a new Eudiometer . 12^ 
was employed ; which (hews, that the error proceeding from 
the difference of purity of the nitrous air is much lefs when it 
is ufed in the full quantity than in a fmaller proportion ; and 
alfo (hews, that if it is ufed in the full quantity it can hardly 
caufe any fenfible error, unlefs it is more impure than ufual. 
One does not ealily fee, indeed, why it fhould caufe any error ; 
for no reafon appears why the mixture of phlogifticated or 
other air, not abforbable by water, and not affe&ed by refe- 
rable air, (hould prevent the nitrous air from diminifliing and 
being diminifhed by the refpirable air in juft the fame manner 
that it would otherwife be. It muft be obferved, however,, 
that if the nitrous air is mixed with fixed air, it will caufe an 
error, as part of the fixed air will be abforbed by the water 
while the teft is trying; for which reafon care fhould be taken 
that the nitrous air fhould not be much mixed with this fub- 
ftance, which it will hardly be, unlefs either the metal it is 
procured from is covered with ruft ; or unlefs the water in 
which it is received contains much calcareous earth fufpended 
by fixed air, as in that cafe, if any of the nitrous acid comes 
over with the air, it will diffolve the calcareous earth, and 
feparate fome fixed air. 
In order to fee whether it is pofiible for nitrous air to differ 
in the fecond refpe<ft,. I procured fome from quickfilver, cop- 
per, brafs, and iron, and obferved the teft of the fame parcel 
of common air with them, on the fame day, making four trials 
with each, when the difference between the tefts tried with the 
three firft kinds of air was not greater than might proceed 
from the error of the experiment ; but thofe tried with the air 
from iron were _^^ths greater than the reft. I then took 
the teft of fome more common air with them in the fame man-* 
ner, only uling four parts of common to one of nitrous air, 
when 
/ 
