VOL. LXXVIII.] PHILOSOPHICAL TRANSACTIONS. 4^5 



green colour. As an additional proof of the production of volatile alkali, he 

 kept in the same tube some paper, which had been dipped in a solution of 

 cupreous nitre, expecting to see its colour changed from green to blue, by the 

 alkali which was to be produced. The green paper became gradually paler, and 

 m a few days the blue colour appeared. This experiment affords a very satisfac- 

 tory demonstration of the formation of volatile alkali. Water and iron filings 

 mixed together yield inflammable air; but if this be given out in contact with 

 phlogisticated air, volatile alkali is produced. In these circumstances a double 

 attraction takes place: one part of the water is attracted by the iron; the other 

 is attracted by the phlogisticated air; and the water seems by these compound 

 affinities to be much more rapidly decomposed, than when iron and water are 

 mixed by themselves. 



Volatile alkali is formed in a very few hours, if nitrous air be used instead of 

 the phlogisticated, all other circumstances remaining as in the former experiment. 

 When Dr. A. used nitrous acid not well freed from its acid, by which the vege- 

 table blue colour has been turned red, a sufficient quantity of alkali has been gene- 

 rated in 24 hours to change it to a green. Jf iron filings and water be exposed to 

 nitrous air for a considerable time, the nitrous air is so altered that a candle 

 burns in it with increased brightness, as was observed by Dr. Priestley. This 

 change is accounted for by the formation of the alkali, which depriving the 

 nitrous air of its phlogisticated part, leaves a greater proportion of dephlogisti- 

 cated air. 



This experiment also succeeds in atmospheric air, though a longer time is 

 necessary to produce a sensible alteration in the colours employed as tests of the 

 alkali; but the change is very evident in a day or two. Hence we may con- 

 clude, that whenever iron rusts in contact with water in the open air, or in the 

 earth, volatile alkali is formed. Phlogisticated air is present in all parts of the 

 terraqueous globe, and operations are constantly going on, by which inflam- 

 mable air is separated from water, and perhaps from other bodies. Thus we may 

 account for the frequent appearance of volatile alkali in the earth, particularly 

 where inflammable matters abound, among coals and volcanic productions, as 

 also in animal and vegetable substances. 



When iron, water, and sulphur act on each other in atmospheric air, volatile 

 alkali is produced. The eudiometer recommended by Scheele is, for this reason, 

 incorrect. Some phlogisticated air disappears, and volatile alkali is formed. 

 This method therefore seems to have misled that great chemist in his analysis 

 of the atmosphere, and induced him to suppose, that the quantity of phlogis- 

 ticated air in the atmosphere is only 24 times that of dephlogisticated air. 



There is a combination of light inflammable air with sulphur forming hepatic 

 air. It has been observed by the celebrated Mr. Kirwan, that if nitrous air be 



