At present, nearly all infants with birth weights of < 1 kg receive blood transfusions for treatment of clinical signs of tissue hypoxia resulting from anemia of prematurity. In contrast to the successful use of recombinant human erythropoietin (rhEp) in adults, treatment of anemic neonates with rhEp to stimulate red cell production and reduce the need for transfusions that pose serious infectious and immunologic risk has not been effective. The present study investigates the pharmacodynamics (PD) of endogenous erythropoietin (Ep) in sheep fetuses to determine possible causes for the poor rhEp response in early development. The dynamic relationship between plasma Ep and plasma iron resulting from spontaneous hypoxemic episodes is investigated by PD system analysis. The erythropoietic effect of Ep is measured in terms of the mobilization of plasma iron needed in the production on new erythrocytes. A hysteresis minimization approach is employed to determine the intrinsic PD dose‐response relationship (transduction) of Ep. The doseresponse relationship shows a well‐defined threshold level that has to be exceeded before Ep begins to show a significant effect on plasma iron. It is postulated that the threshold mechanism may serve a useful purpose during early development by reducing the risk of the fetus developing a pathological degree of polycythemia and hyperviscosity in the relatively hypoxemic fetal environment. At the same time, the threshold serves the purpose of providing a needed response to more severe pathologic hypoxemic episodes. The occurrence of anemia during subsequent postnatal life when PaO2 levels increase markedly may be the inevitable, but unfortunate corollary of a continuation of this mechanism. If the poor response to rhEp in early neonatal life is due to a threshold mechanism, then it will be necessary to administer Ep at significantly larger dosing rates to ensure a sufficient and sustained coverage above the threshold level.