108 



tfiuperature was i-iiised to betweeu 22(J° aud 'I'^O' uud maiutaiued there 

 until the gas was about one-half evolved. The temperature was then 

 lowered to 170° and kept there until the action was nearly complete, when 

 it was again raised to 230° as long as a gas was being evolved. Finally 

 it was cooled to 170° and the burette disconnected. Any oxygen in the 

 burette was absorbed by pyrogullate (usually a small amount) and the 

 nitrogen measured. Nitrogen was passed through the generator to sweep 

 any acid vapors into the water, and the amount of acid determined by 

 titrating against standard alliali. The residue was extracted several times 

 with boiling water and the water evaporated in a platinum basin. The 

 small amount of solid found on evaporation consisted mostly of unchanged 

 ammonium nitrate. 



In three closely agreeing experiments carried out as described the 

 following figures were obtained : 



1. 2.0079 gms. ammonium nitrate gave 500.5 cc. (corr.) (=0.0286 

 gms.) free nitrogen, and 0.7.^52 gms. acid (calculated as nitric). 



2. 2.8955 gms. ammonium nitrate gave 635.6 cc. (corr.) (=0.7983 

 gms.) free nitrogen and 0.9998 gms. acid. 



3. 3.8527 gms. ammonium nitrate gave 820 cc. (corr.) (=1.025 gms.) 

 free nitrogen and 1.1000 gms. acid. 



In these experiments the average ratio of free nitrogen to acid is 

 1 :1.15. This ratio approaches very nearly to that for two molecules of 

 nitrogen to one of the acid, viz., 1:1.125. The equation which best corre- 

 sponds to this ratio is as follows : 



5 NH,N03=2HN03+4N2-h9H,0. 



Reiset and Mellon have shown (Journ. fur Practische Chemie, 29-305) 

 that when ammonium nitrate is mixed with platinum sponge and heated, 

 decomposition begins at 160° and that the products are nitrogen and nitric 

 acid. The equation by which they evpress the reaction is identical with 

 that given above. It is unlikely that the itlatinum enters into the reac- 

 tion, though It is stated that an insoluble platinum compound is produced. 

 I have not been able to confirm this latter statement. If the platinum does 

 not enter intv) the reaction, but acts as a true contact agent, then there 

 seems no reason for believing that the manganese dioxide, in this reaction 

 at least, acts in a chemical way. 



It is possible that an intermediate product containing manganese may 

 be isolated ; and this will be the object of further research. 



