Antibodies to individual bee venom antigens were studied in detail in nine bee sting-allergic patients who received venom immunotherapy without side effects, in two patients who failed to reach maintenance, and in two whose sensitivity returned. The study was confined to patients who had IgE antibodies to at least one of four purified bee venom antigens at the start of treatment. IgE and IgG antibodies to phospholipase A 2 (PLA 2 ), hyaluronidase (HYAL), and acid phosphatase (ACID P) and IgE antibodies to melittin (MEL) were measured, and changes in the antibody levels were followed during bee venom immunotherapy. Two contrasting patterns of antibody response were seen in the nine successfully treated patients. In five patients there was a rise in serum IgG antibodies to the same antigens as the IgE antibodies. In two patients' serum IgE antibody to HYAL or ACID P fell without a marked IgG antibody response to these antigens, although high levels of IgG antibody to PLA 2 were present in both. Although the first pattern is consistent with a “blocking” role for IgG antibody, clearly the second is not. Not all patients can be conveniently divided into these two categories, and two patients did not show any significant change in either IgG or IgE antibody but were nevertheless able to tolerate the maintenance dose of 100 μg of venom. Two patients who failed to reach the maintenance dose of 100 μg because of their allergic reactions to the injections of venom were distinguished by (1) very high serum IgE antibody and (2) a low ratio of IgG/IgE antibody. Passive immunization with IgG antibody from a hyperimmune beekeeper was, however, protective in these patients, although it did not raise their overall serum IgG antibody level very much. We are unable to explain either the failure of conventional therapy or the beneficial effect of passive immunization in these two patients. Two bee sting-allergic beekeepers lost their sensitivity to stings, but later, when their sera contained IgE antibody to another bee venom antigen, they reacted to stings and inhalation of beehive dander. These data suggest that either falling IgE antibody or IgG-“blocking” antibody could be responsible for providing clinical protection to bee venom-allergic subjects. Renewed clinical sensitivity was observed when the IgE response was modulated, with patients making IgE antibody first to one antigen and then to another.