1894.] on Electric Discharge through Gases. 241 



cistern containing the mercury I can stop up the passage by a pellet 

 of mercury. We will first observe the discharge when the side 

 passage is open ; you see that the discharge, instead of passing across 

 the thin piece of platinum leaf, takes the very much longer route 

 round the side tube, so as to avoid crossing the metal. We will now 

 raise the mercury cistern and close the side tube by a pellet of 

 mercury; the discharge now has no alternative but to cross the 

 metal at some part of its course, and you see that the main portion of 

 the discharge goes back into the main tube. 



In the second experiment the metal diaphragm is replaced by a 

 very thin plate of mica ; when the side passage is opened the discharge 

 goes round, but when this is closed by a pellet of mercury the dis- 

 charge prefers to go across the mica than through the mercury. 



A second experiment which shows the same thing is the following. 

 Two long electrodes are fused into a bulb, so that the tip of an elec- 

 trode is a considerable distance 



from the place where it passes ri £ 2> - 



through the glass. We will 

 now send an alternating dis- 

 charge through the tube, and 

 you will see, I think, that the 

 discharge, instead of going 

 straight across the short dis- 

 tance between the ends of the 

 electrodes, goes from the tip of 

 one electrode to the place 

 where the other passes through 

 the glass, thus staying as long 

 as possible in the gas before 

 passing into the metal. The 

 appearance of the discharge 

 shows that the positive electrode is at the end of the wire, the 

 negative at the junction of the wire with the glass. 



Another interesting example of the difficulty the discharge 

 experiences in passing from gas to metal is the discovery made by 

 Professors Liveing and Dewar, that when the discharge passes 

 through a gas containing a large quantity of metallic dust, the light 

 from the discharge, when examined in the spectroscope, does not show 

 any of the lines of the metal. 



The difficulty which the positive electricity finds in passing from 

 the gas to the electrode depends a great deal upon the nature of the 

 gas, as well as upon that of the electrode ; it is influenced by the 

 position of the gas and the electrode relatively to one another in 

 the electro-chemical series. 



I have lately made a series of experiments on this point in the 

 following way. An alternating discharge from a high tension trans- 

 former was made to pass between two electrodes fused into a bulb, 

 which could be filled with the gases under examination. Another 



