COMBUSTIBILITY OF GASES. 
flame. I suspect that the ignited body, before it touches the 
hydrogen, dilates it slowly within the sphere of its activity, so 
as to render it incapable of taking fire, whereas the flame 
exerts no action till it touches the gas. At this moment the 
gas dilates itself suddenly, and with violence, at the part in 
contact, and by that means produces, in the parts adjacent, 
that compression which determines the combustion. This 
explanation is also confirmed by the following facts : I could 
not set fire to hydrogen gas contained in a narrow tube, which 
I gradually and vertically brought near the upper extremity of 
a lighted taper. In this case the gas is dilated in the same 
manner as by an ignited body. 
11. Since in all these experiments the dilatation must be 
attributed in part to the vapour of water, which is inseparable 
from the gases, I was desirous of ascertaining whether other 
very expansible fluids, such as alcohol and ether, would change 
the degree of dilatation, which may be considered as the limit 
of the combustion of hydrogen gas. The experiments I made 
on this subject led to other results. 
The tube, capped with a metallic cylinder, as before described, 
was filled with mercury and a drop of alcohol. I then passed 
to half the metallic cylinder a mixture of two parts air and one 
hydrogen. The drop of alcohol was, therefore, between the 
gases and the mercury. I then heated the tube till the alcohol 
began to be converted into vapour. 
It was in vain that I attempted to inflame the gas dilated to 
four times its original volume, by the electric spark of an inch 
in length. I obtained the same result with the vapor of ether, 
and consequently the limit of inflammability depends solely on 
the degree of dilatation. When the tube was quite cold, the 
gas occupied a larger volume than before, and it was not pos- 
sible to set fire to it by the strongest electric spark, though it 
still burned with a blue weak flame by the contact of a lighted 
taper. Hence I concluded, that the alcoholic vapor was de- 
composed ; that its elements had united with the detonating 
air to form an oxicarburetted hydrogen gas, which is very vari- 
able in the proportions of its components. I was more 
confirmed in this opinion by the sudden expansion I had re- 
marked in gas charged with alcohol by means of the electric 
spark. 
D2 
Water is of 
great impor- 
tance in com- 
bustion. Tri- 
als with alco- 
hol. 
Expansion 
with vapour of 
alcohol pre- 
vents combus- 
tion. 
The alcoholic 
vapor decom- 
posed by eltc- 
tricitv. 
12 . 
