350 Mr. J. B. B. Burke on the 



tap and the direction of the handle. The inference to be* 

 drawn is that when the path of the gas in diffusing from A 

 into B is crooked, even the fast-moving ions we have been 

 considering are unable to make their way through, and are 

 caught just as the slower-moving ions are by the sides of the 

 metal tube, so that a maximum effect is obtained in B when 

 the path is just sufficiently crooked for this to happen. 



(6) The hypothesis that the destructive after-effects of the 

 spark on the glow are due to the ionization of the gas has 

 been further confirmed by the following experiments. 



The tube, in fig. 3, had side tubes and electrodes c, d, e, f r 

 and the effect of a discharge from a small induction-coil 

 between a pair of these electrodes, such as c d, e /, was 

 observed. 



It was found that a discharge between c d destroyed the 

 glow in A completely, and similarly that between ef destroyed 

 the glow in B almost instantaneously whenever the electrode 

 c or e was cathode : but when they were the anodes there 

 was no very marked diminution in the persistency of the 

 glow. When the brass tube a b was anode, and d or/ cathode, 

 and a discharge passed through the bulbs A or B, there was 

 still no very great destructive influence of the discharge on the 

 glow, and when the discharge was a very feeble one it was 

 possible to see the two at once without any marked diminu- 

 tion in the phosphorescence. 



Thus it is evident that the destructive effect is due to the 

 exposure of the gas in the bulb to the radiation from the 

 cathode c or e. 



These experiments throw considerable light upon several 

 observations I had made on the glow when working with the 

 long tubes. For instance, it was often observed when the 

 ring-discharges in the bulb followed each other very rapidly, 

 that after a short while, the glow no longer travelled down 

 the tube ; but that as soon as the discharges ceased, the 

 glow diffused down as usual. 



This phenomenon is doubtless due once more to the de-- 

 structive after-effect of the spark. 



(7) The passage of the ring-discharge then gives rise to 

 the ionization of the gas and to molecules, or groups of ions,. 

 that do not carry a charge of electricity and which have given 

 to them a certain amount of energy that is radiated away in 

 the form of light ; and the lite period, if it might be so called, 

 of these molecules is greatly increased by removing them from 

 the presence of the ionized gas. 



More will be said upon this point presently when discussing. 



