through Exhausted Tubes without Electrodes. 



449 



that a discharge equivalent in its effects to a current can 

 exist in the gas without sufficient luminosity to be visible 

 even in a darkened room. We shall have occasion to mention 

 other cases in which the existence of a discharge non-lumi- 

 nous throughout the whole of its course is rendered evident 

 in a similar way. 



Another experiment by which the screening can be effec- 

 tively shown is to place the primary coil inside a bell-jar 

 which is connected with a mercury pump, the electrical con- 

 nexions with the primary being led through mercury joints. 

 An exhausted bulb is placed inside the primary, the bulb 

 being considerably smaller than the primary, so that there is 

 an air-space between the two. Before the bell-jar is ex- 

 hausted the discharge passes through the bulb, but when the 

 bell-jar is exhausted sufficiently to allow of the discharge 

 passing through the gas outside the bulb the discharge in the 

 bulb ceases, and the only discharge is that outside. I have 

 never been able to exhaust the bulb sufficiently well to get 

 the discharge outside the bell-jar to cease, and that in the 

 bulb to appear again, as in the preceding experiment. In 

 this experiment, as in the preceding one, there was a range 

 of pressure when neither the bulb nor the bell-jar was lumi- 

 nous, showing again the existence of currents in the gas which 

 are not accompanied by any appreciable luminosity. 



A curious bending-in of the discharge which takes place 

 in a square tube provided with a bulb can, I think, be ex- 

 plained by the principle of shielding. The discharge in the 

 bulb does not, unless very long sparks are used, take as its 

 course through the bulb the prolongation of the direction of 



In 



fig- 



12 



Fig-. 12. 



the tube, but is bent-in towards the primary, 

 the dotted line represents the course the dis- 

 charge would have taken if there had been no 

 bulb, the continuous line the course actually 

 taken. This bending-in can be explained by 

 supposing the currents started near the pri- 

 mary to shield off from the outlying space the 

 action of the primary, and thus make the electro- 

 motive intensity along the axis of the tube 

 smaller than it would have been if no discharge 

 had been possible between the axis and the 

 primary circuit. 



Before describing some further experiments on this shield- 

 ing effect, it will be useful to consider the means by which it 

 is brought about. Let us suppose we have a vertical plate 

 made of conducting material, and to the right of the plate a 



Phil. Mag. S. 5. Vol. 32. No. 198. Nov. 1891. 2 H 



