It is not less extraordinary, that a still longer continuance of nitrous air in these circumstances (but _how long_ depends upon too many, and too minute circumstances to be ascertained with exactness) makes it not only to admit a candle to burn in it, but enables it to burn with an _enlarged flame_, by another flame (extending every where to an equal distance from that of the candle, and often plainly distinguishable from it) adhering to it. Sometimes I have perceived the flame of the candle, in these circumstances, to be twice as large as it is naturally, and sometimes not less than five or six times larger; and yet without any thing like an _explosion_, as in the firing of the weakest inflammable air. Nor is the farther progress in the transmutation of nitrous air, in these circumstances, less remarkable. For when it has been brought to the state last mentioned, the agitation of it in fresh water almost instantly takes off that peculiar kind of inflammability, so that it extinguishes a candle, retaining its noxious quality. It also retains its power of diminishing common air in a very great degree. But this noxious quality, like the noxious quality of all other kinds of air that will bear agitation in water, is taken out of it by this operation, continued about five minutes; in which process it suffers a farther and very considerable diminution. It is then itself diminished by fresh nitrous air, and animals live in it very well, about as well as in air in which candles have burned out. Lastly, One quantity of nitrous air, which had been exposed to iron in quicksilver, from December 18 to January 20, and which happened to stand in water till January 31 (the iron still continuing in the phial) was fired with an explosion, exactly like a weak inflammable air. At the same time another quantity of nitrous air, which had likewise been exposed to iron, standing in quicksilver, till about the same time, and had then stood in water only, without iron, only admitted a candle to burn in it with an enlarged flame, as in the cases above mentioned. But whether the difference I have mentioned in the circumstances of these experiments contributed to this difference in the result, I cannot tell. Nitrous air treated in the manner above mentioned is diminished about one fourth by standing in quicksilver; and water admitted to it will absorb about half the remainder; but if water only, and no quicksilver, be used from the beginning, the nitrous air will be diminished much faster and farther; so that not more than one fourth, one sixth, or one tenth of the original quantity will remain. But I do not know that there is any difference in the constitution of the air which remains in these two cases.