306 MR. W. CROOKES OK REPULSION RESULTING EROM RADIATION. 
air currents : as the sloping vanes are eminently favourable for their detection, I 
therefore endeavoured to devise an instrument which should rotate without being 
unduly influenced by air currents, and with which I can get some insight into the 
direction of the lines of molecular pressure after passing and acting on the moving fly. 
Fig- 31. 
345. Fig. 3 1 represents the apparatus. Instead of being open and closed with 
a plate cemented on, the cylinder is now sealed at the top, so as to enable me to 
proceed to the highest exhaustion, which cannot be reached unless all the joints and 
connections are fused together. The platinum wire ring is shown at a, a, a, the 
sloping mica vanes are shown at b b. Above the vanes is a flat disk of clear mica, 
c c, having a glass cap in its centre, and easily rotating on a needle point. The vanes 
and the mica disk are supported independently of each other on separate needle points 
which are held in glass rods, d, d, d. 
346. In air of the ordinary pressure (Bar. = 761 millims.) on igniting the platinum 
ring to redness by a current from two Grove’s cells, both the vanes and disk rotate in 
this direction ; that is, the vanes go in the positive direction, supposing they are 
driven round either by molecular pressure from the wire or by air currents. The 
speed of the vanes is 13 '3 revolutions a minute, and that of the disk 1 a minute. 
347. I continued the exhaustion. At a pressure of 220 millims. on igniting the 
wire the rotations are the same as hi air, but the speed of the vanes has diminished 
to 7 ’5 revolutions a minute, and the disk scarcely rotates at all. 
Pressure 80 millims. — The disk will not rotate, but oscillates a little to and fro. 
The vanes still rotate positively at a speed of 1 ‘5 revolutions a minute. 
Pressure 34 millims. — The disk has ceased to move. The vanes move very slightly 
in the positive direction when the apparatus is tapped. 
Pressure 19 millims. — No movement whatever. The disk and vanes are as still 
when the wire is ignited as when it is cold. 
