In nephrosclerosis (Volhard and Fahr classification of nephritis) there was no decrease in the plasma proteins. With heart failure the ratio of albumin to globulin fell from the normal 1.7 +/- 0.3 to about 1. There is a sharp difference between the two types of glomerulonephritis in the effect on plasma proteins. In the vascular-interstitial type the effect was the same as in nephrosclerosis. There was no decrease in the plasma proteins until shortly before death. In the glomerulotubular or nephrotic type, active or recently active, the total plasma proteins were less than 5 gm. per 100 cc. This decrease from the normal 5.5 to 7.5 occurred whether edema was present or absent. The decrease affected chiefly the albumin, the globulin being usually diminished but little, and sometimes slightly increased. Consequently the ratio of albumin to globulin was reduced to less than 1, and occasionally to 0.6. In nephrosis similar changes were found in the total protein concentration but in severe cases the decrease in albumin was greater than in nephrotic glomerulonephritis while the globulin was either very slightly reduced or was increased. The albumin : globulin ratio was therefore lower than in nephrotic glomerulonephritis, ranging down to 0.26. With the disappearance of edema there was usually an increase in the plasma protein concentration, but this was not invariable, and concentrations of 4.5 per cent or less were compatible with the persistent absence of edema. The ratio of albumin to globulin showed an irregular tendency to rise. Whenever the total concentration was less than 4 per cent there was some edema present, but it was sometimes slight in amount. With recovery in acute cases, and remission with decrease of proteinuria in chronic cases, normal concentrations were regained. The albumin : globulin ratio remained low in some instances for a longer period on account of an absolute increase in globulin concentration. In "functional proteinuria" no change in plasma proteins was found. Low plasma protein concentrations were always associated with considerable losses of protein in the urine, but these losses did not provide an explanation for all our observations. A disturbance in the production of the plasma proteins appears probable.