a new Eudiometer. i i y 
The heft way I know of obviating this inconvenience is to 
be careful always to life the fame kind of water : that which I 
always ufe is diflilled, as being moil certain to be always alike. 
Ifhould have nfed rain water, as being eafier procured, if it had 
not been that this water is fometimes apt to froth, which I have 
never known diflilled water do. 
As I found that the power with which the diflilled water I 
nfed abforbed nitrous air was greater at fome times than others, 
which mull; neceffarily make an error in the obfervation, 1 was 
in hopes that, by obferving the quantity of nitrous air which 
the water abforbed in the fame manner as in the preceding ex- 
periment, together with the heat of the water, as that alfo 
deems to affedt the experiment, one might be able to corredl 
the obferved ted, and thereby obviate the error which would 
otherwife arile from any little difference in the nature of the 
water employed. With this view I made the following expe- 
riment. 
I purged fome diflilled water of its air by boiling, and kept 
one part of it for a week in a bottle along with fome dephlo- 
gifticated air, and fhook it frequently ; the other part was 
treated in the fame manner with phlogifticated air. At the 
end of this time I found, by a mean of three different trials, 
that the telfb of common air tried with the firft of thefe waters 
was 1. 1 39, the diminution which nitrous air fuffered by being 
fhook 2 ' in it in the ufual manner was .285. The teff; of the 
fame air tried with the laft of thefe waters was only 1.054, and 
the diminution of nitrous air only .090, the heat of the water 
in the tub and of the diflilled waters being 45 0 . 1 then railed 
both the water of the tub and the diflilled waters to the heat 
of 6 7°, and found that the teff; of the fame air, tried by the 
firff; water, was then i.ico, and by the latter 1.044; and that 
the 
