iOG MR. W. CROOKES OK REPULSION RESULTING FROM RADIATION. 
to cl, and being obstructed by tlie screen. The present apparatus seemed capable of 
rendering those rays sensible. The indicator was accordingly placed in position 1, 
fig. 15, D, the screen being quite away from it and the candle being at c. The rays 
of pressure are here supposed to start from the hemi-cyfinder, and strike both sides of 
the indicator equally ; the result is that no movement is produced. Without moving 
the indicator, the screen was brought into position s s. The indicator immediately 
rotated rapidly in the direction of the arrow heads. The action of the screen here is 
to strike back the impinging molecules, and thus cause a neutralisation of the pressure 
acting on the left side of the indicator. The line of pressure acting on the right side 
of the indicator passes on unimpeded, and therefore causes rotation. This explanation 
was verified by putting the indicator in position 2, and gradually bringing the screen 
nearer to it. With the screen at s, there was no movement. When the screen got to 
position s' s', movement of the indicator commenced ; and when it reached position s" s", 
the rotation of the indicator was rapid. A glance at the lines of force in fig. 14, B, 
will show how closely theory and experiment agree. 
423. The action of heat applied to the bulb was now tried, the indicator being 
in position 1, and the screen at s s, fig. 15, E. A spirit flame was applied for a 
few seconds to the bulb at a b. There was rapid rotation of the indicator in the 
direction of the arrow heads, with strong negative rotation on cooling. When 
the action had ceased, the part of the bulb between c and d was heated ; the result 
was rapid rotation in the opposite direction to the arrow heads. These movements 
are perfectly explicable on the supposition that lines of pressure radiate from the 
hot internal surface of the glass, and, being partly cut off by the screen, strike 
the left or the right side of the indicator as the case may be. When the bulb is 
heated in such a part that the lines of force are entirely cut off, or are not at all 
obstructed by the screen, as at d e, or b c, no rotation of the fly is produced. 
424. In the hope of deciding whether the force was capable of true reflection 
from a surface on which it impinged, or whether it was only deflected out of its 
course as a current of air would be, the apparatus represented in plan in fig. 16 
was fitted up. A represents a plate of aluminium, lampblacked on each side, firmly 
fixed in the centre of a large glass globe capable of good exhaustion. B and B' are clear 
mica screens, a little larger than the aluminium plate, and capable of being held in 
any desired position by exterior magnets. A little exploring fly, as in the last ex¬ 
periment, shows the direction and strength of the lines of pressure. The positions 
into which this indicator was brought are shown by small circles. The candle was 
at c. It was found that similar results were obtained wherever the candle was 
placed, allowance being made for the interfering actions of the screens, but at c 
there was less interference than at any other part outside the bulb. The force is 
supposed to issue from the plate A in a direction normal to its surface. The dotted 
lines a a' a" a'", b b' b" b'", c c c" c'", show the direction taken by the lines of pres¬ 
sure, supposing specular reflection to take place from the surfaces of B and B', 
