Erythropoietin titers, measured by bioassay of plasma extracts in hypertransfused mice, were determined in 162 patients with absolute erythrocytosis, and the results were correlated with the clinical diagnosis. Fifty-two patients met the diagnostic criteria for polycythemia vera, and all had low or nonmeasurable erythropoietin titers. Of the remaining 110 patients, 62 were suspected clinically as having secondary polycythemia. However, 15 had low erythropoietin titers, casting doubt on the accuracy of the clinical diagnosis. The pathogenesis of the erythrocytosis in the last 48 patients was unknown, and they were designated clinically as having pure erythrocytosis. However, in 20, the erythropoietin titers were increased, and in 28, the titers were low, suggesting that they belonged to at least two different groups. Using erythropoietin titers in the classification of absolute erythrocytosis, the first group should be added to the category of patients with secondary polycythemia as a subgroup with disease due to idiopathic overproduction of erythropoietin (hypererythropoietinemia or essential erythrocytosis). The second group should be added as a subgroup of patients with primary polycythemia under the term erythremia.