MR. W. CROOKES ON REPULSION RESULTING FROM RADIATION. 
305 
on the supposition that the movement is due to a push away from the source of 
radiation. 
In an experiment tried in 1875* 1 used a pendulum with a magnesium weight at the 
end (99), exposed in vacuo to the radiation from a platinum coil ignited by electricity. 
The distance from the pendulum bob and the spiral was 7 millims., and two series of 
observations were given in tabular form (100), showing the direction and amplitude 
of the swing of the pendulum at different barometric heights when the wire was 
ignited. The hot wire gave attraction in air of ordinary density ; this kept pretty 
constant up to a pressure of about 20 millims., when the attraction began to increase. 
At about 1 millim. the attraction was at its maximum. Above this exhaustion the 
attraction suddenly dropped and changed to repulsion, which kept on increasing up to 
the highest exhaustion I was then able to get.t 
In the present experiments the parallel is sufficiently close to show that the same 
causes are at work. Negative rotation is apparent 12 millims. below a vacuum, and 
a negative movement would be detected at much lower exhaustions. This negative 
movement rises to a maximum, suddenly sinks to zero as the exhaustion proceeds, and 
gives place to a positive rotation which keeps up at nearly a maximum speed and at 
an exhaustion which would almost stop an aluminium radiometer (334). 
342. Were the rotation in moderately rarefied air due to air currents rising from 
the hot wire, it is difficult to understand why it should be negative. To eliminate as 
far as possible the action of air currents, the apparatus was modified by raising the 
platinum wire ring so that it was about 5 millims. above the sloping mica vanes, 
instead of under them, everything else remaining the same. 
At low exhaustions, when the wire was ignited, the fly rotated slowly in this 
direction The hot wire being above, this is a negative movement, as it is opposite 
to what would be caused by pressure from the wire acting in the vanes, and is there- 
fore quite in keeping with the results obtained with the hot wire beneath. 
When the rarefaction reached 0'5 mi ll im . the negative rotation ceased. The fly 
oscillated to one side and the other without rotating. At higher exhaustions the fly 
moved positively , and this movement increased in strength as the rarefaction 
increased. 
343. When the exhaustion was good, and the vanes were still, a piece of wet 
blotting-paper was put on the top plate of glass to cool it slightly. Rapid negative 
rotation took place . Making battery-contact immediately reversed this move- 
ment. Touching the top plate of glass with the finger caused positive rotation. 
344. These movements are of the same kind as those given by the apparatus 
in which the wire was beneath the vanes. It is difficult to prevent the action of 
* Phil. Trans., Yol. 165, pt. 2, pp. 528-532, March 20, 1875. 
f In my first paper on this subject (Phil. Trans., Yol. 164, pt. 2, pp. 513-518, pars. 37-46, August 12, 
1873) I also described analogous results with an ignited spiral acting on a balanced brass ball. 
MDCCCLXXVIII. 2 R 
