Nipher — The Nature of Electrical Discharge. 83 



wire was observed. If the wire protrudes too far from 

 the tube at the positive end, it is likely to cause frictional 

 contact with the glass tube. This is particularly the case 

 when the wire has been long in use so that its surface 

 has become irregular from the effects of the discharge. 

 It is also necessary that the terminal potentials shall be 

 high, and that the condenser shall be of capacity not 

 much less than that described. 



Apparently the effect here observed in the creeping 

 of the wire, is what might have been expected, as a logical 

 sequence of Rowland's experiment. His results showed 

 that a wire having upon its outer surface a thin film of 

 negative supercharge, when given a longitudinal motion, 

 produced an external magnetic field. When a thin film 

 of the wire surface is drained of the negative fluid, and 

 the wire is moved in the opposite direction the same ex- 

 ternal field is produced. In both cases the external field 

 is electro-magnetic in character. 



The phenomenon here described appears to indicate 

 that when the external field is imposed upon the wire 

 from an external source, the negative fluid and the solid 

 aggregation of positive ions, namely, the copper wire, are 

 urged in opposite directions. 



The ends of the wire were bent at right angles and 

 dipped into mercury cups. By means of a snap key, the 

 wire was momentarily connected with terminals of a sep- 

 arately excited dynamo. The potential difference thus 

 imposed was 175 volts. In this case very appreciable 

 expansions could be observed with the unaided eye. The 

 wire was not appreciably distorted in form, as was the 

 case in smaller wires through which spark discharges 

 were sent. Some deceptive results were at one time ob- 

 served which appeared to be a creeping of the wire along 

 the tube. They were later traced to contraction effects. 

 The two ends of the wire were found to be unequally free 

 to contract after expansion had taken place. There was 

 no creeping of the wire as a whole under the conditions 

 last described. 



