Although the hormone erythropoietin (Epo) and its receptor (EpoR) are known to play important roles in the regulation of erythropoiesis, several questions remain concerning the developmental role of Epo/EpoR signaling. As the functions of Epo have been defined primarily through studies of definitive erythroid cells, its importance for primitive, embryonic erythropoiesis remains uncertain, as does the significance of EpoR expression in several nonerythroid cell types. To address these questions, mouse embryonic stem cells and embryos lacking a functional EpoR gene were produced by gene targeting. The effects of the mutation were examined in embryos developing in vivo, in chimeric adult mice produced with homozygous mutant embryonic stem cells, and in hemopoietic cells cultured in vitro. No defects were apparent in nonerythroid cell lineages in which the EpoR normally is expressed, including megakaryocytes and endothelial cells. In the mutant yolk sac, primitive erythrocytes were produced in normal numbers, they underwent terminal differentiation, and expressed near normal levels of embryonic globins, although they were reduced in size and their proliferation was severely retarded after E9.5. In contrast, in the fetal liver, definitive erythropoiesis beyond the late progenitor (CFU-E) stage was drastically inhibited by the EpoR mutation, and virtually no definitive erythrocytes were produced in vivo, leading to embryonic death by E13.5. Thus, our results suggest a fundamental difference in the molecular mechanisms stimulating primitive and definitive erythropoiesis. It was also observed that a few mutant definitive erythroid cells could terminally differentiate when cultured with additional cytokines, demonstrating that although Epo/EpoR signaling is important for definitive erythroid cell survival and proliferation, it is not an obligatory step in differentiation.