246 Mr. G. A. Hemsalech : Excitation of Spectra of Carbon* 



be heated are generally clamped in such a way as to leave 

 an effective length of from 26 to 28 mm. between the ends 

 of the clamping rods. The upper surface of the plate is then 

 covered with a |- in. thick layer of carborundum powder, in 

 order to reduce the great loss of heat which w r ould other- 

 wise result through unrestricted radiation and convection. 

 With a heating current of about 250 amperes the tempera- 

 ture of the graphite plate, mounted in this way, exceeds 

 3000° C. 



When interpreting the luminous effects observed beneath 

 the plate, account must be taken of the possible action of 

 air draughts which are occasioned by convection currents 

 rising up from the hot plate. These currents cause air to 

 be drawn up from below and to rise upwards in two wide 

 streams, one on each side of the hot plate. Since the plate 

 represents a flat obstacle, a certain quantity of apparently 

 stagnant gas or vapour will be held in equilibrium between 

 the plate and the rising air currents. Hence, the under- 

 surface of the plate is protected from direct contact with 

 air and it is, therefore, in its vicinity that the luminous 

 phenomena, to be described in this paper, occur. 



As was already suggested in § 1, such a plate may be 

 regarded as equivalent to a tube cut open longitudinally and 

 rolled out flat. But, whereas in a tube the ionization of the 

 enclosed vapours and gases is easily accomplished, thanks to 

 the complete protection from cold air, in the case of the 

 exposed plate, on the other hand, a high degree of ionization, 

 even in the immediate vicinity of the undersurface, is naturally 

 much more difficult to attain. Hence, in order to produce 

 effects similar to those observed in a tube-furnace, the tem- 

 perature of the plate must be higher than that of the former, 

 so as to make good the very rapid loss of heat by radiation. 

 Now, this can only be accomplished by forcing heating 

 currents of great density through the plate — an operation 

 which inevitably entails the application of much higher 

 potentials than are required in order to bring the tube- 

 furnace to corresponding temperatures. Thus, for a tem- 

 perature of 2700° C, the drop of potential along the plate 



is about 5*8 - ■"- , as against about 1 — with the tube- 



„ cm, 7 G cm. 



rurnace. 



As is to be expected, the life of such a thin plate of 

 graphite, raised to a high temperature, is a very short one. 

 At the lower temperatures they will last 30 seconds and 

 longer, but only from 6 to 10 seconds at about 3000° C. 

 The phenomena to be examined are, however, so luminous, 



