of Magnesium, Copper, and Manganese. 301 



a similar band given by calcium in the outer mantles of 

 flames ; the emission of this band is controlled by the 

 temperature of the plate. The lower red fringe extends 

 only to a very short distance from the plate ; but the upper 

 fringe, thanks to the upward rush of relatively great 

 quantities of calcium vapour, attains about four times that 

 distance. Particular attention should be directed to the 

 abrupt extinction of the upper fringe bands in spite of the 

 upward motion of the vapours. This, as has already been 

 explained, is of course quite in accordance with the electrical 

 origin of the fringe; for were the emission of these bands 

 solely controlled by the plate temperature, they would, like 

 the great flame band, rise higher and die out only gradually. 



§ 3. Circulation of the luminous vapour round the plate 

 of graphite. 



It was suggested in § 4 of my preceding paper that the 

 bluish vapour observed beneath a plate of graphite covered 

 with carborundum powder is caused by the cooling of the 

 luminous vapour at the boundary surface of the protected 

 space. "When no carborundum or other substance is placed 

 on the graphite plate the bluish vapour is not observed; 

 hence, in the former case the vapours which rilled the pro- 

 tected space must have in some way passed from the upper 

 surface of the graphite plate into the space below. Numerous 

 experiments were made in order to account for this fact, and 

 the following observations seem to throw some light on the 



Graphite 

 plate 



Striae of 

 bluish vapour 

 Striated appearance of bluish vapour. 



manner in which the displacement of the metal vapour is 

 accomplished. A small quantity of impure carborundum 

 powder was laid upon the negative end of the graphite plate 

 as shown in fig. 1. When the temperature of the latter* was 



