on Combustion and Flame. 95 



position takes place ; this I have not yet tried. TJiis effect of 

 density in cooling the Jlame depends on the excessive diffusion 

 qfthejlame in the denser gas. 



The effect of diffusibility on a flame burning in another 

 gas is obvious enough, for it is visible ; but when the gases 

 are mixed together in one vessel its influence is not so easily 

 perceived. It appears to me, however, that in this case it 

 operates in a similar manner: in passing an electric spark 

 through an explosive mixture, or applying a heated body to 

 it, the combination and combustion take place amongst the 

 particles immediately in contact with the exciting cause, and 

 are thence propagated to the rest of the mixture; when the 

 gases are pure, this takes place so rapidly as to appear in- 

 stantaneous; but if they be diluted with another gas, the pro- 

 gress of the flame may be easily seen. In a mixture which 

 does not explode by the electric spark, I conceive that the 

 particles of the combustible mixture immediately subject to 

 the influence of the spark do combine, but that being diffused 

 to too great a degree, either from excessive dilution or from 

 excessive diffusibility in a denser gas, the temperature is re- 

 duced so much that it does not cause combination of the rest 

 of the mixture. This idea seems to be confirmed by the fact 

 (Turner's Elements, sixth edit., p. 252), that " An explosive 

 mixture diluted with air to too great a degree to explode by 

 electricity, is made to unite silently by a succession of electric 

 sparks," namely, from new particles of the mixture coming to 

 be subjected immediately to the action of the spark. This view 

 is likewise supported by other considerations. 



In the table I have given fluosilicic acid does not seem to 

 be so powerful as it should be, but this probably depends on 

 impurity. These experiments would require to be repeated 

 on a larger scale, as mine were performed with a jar capable 

 of containing only about 4 oz. of water, and a still smaller 

 trough, with 6 or 7lbs. of mercury. 



These researches throw complete lighten the action of the 

 blowpipe : the jet of air thrown into the central cool part of 

 the flame is in precisely the same circumstances, and has ex- 

 actly the same appearance as a jet of air burning in the flask 

 containing a carburetted hydrogen gas, except of course in 

 being surrounded by a hot external flame, and therefore im- 

 proved by this circumstance. In this case we see the two 

 phaenomena at once, the vapour burning in the air, and the 

 air burning in the vapour. Hence also the reason why the 

 flame of the blowpipe is so distinct when thrown through this 



