On so-called " Protagon " 301 



show that, whilst it is possible to retain approximately the composition of 

 " protagon " during recrystallisation, this achievement is in no way due to 

 the prevention of a hypothetical decomposition. The great variations in 

 the analytical figures usually obtained for " protagon " are explained in the 

 simplest manner. There is, therefore, no need t6 take refuge in the far- 

 fetched " decomposition " theory which, besides not being confirmed by our 

 critical re-examination of Wilson and Cramer's statements as to the 

 physical constants of "protagon," is also a priori most improbable in view 

 of the fact that alcohol is not a hydrolytic agent. 



The Isolation of the Constituents of '"Protacon" 



BY MEANS of PYRIDINE. 



A generally recognised method of showing the uniformity of any 

 substance consists in subjecting the same to fractional crystallisation or 

 precipitation.^ Wilson and Cramer considered the results of our previous 

 fractionations as being produced ■ by " decompasition." Although this 

 criticism has now been shown to be unfounded, we thought it nevertheless 

 advisable to repeat our fractionation experiments under conditions which 

 make a decomposition in Wilson and Cramer's sense impossible. This 

 can be easily done, as already stated in a previous communication,^ by using 

 a mixture of inert solvents. 



It is remarkable that Wilson and Cramer refrain from employing 

 fractional crystallisation as a test for the uniformity of their new " pro- 

 tagon," especially as the solvent (pyridine) chosen by them for the deter- 

 mination of the supposed physical constants of " protagon " lends itself 

 admirably to this purpose. A decomposing influence at the low temperature 

 in question is not to be assumed, and pyridine would not have been 

 employed by them as their standard solvent if such an influence had 

 been suspected. 



" Protagon " is fairly soluble in pyridine at 30" to 45° C, and a precipitate 

 is formed on cooling its solution. Evidently, if " protagon " is of a uniform 

 composition, the precipitate must again consist of " protagon " with 1 per 

 cent, phosphorus. Although we hardly expected this from our previous 

 experience with this substance, we were nevertheless surprised to find on 

 analysis that the precipitate contained more than double the amount of 

 phosphorus, namely, 2 5 per cent. This result led us to a systematic 

 examination of the fractions into which " protagon " can evidently be 

 divided by means of pyridine. 



For this purpose a 3 per cent, solution in pyridine of a recrystallised 

 "protagon," prepared by Wilson and Cramer's modification of Gamgee's 

 method, was employed. The solution was effected at 45° C, and the 

 temperature kept at 45" for only one minute. The perfectly clear solution 



' See H. Meyer, Analyse, etc., organ. Verbindungen, p. 13. 

 * Proc Physiol. Soc, xxxvii., 1908, p. 1. 



