126 LECTURE VII. 



fact that the most varied albuminous substances have very similar elemen- 

 tary compositions. It is impossible to recognize the presence of any 

 foreign albumin in a mixture by any such procedure. There is also the 

 added objection, that the same mixture will invariably be obtained by 

 following out a prescribed method. 



It is very instructive in this direction that oxyhemoglobin crystals show- 

 ing no indication of any admixture under the microscope, may, neverthe- 

 less, be impregnated with foreign protein. 1 This fact has been established 

 through the discovery that glycocoll was present in a globulin from serum, 

 whereas oxyhemoglobin does not show the least trace of this ammo- 

 acid. It has also been shown that glycocoll appears in the oxyhemoglobin 

 of horse-blood, after one crystallization. Another recrystallization gives 

 us a preparation entirely free from glycocoll. In this connection we would 

 refer to the various descriptions of albumins containing carbohydrates, 

 even when the observations were made with crystalline preparations. 2 



The crystallization of albumins does not correspond with the ordinary 

 formation of crystals from other sources. Most of the albumin crystals 

 are obtained by withdrawing the solvent. The production of these 

 crystals does not gradually follow the removal of the solvent. The crys- 

 tallization is, on the contrary, a very sudden one. This can be best 

 illustrated by salting out with ammonium sulphate. A very slight excess 

 is sufficient to throw out large quantities of crystals from an otherwise 

 clear solution. It is certainly a matter of great importance that no albu- 

 min as such has yet definitely been isolated in a crystalline state. A 

 possible exception would be that of the globulins, generally called edes- 

 tin, separated from plant seeds. They contain a considerable amount 

 of sodium chloride. It is impossible to remove this entirely, and still 

 retain the crystalline form. Although we are still in the dark, regarding 

 the influence of sodium chloride on the crystallization of edestin we do 

 know that the egg-albumin and serum-albumin do not themselves crystal- 

 lize, whereas their sulphates do, as was shown by K. A. H. Morner. 3 The 

 crystallization of the globih in hemoglobin depends on the presence of 

 hematin. 



When we remember the extreme difficulty experienced in obtaining 

 absolutely pure crystals of substances of even low molecular weight, we 

 can hardly expect to obtain really pure products through any methods 

 which in themselves can give no guarantee of efficiency. Although we 

 thoroughly appreciate the advantage of possessing the albuminous body 

 in a crystalline condition, we likewise find it necessary to state that not 

 the least confidence can be placed in the method of obtaining the crystals, 



1 E. Abderhalden: Z. physiol. Chem. 37, 484 (1903). 



2 E. Abderhalden, P. Bergell and T. Dorpinghans: Z. physiol. Chem. 41, 530 (1904). 



3 K. A. H. Morner: Z. physiol. Chem. 34, 207 (1901). 



