24G DISCHARGE OF ELECTRICITY. 



s] talks are u«ed, take as its course throu<ili tlie bull) tlie prolongation 

 ot the direction of the tube, but is bent-iu towards the primary. In 

 Fig. 12 the dotted line represents the course the discharge 

 would have taken if there had been no bulb, the continu- 

 ous line the course actually taken. This bending-in can 

 be explained by supposing the currents started- near the 

 j)rimary to shield off from the outlying space the action 

 of the ijrimary, and thus make the electro-motive in- 

 tensity along the axis of the tube smaller than it would 

 Fiu. 1.'. have been if no discharge had been possible between the 



axis and the primary circuit. 

 Before describing some further experiments on this shielding 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 i^late a region A which it shields. This region 

 has to be shielded from tubes of electrostatic induction coming from 

 the left, which have to i)ass through the shield before reaching A, and 

 from tubes coming from the right which have to pass thr<mgh A before 

 reaching the field. The action of the shield in the first case is very 

 simple, for when a tube gets inside a conductor it at once attempts to 

 contract to molecular dimensions, and after a time i^roportional to the 

 specific resistance of the conductor it succeeds in doing so. Thus if the 

 shield is made of a good conductor the tubes of electrostatic induction 

 will be triinsformed into molecular tubes before they have time to get 

 through; so that the shield will protect A from all tubes which have 

 to go through it. The way the shield destroys or rather neutralizes 

 the effect of the tubes coming from the right is somewhat different: 

 when a positive tube reaches the shield a negative one emerges from it, 

 travelling at right angles to itself in the opposite direction to the inci- 

 dent tube; thus, when the first few tubes reach the shield from the 

 riglit they will produce a supply of negative ones, and the presence of 

 these negative tubes at A concurrently with the positive ones which 

 continue to arrive there will ^A■eaken the field to a greater and greater 

 extent as A approaches the shield. At the surface of the shield itself 

 the neutralization will be complete. A dielectric whose specific induc- 

 tive capacity is greater than usual will behave in a similar way to a 

 metal plate, but to a smaller extent. It will emit tubes of the opposite 

 sign, but not so numerous as those incident upon it. Thus a metal 

 plate, or even one made of a dielectric of considerable specific inductive 

 capacity, will reduce very considerably the tangential electromotive 

 force on either side of it. 



I have made several experiments in which this effect was very strik- 

 ingly .shown. In one of these, two square discharge-tubes of equal 

 cross section i)laced near and parallel to each other were connected by 

 a cross tube, so that the pressure was the same in both tubes; a fine 

 wire passed round the inside of one of the tubes, its ends being con- 



