Recent evidence suggests that prothymocytes, which occur in a low frequency in murine bone marrow (BM), are already committed to thymocyte differentiation and discrete from precursor B cells as well as pluripotent hematopoietic stem cells. Furthermore, it was suggested that, in rodents, prothymocytes are positive for the nuclear enzyme terminal deoxynucleotidyl transferase (TdT) and a T cell surface antigen. The human prothymocyte has not been identified as yet. We analyzed human BM cells by double immunofluorescence staining for TdT and the T cell surface markers Tp41 (recognized by the monoclonal antibodies WT1 and 3A1), T11, T1, and T6. In the BM samples tested, neither T1+/TdT+ nor T6+/TdT+ cells were detected, but Tp41+/TdT+ and T11+/TdT+ cells were present in low frequencies. In childhood BM, the frequency was about two to five in 10,000, whereas in adult BM and regenerating BM, these cells were not always detectable, but if detected, their frequency was five- to 10-fold lower. In a triple staining, using fluorescein, rhodamine, and colloidal gold particles as labels, it appeared that all Tp41+/TdT+ cells were also positive for HLA-DR. These Tp41+/HLA-DR+/TdT+ cells were also detectable in low frequencies in the thymus, and occasionally Tp41+/TdT+ and T11+/TdT+ cells were detected in the peripheral blood (PB), suggesting a migration from the BM to the thymus via the PB. The malignant counterpart of the Tp41+/HLA-DR+/TdT+ cell was detected in a patient with acute lymphoblastic leukemia with the Tp41+/T11+/HLA-DR+/TdT+/T1-/T6- phenotype and germ-line immunoglobulin heavy chain genes. We postulate that the Tp41+/T11+/HLA-DR+/TdT+/T1-/T6- cell represents a human prothymocyte.