262 HISTORICAL INTRODUCTION TO CHEMISTRY CHAP. 



[azote] is suddenly liberated and the metal is revived" 

 (ibid. p. 321). 



Although the actual conversion into nitre of the nitrogen 

 of the volatile alkali was not realised in this experiment, 

 Berthollet was able to quote two cases in which the converse 

 change had been effected, since ammonia had actually been 

 prepared by the action of zinc on a solution of nitre, and by 

 the action of a fixed alkali on the products of interaction of 

 tin and nitric acid. "It is necessary to note that in each 

 case nitric acid was present which must have provided the 

 [azote]" (ibid. p. 326). 



Berthollet's analysis of ammonia. In a paper published 

 in the Journal de Physique for April, 1786, vol. 28, 

 p. 273, Berthollet compared the weight of ammonia driven 

 off by gentle heat from fulminating gold, with the weight 

 of water produced when the fulminate was detonated ; 

 he " concluded that the inflammable gas formed almost 

 one-sixth by weight ... of the volatile alkali," the pro- 

 portion now accepted as correct being 3/17 instead of 

 3/18. In 1785 J he carried out a volumetric analysis which 

 showed that two volumes of ammonia gas were increased 

 to four volumes by sparking, and that these four volumes 

 consisted of nearly three parts of hydrogen with one part of 

 nitrogen. 



" We made some alkaline gas from a mixture of one part of 

 sal-ammoniac and three to four parts of freshly-burnt lime, 

 so that it should be free from water . . . : we introduced 

 it into a glass tube provided with an exciter; the space 

 which it occupied was ... 1*7 cubic inches. It was 

 electrified until an increase in volume was no longer 

 perceptible ; after that, a little water was introduced into 

 the tube, and no absorption was perceived, although it was 

 shaken, so that the alkaline gas was completely decomposed ; 

 the increase of its volume was i '6 cubic inches ; then the 



1 The paper was not published till 1788. 



