Discharge in Rarefied Gases. 175 



These and similar appearances were inexplicable as long as 

 the production of light on the walls was ascribed to optical 

 radiation from the whole mass of gas or from the whole length 

 of the electric rajs, which could never have produced sharp 

 images, but only uniform illumination of the walls. On the 

 other hand, the action of the ends of the rays as opposed to the 

 rest of their length, which is now made clear, explains the whole 

 of the observed phenomena at once. 



The optical character of the action in question is confirmed 

 by the existence of photochemical action, which, again, is ex- 

 erted by the ends of the rays, and not by their whole length. 

 Substances which are decomposed under the influence of solar 

 rays of high refrangibility are decomposed in the same way 

 when the ends of the rays fall upon them. I succeeded, by way 

 of control of statements (2) and (3), in obtaining direct pho- 

 tographs of the image of a kathode in relief produced on the 

 wall of the tube, by placing dry sensitive paper on the side of 

 the tube and allowing the rays to fall on it. Pictures were 

 obtained with potassium bichromate, silver chloride, and the 

 highly sensitive ferric oxalate. 



Further experiments showed that 



(4) The modification of the end of the ray is produced, not only 

 when the ray impinges on a fluorescent ivall, but also whenever it 

 falls on any solid substance. 



This is seen when the electric rays are caused to fall on 

 substances not capable of shining themselves, such as quartz, 

 or a particular variety of mica. If then, as already described, 

 fluorescent screens are placed at a distance from the plate of 

 mica and ends of the rays, but exposed to any radiation from 

 the mica, they emit light as soon as the electric rays strike the 

 mica, although the mica itself remains dark. 



If the induced current which traverses the tube be employed 

 in the usual way (that is, without including in the circuit any 

 other non-metallic resistance than the tube itself), this differen- 

 tiation of the ends of the rays occurs only when the density is 

 small. It may be shown, however, that 



(5) The differentiation in question is not associated ivith a 

 particular pressure. It may be produced within wide limits 

 of density as soon as the kathode is completely surrounded 

 with light by the interpolation in the circuit of sparks of vary- 

 ing length in air. 



In the same way, 



(6) The phenomenon is not associated with any particular 

 intensity of discharge. 



This is easily seen by including in one circuit several 

 vacuum-tubes of different densities, with attention to the pre- 



