316 
ME. W. CROOKES ON REPULSION RESULTING FROM RADIATION. 
By gently tilting the apparatus, I can make the rotating vanes strike the support. 
They are thus stopped, and the disk immediately commences its positive rotation. 
The drag of viscosity being removed, molecular pressure prevails. 
The deflection of the molecular pressure, by the vanes, on to the disk, turning the 
disk in the direction opposite to that of the vanes, may also be represented by 
fig. 33, par. 364, which shows the similar action of a current of hot air. 
381. Whilst this action was going on, the disk stationary and the vanes in full 
rotation, I turned off the battery current. The vanes continued to move in the posi- 
tive direction, and when their speed had sunk to about 100 revolutions a minute, the 
disk commenced to turn with the vanes, being dragged round by the viscosity, there 
now being no molecular pressure tending to give positive movement. 
382. Pressure 1*4 millionth of an atmosphere. — On first turning on the battery, both 
the vanes and disk turn positively. As the speed of the vanes increases, that of the 
disk diminishes ; but when the vanes are rotating at their uniform maximum speed, 
the disk does not quite stop but moves a little positively. The viscosity of the 
hydrogen at this rarefaction is only *0168 (as against - 057 at the normal pressure), and 
it is therefore not enough to quite neutralise the action of the molecular pressure. On 
turning the battery off, the vanes nearly stop before they exert any drag on the disk. 
At this pressure the spectrum of hydrogen in the vacuum tube only shows one line 
in the green (the F line), and that very faint. 
Fig. 34. 
383. In fig. 34 1 have plotted down the observations taken in air vacua in this appa- 
