906 
Proceedings of Royal Society of Edinburgh. [sess. 
2°. Flame lit. 
(Plate cd to Earth.) 
Time in Minutes. 
Reading. 
Leak per 5 Minutes. 
0 
28*9 
5 
30*8 
1-9 
10 
32-7 
1-9 
15 
347 
2-0 
Average leak per 5 minutes, 1 ’9. 
3°. Flame lit. 
(Plate cd to 360 volts.) 
Time in Minutes. 
Reading. 
Leak per 5 Minutes, j 
, 1 
0 
36-0 
5 
37*9 
1*9 
10 
39*9 
2-0 
15 
41*9 
2*0 
Average leak per 5 minutes, 2'0. 
The maximum deviation of any reading from the mean was 
about 10 per cent., so that, since the leak of the electroscope due 
to the ions from the flame, when the foil was removed, was 50-80 
times the natural rate of leak, the accuracy of the experiments, 
therefore, was such that if 1 per cent, of the ions had gone through 
the aluminium foil, their presence could have been detected. 
With the limits of accuracy in question it is concluded that no 
ions penetrate the aluminium either under no field or under the 
small field applied. 
M. Le Bon* has described an experiment showing that ions 
from a flame traversed a “cage de Faraday” formed of alumin- 
ium, the leak of his electroscope being very great when the 
gas from the flame was not cooled, but appreciably reduced 
when the gas was cooled. Now Campbell f has shown that M. 
Le Bon’s result that radio-activity accompanies chemical change 
* Revue Scientijique, iv. 18, 1902, p. 649. 
t Phil. Mag., vi. 52, 1905, p. 545. 
