In this paper we present data which show (1) a quantitative relationship between plasma prqtein content and the tendency to formation of non-cardiac edema in nephritis, (2) demonstrate that the specific gravity of the blood plasma parallels the plasma protein content with sufficient exactness to indicate the level of the latter, and (3) show that the tendency to edema formation is related no less regularly to fall in plasma specific gravity than it is to fall in plasma protein content.That the concentration of protein in the blood plasma of nephritic patients is low was observed by Bright (1836). CsatAry (1891) noted that the deficiency affected the serum albumin more than it did the globulins, so that the albumin: globulin ratio, normally 1.5 to 2.0, fell below 1. These observations have been confirmed and amplified by other authors, whose work up to 1923 has been reviewed by Linder, Lundsgaard, and Van Slyke (1924). Govaerts (1924) has found that the albumin fraction, which suffers as a rule the entire deficit that occurs in the plasma proteins in the nephritic, is osmotically about four times as active per gram as the globulin fraction. Consequently especial interest attaches to determination of not only the total plasma protein content, but also the proportions in it of albumin and globulin. Nevertheless, as we shall show, the tendency to edema, except in rare cases, appears to be correlated about as closely to the total protein as to the albumin fraction, so that as a rule the simpler estimate of total protein content may be used for clinical purposes instead of the more complicated analyses involving separation of ihe albumin and globulin fractions.