424 Professor J. J. Thomson [April 30, 



on the glass are by no means always the terminations of the brightest 

 streaks of luminosity in the gas; in fact, in some cases a very 

 bright spot on the glass is not connected with the cathode by any 

 appreciable luminosity, though there is plenty of luminosity in other 

 parts of the gas. 



One very interesting point brought out by the photographs is 

 that in a given magnetic field, with a given mean potential dijBference 

 between the terminals, the path of the rays is independent of the 

 nature of the gas ; photographs were taken of the discharge in 

 hydrogen, air, carbonic acid, methyl iodide, i.e. in gases whose 

 densities range from 1 to 70, and yet not only were the paths of the 

 most deflected rays the same in all cases, but even the details, such 

 as the distribution of the bright and dark spaces, were the same ; in 

 fact, the photographs could hardly be distinguished from each other. 

 It is to be noted that the pressures were not the same ; the pressures 

 were adjusted until the mean potential difference was the same. When 

 the pressure of the gas is lowered, the potential difference between 

 the terminals increases, and the deflection of the rays produced by a 

 magnet diminishes, or at any rate the deflection of the rays where 

 the phosphorescence is a maximum diminishes. If an air break is 

 inserted in the circuit an effect of the same kind is produced. In 

 all the photographs of the cathode rays one sees indications of rays 

 which stretch far into the bulb, but which are not deflected at all by 

 a magnet. Though they stretch for some two or three inches, yet in 

 none of these photographs do they actually reach the glass. In some 

 experiments, however, I placed inside the tube a screen, near to the 

 slit through which the cathode rays came, and found that no appre- 

 ciable phosphorescence was produced when the non-deflected rays 

 struck the screen, while there was vivid phosphorescence at the places 

 where the deflected rays struck the screen. These non-deflected rays 

 do not seem to exhibit any of tLe characteristics of cathode rays, and 

 it seems possible that they are merely jets of uncharged luminous 

 gas shot out through the slit from the neighbourhood of the cathode 

 by a kind of explosion when the discharge passes. 



The curves described by the cathode rays in a uniform magnetic 

 field are, very approximately at any rate, circular for a large part of 

 their course ; this is the path which would be described if the cathode 

 rays marked the path of negatively electrified particles projected with 

 great velocities from the neighbourhood of the negative electrode. 

 Indeed, all the effects produced by a magnet on these rays, and some 

 o^ these are complicated, as, for example, when the rays are curled up 

 into spirals under the action of a magnetic force, are in exact agree- 

 ment with the consequences of this view. 



We can, moreover, show by direct experiment that a charge of 

 negative electricity follows the course of the cathode rays. One way 

 in which this has been done is by an experiment due to Perrin, the 

 details of which are shown in the accompanying figure (Fig. 8.) In 

 this experiment the rays are allowed to pass inside a metallic cylinder 



