172 



less nitrogen ; whereas, on heating the same substance in a 

 tube of ^j^a/zww/», the potassium alloyed with the platinum, and 

 the ammonia was given off almost entirely undecomposed. 



How can these experiments be explained except upon 

 the supposition that the potassium and the iron had con- 

 jointly decomposed the nitrogen % The latest experiments 

 which bear upon this subject, and from which I received the 

 idea which led me to this investigation, are those of Doctor 

 Brown, " upon the conversion of Carbon into Silicon," an 

 explanation of phenomena which appears to me most un- 

 reasonable, and contrary to all chemical analogy ; whilst the 

 supposition of the carbon having reduced the nitrogen is not 

 only a simple but an unavoidable conclusion to arrive at, if 

 nitrogen be a compound substance. To determine, by ex- 

 periment, the correctness or incorrectness of this idea, it were 

 only necessary to reduce nitrogen by some other substance 

 than charcoal ; and should silica result from its decomposition, 

 the problem might be considered to be solved. 



Exp. I. — A considerable quantity of ammonia-nitruret of 

 potassium was formed, by passing ammonia over potassium 

 heated in an iron tube ; the part which had not been in con- 

 tact with the tube, having been examined for silica, contained 

 none. 



Exp. II. — Ammonia was passed for several hours over 

 pure iron, heated to a dull red heat ; examined for silica, it 

 contained none. 



Exp. III. — Ammonia-nitruret of potassium was heated 

 with pure iron in an iron crucible, for one half hour, over a 

 large Rose's lamp ; the contents of the crucible, on examina- 

 tion, gave silicon and silica, the weight of which was not re- 

 gistered, as it might have been said to have derived a portion 

 of silica from the inner surface of the crucible. 



Exp. IV. — Twenty grains of ammonia-nitruret of potas- 

 sium were heated with twenty grains of pure iron in the 

 same iron vessel for one half hour ; when treated with nitric 



