MR. W. CROOKES OK REPULSION RESULTING FROM RADIATION. 
107 
In positions 10 and 1 no rotation was produced ; rotation in one direction was 
produced at 9, and in the other direction at 11 ; these results agree with those 
given with apparatus 15. In position 2, rotation was obtained in the direction of the 
arrow heads, and it was still stronger in the same direction at 3. These also agree 
Fig. 16. 
with theory. At 4 and 5 there was slow rotation in the direction of the arrows. 
Were the lines of force to follow the path shown by the dotted lines b b' b" b"", &c., 
there should have been no rotation in these positions, but supposing the line c c" 
to have struck the indicator in position 5, there might be slow rotation in the 
opposite direction to what is here shown. 
The black lines d d' d", e e' e", show the direction which the lines of pressure 
might take, supposing they were merely deflected out of their course by the screen 
B. On this supposition the movements of the exploring fly are quite reasonable. 
At 4 the force, striking one side only, would cause rotation in the direction as 
shown, and at 5 rotation in the same direction would also take place, but not with 
the same speed, owing to its being more surrounded by the stream. On moving 
away the screens, the fly remaining at position 5, the rotation stopped. The 
indicator was successively brought to a great many positions in respect to the 
screens, and the screens were also moved about to different parts of the bulb, and 
the results were in all cases easy of explanation on the de-flection theory, whilst 
they were sometimes contradictory on the re-flection theory. At 8 there was rota¬ 
tion due to the screen B cutting off some of the force from one side of the fly. 
At 6 and 7 there was no rotation, the screen B' having obstructed all the force. 
The molecular pressure at the exhaustion found best for these experiments 
diminishes greatly as the distance from the plate A increases; it is not easy, therefore, 
to say whether a diminished speed, as in positions 2, 4, and 5, is due simply to 
increased distance, or to some action of the screens; but from the indications given 
generally by the exploring fly, I have come to the conclusion that the force is not 
reflected in the sense that light is reflected, but that it is merely deflected from 
a surface against which it strikes, as would be the case with a molar wind. 
