Recently, culture conditions that stimulate the proliferation of primitive hematopoietic cells defined by various phenotypic and functional endpoints in vitro have been identified. However, evidence that they support a high probability of self-renewal leading to a large net expansion in vitro of transplantable cells with lympho-myeloid repopulating ability has been more difficult to obtain. The present study was designed to investigate whether the low overall expansion of human repopulating hematopoietic cells seen in vitro reflects a selective unresponsiveness of these rare cells to the growth factors currently used to stimulate them or, alternatively, whether they do proliferate in vitro but lose engrafting potential. For this, we used a high-resolution procedure for tracking and reisolating cells as a function of their proliferation history based on the loss of cellular fluorescence after staining with (5- and 6-) carboxyfluorescein diacetate succinimidyl ester. The results show that the vast majority of long-term culture-initiating cells and in vivo lympho-myeloid competitive repopulating units present in 5-day suspension cultures initiated with CD34(+) human cord blood and fetal liver cells are the progeny of cells that have divided at least once in response to stimulation by interleukin-3, interleukin-6, granulocyte colony-stimulating factor, Steel factor, and Flt3-ligand. Thus, most human repopulating cells from these two sources are stimulated to undergo multiple divisions under currently used short-term suspension culture conditions and a proportion of these retain engraftment potential.