MR. W. CROOKES ON REPULSION RESULTING FROM RADIATION. 
279 
disk moving positively* and the other negatively, according to the one on which the 
light was allowed to shine. The disk which moved the negative way was somewhat 
crumpled, and on examining the action more closely I found that the slight curvature 
of the gold leaf thereby produced was the cause of the anomalous action. By means 
of screens with small apertures in front of the candle, and by concentrating a small 
image of the candle flame on various parts of the crumpled disk by means of a lens, 
motion couid be produced in either direction ; repulsion being produced when the light 
fell on a conveXj and attraction when it shone on a concave portion of the disk. The 
gold disk which behaved abnormally happened to be slightly turned up at the edge 
next the tube, so as to present a bright convex surface to the candle. 
265. Experiments were now commenced with a view to clear up this interfering 
action due to the shape of the vanes. As the experiments are very numerous and only 
differ from one another in small modifications or details, which have been successively 
introduced in order to ascertain what are the accidental and what the necessary 
accompaniments of the phenomena, I will only describe the first in detail, and will 
give the others of the series as briefly as possible consistent with clearness. 
Plates 12 millims. square, cut from thin aluminium foil, were mounted diamond- 
wise on arms and supported on the needle point inside the bulb shown in fig. 5. The 
plates were lampblacked on sides facing opposite ways, and the apparatus was well 
exhausted. The vanes behaved like an ordinary metal radiometer in respect to light 
Fig. 6. Fig. 7. 
and radiant heat. Fig. 6 shows the elevation and plan of the fly, the dotted side 
representing the one which was lampblacked. The arrows show the direction of positive 
rotation! when exposed to the light of a standard candle 3‘5 inches off. The 
exhaustion in this and in all cases, except where otherwise specified, was carried to the 
point at which the fly moved most rapidly to the candle. 
266. The outer corners of the aluminium plates were now turned up at an angle of 
45°, 4 millims. of the two sides being turned up, leaving 8 millims. flat, as shown in 
fig. 7. They were lampblacked on the inside, as shown in the figure by dots. 
A lighted candle, 3 '5 inches off, caused very slow and feeble positive rotation. On 
* I call the rotation, positive when the black or driving side is repelled, and negative when the side 
which, under ordinary circumstances, would be the driving side, moves towards the light. 
f In the experiments illustrated by figs. 6 to 11 (pars. 265 to 272), the direction of positive rotation is 
shown by the arrows. 
