36 PROCEEDINGS OF THE AMERICAN ACADEMY 



little force. On the other hand, we have not been able to find any 

 explanation for the stability of the compound derived from ammonia, 

 as the only one we could think of — viz. that the ammonia being a 

 stronger base than aniline would attach itself more firmly to the 

 slightly acid molecule — is rendered inadmissible by the slight stability 

 of the compound made from dimethylamine, a base nearly, if not quite, 

 as strong as ammonia itself. It is possible that the ammonia com- 

 pound has an entirely different constitution from the compounds of the 

 organic bases, but we have no experimental material for testing the cor- 

 rectness of this hypothesis except Mixter's determination * of the vapor 

 density of this substance, which showed that it was dissociated into 

 four volumes of ammonia and two of fluoride of silicon, and therefore 

 that the simplest formula of the ammonia compound must be doubled, 

 which would look as if it had a constitution similar to the organic 

 compounds. 



We may add that Harden f obtained by the action of chloride of 

 silicon on pyridine, or chinoline compounds (C 5 H 5 N) 2 SiCl 4 , and 

 (C 9 H 7 N) 2 SiCl 4 ; but as they give up chloride of silicon spontaneously, 

 it is probable that they are not analogous in constitution to our sub- 

 stances. 



The study of the action of fluoride of silicon on organic bases will 

 be continued by one of us in this Laboratory. 



* Am. Chem. J., ii. 153. 



t J. Lond. Chem. Soc, 1887, i. 40. 



