178 



Professor Sir J. J. Thomson 



[April 10, 



streamed off from both the corners and the middle points of the sides, 

 there were but few in any intermediate position ; the most abundant 

 stream came, as was the case for cathode rays, from the middle points 

 of the sides, but the disproportion between the streams from the 

 corners and from the sides was nothing like so great for the positive 

 as it was for the cathode rays ; so that the ratio of the quantity of 

 the positive rays to the quantity of cathode rays was greatest at the 

 corners of the triangle. 



I also measured the velocity and the value of elm for the positive 

 rays. I found that this was the same whether the rays came from tlie 

 corners or from the sides, and the same as that of one type of Canal- 

 strahlen (the type for which ejm = 10^), which went down the tube 

 when A was made cathode. In the case of the positive rays coming 

 from the triangular cathode, often only a small fraction were at all 

 affected by electric and magnetic forces, by far the larger portions 

 were quite undeflected by these forces ; so that the phosphorescence 

 of the screen when the magnetic force was applied to the tube pre- 

 sented the appearance of a bright central undeflected patch with a 

 faintly luminous tail. 



We can, I think, explain this distribution of the rays from the 

 triangular ctithode on the view that there is a reciprocity between the 

 cathode rays and the positive rays of tlic following kind. The cor- 

 puscles in the cathode rays are due to the impact of positive ions at 

 or near the cathode, while the positive rays are due to the impact of 

 the corpuscles at some distance from the cathode. Let us consider now 

 what happens when the lines of electric force in the neighbourhood 



Fig. 4. 



of the cathode are curved as in Fig. 4. A corpuscle starting along 

 the normal from P would on account of its inertia not follow the line 

 of force P Q but some path P Q' between P C and P Q, P C being 

 the normal to the cathode. If, now, the corpuscle produces a posi- 



