the Luminosity of Gases. 253 



oxygen or to the dissociation and recombination of the atoms 

 of the iodine molecule. 



The result of this experiment is in conformity with the 

 views already expressed as to the luminosity of flame and 

 enables us to see that, whilst colourless gases like those forming 

 the products of ordinary combustion are not likely to glow 

 at the average temperature of a flame, they would at some 

 higher temperature emit visible radiations. 



The chief conclusions arrived at in this part of the paper 

 are : — 



1. That in speaking of the temperatures of gases regard 

 must be paid to the mode of heating. 



2. That external heating produces an average temperature 

 compounded in a different way from the same temperature 

 as developed by internal chemical reaction. 



3. That the experiments of Hittorf, Siemens, and others 

 showing that gases do not glow at an average temperature 

 of 1500° C. produced by external heating are no evidence that 

 the same gases when produced in flames of the same average 

 temperature are not veritably incandescent. 



4. That if the heat of chemical combination be supposed 

 to be stored in the products of combustion in a flame, the 

 temperature of these products must be exceedingly high and 

 probably high enough to produce luminosity. 



5. That iodine vapour can be made incandescent by heating 

 to temperatures below that at which glass is red-hot. 



* 



Part II, — The Genesis of Flame- Spectra. 



I propose now to discuss some cases of flame-coloration in 

 respect to the chemical reactions by which they are accom- 

 panied in the flame. 



Sodium. — The characteristic yellow flame of sodium with 

 the D line-spectrum represents, as is well known, the most 

 easily obtainable flame-spectrum. It is yielded by all salts 

 of the metal in the flame of coal-gas or hydrogen. It is 

 quenched, according to Mitscherlich *, if hydrochloric acid 

 be supplied in excess along with the sodium salt, which may 

 be done by adding an excess of ammonium chloride to the 

 bead. 



The yellow tint is also, according to Mitscherlich, not 

 yielded in the flame of hydrogen burning in chlorine when 

 these gases are burnt in an oxy-hydrogen burner. 1 have 

 repeated Mitscherlich's experiments with great care, using 



* Pogg. Ann. cxvi. p. 499 ; also cxxi. p. 459. 



