anti-a3. Anti-TMA PFC, on the other hand, appeared to consist mostly of cells making antibody of allotype a3, since less than 1% of them were inhibited by anti-a1 but 47% by anti-a3. Antibody allotype for spleen PFC was also determined on day 6 and was similar to that found for blood PFC. Anti-AA PFC were inhibited 74% by anti-a1 serum and 15% by anti-a3 whereas anti-TMA PFC were inhibited 19% by anti-a1 and 43% by anti-a3. Serum hemolysin specific for AA hapten from a1a3 animals was also strongly inhibited by anti-a1 serum but not by anti-a3 whereas the converse was true for hemolysin against TMA hapten. The a1a3 rabbits, in whcih the anti-AA was restricted to allotype a1, were mated to produced homozygous a3a3 animals. When the PFC and serum antibodies of these a3a3 offspring were examined by specific inhibition, the anti-AA activity was found to be of allotype a3 rather than being a-negative. The number of anti-AA PFC in the blood of a3a3 rabbits was lower than that in blood of a1a3 or a1a1 animals. In addition, the TMA hapten appeared to inhibit the response to the AA hapten. Thus a1a3 rabbits immunized with AA-Prbc alone had 14-fold more anti-AA PFC or 18-fold higher anti-AA hemolysin titer than a3a3 animals immunized with both AA-Prbc and TMA-Prbc. Our results are discussed in relation to various explanations which have been offered for an imbalance of allotypes in a given antibody.