These data indicate that the increment in the anti-leukemia effect, as expressed as FFR, is comparable for transplants for AML in first remission, advanced leukemia, and in persons never achieving remission. These data are consistent with the notion that the major anti-leukemia effect of HLA-identical bone marrow transplantation in AML results from an immune-mediated graft-versus-leukemia effect rather than from high doses of chemotherapy and radiation. Of course, other factors might explain these results. The superior outcome observed for transplants in first remission versus more advanced disease results not from increased anti-leukemia efficacy of transplants but rather that more persons already cured by chemotherapy receive transplants. Otherwise stated, a substantial portion of the persons cured following transplantation for AML in first remission were cured before receiving a transplant. These data have implications for other aspects of bone marrow transplantation. For example, it is suggested that transplants should be performed earlier in solid tumors when these diseases are more likely to respond to high-dose chemotherapy and radiation. Although this hypothesis may be correct, it need not necessarily be so as evidenced by these data in AML. The data we review show that bone marrow transplants in AML are of comparable anti-leukemia efficacy when performed in first remission, advanced leukemia, and initial resistant disease. Similar conclusions may apply to transplants in CML and ALL. The superior overall outcome observed with transplants in earlier leukemia results from transplanting a greater proportion of subjects already cured by chemotherapy. The increased anti-leukemia efficacy of transplants when compared with chemotherapy is compatible with an anti-leukemia effect other than that of high-dose chemotherapy and radiation. An immune-mediated graft-versus-leukemia effect is a likely explanation. Caution in predicting results of autotransplants in solid tumors is likewise necessary.