88 MR. W. CROOKES ON REPULSION RESULTING FROM RADIATION. 
time to time during the last few years—experiments which at the time were isolated 
in then- bearings, but which now fit into then places. 
388. My first experiments were directed to ascertain whether the fines of pressure 
could be deflected from then- path by a thin plate of mica fixed on the fly of an 
ordinary radiometer. A two-disk radiometer was made, having flat roasted mica 
disks, lampblacked on one side (fig. 1, a). The lampblack surface is represented in 
Fig. 1. 
Fig. 2. 
.the figure by a row of dots close to the surface. In front of the black surface, and 
separated from it by a space of 1 milfim., is fixed a thin disk of clear mica, b, about 
1 milfim. larger in diameter than the blacked disk. When properly exhausted and 
exposed to the radiation from a candle, the normal direction of movement would be 
opposite to that shown by the arrows, supposing no screen were present. The screen 
has, however, changed the direction ; the pressure, instead of reacting between the 
black surfaces and the sides of the glass case opposed to them, is deflected back, as 
shown at c, and the result is rapid rotation in the direction shown by the arrows, the 
black side now approaching the fight. 
389. To test this action more completely, another thin mica disk was fixed on the 
plain side of the blacked disk, the latter being now guarded on each side, as in fig. 2. 
The effect of this is to cause an almost total loss of sensitiveness, the fly now only 
moving very slowly in full sunshine. 
390. Two explanations of the action of the screen in fig. 1 suggest themselves. 
According to one, the radiation passes through the clear mica screen, and generates 
molecular disturbance on the black surface. The direction of greatest stress being- 
prevented by the screen from exerting itself between the black surface and the glass 
bulb, is reflected back in the direction of the arrow, c, and produces negative rotation. 
Another view of the action is this. The radiation, falling on the compound fly, warms 
it in proportion to the amount absorbed. Molecular pressure is exerted on that side 
which is most easily warmed, and winch most readily parts with its heat to the adja¬ 
cent gaseous molecules. The roasted mica, a, coated on the inner side with lampblack, 
answers these conditions better than the clear mica ; it warms up as a whole, and gives 
up its heat to the gaseous molecules on the outside, because there they are free to 
carry off the heat. This side, therefore, becomes the driving surface. 
This view is rendered less probable on referring to the behaviour to a standard fight 
