198 Prof. J. A. Cunningham on the 



& 



discharge so long as the cathode was below a bright yellow 

 heat, and apparently most rapidly at a dull red. When, how- 

 ever, the filament was heated to whiteness, the gas began 

 slowly to come out, only to go back again as soon as the 

 temperature fell. The pressure remained practically quite 

 constant over a range of about 1250° to 1360° with currents 

 varying from 1 to 2 milliamperes. At lower temperatures 

 fresh gas had to be admitted three times ; and at higher 

 temperatures it had to be pumped out before equilibrium was 

 established and a fairly steady pressure maintained. If, on 

 the other hand, the discharge was stopped, the gas began to 

 be evolved at all temperatures above redness. If now this 

 filament, after being used for some time as cathode, was made 

 the anode, the rate of giving off gas was greatly increased at 

 all temperatures down to coldness. And it was further re- 

 markable that the power at my disposal (1050 volts) was able 

 to send a discharge through this evolved gas at a pressure 

 three times as high as it could when the gas was freshly 

 prepared. It would therefore appear that it was in the 

 process of helping out the corpuscles * that the positive 

 atoms of the gas got entangled in the carbon cathode, and 

 that a certain number of them even retained their ionic 

 character on getting released. 



A large number of measurements were made with cold 

 electrodes and small current-densities, and the results obtained 

 were in substantial agreement with those of the numerous 

 workers (especially Hittorf, Warburg, Graham, H. A. Wilson, 

 and C. A. Skinner) who have investigated the various phe- 

 nomena associated with the discharge in a vacuum-tube. 

 Such results need not be related in detail, but they served to 

 show that neither the carbon electrodes nor any other part 

 of my apparatus differed in any essential point from those used 

 by other experimenters. 



One point of difference, on the other hand, I might perhaps 

 mention, viz., that in my experiments with a striated positive 

 column I invariably found that a sudden rise of potential 

 took place to the cathode side of each bright band, succeeded 

 towards the anode by a region of very small potential-gradient 

 until the next stria was reached, and accompanied by another 

 sudden step up of potential. The same sort of thing is to be 

 noticed in one of Wilson's f curves for a striated positive 

 column in nitrogen; but in general, both his and Graham's J 



* Cf. H. A. Wilson, Phil. Trans. A, vol. ccii. p. 243, and J. J. Thomson, 

 1 Conduction of Electricity through Gases,' p. 480 (Cambridge, 1903). 

 t Phil. Mag. June 1900, p. 611. 

 % Wied. Ann. vol. lxiv. p. 69 (1898). 



