Light Radiations from Metallic Vapours. 297 



In order to verify the results of these considerations experi- 

 mentally, it is necessary to till the protected space with the 

 vapour of the metal to be examined. This was accomplished 

 by means of a very simple method based upon the observed 

 fact that the vapours of any substance placed on the upper 

 surface of the incandescent graphite plate pass automatically 

 into the protected space beneath. In this way it has been 

 possible to examine the spectra of a great many substances 

 and study the relative effects of thermo-chemical and thermo- 

 electrical excitations upon their bands and lines. The results 

 of this investigation have rlso completely confirmed my 

 anticipations with regard to the similarity between the 

 characteristic emissions caused respectively by chemical and 

 thermo-electrical excitations. 



The present paper deals briefly with the observations and 

 experiments which have led to the establishment of the new 

 method, and with the application of the latter to the spectra 

 of magnesium, copper, and manganese. 



§ 2. Observation of fringe emission with an unprotected 

 plate of graphite. 



The experiments described in an earlier paper were made 

 with graphite plates protected from excessive loss of heat 

 through radiation by means of a cover of carborundum 

 powder. Under these conditions the temperature of the 

 under surface was found to be from 200 c to 300° C. higher 

 than when the plate was unprotected. Further, the car- 

 borundum powder employed contained many impurities, 

 notably Ca, Al, K, Na, Sr, Mn, Fe, &c, and the protected 

 space beneath the plate became filled with a great variety of 

 highly ionized vapours in addition to those expelled by the 

 graphite. The presence of these vapours was, however, 

 found to be of great advantage, because it provided the 

 opportunity of comparing the spectra of these elements with 

 those emitted by them in the tube-furnace. The carbon 

 tubes employed in my furnace work contained as impurities 

 nearly all the same elements as the carborundum powder, 

 and this fact proved to be invaluable for establishing the 

 identity of the character of the emissions given by these 

 substances in the carbon tube-furnace and in the luminous 

 vapour beneath an electrically heated plate of graphite. 

 But for the detailed investigation of the spectrum of some 

 definite metal or other substance, it is of course essential to 

 reduce the disturbing influence of the presence of impurities 

 to a minimum. For this reason it is preferable to dispense 



Phil. Mag. S. 6. Vol. 40. No. 237. Sept. 1920. X . 



