MR. W. CROOKES ON REPULSION RESULTING FROM RADIATION. 
95 
from the light, and the latter when the convex side is screened from the light. The 
screen in these cases, being interposed between the cups and the light, could not assist 
rotation. In other ca.ses the mica screen appears to intensify the action of repulsion. 
400. It was now necessary to ascertain what effect would be produced by varying 
the distance of the screens from the metallic cup-shaped vanes. The simplest form of 
cups, with no lampblack on them, was chosen, and to avoid the difficulty which was 
sometimes found in starting the rotation and keeping it uniform, four vanes were used 
instead of two. The instrument had the form shown in elevation in fig. 8, the 
aluminium cups are affixed to rigid arms pivoted on a glass cap and needle point, 
and the four clear mica disks, acting as screens, are also pivoted on a quadruple 
arm capable of adjustment in any relative position in respect to .the arms carrying the 
cups. 
Fig. 8. 
The candle was kept at a distance of 4 inches from the bulb. The first experiment 
was tried with the disks as close as possible to the convex side of the cup, but not 
touching, as in position A. Rotation took place in the direction of the arrow, at the 
rate of 6 revolutions a minute. 
In position B the screens were adjusted 2 millims. from the convex surface. The 
direction of rotation was as before, but the speed was reduced to 3 revolutions a 
minute. 
The space between the convex side and the screens was now increased to 3 millims., 
as in position C. No rotation took place. 
The vanes were still further separated from the cups, being now 6 millims. from the 
convex sides, as in position D. The direction of movement changed, rotation being 
produced, in the direction of the arrows, at the rate of 6 ‘5 revolutions a minute. 
When the screens were midway between the cups, i.e., 10 millims. from the convex 
side and 10 millims. from the plane of the concave side, as in position E, the revolu¬ 
tions in the direction of the arrow were 12a minute. 
In position F, with the screens 5 millims. from the convex sides of the vanes, the 
speed in the direction of the arrows was 30 a minute. 
When the disks were brought as close as possible to the concave sides of the cups, 
as in position G, the speed was 50 a minute. 
401. These experiments, taken in conjunction with those immediately preceding, 
prove that, under the influence of radiation, a pressure is exerted between each side of 
the metal cup and the side of the glass case facing it, the pressure from the convex side 
