52 Prof. D. N. Mallik and Mr. A. B. Das on Electric 

 If we assume /3' to be of the form - +b, the X, p curve 



is a straight line, and if a is small,, — = const. 



As X and probably also f and ft depend on the distance 

 between the electrodes and the voltage of the induction- 

 coil, there must be an exact relation between pressure &c. 

 determining the condition under which the discharge be- 

 comes rotatory. This exact quantitative relation is under 

 investigation. 



6. Starting now from the stage at which the discharge is 

 rotatory, we reach, when the pressure is gradually reduced, 

 a stage at which the character of the discharge changes as in 

 figs. 5, 6, 7. 



7. Comparing the figs. 5 and 6, which give the initial and 

 the final stages of a discharge at the same pressure, we 

 observe that the illumination is very slight when the dis- 

 charge is first passed, but that after a time it becomes much 

 more marked. The effect is evidently due to the fact that 

 when the discharge is first passed there is only slight 

 ionization, but that, as the discharge is continued, ionization 

 increases and the consequent illumination. 7 corresponds to 

 a lower pressure than 6. 



8. When this stage is reached, the ring-end of the dis- 

 charge is found to spread over a finite length of the ring- 

 electrode, instead of being confined practically to one point 

 in it. 



9. As the pressure is further reduced, there is a further 

 spreading out as in fig. 7. 



10. If at these stages the electromagnet is excited, there 

 is dispersal of the streams constituting the discharge, which 

 is, at the same time, twisted about the Faraday dark spare 

 on either side of it. 



Figs. 8 asid 9 show the effect of the magnetic field on 

 fig. 6, which is a photograph of the discharge before the 

 electromagnet is excited. 



Similarly figs. 10 and 11 correspond to fig. 7 of the non- 

 magnetic field. In 8 and 11 the disk is the cathode, and in 

 9 and 10 the ring is the cathode. 



11. If we admit that the twist is due to an angular dis- 

 placement of ions about the axis of the electromagnet, this 

 behaviour of the Faraday dark space [10] must be due to the 

 fact that it is a region practically devoid of ions. 



12. Fig. 12 a, curve I. gives the amount of twist for 

 different current strengths in the coil of the electromagnet. 

 It is thus easily seen that since the magnetic force is 



