304 Mr. G. A. Hemsalech on the Luminous Vapours 



downward drift would simply be caused mechanically by 

 the continual expulsion of fresh ions from above. But 

 against this explanation stands the fact that even with very 

 small quantities of luminous vapours, such as are obtained 

 with an unprotected plate of graphite, the emission centres 

 still drift away from the plate downwards against strong 

 magnetic forces. Since there can be no question in this case 

 of a mechanical push due to overcrowding of the space, it 

 seems to me that the only forces which can act upon the 

 molecules of the luminous vapour are gravity and radiation 

 pressure. In any theory concerning the downward drift of 

 the emission centres in the protected space of a single plate 

 resistance furnace the possible action of either of these two 

 forces or of both will undoubtedly deserve serious con- 

 sideration. 



§ 5. Practical application of the foregoing observations 

 to the spectroscopic analysis of metal vapours. 



The fact that metal vapours originating from substances 

 laid upon the top surface of the electrically heated plate of 

 graphite pass automatically into the protected space beneath, 

 at once provides a most convenient method of studying the 

 spectroscopic or other properties which characterize these 

 vapours at high temperatures. The experimental arrange- 

 ments employed for this purpose may, on account of the 

 role played by the graphite plate, be regarded as a single 

 plate resistance-furnace, and will be referred to as such 

 hereafter. Exhaustive trials have shown that practically 

 any substance may be examined under these conditions by 

 placing small quantities of it upon the plate, preferably in 

 two places, one on either side of the slit image as indicated 

 in fig. 2. The two portions of the substance should be not 



Substance 



Graphite plate 



Image of sift 



Method of observing spectrum of a substance with single plate 

 resistance-furnace. 



less than 10 mm. apart, in order to prevent the formation of 

 a conducting deposit on the graphite surface between them. 



