To render the success of this experiment certain, it is necessary that the tube EF be made of well annealed and difficultly fusible glass, and that it be coated with a lute composed of clay mixed with powdered stone-ware; besides which, it must be supported about its middle by means of an iron bar passed through the furnace, lest it should soften and bend during the experiment. A tube of China-ware, or porcellain, would answer better than one of glass for this experiment, were it not difficult to procure one so entirely free from pores as to prevent the passage of air or of vapours.
When things are thus arranged, a fire is lighted in the furnace EFCD, which is supported of such a strength as to keep the tube EF red hot, but not to make it melt; and, at the same time, such a fire is kept up in the furnace VVXX, as to keep the water in the retort A continually boiling.
In proportion as the water in the retort A is evaporated, it fills the tube EF, and drives out the air it contained by the tube KK; the aqueous gas formed by evaporation is condensed by cooling in the worm SS, and falls, drop by drop, into the tubulated bottle H. Having continued this operation until all the water be evaporated from the retort, and having carefully emptied all the vessels employed, we find that a quantity of water has passed over into the bottle H, exactly equal to what was before contained in the retort A, without any disengagement of gas whatsoever: So that this experiment turns out to be a simple distillation; and the result would have been exactly the same, if the water had been run from one vessel into the other, through the tube EF, without having undergone the intermediate incandescence.
Experiment Second.
The apparatus being disposed, as in the former experiment, 28 grs. of charcoal, broken into moderately small parts, and which has previously been exposed for a long time to a red heat in close vessels, are introduced into the tube EF. Every thing else is managed as in the preceding experiment.
The water contained in the retort A is distilled, as in the former experiment, and, being condensed in the worm, falls into the bottle H; but, at the same time, a considerable quantity of gas is disengaged, which, escaping by the tube KK, is received in a convenient apparatus for that purpose. After the operation is finished, we find nothing but a few atoms of ashes remaining in the tube EF; the 28 grs. of charcoal having entirely disappeared.
When the disengaged gasses are carefully examined, they are sound to weigh 113.7 grs.[16]; these are of two kinds, viz. 144 cubical inches of carbonic acid gas, weighing 100 grs. and 380 cubical inches of a very light gas, weighing only 13.7 grs. which takes fire when in contact with air, by the approach of a lighted body; and, when the water which has passed over into the bottle H is carefully examined, it is found to have lost 85.7 grs. of its weight. Thus, in this experiment, 85.7 grs. of water, joined to 28 grs. of charcoal, have combined in such a way as to form 100 grs. of carbonic acid, and 13.7 grs. of a particular gas capable of being burnt.
I have already shown, that 100 grs. of carbonic acid gas consists of 72 grs. of oxygen, combined with 28 grs. of charcoal; hence the 28 grs. of charcoal placed in the glass tube have acquired 72 grs. of oxygen from the water; and it follows, that 85.7 grs. of water are composed of 72 grs. of oxygen, combined with 13.7 grs. of a gas susceptible of combustion. We shall see presently that this gas cannot possibly have been disengaged from the charcoal, and must, consequently, have been produced from the water.
I have suppressed some circumstances in the above account of this experiment, which would only have complicated and obscured its results in the minds of the reader. For instance, the inflammable gas dissolves a very small part of the charcoal, by which means its weight is somewhat augmented, and that of the carbonic gas proportionally diminished. Altho' the alteration produced by this circumstance is very inconsiderable; yet I have thought it necessary to determine its effects by rigid calculation, and to report, as above, the results of the experiment in its simplified state, as if this circumstance had not happened. At any rate, should any doubts remain respecting the consequences I have drawn from this experiment, they will be fully dissipated by the following experiments, which I am going to adduce in support of my opinion.