104 MR, W. CROOKES OK REPULSION RESULTING FROM RADIATION. 
aluminium. The indicator i, which can be put into any part of the bulb, is a very 
small and light fly of a radiometer, the vanes being of clear mica not blacked on 
either side, and each being about 2 millims. square. They rotate under the influence 
of a very slight force, and the direction of rotation gives the direction of the force, and 
also shows on which side it is increasing or diminishing. 
Pig. 15. 
419. Fig. 15, A, is a plan of the hemi-cylinder a, with the mica screen s s put as far 
from it as possible, so as to exert no appreciable influence on the movement of the 
indicator. The small circles, 1, 2, 3, &c., show the different positions in which the 
indicator was placed, the candle being at c when the indicator was (apparently) below 
the hemi-cylinder, and at c when it was above—the object being to prevent the light 
from the candle, or the force from the heated glass near it, from falling direct on the 
indicator ; my wish being that the indicator should be moved as much as possible by 
the force generated by the hot aluminium. 
The direction of the arrow heads on the circles shows the direction of rotation taken 
by the indicator, and the number of arrow heads show the relative velocity. In 
positions 1, 2, 5, 7, 12, 10, there was no rotation. A reference to the hypothetical lines 
of force given in fig. 14, B, shows that along the line of positions 7, 12, 10, there 
would be equality of force on each side. At 5 also is another point of neutrality, the 
indicator being midway between the positive rays of force a h, fig. 14, B, and the 
negative rays c cl, in the same figure. Positions 8 and 6 are those of greatest move¬ 
ment ; this also is intelligible, as the indicator is struck almost entirely on one side. 
Position 8 is slightly the most favourable to sensitiveness ; this is probably owing to 
that end of the aluminium hemi-cylinder being a little warmer than the other, from 
