185 
ammonia, the residue will be absorbed, which is an addi- 
tional characteristic property of oxide of carbon, as indicated 
by M. Le Blanc. 
4. If one volume of this residual gas be introduced into a 
eudiometer, and after being mixed with one volume of oxygen 
the mixture be exploded by an electrical spark, it is found 
that half the volume of oxygen employed has disappeared, 
and that the volume of carbonic acid formed is equal to that 
of the oxide of carbon introduced into the eudiometer. 
As to the quantity of oxide of carbon produced during 
these experiments, it depends firstly upon the concentration 
of the pyrogallate employed, and secondly upon its neutral 
or alkaline state, the minimum being with the acid liquor, 
and the maximum with an alkaline. 
In taking the mean of ten distinct experiments, I found 
that 100 volumes of oxygen gave two volumes of oxide of 
carbon, but in some instances the proportion was as high as 
four per cent. 
The production of oxide of carbon also occurs when 
atmospheric air is substituted for oxygen, only the oxide of 
carbon is so much diluted by the nitrogen that its presence 
cannot be detected by several of the above methods. But 
if, on the other band, several quarts of air, carefully deprived 
of carbonic acid, are passed through an alkaline solution of 
pyrogallate, or gallate of soda, and from thence over some 
heated oxide of copper, carbonic acid will be found issuing 
from the apparatus, and can be easily characterised, and even 
weighed, as I have done in some experiments. 
It is no doubt owing to the difficulty of detecting a small 
quantity of oxide of carbon when diluted with nitrogen, 
that the presence of that gas has not been observed before 
by chemists, who have employed Chevreul’s and Liebig’s 
method of analysing the atmosphere. 
M. Boussingault having observed the presence of a light 
carburetted hydrogen- in the residuum which he obtained 
