ME. W. CROOKES ON REPULSION RESULTING FROM RADIATION. 
105 
the conduction of' heat away along the arm. The direction of movement in 8 and 6 is 
also in accordance with theory. The slight action on the convex side of the curved 
metal plate is probably due to the lines of force acting here in both directions to and 
fro between the metal and the glass. This would cause the pressure here to be much 
greater than on the concave side, but at the same time would cause much less rotation. 
The slight movement in position 4, and its direction, is caused by the molecular 
disturbance generated on the surface of the aluminium, where it is nearest the glass, 
being struck back in somewhat greater quantity than at position 3, where the interval 
between the metal and glass is greater. In positions 1 and 2 the lines of force 
issuing from the convex surface strike each side of the indicator equally, the opposing 
face of the glass being too far off to cause a reaction. 
420. Fig. 15, B, represents the position in which the mica screen, s s, was next 
placed, touching the convex side of the hemi-cylinder. The candle was at c, and the 
circles represent the successive positions of the indicator. The screen has now entirely 
altered the disposition of the lines of pressure. Position 3, where in the last experi¬ 
ment the indicator was almost stationary, is now the position of greatest speed. The 
indicator, in fact, has the rays which, in the absence of the screen, struck it on both 
sides, now acting on one side only, and the result is a rapid rotation in the direction of 
the arrows. No. 1 is the next best position, and then 2. The velocity and direction 
of movement in these two positions are clearly due to rays of force not acting so 
strongly on the side next the screen. As the indicator is moved in the other direction, 
away from the most sensitive position 3, the influence of the screen diminishes; at 4 
the rotation is slight; and at 5 there is no movement. In position 6 the influence of 
the screen has almost disappeared, the motion now being opposite to that at No. 4, 
At 7 there is another position of neutrality, and at 8 the motion is again reversed. 
The movements in the last three positions correspond w T ith those in similar positions 
when the screen was absent (fig. 15, A). 
421. The screen was now moved round till it touched the hemi-cylinder on the 
concave side, as shown at ss in fig. 15, C. On comparing the direction taken by the 
indicator with that in fig. 15, A, it will be seen that no change is produced in positions 
1, 2, and 3 (corresponding to 6, 8, 9, fig. 15, A), the speed only being reduced. In 
positions 4 and 5 the interference of the screen has caused the direction of rotation to 
change. The lines of pressure, being deflected by the mica, now radiate towards the 
centre of the bulb, and being stronger the nearer they are to the generating surface 
of aluminium, cause the indicator to rotate, as shown at 4 and 5. In position 6 there 
is no movement when the screen is close to the hemi-cylinder, but when it is brought 
into the position shown by the dotted line s' s', rays are deflected, and the indicator is 
rotated in the direction shown by the dotted arrow. 
422. The screen thus renders evident the existence of active lines of force at great 
distances from the generating surfaces. A reference back to fig. 14, B, shows certain 
hypothetical lines of force stretching across the bulb from the generating hemi-cylinder 
MDCCCLXXIX. 
p 
