[M'LENNAN] ELECTRIC SCREENING IN VACUUM TUBES 89 
down over them. Joints made in this way worked quite satisfactorily 
and besides being airtight they also afforded a means of readily taking the 
tube apart in order that plates of different 
thicknesses might be used for the end of the 
cylinder. 
As before, the electrode C was insulated 
by melting sealing wax on the glass at a and 
a, and the cylinder was kept connected to 
earth through the socket 77. 
Experiments were first made by exhaust- 
ing the upper chamber, while the lower one 
was kept at atmospheric pressure. The disc 
d was taken from aluminium ‘04 mms. in 
thickness, and the coil was kept running 
during the exhaustion. 
Various selections of the terminals À, 
B and H were made, but in no case, even 
with the lowest pressure obtainable, did the 
electrometer attached to the electrode C give 
the slightest evidence of any electric action. 
The two compartments of the tube were 
then united, as shown in Fig. III, and pro- 
be vided with the Fie. II. 
taps # and Ff. 
On then re-exhausting the upper 
chamber and keeping the tap F 
closed, no action was observed until 
a pressure somewhat lower than a 
millimetre of mercury was reached. 
At this point the electrometer hegan 
to gain a negative charge, and the 
action became more intense as the 
exhaustion proceeded. It was very 
strongly marked when A and H 
were the terminals, and was measur- 
able, though feeble, when A and B 
were the discharging electrodes. 









The effect appeared to be slight- 
ly greater when H was the cathode 
and A the anode, than when the 
opposite arrangement was made. 


As the air in the lower chamber 
was at atmospheric pressure dur- 
