Discharge of Electricity from Glowing Carbon. 367 



circuit arranged to include an air-gap, of a few millimetres, 

 between the middle point o£ the heated cylinder and the end 

 of a comparatively cool carbon rod. The currents in the 

 circuit have been measured for different temperatures of the 

 carbon cylinder, and for various voltages across the air-gap 

 up to the point at which an arc forms, with a view to finding 

 the conditions under which the change from the non-luminous 

 to the luminous discharge occurs in the case of hot carbon in 

 air at natural pressure. The conditions for the similar change 

 in connexion with a hot lime cathode in air at low pressure 

 have been investigated by Professor J. J. Thomson*. 



2. Experimental Detail. 



The apparatus is shown in plan and elevation in fig. 1 

 (PL VI.), C, the hot carbon cylinder, is held by large blocks 

 of carbon A & B (not shown in elevation), through which the 

 heating current passes. The cool carbon rod D is fixed in a 

 holder E which can be moved in the direction of its length by 

 turning the nut F, the latter being prevented from longi- 

 tudinal movement by the stop G. The holder E is attached 

 to a cross piece provided with an adjustable stop H, the 

 cross piece working on pivots so that the cool carbon rod D 

 can be very quickly placed in position before the heated cylinder 

 C or readily removed. 



A scheme of the connexions is shown in fig. 2, where A 

 and B are storage-cell batteries, G a galvanometer, and P a 

 potentiometer; by this arrangement any difference of potential 

 desired could be set up between H and C, allowance being- 

 made, of course, for the potential gradient in the hot rod 

 due to the heating current. A fine wire fuse was used to 

 protect the galvanometer from excessive currents. 



When the cold carbon is close to the hot one its tempera- 

 ture rises, and the current varies with time. To secure 

 uniformity measures have been taken in the following way :— « 

 with the cool carbon removed, the hot cylinder was brought 

 to its full temperature; then by tilting the holder E, fig. 1, 

 the cool carbon was put quickly into position at the required 

 distance from the cylinder ; in 6 to 10 seconds after this the 

 galvanometer deflexion became fairly steady, and all obser- 

 vations have been taken within those limits of time. 



The heated carbon cylinder was a piece of the carbon rods 

 supplied by Messrs. Siemens Brothers for use with their 

 Lilliput arc-lamps, while the cooler carbon was a portion of 



* Thomson, < The Conduction of Electricitv through Gases,' 2nd ed. 

 p. 477 (1900) ; see also Horton, Phil. Trans. A. 207. p. 149 (1907), 



2C2 



