1881.] on the PhemnHi'iut of thr Elrrtrir Dkchanji: 479 



8000 cells, but with 11,000 cells \vc obtained a further confirmation 

 of them, the effects being shown in another figure on the diagram (Fig. 

 17), which represents a streamer discharge between two points. Tlic 

 negative discharge is a brush which is seen continuously on the lower 

 terminal, while the positive consists of a series of intermittent, ever- 



FiG. 10. 



changing spiral streamers which envelop the negative brush discharge 

 without in the least disturbing its form. They go past the negative 

 point and then curl upwards towards it. 



If I insert a very high resistance between the battery and the 

 terminals the streamer discharge ceases, and a static spark passes 

 from time to time which is exactly like that from an ordinary 

 frictional machine ; it pierces a thin strip of paper just as a static 

 charge would do. The battery gathers up at intervals a charge at 

 the terminals, and the discharge occurs as soon as the potential is 

 sufficient to force its way across the obstacle opposed by the inter- 

 vening air. 



The same thing occurs if I attach a condenser to the terminals of 

 the battery : it takes a longer time for the battery to charge it, and 

 consequently the discharge occurs at longer intervals, shorter or 

 longer according as the terminals are adjusted to a less or greater 

 distance. The condenser I am now using has a capacity of 1 • 5 micro- 

 farad, and hence the accumulated charge is very considerable, and 

 the discharge is like that of a powerful electric battery. But whether 

 the capacity of the accumulator be large or by comparison infinitesi- 

 mally small like that of the points in the discharging micrometer, 

 there is always an interval of time which elapses between successive 

 discharges ; the interval may be so extremely minute that thousands 

 of millions of discharges may occur in a second, but the flow is never- 

 theless discontinuous, like the drops constituting the stream of water 

 before referred to. 



I will endeavour to prove to you that an apparently perfectly 

 steady discharge through a vacuum tube, in which there is no apparent 

 motion of the strata, and in which even the rotating mirror would fail 

 to detect any intermittence, is nevertheless discontinuous. It is truo 

 that the period of pulsation must be of a very high order, millions in 



2 L 2 



