Determination of the serum protein content is supposed to give a fairly accurate estimate of the colloid osmotic pressure of the blood serum. However, the frequent finding of normal levels for serum protein in the examination of patients with marked ascites and edema suggested that the usual protein determination is an inadequate measurement of the osmotic conditions at the capillary membrane and cannot be used to predict the effective osmotic pressure. It was found that measurement of the colloid osmotic pressure gave better knowledge of these physical processes and in many instances explained the presence of indeterminable edema and ascites. Variations in the values for serum albumin and globulin in hepatic disease have been recognized for more than thirty years. Grenet1 and Gilbert and Chiray2 were among the first to note decrease in concentration of total protein among patients with cirrhosis of the liver, and their observation was