2. All the kinds of factitious air on which I have yet made the experiment are highly noxious, except that which is extracted from saltpetre, or alum; but in this even a candle burned just as in common air[10]. In one quantity which I got from saltpetre a candle not only burned, but the flame was increased, and something was heard like a hissing, similar to the decrepitation of nitre in an open fire. This experiment was made when the air was fresh made, and while it probably contained some particles of nitre, which would have been deposited afterwards. The air was extracted from these substances by heating them in a gun-barrel, which was much corroded and soon spoiled by the experiment. What effect this circumstance may have had upon the air I have not considered. November 6, 1772, I had the curiosity to examine the state of a quantity of this air which had been extracted from saltpetre above a year, and which at first was perfectly wholesome; when, to my very great surprize, I found that it was become, in the highest degree, noxious. It made no effervescence with nitrous air, and a mouse died the moment it was put into it. I had not, however, washed it in rain-water quite ten minutes (and perhaps less time would have been sufficient) when I found, upon trial, that it was restored to its former perfectly wholesome state. It effervesced with nitrous air as much as the best common air ever does; and even a candle burned in it very well, which I had never before observed of any kind of noxious air meliorated by agitation in water. This series of facts, relating to air extracted from nitre, appear to me to be very extraordinary and important, and, in able hands, may lead to considerable discoveries. 3. There are many substances which impregnate common air in a very remarkable manner, but without making it noxious to animals. Among other things I tried volatile alkaline salts, and camphor; the latter of which I melted with a burning-glass, in air inclosed in a phial. The mouse, which was put into this air, sneezed and coughed very much, especially after it was taken out; but it presently recovered, and did not appear to have been sensibly injured. 4. Having made several experiments with a mixture of iron filings and brimstone, kneaded to a paste with water, I had the curiosity to try what would be the effect of substituting _brass dust_ in the place of the iron filings. The result was, that when this mixture had stood about three weeks, in a given quantity of air, it had turned black, but was not increased in bulk. The air also was neither sensibly increased nor decreased, but the nature of it was changed; for it extinguished flame, it would have killed a mouse presently, and was not restored by fixed air, which had been mixed with it several days.