so 



Mr. A. A. C. Swinton. 



means of a magnet, the luminescent spot on the carbon moves with 

 no perceptible lag. Further, though, as is also the case with glass, 

 the whole of the carbon becomes gradually heated to a considerable 

 extent if much power be employed for a long period of time, these 

 luminescent spots are instantaneously produced on carbon of very 

 considerable brilliancy with but a comparatively low power. Again, 

 just as glass is known to become fatigued under the influence of 

 cathode rays, so that after a time it refuses to fluoresce so brightly 

 as before, so carbon is similarly fatigued, though only after having 

 been very strongly acted upon. Carbon, like glass, also recovers it3 

 property of giving a surface luminescence to some extent, though it 

 does not seem to entirely recover, at any rate at all rapidly. 



That the rays which produce the luminescence of the carbon are 

 the same rays that cause fluorescence of the glass can be proved by 

 deflecting the rays from the carbon on to the glass by means of a 

 magnet. 



As it is exceedingly difficult, if not impossible, to obtain carbon 

 plates which do not contain hydrocarbons and other volatile matter 

 which are rapidly given off and reduce the vacuum very quickly 

 when the carbon becomes at all heated, it is necessary to keep the 

 tube connected to the mercury pump, so that the vacuum can be 

 restored after each experiment. This arrangement was followed in 

 all the experiments described below, except where specific mention is 

 made to the contrary. 



Apparent form of the Cathode Bay Discharge in a Focus Tube. 



As is well known, in tubes of the ordinary focus type with a single 

 spherical concave cathode, the rays coming off normally to the 

 cathode surface appear to converge in more or less of a cone to a 

 focus, and if the vacuum be not too high, to diverge again immedi- 

 ately in another cone upon the other side of the focus. At higher 

 vacua the rays, after passing the focus, do not appear to diverge 

 again at once, but seem to form themselves into a description of 

 thread which connects the convergent and divergent cones, and is 

 longer or shorter according as the vacuum is higher or lower. The 

 angle of the divergent cone appears, however, to be always propor- 

 tional to that of the convergent cone. The focus, or perhaps more 

 correctly, the point at which this thread commences, seems always to 

 be more distant from the cathode than the centre of curvature of the 

 latter, but the variation in this respect seems to be less and less the 

 higher the exhaustion. This is no doubt due to the mutual repulsion 

 of the rays, and accords with the assumption that the rays consist 

 of charged particles, which travel more and more rapidly the higher 

 the exhaustion. Probably for the same reason, cathodes that are 



