EDWIN JOSEPH COHN 715 



In the first place freeing the protein from all but one salt, ammonium sulfate. 

 In the second place freeing the protein from the sulfate ion. The first step was 

 accomplished by reprecipitating with ammonium sulfate at ditTerent hytlrogen 

 ion concentrations. The protein was freed from anions other than the sulfate by 

 reprecipitation with c.p. ammonium sulfate from dilute sulfuric acid solution, 

 approximately 10""'n, and from cations other than ammonia by reprecipitation 

 from solution in tlilute ammonia. The globulin was then clialyzed, isosmotically, 

 in the manner described by Sorensen (13). A slight excess of ammonia was 

 maintained in the dialyzers until the globulin was completely freed from sulfate. 

 The transformation of the soluble ammonium-globulin compound into the very 

 slightly soluble uncombined isoelectric globulin has already been described. 



A somewhat different and more exact procedure was adopted in 

 studying the isoelectric sokibility of serum glol)ulin Va. Not only were 

 different amounts of the globulin preparation used to saturate the 

 same volume of water, but the same precipitate was used to saturate 

 successive additions of water. The dates are recorded on which 

 samples were removed for analysis. On these dates a further amount 

 of fresh water was added and saturation recommenced. The results 

 obtained in this way were far more concordant. They are tabulated 

 in Table IV. As in preparation IV, more protein dissolved in the 

 flasks which contained most protein. That this was for the most 

 part due to the fact that all of the soluble protein had not been re- 

 moved, is suggested by the fact that solubility soon fell to a lower 

 and constant value in the flasks containing the three smallest amounts 

 of globulin. 



The persistent higher solubility in the flasks containing most pre- 

 cipitate is probably to be ascribed to another phenomenon that need 

 not be considered at this time. 



Twenty-four nitrogen detenninations, made upon the filtrates of 

 solutions saturated with three different amounts of protein, on 4 

 different days, after their solubility had become constant, agree within 

 the errors of measurement. The average of these determinations is 

 0.30 mg. protein nitrogen in 25 cc, or 12.0 mg. protein nitrogen in 1 

 liter. This result is not very different from the lower results obtained 

 with preparation IV. If we adopt 15.85 as the percentage nitrogen 

 in this protein the most probable solubility of the pseudoglobulin 

 fraction of serum becomes 0.07 gm. per liter at 25°C. 



