180 Dr. E. Goldstein on the Electric 



return current does not cause the least appearance of light, 

 although it is present in the same medium, and certainly not 

 of greater section than the direct current which fills the whole 

 width of the tube. Any manifestation of light due to it must 

 become visible when the direct current is diverted by the 

 action of the magnet to one side of the tube. In the space 

 thus rendered free, any possible luminous effect of the return 

 stream would show itself. Experiment proves, however, that 

 this space is dark. 



(4) Let the kathode a be again a plane, whose direction is 

 parallel to and contains the axis of the cylinder. Then the 

 negative rays, as is always the case, are almost entirely at right 

 angles to the radiating surface, and directed towards the sides 

 of the tube. When the pressure is somewhat high, the rays 

 end in free space before they reach the wall ; but with smaller 

 densities, as soon as they strike on the solid wall. 



The phenomena are altogether correspondent when (5) the 

 kathode, as is most frequently the case, is formed by a wire in 

 the axis of the tube. Here also the rays are directed towards 

 the sides of the tube, and in a particular case are radial in each 

 section of the cylinder. 



The electricity must therefore first traverse the direction of 

 the negative rays to their termination, and then take a path at 

 right angles to it in order to reach the anode; whilst, again, 

 both positive and negative light have the same properties as 

 in the preceding cases, when we assumed either direct trans- 

 ference or a direct and return current. 



The complexity of the assumptions necessary in supposing 

 that the current (following always the direction of the negative 

 current) propagates itself from the negative light into the first 

 stratum of the positive, from this into the second, and so on 

 till the anode is reached, becomes still greater when account 

 is taken of the existence of the dark space between the posi- 

 tive and negative light. 



In the preceding cases the dark space has not been men- 

 tioned ; it invariably disappears at a certain exhaustion ; and 

 for the sake of simplicity I have described the phenomena cor- 

 responding to that exhaustion. 



If the kathode is again a plane (a) at right angles to the 

 axis of the cylinder, the anode an electrode (b) of any form at 

 the opposite end of the tube, the phenomena of discharge 

 when the dark space is 'present correspond to fig. 4*. 



* The variously coloured layers of the negative light are represented in 

 the figures by different shading : the layer nearest the kathode is chamois- 

 yellow, the next sky-blue ; and the third, which forms the chief portion of 

 the light, is blue with a shade of violet. Between the layers of positive 



