Serum IgE levels were followed longitudinally twice a week for up to 100 days in 60 children treated for leukemia, severe aplastic anemia or severe combined immunodeficiency: 52 underwent allogeneic bone-marrow transplantation and 8 immunosuppressive treatment. In 55 of 58 treatment periods which could be analyzed (95%), a transient sharp increase of serum IgE was detected, irrespective of the initial diagnosis and mode of treatment. A second IgE peak was recorded in 16% of evaluable treatment periods. In the transplanted leukemia and aplastic anemia patients, the rise of serum IgE levels occurred at the same time, i.e. at a mean of 14 days after transplantation; it occurred significantly later in children grafted for severe combined immunodeficiency. In children who received immunosuppression for the treatment of severe aplastic anemia, IgE elevations were always seen within 2 weeks after institution of therapy. No relation was found between either the occurrence of clinically acute graft-versus-host disease or infections after treatment, and the time of onset of IgE elevations. It is suggested that the phenomenon of IgE peaks in the population of patients investigated was due to disturbance of T-cell regulation, i.e. a temporary impairment of T-suppressor cell activity.