Phosphorescent Glow in Gases. 341) 



the discharge has passed to diffuse through very narrow metal 

 tubing. This destructive after-effect of the spark may be 

 attributed to the ionized state of the gas in A, but which is 

 not communicated to the gas in B ; since it is quite impossible 

 that the ions could diffuse from A to B, except a very 

 small number that move with a great velocity, and the direc- 

 tion of whose motion would just enable them to shoot 

 straight through the opening in the tube. The brass tube 

 will thus act somewhat as a Maxwell's Demon, and if they are 

 negative ions they will tend to multiply as they collide with 

 ithe molecules in their path in the bulb B. (See J. J. 

 Thomson's ' Recent Researches,' p. 127, and further discussed 

 in Phil. Mag. Sept. 1900; also J. S. E. Townsend, 'Nature,' 

 Aug. 9th, 1900, " On the Effect of Negative Ions in Producing 

 Ionization at Pressures of about or below 1 mm.") The small 

 number of such rapidly moving ions as may make their way 

 through the opening in the tube may in virtue of their high 

 velocity produce a considerable effect in ionizing or breaking 

 up the molecules in B. 



It is quite conceivable that an ion moving with a velocity 

 .approaching that of light may make its way through a 

 hundred thousand molecules. The comparatively small quan- 

 tity of electricity carried by cathode-rays when compared with 

 the large conductivity produced by them, indicates that each 

 corpuscle or cathode-ray gives rise to the ionization of a 

 number of molecules of this order, so that one may still 

 expect that some of the destructive after-effects of the spark 

 may still extend to the gas in the bulb B, although very 

 greatly reduced. This would of course be a residual effect, 

 but it affords an explanation of a somewhat curious observa- 

 tion of the way of obtaining the maximum persistency of 

 the glow in B, which was very puzzling for some time until 

 this idea gave a clue as to its nature. 



I have observed that the maximum duration of brightness 

 of the glow in B is not obtained when the tap t is fully 

 turned on, but rather when the handle makes an angle of 

 about 20 or 30 degrees with the axis of the tube. At first 

 it seemed most probable that this was due to some mechanical 

 effect, such as an impediment to the diffusion of the gas by 

 some obstacle that may have stuck in the tube, or possibly 

 to the handle and the axis of the tube not being in a straight 

 line. But on taking the apparatus to pieces and examining 

 the tube and tap it was found that there was no obstacle 

 whatsoever that could have hindered the free passage of the 

 gas through the tube ; and also that there was no appre- 

 ciable difference between the direction of the hole in the 



