92 
MR. W. CROOKES ON REPULSION RESULTING FROM RADIATION. 
described instruments (392), whilst they retain any position into which they are put 
by tapping. 
Fig. 5. 
The candle in these experiments was 3 inches from the bulb. In position A the 
fly revolved in the direction of the arrows at a speed of 40 revolutions a minute. 
When the edge c of the screen was 3 millims. from the vane a, the revolutions were 30 
a minute. When the distance was increased to 7 millims. the speed was reduced to 1 3 
revolutions a minute, and at 15 millims. distance the speed was only 4 revolutions a 
minute. A little beyond this distance the rotation stopped. 
The screens were now turned round till they touched the other sides of the vanes, 
as shown in position B. The speed was thereby greatly increased, being more than 60 
revolutions a minute. When the edge b was separated 3 millims. from the vanes, the 
speed was 44 revolutions a minute; at 7 millims. apart the speed was 21 revolutions a 
minute ; and at 15 millims. there were 10 revolutions a minute. At greater distances 
the speed rapidly diminished, and when the arms were at right angles to each other, 
as nr position C, no rotation could be obtained. 
395. The explanation of these movements is not difficult. At position C, when the 
vanes are not screened, the molecular pressure reacts equally on each side between the 
bright aluminium and the glass bulb, and no movement takes place. When the mica 
screen is put in position A, it offers obstruction to the lines of pressure between the 
inner half of one side of the aluminium plate and the side of the bulb, whilst the 
corresponding pressure is free to act on the other face. In position B the same occurs, 
but more of the aluminium plate is obscured by the mica screen. Hence, in each 
position the balance of pressure is exerted on the unscreened face of the plate, which 
retreats from the light, the available force being greater in position B than position A. 
396. The screens were put into position A, and the whole bulb was heated with a 
spirit lamp. Strong negative rotation took place (screened side retreating); this kept 
on for 2'5 minutes; the fly now changed its direction and rotated positively, keeping 
on slowly for 12 minutes. 
When the screens were in position B, the same phenomena occurred on heating the 
bulb and allowing it to cool, the first negative rotation being very rapid, and lasting 
3 minutes, and the succeeding positive rotation being slower, and lasting 13 minutes. 
The negative rotation lasts whilst the glass bulb is giving heat to the fly ; when the 
