96 Prof. E. Wiedemann on the 



which must be reserved for a later paper. We see thus that 

 kathodes formed of poor conductors do become disintegrated. 



10. Dependence of the Potential of Discharge upon the Form 

 of the Electrode. — A series of experiments were made to de- 

 termine in particular tubes how the potential necessary to 

 initiate the discharge depends upon the form of the electrodes. 

 All the tubes were of the same size, 25 millim. wide and 

 120 millim. long. The first tube contained two plates, the 

 second two points, and the third a point and a plate at a dis- 

 tance apart of 58 millim. 



The first result obtained was that, if the three tubes are 

 connected with the positive and negative poles of the machine 

 side by side at pressures exceeding 1 millim., the discharge 

 passed only in the tube with the points; at lower pressures 

 only in that with the plates. 



A second mode of experiment consisted in using one tube 

 only, but including a spark-micrometer in a second parallel 

 circuit, and then gradually separating the balls of the micro- 

 meter until the discharge passed through the tube, or by 

 bringing the balls nearer to each other from a considerable 

 distance until the discharge took place between the balls. 

 In the last case the distance is always greater than in the first, 

 partly because the discharge between the balls of the micro- 

 meter takes place more easily when the air between them has 

 been heated and rarefied by the stream of sparks, and partly 

 because the walls of the glass tube become charged; and it 

 therefore requires a higher potential to produce the first dis- 

 charge through the tube than is necessary for the following 

 ones. 



The experiment showed next that connecting either of the 

 electrodes with the earth is without any important effect. 



At very low pressures the following results were obtained : — 



(1) With the tube with two points the spark-distance might 

 be increased to 9 millim. without the discharge taking place. 



(2) With the tube with one plate and one point, if the plate 

 was negative, no discharge took place in the tube, even when 

 the spark-distance amounted to 11 millim.; but if the plate 

 was positive, the discharge occurred at a distance of 3 millim. 



(3) If both electrodes were plates, discharges took place at 

 a spark-distance of 2 millim. 



Hence it follows that a greater potential is necessary for 

 the discharge between two points than between two plates, 

 and that part is taken not only by the negative electrode, but 

 also by the positive. 



An altogether peculiar phenomenon manifests itself if we 

 employ a tube of the form shown in fig. 14, making the small 



