It being evident that the ether occasioned the difference that was observable in these two cases, I next proceeded to take the electric spark in a quantity of ether only, without any air whatever; and observed that every spark produced a small bubble; and though, while the sparks were taken in the ether itself, the generation of air was slow, yet when so much air was collected, that the sparks were obliged to pass through it, in order, to come to the ether and the quicksilver on which it rested, the increase was exceedingly rapid; so that, making the experiment in small tubes, as fig. 16, the quicksilver soon receded beyond the striking distance. This air, by passing through water, was diminished to about one third, and was inflammable.
One quantity of air produced in this manner from ether I suffered to stand two days in water, and after that I transferred it several times through the water, from one vessel to another, and still found that it was very strongly inflammable; so that I have no doubt of its being genuine inflammable air, like that which is produced from metals by acids, or by any other chemical process.
Air produced from ether, mixed both with common and nitrous air, was likewise inflammable; but in the case of the nitrous air, the original quantity bore a very small proportion to the quantity generated.
Concluding that the inflammable matter in this air came from the ether, as being of the class of oils, I tried other kinds of oil, as oil of olives, oil of turpentine, and essential oil of mint, taking the electric spark in them, without any air to begin with, and found that inflammable air was produced in this manner from them all. The generation of air from oil of turpentine was the quickest, and from the oil of olives the slowest in these three cases.
By the same process I got inflammable air from spirit of wine, and about as copiously as from the essential oil of mint. This air continued in water a whole night, and when it was transferred into another vessel was strongly inflammable.
In all these cases the inflammable matter might be supposed to arise from the inflammable substances on which the experiments were made. But finding that, by the same process I could get inflammable air from the volatile spirit of sal ammoniac, I conclude that the phlogiston was in part supplied by the electric matter itself. For though, as I have observed before, the alkaline air which is expelled from the spirit of sal ammoniac be inflammable, it is so in a very slight degree, and can only be perceived to be so when there is a considerable quantity of it.
Endeavouring to procure air from a caustic alkaline liquor, accurately made for me by Mr. Lane, and also from spirit of salt, I found that the electric spark could not be made visible in either of them; so that they must be much more perfect conductors of electricity than water, or other fluid substances. This experiment well deserves to be prosecuted.
I observed before that inflammable air, by standing long in water, and especially by agitation in water, loses its inflammability; and that in the latter case, after passing through a state in which it makes some approach to common air (just admitting a candle to burn in it) it comes to extinguish a candle. I have since made another observation of this kind, which well deserves to be recited. It relates to the inflammable air generated from oak the 27th of July 1771, of which I have made mention before.
This air I have observed to have been but weakly inflammable some months after it was generated, and to have been converted into pretty good or wholesome air by no great degree of agitation in water; but on the 27th of March 1773, I found the remainder of it to be exceedingly good air. A candle burned in it perfectly well, and it was diminished by nitrous air almost as much as common air.
I shall conclude this section with a few miscellaneous observations of no great importance.