A simple, reproducible method for the separation of human erythrocytes, described recently (Murphy, J. R. (1973) J. Lab. Clin. Med. 82, 334-341) has been utilized for the purpose of obtaining a wide range of biochemical data on these cells. Using phthalate ester density centrifugation of the fractions obtained by Murphy's method, we established that the cells were separated exclusively on the basis of their densities. Data on a wide range of biochemical and hematological parameters, when compared with previously reported density separation procedures showed that this simple technique can be used to fractionate the cells according to their densities (age) in their own plasma. Cells of increasing density consistently and reproducibly exhibited an increase in hemoglobin concentration, a moderate elevation in Na+ and a decrease in the following: K+, acetylcholinesterase, sialic acid, membrane protein, 2,3-diphosphoglycerate, ATP, cholesterol, phospholipid, mean corpuscular volume and critical hemolytic volume, However, no change in mean corpuscular hemoglobin was evident. The observed differences were not artifacts of the centrifugation process. This was determined in recentrifuged top fractions from which new top and bottom cells were obtained. The latter cells resembled the top fraction from which they were obtained, rather than the original bottom fraction. Whereas the parameters mentioned above exhibited consistency and reproducibility, such was not the case with the ATPase values. Depending on the cell density group examined and/or buffer as well as other conditions, significant variability in the activity levels of the ouabain sensitive, as well as the Ca2+ -stimulated ATPase, was observed. Use of these enzyme activities as indicators of cell age must be viewed with caution.