100 
MR. W. CROOKES ON REPULSION RESULTING FROM RADIATION. 
The vanes were now set touching the convex side as at fig. 13 a, B. The exhaus¬ 
tion continued, and when it had reached 117 M. permanent negative rotation com¬ 
menced in the direction of the arrow. As the position of the screen was gradually 
shifted from B to A, the exhaustion remaining 117 M, the rotation got quicker and 
quicker, showing that for this exhaustion, position A is most sensitive, and position 
B least sensitive. 
410. The screen was put into position 13 a, B, touching the convex side of the 
vane, and the pumping was continued. The positive rotation got a little quicker, 
until at 82 M the speed was 5 revolutions a minute. The screen was moved over 
to position A, the pressure remaining the same, when the speed in the positive 
direction was 13 a minute. As the screen was moved round from touching the 
convex side to touching the concave side, the speed of positive rotation of the fly 
increased. 
411. The screens were now replaced in position B, touching the convex side of 
the vanes, and pumping was continued. As the exhaustion increased the rotation got 
slower and slower, until at a pressure of 87 M, the vanes stopped. Before the 
movement ceased the vanes acted as if, in revolving, they came successively under 
the influence of a strong attractive force. Each vane went quickly up to the spot 
of warm glass opposite the candle, then hesitated, and got past with difficulty. This 
hesitation became more and more decided, till at last a vane refused to pass the dead 
centre, when it swung back and the fly set equidistant from the candle. 
On continuing the exhaustion a tendency to negative movement was soon observed, 
and at an exhaustion of 12 M, permanent negative rotation commenced, the screens 
all this time remaining in position B. At this pressure of 12 M, the speed in 
position A was 46 times a minute, positive. As the screen was moved from the 
concave side the speed got less. When it was 6 millims. from the convex surface 
(position C, fig. L3 b), the speed was 3 revolutions a minute, positive ; when they were 
3 millims. apart (position D), the rotation stopped; and when the screens and convex 
surfaces touched (position B) there was negative rotation, at the rate of about 1 turn 
a minute. 
Exhaustion 12 M. 
412. When in position 13 6, A, no change in direction of rotation takes place, what¬ 
ever the exhaustion. There is an increase of speed up to a rarefaction of about 
15 M, after which further exhaustion has little or no effect on the speed. 
