Although hematopoiesis is known to originate in a population of very primitive cells with both lymphopoietic and myelopoietic potential, a procedure for enumerating such cells has to date not been available. We now describe a quantitative assay for long-term repopulating stem cells with the potential for reconstituting all hematopoietic lineages. This assay has two key features. The first is the use of competitive repopulation conditions that ensure not only the detection of a very primitive class of hematopoietic stem cells but also the survival of lethally irradiated mice transplanted with very low numbers of such cells. The second is the use of a limiting-dilution experimental design to allow stem cell quantitation. The assay involves transplanting limiting numbers of male "test" cells into lethally irradiated syngeneic female recipients together with 1-2 x 10(5) syngeneic female marrow cells whose long-term repopulating ability has been compromised by two previous cycles of marrow transplantation. The proportion of assay recipients whose regenerated hematopoietic tissues are determined to contain greater than or equal to 5% cells of test cell origin (male) greater than or equal to 5 weeks later is then used to calculate the frequency of competitive repopulating units (CRU) in the original male test cell suspension (based on Poisson statistics). Investigation of this assay system has shown that all three potential sources of stem cells (test cells, compromised cells, and the host) can under appropriate circumstances contribute to long-term hematopoietic regeneration, thus establishing both the competitive pressure of hematopoietic stem cells in the cotransplanted compromised population and in the host, and the need to use genetic markers to track the specific contribution of the injected test cells. Analysis of the frequency of CRU in test marrow suspensions that varied widely in their CRU content gave similar values when endpoints of either 5 or 10 weeks posttransplantation were used and when either recipient marrow or thymus was used to identify progeny populations. In addition, repopulation of marrow and thymus was found to be associated in most mice injected with limiting numbers of test cells. These findings are consistent with the conclusion that the assay is highly selective for a very primitive, totipotent, reconstituting hematopoietic stem cell and should therefore be particularly useful in future gene therapy-oriented research as well as for more basic studies of hematopoietic stem cell regulation and differentiation.