I have repeated this experiment two or three times, with similar results; whenever the air was agitated in water immediately after its production, it gave _almost_ the same diminution with nitrous gas as common air; when, on the contrary, it has been suffered to remain for some time in contact with the phlogisticated nitrous acid suspended in it, the condensation has been less with nitrous gas by five or six hundred parts. Hence I am inclined to believe, that if it were possible to condense all the nitrous acid formed, immediately after its generation, so as to prevent it from absorbing oxygene from the permanent gas, this gas would be found identical with the air of the atmosphere. The changes effected by fire on nitrous oxide are not analogous to those produced by it in other bodies; for the power of this agent seems generally _uniform_, either in wholly separating the constituent principles of bodies from each other, or in making them enter into more intimate union.[169] [169] On the one hand, it decomposes ammoniac into hydrogene and nitrogene, whilst on the other, it converts free oxygene and nitrogene into nitrous acid. It likewise converts nitrous gas into nitrous acid and nitrogene. Till we are more accurately acquainted with the nature of heat, light, and electricity, we shall probably be unable to explain these phænomena. It is a singular phænomenon, that whilst it condenses one part of the oxygene and nitrogene of nitrous oxide, in the form of nitrous acid; it should cause the remainder to expand, in the state of atmospheric air. Does not this fact afford an inference in favor of the _chemical_ composition of atmospheric air? III. _Decomposition of Nitrous Oxide by Hydrogene, at the temperature of Ignition._ In the following experiments on the decomposition of nitrous oxide by hydrogene, the gases were carefully generated in the mercurial apparatus, and their purity ascertained by the tests mentioned in Research I. They were measured in small tubes graduated to grains, and then transferred into the detonating tube, which was eight tenths of an inch in diameter, and graduated to ten grain measures. The space occupied by the gases being noted after the inflammation by the electric shock, green muriate of iron, and prussiate of potash, were successively introduced, to ascertain if any nitrous acid had been formed. The absorption, if any took place, was marked, and the gases transferred into a narrow grain measure tube, and their bulk and composition accurately ascertained. _b._ The hydrogene employed was procured from water by means of zinc and sulphuric acid. 50 grain measures of it fired by the electric spark, with 30 grain measures of oxygene containing one eleventh nitrogene, gave a residuum of about 4. Nitrous gas mingled with those 4, indicated the presence of rather less than 1 of unconsumed oxygene. In another experiment 23 of it, with 20 of the same oxygene left rather more than 6 residuum.