COMBUSTIBILITY OF GASES. 
31 
extremity was defended by a metallic cylinder) to one-sixth of though most 
its capacity of e.qual parts of oximuriatic and hydrogen gas. ^ 
The tube was inverted in a bason, and filled to one-sixth of oily deprived 
water deprived of air. The smallest electric spark inflamed °^^ iat ^P ro " 
this gas with such violence, that several tubes were reduced to pansioin ° 
powder. I repeated the experiment in a stronger tube ; and, 
after having placed the apparatus under the receiver of an air 
pump, I exhausted the air, till the gas occupied nearly the 
whole of the tube, that is to say, a space six times greater than 
before. 
The electric spark was then passed through it, and I saw its 
course, but it was not possible to inflame this dilated gas. 
After having restored the ordinary pressure of the atmosphere, 
a slight diminution of volume was observed, which indicated 
the slow formation of water. There was, nevertheless, enough 
of the undecomposed oximuriatic gas to be inflamed by the 
electric fluid. I could not exactly determine by this means 
the degree of dilatation at which a mixture of these two gases 
loses the faculty of inflaming, because the receiver was not 
hermetically closed. But I perfectly accomplished the inten- 
tion of this experiment, namely, to prove that this very in- 
flammable gaseous mixture loses its faculty of burning at a 
much less degree of dilatation than a mixture of oxigen and 
hydrogen gas, which does not cease to be inflammable, until 
dilated more than sixteen times its primitive volume. 
May not the cause of this anomaly consist in the circum- Speculation 
stance that, in the oximuriatic acid the oxigen is in a more con- the cause, 
crete state, and, consequently, contains less caloric than oxigen 
gas does j so that, when at a certain dilatation, some of the 
parts of the oxigen of the oximuriatic gas are applied to a part 
of the hydrogen, the heat developed is not sufficient to effect the 
continuation of union in the two principles of the rest of 
the gas ? 
IV. We have here an additional proof that the varied pressure Deduction 
of the atmosphere remarkably changes the effects of chemical 
affinity. The inflammable property of hydrogen gas, and of 
all the combustible bodies, would certainly be unknown at a 
suitably diminished pressure of the air ; and we may suppose, 
on the contrary, that many bodies would appear to us very in- 
flammable, at a pressure of from two to ten times as great. The 
nitrous 
