AFTER completing my residency in internal medicine in 1957, I went to the New England Medical Center in Boston as a trainee in the Hematology Division of Dr William Dameshek. Dameshek was a pre-eminent hematologist—the editor of Blood, a founder of the American Society of Hematology, and the author of hundreds of papers and three books on hematologic topics. His huge clinical practice gave us trainees the opportunity to deal with every aspect of hematologic diseases in all their variations. It was an exciting time, because clinical investigations of bone marrow transplantation were just beginning. Dameshek thought that bone marrow transplantation would be a logical treatment for aplastic anemia: replace the patient’s injured marrow with normal marrow from a relative. These attempts were spurred by the nuclear reactor accident at the Boris Kidrich Institute of Nuclear Sciences in Belgrade, where six laboratory workers were heavily irradiated with neutron and gamma rays. They were flown to Paris and treated by Georges Mathe, a pioneer in bone marrow transplantation, with fetal spleen and liver cells or allogeneic bone marrow. All six patients recovered, but whether the restoration of hematopoiesis was due to the hematopoietic cell grafts was never clear. In 1957, we began to use allogeneic bone marrow to treat patients with aplastic anemia. The conditioning regimen was total body radiation, supplemented with high doses of corticosteroids. It was clear from work in mice that the rejection of allografts, including allogeneic bone marrow cells, is mediated by an immune response against foreign histocompatibility antigens. This work in rodents, especially by investigators at the Oak Ridge National Laboratories in the United States and the Radiobiological Institute in Rijswijk, Netherlands, was the basis for using total body radiation to prevent rejection of bone marrow allografts in humans (reviewed in Van Bekkum and De Vries.) The outcome in almost all the patients we treated were disastrous, with death due uniformly to overwhelming infection. It is well to remember that in 1957 the HLA system was unknown, the hematopoietic stem cell was only a theoretical idea, and the management of sepsis in neutropenic patients was at best primitive. Dameshek blamed the poor results on the total body radiation, which he considered to be the main element in the treatment that was beyond his control. He thought that we would be better off with a chemical substance that could be controlled by the physician on a day-to-day basis. Dameshek assigned to me the task of investigating how the immune response could be controlled by chemicals. At that time, there were indications in the literature that lymphocytes proliferated in response to antigenic stimuli. Moreover, hematologists had been using two drugs, 6-mercaptopurine and methotrexate, to treat acute lymphoblastic leukemia, a disease characterized by uncontrolled proliferation of primitive lymphocytes. For these reasons, I wrote two letters, one to the Lederle Company for a sample of methotrexate powder, the other to the Burroughs Wellcome Company for some 6-mercaptopurine. I never heard from Lederle, but George Hitchings of Burroughs Wellcome kindly sent me 25 g of yellow 6-mercaptopurine in powder form, with instructions for dissolving it. I had in mind to test these drugs on the immune responses of rabbits to a foreign protein, so the lack of an answer from Lederle was a stroke of luck—it turned out that methotrexate has virtually no effect in rabbits, because they metabolize the drug very quickly. By contrast, 6-mercaptopurine suppressed the response of rabbits to bovine serum albumin. Had I received methotrexate from Lederle, but no 6-mercaptopurine from Hitchings, it is likely that I would have given up the idea as unworkable. After obtaining these results with 6-mercaptopurine, the question arose whether the drug could induce specific tolerance to a foreign protein. By timing the injections of the drug and antigen, we found that rabbits could be rendered incapable of responding to bovine serum albumin, but still respond normally to bovine gamma globulin, another foreign protein that was administered after treatment with 6-mercaptopurine was discontinued. We thought that this effect was due to a preferential effect of 6-mercaptopurine on antigen-stimulated lymphocytes. In retrospect, those results were among the first experimental indications of clonal selection during the immune response. The next logical step was to test the ability of 6-mercaptopurine to suppress the allograft reaction. At that time, I