xy AND OF ORDINARY MATERIALS 373 
has been deduced from the study of the conductivity of gases in 
closed vessels. The conductivity is extremely minute, and special 
methods are required to determine it with accuracy. A brief 
account will now be given of the gradual growth of our knowledge 
on this important question. 
218. Conductivity of air in closed vessels. Since the 
time of Coulomb onwards several investigators have believed that 
a charged conductor placed inside a closed vessel lost its charge 
more rapidly than could be explained by the conduction leak 
across the insulating support. Matteucci, as early as 1850, observed 
that the rate of loss of charge was independent of the potential. 
Boys, by using quartz insulators of different lengths and diameters, 
arrived at the conclusion that the leakage must in part take place 
through the air. This loss of charge in a closed vessel was believed 
to be due in some way to the presence of dust particles in the air. 
On the discovery that gases became temporary conductors of 
electricity under the influence of Rontgen rays and the rays from 
radio-active substances, attention was again drawn to this question. 
Geitel? and C. T. R. Wilson? independently attacked the problem, 
and both came to the conclusion that the loss of charge was due 
to a constant ionization of the air in the closed vessel. Geitel 
employed in his experiments an apparatus similar to that shown 
in Fig. 59. The loss of charge of an Exner electroscope, with the 
cylinder of wire netting Z attached, was observed in a closed vessel 
containing about 30 litres of air. The electroscope system was 
found to diminish in potential at the rate of about 40 volts per 
hour, and this leakage was shown not to be due to a want of 
insulation of the supports. 
Wilson, on the other hand, used a vessel of very small volume, 
in order to work with air which could be completely freed from 
dust. In the first experiments a silvered glass vessel with a 
volume of only 163 cc. was employed. The experimental arrange- 
ment is shown in Fig. 60. 
The conductor, of which the loss of charge was to be measured, 
was placed near the centre of the vessel A. It consisted of a 
1 Phys. Zeit. 2, p. 116, 1900. 
2 Proc. Camb. Phil. Soc. 11, p. 52,1900. Proc. Roy. Soc. 68, p. 152, 1901. 
