406 Proceedings of the Royal Society 
distance without experiencing a collision. In the case of hydrogen, 
at atmospheric pressure, this distance, called the mean free path of 
a molecule, is the ^ Jowth part of an inch. If the hydrogen in a 
tube at atmospheric pressure be gradually removed by an air-pump 
until only a small fraction of that quantity of gas remains in the 
tube, then each molecule can on an average traverse a distance much 
greater without having a collision with another molecule. During 
the time of passing over this space their vibrations are more or less 
given off to the other. Thus they soon cease to be luminous, even 
if incandescent when they last collided with a molecule. These 
facts explain the action of the radiometer.* They also explain the 
fact observed in very high vacua, that there is no luminosity round 
the negative poleb of a vacuum tube (highly exhausted) when two 
poles inside it are connected with the poles of an induction coil. 
[The first experiment was here shown, when it was seen that in a 
very perfect vacuum no light was seen in the tube except a phos- 
phorescence of the glass tube, due to the impulse of the molecules of 
gas upon it. When water-vapour w^as admitted by heating a sub- 
sidiary bulb containing caustic potash, the ordinary phenomena of 
vacuum tubes were gradually developed.] 
The impact of these molecules on German glass or other phosphor- 
escent substance produces intense phosphorescent light. [This was 
shown by placing in these vacuum tubes phosphorescent sulphides 
and rubies, and by directing the discharge upon German glass.] 
When the electric density of a considerable portion of the nega- 
tive electrode is uniform, we should be led to expect that the mole- 
cules of gas which strike it should be driven off perpendicular to the 
surface, that being the direction in which the repulsive action 
between the electrified electrode and the similarly electrified mole- 
cules should take place. This is exactly what happens. If the 
negative electrode be a circular disc, the molecules are driven off to 
the opposite side of the tube or bulb. This direction depends in 
no way upon the position of the positive pole. In this case the 
author remarked on the existence of a black mark in the centre of 
* See Tait and Dewar, Nature, 1875. But also see Clerk Maxwell, R.S. , 1879. 
t The molecules of gas certainly fly off from the negative pole in obedience 
to the above laws. At present we know little about what happens at the 
positive pole, 
