130 On the Behaviour of Electricity in Rarefied Air. 
tap, the pump was removed and the spherical conductor of an 
ordinary electrical machine feebly charged was brought near 
to it. When the tube was rapidly moved to and fro in the 
neighbourhood of the conductor, it was illuminated by an in- 
tense light, but remained dark if held still or if made to 
describe a circle round the conductor. This shows clearly 
that it was an induced current produced in the interior of the 
tube which caused the light ; in fact, no induction-current 
was produced in the second case, and consequently there was 
no light. Then the pressure of the air in the tube was in- 
creased to about 350 millim. The current of the induction- 
coil now produced a spark between the electrodes, but no 
induced light was to be seen on approaching the tube to a 
charged conductor or on withdrawing it. In this experiment 
the tube had no coatings of tinfoil, and the portions of the 
electrodes outside the tube were covered with insulating 
material. 
Corresponding experiments with the ring-shaped tube 
yielded the same result. 
Fifth Hxperiment.—In a tube like the first described the 
air was exhausted so far that the current from the induction- 
coil would not pass between the electrodes. If, then, the one 
half of the tube was rubbed with a suitable cushion, it began 
to light up altogether, the electricity contained in the tube 
being put into motion by means of the electricity produced by 
the external friction. If, on the other hand, the tube contained 
air ata high pressure it was impossible to produce light, 
although the current of the induction-coil easily passed from 
the one electrode to the other. The same result was obtained 
with the ring-shaped tube if the air contained in it was suf- 
ficiently rarefied. The whole tube became luminous when 
only one fourth of it was rubbed. These experiments thus 
completely confirm the previous ones. 
We see thus that the admitted fact that an induced current 
passing through a gas increases in intensity in proportion as 
the gas is exhausted, but that, if the exhaustion is continued, it 
decreases, is not explained by supposing that from the moment 
when a certain limit is reached the resistance increases as the 
exhaustion is continued. On the contrary, experiments show 
that the fact in question is to be explained by the increase 
which takes place in the resistance offered to the electricity in 
its passage from the electrodes into the gas as the exhaustion 
of the gas increases, until it becomes so great as to present an 
insurmountable obstacle to the passage of the electricity. 
There is thus no experimental foundation for the assumption 
that an absolute vacuum is an insulator. Since the conducting 
