478 PROTOPLASM 



The following molecular weights of proteins are fairly exact 

 multiples of the first: egg albumin, 34,500; hemoglobin, 68,100; 

 serum globulin, 103,600; phycoerythrin, 208,000. 



The problem of determining the molecular weights of proteins 

 is difficult not only because of the large size of the molecules and 

 our comparative ignorance of their constitution but also because 

 the molecular weight is apparently not constant, differing under 

 different conditions. Svedberg believes that this is true for acid 

 and alkaline solutions and that the molecular weight is lowest in 

 solutions of high alkalinity. Burk finds hemoglobin in urea to 

 have one-half the molecular weight that it has in water, while 

 albumin is the same in both solutions. Apparently, depoly- 

 merization (a breaking down of the molecule) of hemoglobin 

 takes place in urea. Possibly, the situation is as in the case of 

 cellulose, where the molecule is a chain of anhydrous glucose 

 links the number of which may vary. The chain, and therefore 

 the molecule, is of indefinite length and weight. This concept, 

 fairly well established in the case of cellulose, is contrary to 

 classical laws, which demand constancy in size, weight, and con- 

 stitution of molecules; but classical laws have often had to give 

 way to newer ideas. 



The situation in regard to the relationship between pH and 

 molecular weight of proteins appears to be that only within a 

 definite pH range, which is characteristic for every protein, is 

 there a well-defined molecular weight. The isoelectric point is 

 always within this range. Outside the range (acid or alkaline), 

 the molecules split up and give other " molecular- weight " 

 values. 



The suggestion of Svedberg that the molecular weights of 

 proteins are all multiples of 34,500 and that if the higher weights 

 are regarded merely as polymerized forms of the lower unit, then 

 "all proteins have the same molecular weight" is not as yet 

 generally accepted, even though his measurements support the 

 idea. Determinations of molecular weights from chemical data 

 (as opposed to Svedberg's physical centrifuge method) give the 

 values^ shown in the table on page 479. 



Crystalline Protein. — One of Graham's criteria of the colloidal 

 state was the absence of a crystalline condition; this distinction 

 has proved to be erroneous, for many proteins can be crystallized 



1 From Burk. 



