174 Mr Richardson , On an attempt to detect radiation, etc. 
The following are typical of the observations taken : 
Conditions of 
Experiment 
Initial 
Beading of 
Electroscope 
Final 
Beading 
Leak in Scale 
Divisions 
Time of 
Experiment 
Stream of air 
very quick 
58-0 
57-0 
1-0 
30 mins. 
Current not pass- 
ing through wire 
(Test Experiment) 
Stream of air 
very quick 
57-0 
56-2 
0*8 
o 
CO 
Current passing 
through wire 
Using small 
Leyden jars 
56-0 
55*0 
10 
30 „ 
Using large 
Leyden jars 
58-0 
56*8 
1-2 
30 „ 
Pressure inside 
apparatus = 4 
cm’s, of mercury 
54-7 
54-0 0-7 
o 
CO 
?> 
It is thus seen that within the limit of the experimental error 
there is no extra leak produced by the current through the 
wire. 
But it seemed possible that, even if ionisation took place near 
the wire, all the ions might have time to recombine before they 
reached the leaking plates. A new apparatus was therefore con- 
structed in which the conductivity of the air in the immediate 
neighbourhood of the wire could be examined. The leaking part 
now took the form of a wire spiral forming a cylinder round the 
wire to be tested (Fig. 2 b). This was supported on a con- 
tinuation (a) of the same stiff wire round which a sulphur cylinder 
was cast. 
The wire was held in the end of the tube by sealing-wax i, 
which was carefully treated on the outside to form a good insulat- 
ing surface. The exterior insulation was therefore of sealing-wax, 
the interior of sulphur. The end of the wire g could be put in 
contact with the electroscope used before. The wire through 
which the discharge passed was fastened (soldered when possible) 
to the stout copper wires ff Y . These were fixed air-tight into 
small tubes at opposite sides of the bulb l by means of sealing 
wax. To prevent a leak from the comparatively high potentials 
