Formation of Volatile Alkali . 385 
air the parts of inflammable air are brought nearer to each 
other than they are in their Ample aeriform date*, and there- 
fore the phlogidlcated air of the nitrous air combines with 
them, and generates volatile alkali. 
From all thefe experiments it follows, that whether phlo- 
gifticated air be in a date of purity, or mixed with dephlo- 
gifticated air, as in the atmofphere, or combined with it as in 
nitrous air, it will in either cafe unite with the gravitating mat- 
ter of light inflammable air, provided this fubflance he pre- 
fented to it in a date of condenfation ; hut if the circum- 
dances be reverfed, the fame combination does not take place. 
No union is formed between inflammable air and the pblogidi- 
cated part of nitrons air, even though marine acid be added, 
which, by its attra&ion to dephlogidicated air, would contri- 
bute to decompofe the nitrous air, and by its attraction to 
volatile alkali would tend to unite its condituent parts : or if 
to light inflammable air we add nitrous air and iron filings, 
no combination enfues ; though it has been often obferved, 
that volatile alkali is readily generated, when nitrous air is 
prefented to the inflammable at the inflant of its extrication 
from water and iron. 
The proportions of the phlogiflicated and inflammable airs in 
volatile alkali, as difcovered by calculation, approach very near 
* Since thefe experiments were made, I have found that this is not the cafe. 
The eledtric fpark decompofes hepatic air, and leaves a quantity of inflammable 
air equal in bulk to the hepatic air very nearly. However, as the inflammable 
air leaves the fulphur upon the application of the eledtrical fpark, it fliould feem, 
that the proper matter of inflammable air is more difpofed to combine with fire 
than with fulphur; which may be the reafon why hepatic air is decompcfed by 
nitrous air, while pure inflammable air is not affedted by it. 
7 
to 
