EDWIN JOSEPH COHN 709 



wash water.^1 This was for the most part globulin like the rest, 

 dissolved by the ammonium chloride formed in the neutralization of 

 the ammonium globulinate by hydrochloric acid. The wash waters 

 containing this protein were accordingly dialyzed further and 

 concentrated. 



When the isoelectric protein had finally ceased to give off impurities 

 and readily soluble protein, it was suspended in the desired volume, 

 and aliquot parts used for analyses. It has been found impracticable 

 to dry proteins, since it is almost impossible uniformly to wet their 

 surface after they have been dried by alcohol and ether. As a result, 

 there can be no guarantee that a true equilibrium is subsequently 

 reached between all of the protein and its solvent. Moreover, it has 

 been thought preferable, for the purposes of this investigation, to 

 retain any lipoid that might be in fixed combination with our proteins, 

 rather than risk the danger of denaturing them by alcohol or ether. 



The Measurement of Solubility. 



In order to obtain reproducible measurements of the solubility of a 

 protein a great many precautions must be observed. To saturate 

 the solution with protein it is necessary to bring the heavy flocculent 

 precipitate into intimate contact with the solvent for a long period 

 of time. This can only be accomplished by protracted mechanical 

 agitation, since the diffusion velocities of the proteins are exceed- 

 ingly low because of their large molecular weights. If the protein 

 precipitate were not enormously subdivided, and not brought into 

 intimate contact with every part of the solvent, the latter would 

 remain unsaturated. ^^ jf the ordinary methods of agitation are used, 

 however, the solution foams, its concentration changes, and a 

 portion of the protein is hkely to become denatured. Moreover, 

 soft glass containers cannot be used, since in them the protein combines 

 with enough alkali to increase its solubility appreciably. 



" The isoelectric solubility in my earlier preparations of serum globulin, which 

 had not been washed at the isoelectric point, was as high as 75 per cent of the total 

 protein present. 



'^ Under these circumstances the solvent may enter the protein-rich pha.se, 

 because of its greater diffusion velocity, and give the appearance of swelling. 



