the Radiation of Heated Gases. 461 



radiations of certain elementary substances, particularly iodine 

 and sodium, when their vapours are subjected to external 

 heating. The important article published in the Philoso- 

 phical Magazine (March 189-1), by Prof. Smithells, may be 

 said to have furnished the main incentive in this work, as it 

 likewise determined the direction followed in the experiments. 

 I am further indebted to Prof. Smithells, to whom I sub- 

 mitted a statement of my results, for valuable advice and 

 criticism, which has enabled me to anticipate certain 

 objections that might have been raised to my conclusions. 



The experiments were undertaken primarily with a view 7 to 

 obtaining a closer personal acquaintance with what I may 

 call the " spectroscopic behaviour " of heated gases. I had 

 no idea that they would lead to results which could throw 

 any new light on the questions discussed in Prof. Smithells 

 paper. Nevertheless, after continuing the experiments in a 

 somewhat desultory fashion for some six months, and with 

 very simple appliances, certain facts have presented them- 

 selves in a prominent way, which, so far as I can discover, 

 have not been sufficiently noticed by previous workers, and 

 which would appear to lend a strong support to the view there 

 advocated, namely, that the luminosity of gases in flames 

 may be directly due to high temperature, and that external 

 heating is in itself sufficient to make a gas emit visible rays. 

 The fact that iodine vapour can be easily made incan- 

 descent by external heating was noticed by Salet [Spectro- 

 scopies p. 173) ; and in the article above mentioned Prof. 

 Smithells concludes that this luminosity can only be due to 

 heat, and that chemical action plays no part in the phe- 

 nomenon. It seemed, therefore, of interest to determine, 

 first, the character of this iodine emission, and then to dis- 

 cover whether the property of glowing by mere heating was 

 peculiar to iodine, or was shared in greater or less degree by 

 other allied elements when vaporized, particularly by those 

 which exhibit strong absorptive properties on light. After 

 having thus gained some experience with the less easily 

 oxidizable metalloids, I proposed finally to approach the 

 question of the radiation of heated metallic vapours, to 

 determine whether or not their characteristic spectra can be 

 produced by mere heating. 



It will perhaps be worth while at the outset to describe 

 briefly a few simple experiments made to determine whether 

 the iodine glow r is affected in any way by the nature of the 

 gas in which it is heated. I will describe them as nearly as 

 possible in the order in which they w^ere performed. 



Experiment I. — A little iodine was placed near the closed 

 end of a piece of hard glass tube of about 7 mm. bore, the 



