at Low Pressures. 
251 
Two small electrodes E and E' were fixed in the tube K by 
means of sealing-wax at its upper end. These electrodes were 
made of platinum wire about OT mm. in diameter, enclosed in 
fine glass tubes 05 mm. in external diameter. At EE' the ends 
of the wires were left bare for two millimetres. The small glass 
tubes were fixed to the inside of the tube K near its lower end 
with sealing-wax at opposite ends of a diameter of its bore. The 
bare wires at E and E' were fixed in this way 6'5 mms. apart and 
extended to the centre of the discharge tube AB. By turning K 
the angle between the line joining E to E' and the axis of the 
tube AB could be made to have any desired value. 
The upper part of the tube K was fixed with wax into a brass 
tube which carried a graduated circle WW 15 cms. in diameter. 
A vernier reading to minutes of arc on this circle was held in a 
fixed position relatively to the tube AB so that the angular 
position of K and consequently of the line EE' could be read off 
on it. 
The small glass tubes enclosing the wires E and E' were 
embedded at their upper ends in sealing-wax, and the wires were 
brought out through holes in the brass tube at TT. These wires 
were connected through a reversing commutator C to an insulated 
quadrant electrometer V. The electrometer and connecting wires 
were enclosed in metallic screens S S S S all well soldered together 
and connected to a ‘good earth.’ In this way all variations in 
the electrometer readings due to outside influences were com- 
pletely stopped. 
The electrodes E and E' were found to be very well insulated 
from each other and from the screens. Either would hold a 
charge for several minutes without appreciable loss as indicated 
by the electrometer. The electrometer was extremely well in- 
sulated in all parts. One pair of the quadrants was connected 
to the case, which latter was insulated by paraffin blocks. The 
quadrants were supported on ebonite legs coated with sulphur, 
and the electrometer ‘jar’ was made of a brass plate coated with 
a thin layer of sulphur on which another brass plate w T as placed. 
All four quadrants were provided with adjustable supports so 
that each could be separately levelled and raised or lowered. 
The quadrants were all very carefully levelled and arranged 
symmetrically about the needle, and when the needle was charged 
no appreciable deflection of the needle occurred. The sensibility 
of the instrument usually only fell off 2 or 3 per cent, in 48 
hours. If the needle was charged up to 2000 volts a sensibility 
of about 250 millimetre divisions (scale at one metre distance) 
for one volt was obtained. When measuring the Hall effect a 
sensibility of about 70 mms. per volt was generally used. 
The tube AB was connected to a Toepler pump and M c Leod 
19—2 
