To assess the thrombopoietic activity of erythropoietin (Epo) in vivo in mice, we consecutively administered recombinant human Epo (rEpo) i.p. to intact and splenectomized mice for varying time intervals. Recombinant human Epo increased platelet counts in a dose-dependent fashion in a short-term trial for 5 days. An 18% +/- 12% increase in platelet counts was observed in intact mice injected with a total of 50 U rEpo over 5 days, whereas an increase of 31% +/- 15% was found in splenectomized mice treated in the same manner. The factor also elicited a significant increase in bone marrow megakaryocytic size of the same magnitude in both groups. The numbers of megakaryocytes and megakaryocytic progenitors in bone marrow were not altered by the treatment. In addition, rEpo induced a dose-related increment in spleen weight. Ultrastructural analysis of spleens in mice injected with rEpo for 5 days revealed no significant increase in sectional platelet counts and in the numbers of vacuolar and platelet-phagocytoid cells when compared with control mice treated with heated rEpo. These findings suggest that Epo has an effect on thrombocytopoiesis, functioning as a late stimulator in short-term administration, they also suggest that its ability is partially masked by splenic pooling by enlarged spleens, giving rise to the discrepancy in platelet counts between and splenectomized mice. However, the induced increases in both parameters, platelet counts and megakaryocytic size, declined after 5 days. Platelet counts returned to normal levels in intact mice 15 days after initiation of the injection, whereas no significant difference in megakaryocytic size was noticed on day 10 compared with controls. The same results came out in splenectomized mice. Anti-human rEpo antibody was not detected in the sera of treated mice in a series of experiments. The splenic weight remained at plateau levels for up to 30 days. Thus, the substantial effects of rEpo on megakaryocytopoiesis and platelet production are transient, neither due to the appearance of the antibody to human rEpo nor the concealment by splenic pooling. The mechanisms of the transient action of Epo on thrombocytopoiesis remain to be further studied.