That the hydrogen gas and oxygen gas may be as much as possible deprived of water, they are made to pass, in their way to the baloon A, through the tubes MM, NN, of about an inch diameter, and filled with salts, which, from their deliquescent nature, greedily attract the moisture of the air: Such are the acetite of potash, and the muriat or nitrat of lime[18]. These salts must only be reduced to a coarse powder, lest they run into lumps, and prevent the gasses from geting through their interstices.
We must be provided before hand with a sufficient quantity of oxygen gas, carefully purified from all admixture of carbonic acid, by long contact with a solution of potash[19].
We must likewise have a double quantity of hydrogen gas, carefully purified in the same manner by long contact with a solution of potash in water. The best way of obtaining this gas free from mixture is, by decomposing water with very pure soft iron, as directed in Exp. 3. of this chapter.
Having adjusted every thing properly, as above directed, the tube H h is adapted to an air-pump, and the baloon A is exhausted of its air. We next admit the oxygen gas so as to fill the baloon, and then, by means of pressure, as is before mentioned, force a small stream of hydrogen gas through its tube D d', which we immediately set on fire by an electric spark. By means of the above described apparatus, we can continue the mutual combustion of these two gasses for a long time, as we have the power of supplying them to the baloon from their reservoirs, in proportion as they are consumed. I have in another place[20] given a description of the apparatus used in this experiment, and have explained the manner of ascertaining the quantities of the gasses consumed with the most scrupulous exactitude.
In proportion to the advancement of the combustion, there is a deposition of water upon the inner surface of the baloon or matrass A: The water gradually increases in quantity, and, gathering into large drops, runs down to the bottom of the vessel. It is easy to ascertain the quantity of water collected, by weighing the baloon both before and after the experiment. Thus we have a twofold verification of our experiment, by ascertaining both the quantities of the gasses employed, and of the water formed by their combustion: These two quantities must be equal to each other. By an operation of this kind, Mr Meusnier and I ascertained that it required 85 parts, by weight, of oxygen, united to 15 parts of hydrogen, to compose 100 parts of water. This experiment, which has not hitherto been published, was made in presence of a numerous committee from the Royal Academy. We exerted the most scrupulous attention to its accuracy; and have reason to believe that the above propositions cannot vary a two hundredth part from absolute truth.