22 PROCEEDINGS OP THE AMERICAN ACADEMY 



glass bottle warmed in a water-bath, which lasted through a number 

 of preparations. The delivery-tube should not dip below the surface 

 of the aniline, as in that case there is danger that it will be stopped 

 by the product ; but if it is brought near the surface of the base, the 

 action takes place so rapidly that very little of the fluoride of silicon 

 is lost. A good deal of heat is given out during the reaction, and the 

 aniline is converted into a loose white solid, which was washed with 

 hot ligroine until free from aniline, and then its purification finished 

 by two sublimations. The yield was essentially quantitative, 30 grams 

 of aniline giving after treatment with fluoride of silicon for 24 hours 

 51 grains of product, instead of 52 grams, the amount which should be 

 obtained for the formula* (C c H.NH 2 ) 3 (SiF 4 ) 2 . The same substance 

 is formed when aniline fluosilicate is sublimed, and the preparation, 

 the analysis of which is numbered I., was made in this way. It is to 

 be observed that the substance analyzed by Laurent and Delbos was 

 really prepared in this way, since by boiling their original product 

 with alcohol they converted it into aniline fluosilicate, which was after- 

 ward reconverted into the silicotetrafluoride by sublimation. Also the 

 substance obtained by Knop by sublimation of his auiline fluosilicate 

 was the trianiline disilicotetrafluoride. 



The method of analysis used for all these substances consisted in 

 neutralizing a weighed quantity of the substance dissolved in hot 

 water in a platinum dish with a standard solution of sodic hydrate, 

 using a solution of litmus as the indicator. The liquid was then 

 heated to boiling, more of the sodic hydrate added, if the reaction had 

 become acid, and evaporated to dryness on the water-bath, the residue 

 treated with water, and the silicic dioxide filtered out. The filtrate, 

 after neutralizing once more with the sodic hydrate, which is usually 

 necessary when the organic base is one with an alkaline reaction, 

 (NH 3 or (CH 3 ) 2 NH), is treated with a solution of zincic oxide in 

 amnionic carbonate, evaporated once more to dryness on the water- 

 bath, treated with water, and filtered ; the precipitate is dissolved in 

 strong nitric acid, evaporated to dryness, the residue after treatment 

 with strong nitric acid extracted with water, and the silicic dioxide 

 thus obtained added to that from the residue of the first evaporation, 

 ignited, and weighed. The fluorine was usually calculated from the 

 amount of the standard solution of sodic hydrate necessary for the 

 neutralization of the hydrofluoric acid present, but it was also occa- 



* Laurent and Delbos found that 59.5 grm. of aniline absorbed 40.5 grm. 

 of fluoride of silicon. Our formula requires 44.3 grm. 



