94 
MR. W. CROOKES OK REPUESIOK RESULTING- FROM RADIATION. 
Following out this idea, a series of eight radiometers were made, having flies as shown 
in plan in fig. 7, A, B, C, D, E, F, G, H. The cups were of thin aluminium, and clear 
mica disks were attached in the positions shown in the drawings. In the pairs A and B 
the aluminium cups were bright on each side, and the mica screens were, in A, facing 
the concave side, and in B facing the convex side. In the pairs C and D the aluminium 
cups were blacked on the concave side; in E and F they were blacked on the convex 
side ; and in G and H on both sides. The screens in each pair followed the arrange¬ 
ment in A and B. The whole series compares with the similar series without screens 
given in Part V., par. 322. 
Experiments were tried with each of these in succession ; the radiometers being 
exposed—1, to the direct rays of a candle, 3 inches from the bulb; 2, to the same 
candle, after having cut off the light from the concave side by an opaque screen ; and 
3, after screening the light from the convex side. 
I will not describe each experiment in detail, but will briefly record that in each 
case, when the eight radiometers were exposed to the candle shining on both sides of 
the fly, the movement was one of rotation, being very strong with G; less so with 
B, C, F, and H; moderate with E; and only slight with D. 
When the opaque screen was placed so as to shade the candle light from the concave 
side of the fly, there was strong rotation in the case of E, F, G, and H ; moderate rota¬ 
tion with A, B, and C ; but no rotation with D. 
When the light was cut off from the convex side of the fly, there was strong rotation 
with B, C, G, and H ; moderate with D and F ; and no rotation with E. 
In every case where rotation was produced the direction of movement was the same, 
the mica screen approaching the light and the aluminium cups being repelled, irre¬ 
spective of the side turned to the light or of the position of the black coating. 
399. On reference to the previously-quoted experiments on similar radiometers 
without mica screens in front (322), it will be seen that in two instances only was 
there no rotation, viz., 1, when the cups were blacked on the concave side, and the 
light was only allowed to shine on the convex side ; and 2, when the cups were 
blacked on the convex side, and the light was allowed to shine only on the concave 
side. It will be seen in the present experiments that rotation is produced in all but 
the two analogous forms, viz., D and E, the former when the concave side is screened 
