xni THE DECOMPOSITION OF THE ALKALIS 263 



gas was introduced into a Volta eudiometer, mixing each 

 measure successively with vital air, and detonating after each 

 addition, in the following manner : 



" (i) Two measures of vital air. 



One measure of decomposed alkaline gas. 



(2) One measure of vital air. 

 One measure of gas. 



(3) Two measures of vital air. 

 One measure of gas. 



(4) One measure of gas. 



Total: five measures of vital air and four measures 

 of gas. 



"These nine measures were reduced to 4*6 measures, so 

 that 4 '4 measures were destroyed. Care was taken, in this 

 experiment, to put an excess of vital air in order to be sure 

 of destroying all the combustible part ; but one knows that 

 combustion causes 145 measures of inflammable gas to 

 disappear for 74 of vital air. It follows then that the 4*4 

 measures destroyed represent 2*9 measures of inflammable 

 gas and 1*5 measures of vital air. The four measures of 

 electrified gas which were tested, would then represent 2*9 

 measures of inflammable gas and i'i of [azote]" (Mem. 

 Acad. Set., Paris, 1785, 99, p. 324). 



St. Claire Deville in 1865 showed that the decomposition 

 of ammonia into nitrogen and hydrogen is not quite complete 

 (see Chapter XX). The decomposition is, in fact, another 

 example of a balanced action, as may be indicated by the 

 reversible equation, 



Ammonia ^ nitrogen 4- hydrogen. 



The proportion of ammonia which persists is so small 

 that Berthollet failed to detect any contraction of volume 

 on adding water to the gas. 1 



The composition of sal-ammoniac. When Priestley had 

 prepared " acid air " and " alkaline air," he tried the experi- 



1 The evaporation of water into the dry gas would produce an 

 increase of volume which would compensate for the decrease of volume 

 due to the absorption of a little undecomposed ammonia. 



