204 New Determination of the Proportions 
every possible facility in our design, by putting at our disposal 
the laboratory of Arcueil. 
She apparatus employed in the first experiments on the com~- 
posidion of water, did not allow of that precision which is now 
required in chemical analysis. But the fact being once esta~ 
blished, that water is the result of the combination of oxygen and 
hydrogen, the knowledge of the precise proportions required only 
two facts to be determined; namely, the relative volumes of the 
two elements, and their specific gravities. The latter, being in- 
dispensable in a variety of researches, were already known ; and 
for the former, Volta’s eudiometer was sufficient. The greatest 
confidence was justly reposed on this method, after Messrs. Gay- 
Lussac and Humboldt had shown, in their masterly Memoir on 
Eudiometry, the true proportions, in volume, of the constituents 
of water; and after Messrs. Biot and Arrago had applied the 
most minute attention to the determination of the specific gra- 
vity of the greater number of the gases. 
If the proportion of hydrogen deduced from these results, 
namely, 13.27 per cent., was erroneous, the error was in esti- 
mating the specific gravity of either the hydrogen or the oxygen, 
or both; for the computation of respective volumes has this re- 
markable advantage, that, being dependent on a general law, it 
is incapable of error. Before we entered on new observations on 
the densities of oxygen aud hydrogen, we wished to obtain, bya 
simple method, the confirmation of our doubts. The decompo- 
sition of an oxide by hydrogen appeared to us the most accurate 
and convenient way ; and we adopted the following precautions 
to render the experiment conclusive : 
We first procured perfectly pure hydrogen gas. Distilled zine 
is not preferable for this object to common zine, fer they both 
contain the same impurities, namely, lead, tin, copper, iron, cad- 
mium, and sulphur; but on passing the hydrogen through a tube 
containing fragments of caustic potash slightly moistened, the 
gas loses its odour completely, and comes out perfectly pure. 
Hydrogen gas, obtained by the action of diluted sulphuric acid 
on zine, was purified by sending the current through moistened 
fragments of caustic potash. It was then dried by being trans- 
mitted through muriate of lime; after which it was placed in 
contact with oxide of copper dried and inclosed in a tube, which 
was united to the apparatus by two tubes of elastic gum, which 
allowed us to weigh it accurately both before and after the ex- 
periment. When the hydrogen had passed in sufficient quantity 
to expel the atmospheric air, the oxide of copper was heated by 
a spirit lamp. In the first experiments, the greater part of the 
newly generated water was reccived in a liquid state in a small 
reci- 
