on Combustion and Flame. 91 



plosive mixture, 2, and of the external envelope and tail of 

 unconsumed matter or products of combustion, 3, often of a 

 greenish colour. 



Olefiant gas issuing with the same degree of force and 

 from the same jet as hydrogen burns with a much larger 

 flame : this is easily explained, as it requires 1 5 times its 

 volume of air to burn it, whereas hydrogen requires only 2| 

 times. This shows the influence of quantity on the size of 

 flames. But hydrogen requires only its own bulk of chlorine 

 or of nitrous oxide for combustion, yet its flame in these 

 gases is much larger than it is in air ; these burn on the other 

 hand with a very small flame in hydrogen. The only cause 

 to which this can be attributed appears to me to be the dif- 

 ference of their diffusibility. From Mr. Graham's researches 

 we now know the law of the diffusibility of gases, — that it 

 varies inversely as the square root of their density ; so that 

 hydrogen, a light gas, diffuses itself much more rapidly through 

 chlorine, than chlorine, a heavy gas, does through hydrogen. 



These then are the two causes by which I would explain 

 the different appearances of different flames, and they are 

 confirmed by all the experiments I have made. Probably 

 they are not the only causes, but they are the primary ones, 

 particularly with the simple gases. A great number of experi- 

 ments were made to ascertain, if possible, the exact propor- 

 tional effect of these two causes, but with the apparatus I was 

 in possession of, it was a hopeless task. The only way in 

 which I could regulate the jet of gas was by the degree of 

 opening of the stop-cock by which the water was supplied to 

 the gas-holder, a very imperfect method where accuracy is 

 required. With some of the gases produced in the method 

 already described there was not even this resource. The 

 peculiarities of size, &c. which I notice here, are, however, so 

 great and decided, as to leave no doubt about them, allowing 

 freely for inaccuracy of the apparatus. 



A few examples may be given in illustration of these re- 

 marks. The diffusibility of oxygen to hydrogen is as 1 to 4, 

 or the inverse of the square root of their respective specific 

 gravities (sp. gr. of hydrogen = 1). The difference of size 

 however of their flames is not so great as might be supposed 

 from this compared with chlorine : to account for this it is 

 to be kept in mind that the oxygen requires twice its volume 

 of hydrogen for combustion, so that it must spread further 

 through the hydrogen than it would otherwise require to do, 

 and thus make its flame larger ; the flame of hydrogen in 

 oxygen being smaller than it otherwise would be from the 

 same reason. Let us suppose it set down in this way : 



