the Luminosity of Gases. 251 



combined hydrogen and oxygen molecules in the flame. Our 

 theoretical temperature of flame is no more an indication of 

 its average temperature than its average temperature is a 

 measure of the temperature of the steam which it contains. 

 A gas may in fact exhibit a certain average temperature in 

 two ways — it may have been raised to the average tempera- 

 ture by external heating or by internal chemical combina- 

 tion. In both cases the velocity of the molecules shows the 

 same average, but this average is very differently compounded 

 in the two cases. In the first case it can be shown theo- 

 retically that the number of molecules having velocities four 

 times the average is extremely small. In the second case 

 (flames) we have a large number of molecules of intensely 

 high temperature and a large number of low temperature. 

 The average temperature is in this case compounded of two 

 extremes, and there is every reason to anticipate a difference 

 of physical qualities. This difference we might reasonably 

 expect to manifest itself in the emission of light by the gas 

 which contains the intensely-heated molecules. 



Similar considerations apply to the temperature of carbonic 

 acid produced in flames. The theoretical temperature for the 

 combustion of carbonic oxide in oxygen is 7180° C. 



As neither steam nor carbon dioxide has been raised by 

 external heating to anything approaching the temperatures we 

 have indicated as accruing to them theoretically in the act of 

 formation, it is still possible that the luminosity of flames in 

 which they are produced may be due to high temperature in 

 the ordinary sense of the term. The remark of Siemens 

 {supra) that flames become smaller on increasing or heating the 

 air-supply does not affect the conclusion arrived at, for it is 

 not claimed that the luminosity of the products of combustion 

 is of appreciable'duration. Admixture with more air produces 

 a smaller region of combustion, for it spares the combustible 

 gas the necessity of wandering outwards to find the necessary 

 oxygen. There would be in consequence a greater number of 

 hydrogen and oxygen molecules uniting within a given space, 

 the flame would be brighter but not necessarily larger, and 

 this is certainly the case. 



It may be stated further, in relation to the luminosity of 

 air, steam, and carbon dioxide, that having regard to their 

 extremely small optical absorptive power it is not to be ex- 

 pected that they will show any great emissive power even 

 when raised to very high temperatures. The degree of lumi- 

 nosity shown by a Bunsen gas-flame, assuming the products 

 to have the temperature calculated theoretically, appears to be 

 just about what we might expect from a priori considerations. 



