Mouse marrow cells were exposed to 5-fluorouracil (FU) either in vivo or in vitro and the effects on the hematopoietic stem cell compartment were studied. The drug was highly toxic to bone marrow cells including the spleen colony-forming unit (CFU-S) population. The small population of stem cells surviving FU, however, caused a different pattern of spleen colony growth when injected into lethally irradiated mice. Whereas numbers of spleen colonies caused by normal marrow cells remained constant during an 8-14 d period after transplantation, spleen colonies derived from FU-treated marrow cells increased by as much as 100-fold during this time. This effect on stem cells was dose dependent both in vitro and in vivo. When FU was given in vivo, the day 14/day 8 ratio of colonies was greatest 1 d after injection and, over the next 7 d, returned to a near-normal value, that is, unity. A number of studies have shown that the stem cell compartment is heterogeneous with respect to self-replicative capacity and developmental potential. An age structure for the stem cell compartment has been proposed wherein cells with a short mitotic history are more likely to self-replicate than they are to differentiate; hence they are more primitive. 'Older' stem cells with a longer mitotic history are, according to the hypothesis, more likely to differentiate. 5-fluorouracil may be toxic to the older stem cells and selectively spare the more primitive subpopulation. Although the surviving cells may not themselves be able to form spleen colonies, they may give rise to an older cohort of cells more likely to differentiate and form spleen colonies. It is the requisite developmental maturation within the stem cell compartment that may be responsible for the delay in appearance of spleen colonies derived from FU-treated marrow. Our results support this explanation and identify the locus of at least part of this activity as the bone marrow. We found that the FU-treated marrow did not cause an increase in spleen colony numbers between 8 and 14 d in hosts with a long-standing marrow aplasia, due to the incorporation of 89Sr into bone. I propose that the delayed spleen colony appearance in normal hosts is the result of developmental maturation of the primitive stem cell compartment that survives FU and is responsible for spleen colonies arising around day 14. This maturation, at least initially, occurs in the marrow and leads to the replenishment of the more differentiated CFU-S subsets ablated by FU, which are normally responsible for spleen colonies appearing earlier after transplantation.