298 Mr. Gr. A. Hemsalech on the Luminous Vapours 



with the carborundum altogether whenever possible ; the 

 loss of heat entailed thereby is compensated for by increasing 

 the intensity of the heating current — an operation which 

 involves the raising of the value of the potential gradient 

 along the plate. It has not been possible to prepare a new 

 set of temperature determinations for the unprotected plate, 

 bat it will be perfectly permissible for the purpose of the 

 present investigation to assume a reduction of about 250° (J. 

 at the highest temperatures as compared with the protected 

 plate. The dimensions of the graphite plates used in these 

 experiments and the method of mounting them were the 

 same as those already given in a preceding paper. A fresh 

 plate was used for every observation. 



It will be remembered that with a heating current of about 

 205 amperes the under surface of: the protected plate gave 

 a temperature of about 2700° C, and in its vicinity was seen 

 the carbon band X '3883. Now, with a heating current of the 

 same intensity passing through an unprotected plate no si^n 

 of this band is observed beneath the plate, but a trace of it 

 is found in the immediate vicinity of the upper surface. 

 When the intensity of the heating current is raised to about 

 250 amperes, the red fringe emission is well developed on 

 both sides of the plate ; in fact, the upper fringe passes to 

 a greater distance from the surface than the lower one. 

 In like manner the lines and bands of the luminous vapour 

 which originates from the impurities expelled by the graphite 

 (principally Ca, Sr, Li, Na, Fe) pass to a greater distance 

 above the plate than below. This is of course caused by the 

 convexion currents which rise up from the plate and carry 

 these vapours through a certain distance before their 

 luminous vibrations have ceased. About half of the total 

 amount of luminous vapour which appears above the plate 

 emanates from the under surface, whence it passes first into 

 the protected space and then upwards round the edges. 

 This latter fact may possibly explain the greater extension 

 of the upper fringe ; for these strongly ionized vapours 

 coming direct from the protected space would by virtue of 

 their upward drift enable the thermelectronic current to 

 spread over a greater vertical distance near the edge of the 

 plate. This explanation for the greater extension of the 

 upper fringe emission receives support by the observation 

 that, when appreciable amounts (as indicated by the degree 

 of development of the spectrum) of iron vapour are expelled 

 from the graphite, the fringe emission always passes to a 

 greater distance from the plate, as though vapour of this 

 metal possessed a relatively high degree of conductivity for 

 the thermelectronic current. In this connexion it is well to 



