and the feeming Converjion of Water into Air . 427 
it, and not what it might have imbibed after its formation, I 
made a quantity of both dephlogifticated and inflammable aii- 
in fuch a manner as that neither of them fhould ever come into 
contact with water, receiving them as they were produced in 
mercury ; the former from nitre, and in the middle of the 
procels (long after the water of cryftallization was come over), 
and the latter from perfectly- made charcoal. Idle two kinds 
of air thus produced I decompofed by firing them together by 
the electric explofion, and found amanifeflr depofition of water, 
and to appearance in the fame quantity as if both the kinds of 
air had been previoufly confined by water. 
I11 order to judge more accurately of the quantity of water 
fo depolited, and to compare it with the weight of the air de- 
compofed, I carefully weighed a piece of filtering paper, and 
then having wiped with it all the infide of the glafs veflcl in 
which the air had been decompofed, weighed it again, and I 
always found, as nearly as 1 could judge, the weight of the 
decompofed air in the mo i flu re acquired by the paper. 
As there is a fource of deception in this experiment, in the 
Imall globules of mercury, which are apt to adhere to the infide 
of the glafs veflel, and to betaken up by the paper with which 
it is wiped, I l'ometimes weighed the paper with the moifture 
and the mercury adhering to it ; and then ex poking it in a warm 
place, where the water would evaporate, but not the mercury, 
weighed it again, and hill found, as nearly as 1 could pretend 
to weigh fo final! a matter, a lofs of weight equal to that of 
the air. I wlfhed, however, to have had a nicer balance fu- 
tile purpofe ; the refill t was fuch as to alford a firong prcfinnp- 
tion that tire air was re-con verted into water, and therefore, 
that the origin of it had been water. 
O 
Another 
