Nipher — On the Nature of the Electric Discharge. 69 



will result in a change from the noisy crackle of a multi- 

 tude of small sparks, to silence and darkness. The same 

 statement may be made concerning the motion of the 

 large knob A 1 , if the plate be removed and the large 

 knob be placed at the end of the positive column. In this 

 case, however, the disruptive discharge does not begin 

 until the large knob has reached the positive column. A 

 slight movement of the knob away from the positive 

 terminal then causes the disruptive discharge to cease. 

 When the copper plate is placed at the end of the posi- 

 tive column and is then moved towards the anode, the 

 "resistance" of the gap appears to be increased. When 

 moved towards the cathode it is diminished, and the 

 positive column, from which Franklin's fluid has been 

 drained into the anode, is made longer. It appears to 

 follow the plate. 



The copper plate does not obstruct the discharge if 

 moved into the dark space. The positive column follows 

 it, and acts as a conductor. 



If the copper plate be placed in contact with the cathode 

 knob, the negative glow passes to the corners and edges 

 of the plate, where it ceases to be effective or visible. If 

 placed in contact with the anode knob, the drainage col- 

 umn (positive brush discharge) into the anode then ap- 

 pears at the corners and edges of the plate. It then also 

 ceases to be effective. 



If the preceding explanation is valid it probably ex- 

 plains the behaviour of the Hittorf tube referred to by 

 J. J. Thomson. 3 In the shorter branch of the tube, the 

 dark convection discharge across the Faraday dark space 

 involves a transfer of super-charged gas molecules from 

 cathode to anode. In the longer branch, the electricity is 

 passing by transfer from molecule to molecule, from 

 cathode to anode. The molecules of gas are, however, 

 moving in the opposite direction. The flow of gas in the 

 two branches thus forms a continuous circulation around 

 the circuit of the two branches. 



3 Conduction of Electricity through Gases, 2d ed., p. 443. 



