on the fixed Alkalies. 
37 
This circumstance seems distinctly to prove the formation 
of water in this operation for the decomposition of ammonia ; 
unless indeed it be asserted that the hydrogene and nitrogene 
gasses evolved hold less water in solution or suspension than 
the ammonia decomposed, an idea strongly opposed by the 
conclusions of Mr. Dalton* and the experiments of Messrs. 
DESORMES and CLEMENT.-f 
After the gas had been passed several times through the 
ignited tube from one gazometer to the other, the results were 
examined. The iron wire became converted superficially into 
oxide, and had gained in weight AA. parts of a grain, about 
of a grain of water were collected from the cooled glass 
tubes by means of filtrating paper, and 33.8 cubic inches of 
gas were expanded into 55.3 cubic inches, and by detonation, 
with oxygene it was found that the hydrogene gas in these 
was to the nitrogene as 3.2. to 1 in volume, 
It will be useless to enter into the more minute details of 
this experiment, as no perfectly accurate data for proportions 
can be gained from them ; for the whole of the ammonia was 
not decomposed, and as the gas had been prepared by being 
sent from a heated mixture of sal ammoniac and quicklime, 
into the airholder, it was possible that some solution of 
ammonia might have been deposited, which, by giving out new 
gas during the operation, would increase the absolute quantity 
of the material acted upon. 
In examining the results of M. Berthollet’s J elaborate 
experiments on the decomposition of ammonia by electricity, I 
# Manchester Memoirs, Vol. V. Part II. page. 535, 1785. 
f Annales de Chemie, Vol. XLII. p. 125. 
t Memoir es de V Academic, 1785, page 324. 
