72 DR MATTHEW HAY ON THE 



These analyses are amply sufficient to show that the amount of nitrous acid 

 formed during the alkaline decomposition of nitroglycerine is neither small 

 nor variable ; and, assuming that nitroglycerine is a trinitrate of glyceryl, it 

 corresponds remarkably with the supposition that two out of tLe three parts 

 of nitric anhydride, which nitroglycerine contains, are reduced to nitrous 

 anhydride ; for the trinitrate of glyceryl ought theoretically to yield, if so 

 reduced, 33 48 per cent., an amount which agrees very closely with that 

 actually obtained, if clue allowance be made for experimental error in a method 

 which, although the best available, cannot claim to be exact. 



It is open to suggestion, reasoning alone from these estimations of nitrous acid, 



that nitroglycerine is perhaps a di-nitrite of glyceryl, or a mono-nitrate di-ni trite 



of glyceryl. As opposed to its being one or other of those bodies, the fact 



that specimens of nitroglycerine, as B and D, prepared in presence of urea, were 



found to yield the same proportion of nitrous acid as the others, is of importance. 



Another weighty objection to its being a nitrite is that on passing nitrous 



anhydride gas into glycerine no substance at all resembling nitroglycerine was 



obtainable ; although there was formed an oily liquid containing nitrous acid 



in combination, which, however, quickly decomposed in contact with water. 



This body has recently been investigated by Mr Masson,* and he believes it to 



be the tri-nitrite of glyceryl. It is not probable that the di-nitrite will possess 



greatly different properties. Were nitroglycerine the di-nitrite, it ought to 



yield a much higher proportion of nitrous acid than was actually obtained. 



For these and various other reasons, it is not the di-nitrite, and much less the 



tri-nitrite. As to the possibility of its being a mono-nitrate di-nitrite, the 



objection as to the yield of nitrous acid is not by any means strong. For 



N 

 such a body ought theoretically to yield 30 - 89 per cent, of 2 9 3 (nitrous 



anhydride), an amount tolerably close to what was actually obtained. And 

 amongst the alkaline decomposition products of nitroglycerine, it is not difficult 

 to separate nitrate of potassium by crystallisation. But, on the other hand, 

 the mono-nitrate di-nitrite ought to yield 21 - 5 per cent, of nitrogen, whereas, 

 by actual experiment, Mr Masson and myself have found that nitroglycerine 

 contains a much lower percentage of nitrogen. 



From these and other considerations, which will be referred to later on, it 

 is impossible to avoid concluding that nitroglycerine is a tri-nitrate of glyceryl, 

 and that two-thirds of its nitric acid is reduced to nitrous acid during the 

 decomposition of the ether.t 



I shall now give a brief account of the other substances, besides nitrite of 

 potassium, which are formed when an alcoholic solution of potash acts on an 

 alcoholic solution of nitroglycerine. 



* Masson, Journ. Chem. Soc, August 1 883. 



+ Vide accompanying communication, on " The Elementary Composition of Nitroglycerine." 



