In vivo observations on the kinetics of F cells and of fetal hemoglobin (HbF) synthesis and in vitro studies of erythroid progenitors, their number, and the gamma-gene expression in their progeny were carried out in baboons (Papio cynocephalus) treated with 5-azacytidine. Maximum effect on the increase of HbF production in vivo was observed only when an expanded erythroid marrow population was present. In these animals, as well as in normal animals, treatment resulted in a significant reduction of the late erythroid progenitor cell pools (erythroid clusters and erythroid colony-forming units, CFU-E) in the marrow. This reduction was more pronounced among those progenitors grown in the absence of added erythropoietin, and it was followed by a rebound a few days after treatment cessation, reflecting the accumulation of regenerating progenitors. An early increase in the in vitro synthesis of HbF in erythroid clusters and CFU-E colonies was observed. This increase was further documented at the cellular level, with immunofluorescent labeling of colonies with monoclonal anti-gamma- globin chain antibodies. In contrast to the findings in late progenitors, the number of erythroid burst-forming unit (BFU-E) colonies and the synthesis of HbF in these colonies was not influenced significantly by 5-azacytidine treatment. It is proposed that the toxic effects of 5-azacytidine on late progenitors, leading to faster mobilization of earlier progenitors to the next more mature compartment, play a role in the in vivo augmentation of HbF synthesis by this drug. This perturbation in the progenitor cell population kinetics and the presumed hypomethylation of the surviving differentiating cells may act synergistically to produce a maximum HbF response after 5-azacytidine treatment.