526 
MR. 0. W. RICHARDSON ON THE ELECTRICAL CONDUCTIVITY 
The numbers in columns I. and III., when plotted against one another, }deld the 
following curves (fig. 13) for the variation of the negative leak from carbon with the 
temperature. The various curves represent successively greater units of current as 
in the case of the curves for platinum (fig. 4). 
The further consideration of these results on the temperature variation of the 
negative leak from carbon will be postponed till § 5. 
§ 4. The Relation between the Negative Leak and the Current used to Lleot the 
Fdaments. 
A series of experiments was made in wdiich the saturation current and the 
corresponding current required to heat the filament were measured; since it was 
thought that these measurements would be of especial interest at temperatures so 
high that they could not be determined, at any rate by the methods used in the 
present paper. In these experiments the portion of the ajjparatus used to measm-e 
the leak was unchanged, while the whole of the arrangement used to measure the 
