Preduilion of Air by freezing Water. 195 



After the breaking of my glafs veflels, I got other cylindrical ones made of iron, feven or 

 eight inches in height, and near three inches wide at the bottom, the upper orifice being 

 clc/"cd with a cork and cement, in the centre of which was a glafs tube, the diameter of 

 •which was about the fifth of an inch. And as this glafs tube was in the greateft danger of 

 breaking by the freezing of the water, and this had happened feveral times in the former 

 experiments, notwithftanding all my care to guard them from the froft, I now made ufe of 

 fnow and fait to freeze the water in the iron veflll only, placed in a veffel of rtiercury, 

 having been previoufly filled as the glafs veflels had been. 



The water on which I now operated was about three ounces, and it had been made as 

 free as poflible from air by previous freezing. With this apparatus I repeated the procefs 

 of freezing nine times, without changing the water ; and the laft portion of air that I 

 procured in this manner was as great as any of the preceding ; fo that there remained no 

 reafonable doubt, but that air might be produced from the fame water, in this manner, ad 

 libitum ; and having got near two inches of air in the glafs tube, I put an end to the ex- 

 periment; and examining the air, I found it to he wholly pklogi/iicated, not being afTedled 

 by nitrous air, and having nothing inflammable in it. 



During the procefs of freezing, the air in the tube was generally comprefled Into one- 

 fiftli of its ufual bulk-; but when I began to thaw the ice, which I did by means of hot 

 water in the place of the freezing mixture, it foon expanded to its former dimcnfions, and 

 no fenfible portion of it was abforbed during the whole procefs, which was about a week. 

 Sometimes the violence of the preflure, occafioned by the expanfion of the water in 

 freezing, would force a little water out of the veflel between the cork and the glafs tube, 

 or the Iron veflel, which prefently became ice. This I always carefully removed, and ap- 

 plied frefh cement to the place, to prevent the introduction of any air from without, before 

 I began to melt the ice. And that no external air had entered, was evident both from the 

 manner in which the air was produced after the water recovered Its fluidity, and from the 

 quality of it when examined after the procefs. 



In the courfe of the experiments with the glafs veflels, a phenomenon occurred which 

 was wholly unexpeded by me, and which was very amufing. Having left the veflels filled 

 In part with water, and in part with mercury, in the evening, I generally found them in 

 the morning feemingly quite full of mercur/, every part of the Ice within the veflel being 

 covered with it. This mud have been occafioned by a vacuum having been formed 

 between the glafs and the ice, and into this fpace mercury had been drawn up on the 

 principle of the capillary tube. When this was not the cafe, the interftlces of the Ice to- 

 wards the centre were filled with thin laminae of mercury, which alfo exhibited a curious 

 appearance. 



Sometimes, when there was no mercury between the glafs and the ice, an mterillee 

 was made between them when they were placed within the influence of the fire. In thefc 

 circumftances 1 have feen the mercury drawn up to the height of feveral inches. As this 

 fpace became enlarged by the increafc of the heat, the laijiinje of mercury were con- 



Cca traftcd. 



