Ignition ft om compressed Air. visg 



tube; and this Tight is so much the more vivid, in propor- 

 tion-as thf compression is more rapid. If the compression 

 be less powerful, the spunk dtjes not kindle; but we per- 

 ceive in the upper part of the tube a light vapour, that falls 

 in undulations on the piston. When ihis has disappeared, 

 JF we draw back the piston, the vapour will re-appear, as 

 long as ihere is any air in the tube. These effects may be 

 produced several times in succession, merely by pushiiicr 

 the piston with the hand. "This vapour is so thin and dia- 

 phanous, that it is not perceptible in a strong lio-ht. It 

 requires a sort of twilight to sec it well. 



But whence arises this vapour, and wliat is its nature ? 

 Assuredly it is not furnished by the materials of the in- 

 strument ; it can only proceed, therefore, from what it 

 contains, from the atmospheric air. Now, accordinc; to 

 the present state of our knowledge, the air contains only 

 nitrogen, oxygen, and a very small portion of carbonic 

 acid ; all gasiform substances, which are kept in this state 

 by tlie great quantity of caloric that penetrates them, and 

 arc consequently heavier than it*. But in compressing 

 the air contained in the tube, what is the substance that 

 must first give wav ? Is it not that wiiicb is lightest, the 

 pdloric, thai general solvent, that principle of fluidity and 

 volatilization, which gives wings even to metals to raise 

 themselves in the air ? Is then the vapour in question ca- 

 loric, rendered visible by approximation of its panicles, 

 which are compressed by the surrounding air, as air becomes 

 visible in passing through liquids? 'This idea, which I am 

 far from presenting as a thing proved, acquires more pro- 

 bability from the following experiments. 



I substltultil hvdrogen for common air, and the vapour 

 .showed itseU' as betore; but the spunk did not take fire. 

 Wiih carbonic acid gas, and with nitrogen, the effects were 

 the same. 'The latter, which contained a little nitrous gas, 

 gave a somewhat denser vapour. Oxygen, lightly com- 

 pressed, yielded a vapour more rare and transient than that 

 f)f the common air. It had sc.ircely fallen on the piston, 

 when it rebounded and disappeared. When I compressed 



• The air like'.vi<.e, in its ordinary state, contain'! twelve grains of water 

 in a cubic foot. 'J'his ii:iiall qiiaiuity of water, reduced to the proportion of 

 f!ie quantity of air contained in tl)e apparaius, contiibutcs nottiing to iit 

 fHcct ; for the lie.it produced by the fiiciion could at most reduce it to v.i- 

 pour, and in this state it would not Icindle the spunk. If ihc vapour seen in 

 ihe tube wcic water in a slate of expansion, when it fell on the surface of 

 fhe pi .Ion it would condense there, and appear in the stale of a liquid. But 

 tilt surface of the piston always rmiains dry, though on moviin; it the va- 

 iii-ur appiais iiA diwpj.iarj several times. 



oxygen 



