MR. W. CROOKES ON REPULSION RESULTING FROM RADIATION. 
Ill 
also at b" 1, no movement took place after exhaustion, when the platinum wire was 
ignited. In the reverse position, the concavity being downwards, as at b" 5, there was 
also no movement. 
With the cup in position 2, there was no movement on igniting the platinum wire 
and keeping it hot. When the current was cut off, and the wire and inside of the 
apparatus allowed to cool, good rotation took place in the direction shown , A), . 
When the cup was turned over a little less, so as to get it about halfway between 
positions 1 and 2, no movement was produced either on heating or cooling. 
With the cup in position 3, heating the wire caused the fly to make 18 revolutions 
in the direction [) ; . It then oscillated and came to rest, the wire remaining hot. On 
turning off the battery current no further movement took place. 
In position 4 the hot wire gave a rotation in the direction _\) i , at the rate of 150 a 
minute. When still further turned over, till the cup was halfway between positions 
4 and 5, the speed was 120 a minute in the same direction. 
433. The other fly experimented with also consists of a single arm, with cap in the 
centre and counterpoise at one end. Instead of an adjustable cup, the other end 
Fig. 19. 
supports rigidly a right-angled hollow mica prism, shown at tig. 19 in end view and 
perspective. 
After exhaustion the current was turned on, and the fly immediately rotated in the 
direction shown by the arrows, keeping up continuously as long as the wire remained 
hot. 
This experiment, I think, proves that the direction of the pressure is not wholly 
normal to the surface on which it is generated. Were it so, no movement would have 
taken place, as the base of the prism was in a parallel plane to the blacked mica disk. 
A tangential direction of the lines of force will, however, exert more pressure on the 
perpendicular face of the prism than on the hypotenuse, and will drive the fly round, 
as shown at fig. 19, A, in the direction given by experiment. 
THE TURBINE RADIOMETER, 
434. Experiments tried with the apparatus described in a former paper (336, 345, 
354), where thin mica vanes, inclined at an angle of 45° to the horizontal plane, form 
the fly of the radiometer, show that this form of fly possesses advantages in cases 
where the incident radiation falls at an angle on the instrument; and the favourably- 
