Whatever methods I took to promote the mixture of fixed and inflammable air, they were all ineffectual. I think it my duty, however, to recite the issue of an experiment or two of this kind, in which equal mixtures of these two kinds of air had stood near three years, as they seem to shew that they had in part affected one another, in that long space of time. These mixtures I examined April 27, 1771. One of them had stood in quicksilver, and the other in a corked phial, with a little water in it. On opening the latter in water, the water instantly rushed in, and filled almost half of the phial, and very little more was absorbed afterwards. In this case the water in the phial had probably absorbed a considerable part of the fixed air, so that the inflammable air was exceedingly rarefied; and yet the whole quantity that must have been rendered non-elastic was ten times more than the bulk of the water, and it has not been found that water can contain much more than its own bulk of fixed air. But in other cases I have found the diminution of a quantity of air, and especially of fixed air, to be much greater than I could well account for by any kind of absorption. The phial which had stood immersed in quicksilver had lost very little of its original quantity of air; and being now opened in water, and left there, along with another phial, which was just then filled, as this had been three years before, viz. with air half inflammable and half fixed, I observed that the quantity of both was diminished, by the absorption of the water, in the same proportion. Upon applying a candle to the mouths of the phials which had been kept three years, that which had stood in quicksilver went off at one explosion, exactly as it would have done if there had been a mixture of common air with the inflammable. As a good deal depends upon the apertures of the vessels in which the inflammable air is mixed, I mixed the two kinds of air in equal proportions in the same phial, and after letting the phial stand some days in water, that the fixed air might be absorbed, I applied a candle to it, but it made ten or twelve explosions (stopping the phial after each of them) before the inflammable matter was exhausted. The air which had been confined in the corked phial exploded in the very same manner as an equal and fresh mixture of the two kinds of air in the same phial, the experiment being made as soon as the fixed air was absorbed, as before; so that in this case, the two kinds of air did not seem to have affected one another at all. Considering inflammable air as air united to, or loaded with phlogiston, I exposed to it several substances, which are said to have a near affinity with phlogiston, as oil of vitriol, and spirit of nitre (the former for above a month), but without making any sensible alteration in it.