92 ROYAL SOCIETY OF CANADA 
The upper end of the cylinder was again made air-tight in the 
manner already described and was closed by a thin dise of aluminium. 
A platinum wire attached to the cylinder made electrical connection 
with the mercury and an iron wire dipping into the mercury in the re- 
ceiver ‘£ connected the whole to earth. 
As before, the electrode ( was carefully insulated by melting seal- 
ing wax on the tube at a and a!. To prevent conduction along the 
surface of the glass the part of the tube below the bulb was tightly bound 
with tin foil connected to earth. 
The highest vacua could be obtained with this tube and the use of 
the ground joint gave easy access to the interior without causing any. 
derangement of the apparatus. 
When ready for use, the tube was placed in position with its lower 
end extending into the earth-connected metal box that contained the 
electrometer. The air was then exhausted through the tube F until 
any desired pressure was reached and the mercury raised until it made 
contact with the steel cylinder. 
By this arrangement the Faraday cylinder consisted of the steel 
tube D, the mercury, and the metal box containing the electrometer, 
the whole practically constituting a single earth connected metallic 
conductor. 
In the first experiments with this tube aluminium ‘04 millimetres 
was used for the plate d. The air was then exhausted until the cathode 
dark space extended completely across the bulb of the tube and its walls 
were covered with a green phosphorescence. 
With these conditions A, B and D, were in turn selected as the 
negative terminal of the tube, but in no case was there obtained any 
evidence of electrical action within the cylinder. Charges, either posi- 
tive or negative, given independently to the exploring electrode were 
maintained without loss when the tube was excited. 
The aluminium plate d was then replaced by one ‘004 millimetres 
in thickness and the tube re-exhausted. 
With B or D now acting as a cathode there was again no indication 
of electrical action within the cylinder, even with the lowest pressures 
obtainable. But, when À was taken as the negative terminal, it was 
found that just as soon as the walls of the tube in the neighbourhood of 
d began to phosphoresce, the electrode C slowly acquired a negative 
charge. With still lower pressures this charging became more rapid 
until finally with very high exhaustions a momentary discharge in the 
tube sufficed to charge the electrode beyond the range of the electro- 
meter. 

