78 CORRELATION OF PHYSICAL FORCES. 



of extending to a much greater distance than when it takes 

 place in air of the ordinary density. Thus, in highly-attenu- 

 ated air a discharge may be made to pass across six or seven 

 feet of space, while in air of the ordinary density it would not 

 pass across an inch. An observer regarding the beautiful 

 phenomena exhibited by this electric discharge in attenuated 

 gas, which, from some degree of similarity in appearance to 

 the Aurora Borealis, has been called the electric Aurora, 

 would have some difficulty in believing such effects could be 

 due to an action of ordinary matter. The amount of gas 

 present is extremely small ; and the terminals, to a cursory 

 examination, show no change after long experimenting. It is 

 therefore not to be wondered at that the first observers of this 

 and similar phenomena regarded electricity as in itself some- 

 thing as a specific entity or fluid. Even in this extreme 

 case, however, upon a more careful examination we shall find 

 that a change does take place, both as regards the gas and as 

 regards the terminals. Let one of these consist of a highly- 

 polished metal a silver plate is one of the best materials for 

 the purpose and let the discharges in attenuated atmospheric 

 air take place from a point, say a common sewing-needle, 

 to the surface of the polished silver plate ; it will be found 

 that this is gradually changed in appearance opposite the 

 point it is oxidated, and gradually more and more corroded 

 as the discharge is continued. 



If now the gas be changed, and highly-rarefied hydrogen 

 be substituted for the rarefied air, all other things remaining 

 the same, upon passing the discharges as before the oxide 

 will be cleared off the plate, and the polish to a great extent 

 restored not entirely, because the silver has been disinte- 

 grated by the oxidation and the portion which has been 

 affected by the discharge will present a somewhat different 

 appearance from the remainder of the plate. 



A question will probably here occur to the reader : What 

 will be the effect if there be not an oxidating medium pre- 

 sent, and the experiment be first performed in a rarefied gas 

 which possesses no power of chemically acting on the plate ? 

 In this case there will still be a molecular change or disinte- 

 gration of the plate ; the portion of it acted on by the discharge 



