298 Rosenheim and Tebb 



explained fully in the other papers just referred to, has not been mentioned 

 by Wilson and Cramer. It is a property of any " protagon," whether 

 the same has been subjected to the supposed decomposing influence of 

 boiling alcohol or not. The value [a]i, = (+ or — )? 133° of Wilson and 

 Cramer is therefore in no way characteristic for " decomposed protagon," 

 and this fact alone deprives the whole decomposition theory of its 

 foundation. 



During these investigations we have also obtained further chemical 

 evidence, if such should be still wanted, which proves clearly that our 

 previous conclusions with regard to the composite nature of " protagon " 

 also apply to the new standard "protagon " of Wilson and Cramer. 



The Composition of "Protagon" is completely changed by 

 Recrystallisation from Alcohol. 



The main chemical fact on which Wilson and Cramer rely for their 

 view that " protagon " is of a definite composition is the possibility (which 

 has never been denied) of its recrystallisation from alcohol without change 

 in its composition. The one and only condition which must be fulfilled 

 to attain this end consists, according to them, in limiting the exposure to 

 the hot solvent to a short period of time. 



It is clear, however, that this fact, even if the above condition was the 

 correct one for its achievement, would in no way prove the definite com- 

 position of "protagon"; for a mixture of substances which possess 

 approximately the same solubility in a given solvent, may also retain its 

 relative composition under these conditions. This is well illustrated by 

 the case of phytosterin, which since its discovery by Hesse ^ has also 

 been assumed to be a substance of definite chemical and physical properties, 

 a -product of constant composition, melting point, and optical activity being 

 always obtained by recrystallisation from alcohol. Windaus and Hauth^ 

 recognised, however, that Hesse's product was a mixture of two substances. 



In the recrystallisation of " protagon " there is, however, another factor 

 which has been completely neglected by Wilson and Cramer. We have 

 looked in vain to find in their paper any statement as to the proportion of 

 " protagon " to alcohol employed for the recrystalli.sation. This proportion 

 is evidently of no importance if one is dealing with a definite chemical 

 substance such as cholesterin. Whatever the amount of solvent used, the 

 recrystallised product will always be cholesterin. But in the case of 

 a mixture of substances the amount of solvent used for recrystallisation is 

 obviously of the greatest importance. 



We found that, by simply varying the proportion of " protagon " to 

 alcohol, we obtained variations in the phosphorus percentage of " protagon " 

 of over 50 per cent., whilst taking care at the same time to avoid the 



' Anil. d. Chem., 192, 1878, p. 175. 



' Ber. d. d. chem. Ges., 39, 1906, p. 4.378. 



