Abstract Mice with transfusion-induced polycythemia have been used to assay erythropoietic activity in plasma derived from mice and rats subjected to various stimuli and experimental procedures. 1. Anemic, anoxic, or cobalt mouse plasma increased erythropoiesis from the zero baseline to about the normal range after four 0.5 cc. injections. No significant effect is observed from the same number of injections of normal rat or mouse plasma or normal saline. Similarly, plasma from rats starved for 4 days had no erythropoietic activity, but erythropoietic activity was present in the plasma of pregnant mice during the last trimester of pregnancy. 2. Plasma from mice made anemic by phenylhydrazine and then subjected to bilateral nephrectomy contained slight erythropoietic activity (less than 0.2 per cent reticulocytes). The plasma of sham-operated anemic animals had considerably more erythropoietic activity (2.0 per cent). The plasma was collected for assay 10 hours after the operative procedures. 3. Plasma from nephrectomized rats exposed to hypoxic anoxia for 8, 16, or 24 hours, had slight erythropoietic activity when measured by the peripheral reticulocyte response of the recipients (0.18, 0.11, and 0.00 per cent reticulocutes, respectively). The same plasma had no erythropoietic activity when judged by the Fe59 red cell incorporation response in starved rats. Plasma from rats subjected to bilateral ureter ligation and similarly exposed had an erythropoietic activity (0.7, 1.38, and 1.26 per cent reticulocytes, respectively), comparable with that of the plasma from normal rats exposed to hypoxic anoxia (0.9, 1.71, and 1.78 per cent, respectively). The low and inconstant erythropoietic activity in the plasma of nephrectomized mice and rats subjected to anemic or hypoxic anoxia is more or less comparable with the response occasionally produced by normal plasma. We are, therefore, disinclined to consider the data as evidence against the renal origin of erythropoietins, but the alternative possibilities are discussed.