ii PERCENTAGE-COMPOSITION 69 



with clupein. The differences which used to be brought forward of old 

 have rapidly become less and less marked with the improvements in 

 our methods. 



Quite apart from a certain unreliability in our methods, we have 

 to deal with another much greater difficulty in making quantitative 

 determinations, namely, the difficulty of obtaining a pure uniform 

 material on which we can experiment. In organic chemistry the 

 most important characteristic of a pure and uniform substance is its 

 power of crystallisation. The only crystalline albumins we know of 

 are haemoglobin, serum-albumin, egg-albumin, ichthulin, and a number of 

 vegetable albumins. It will be pointed out, however, in Chapter VIII. 

 that we cannot consider crystalline albumins in every case to be pure 

 substances, inasmuch as Wichmann 1 and Schulz and Zsigmondy 2 

 have shown how readily albumins absorb impurities, and with what 

 tenacity they retain them. According to Schulz and Zsigmondy it is 

 necessary to recrystallise egg-albumin five to seven times to obtain it in 

 a pure state. The impurities in question are by no means small in 

 amount, for Abderhalden 3 found that haemoglobin which was crystallised 

 twice contained no glycocoll, while after the first recrystallisation 

 there was still present 0*62 per cent of glycocoll. This difference he 

 explains as due to an admixture of serum-globulin. As serum-globulin 

 contains 4 per cent of glycocoll, there must be present in oxyhaBma- 

 globin after the first recrystallisation 1 5 per cent of serum-globulin. 



Amongst non- crystalline albumins those precipi table by acids, 

 namely, casein, mucins, etc., are the purest, while the others can only 

 be classified according to differences in solubility and in precipitability 

 by the salting-out method. For this reason serum-globulin is believed 

 to contain 1, 2, 4, and 6, and wheat-glutin 1, 3, and 4 different 

 substances. 



What difficulties are met with in using the salting-out method is 

 described on p. 184. 



In Cohnheim's tables on pp. 70-75 only well-ascertained figures are 

 given, and when these were not available, the occurrence ( -f ) or absence 

 ( - ) of a substance is indicated. For the mono-ammo-acids only those 

 values have been given which have been obtained with pure substances ; 

 in no case have values been included which were determined from crude 

 substances (' Rohfraktion '). If several reliable estimates were available, 

 then the one giving the higher value received preference. 



1 A. Wichmann, Zeitschr. f. physiol. Chem. 27. 575 (1899). 



- F. N. Schulz and K. Zsigmondy, Hofmeisters Beitr. 3. 137 (1902). 



3 E. Abderhalden, Zeitschr. f. physiol. CJiem. 37. 484 (1903). 



