MU. W. CKOOKES ON REPULSION RESULTING EROM RADIATION. 
349 
167. On removing the candle a remarkable phenomenon was observed. The arms 
stopped and immediately commenced revolving the negative way, keeping up rotation 
for more than ten minutes, and being little inferior in speed to what it was when the 
candle shone on it. 
The whole of the bulb was heated with a Bunsen burner ; whilst it was getting hot 
the aluminium arms revolved rapidly in the normal direction ; but as soon as the source 
of heat was removed and cooling commenced, negative rotation set up, and continued 
with great energy till the whole thing was cold. It appeared as if the negative 
movement during cooling was equal in amount to the positive movement as it was 
being heated. 
168. The very sensitive pith radiometer used in experiments 152 et seq. was now 
experimented with. A little ether was dropped on the bulb as it was standing still in 
a faint light. The evaporation of the ether caused a chilling of the instrument and a 
rapid abstraction of heat from the fly. It commenced to move in the positive 
direction, and increased quickly in speed until it revolved at a rate of one in four 
seconds. This movement kept up for several minutes, and as it slackened it could at 
any time be revived by a few drops of ether on the bulb. 
When in rapid positive movement, produced in the above manner, a hot glass shade 
(161, 165) was placed over the radiometer. The movement slackened, the arms quickly 
came to rest, and then immediately revolved in the negative direction, acquiring a speed 
of about two revolutions a minute, and keeping up this negative movement for more than 
ten minutes. 
169. The radiometer was again set in rapid positive rotation by dropping ether on the 
top of the bulb. I applied the tip of one finger to the side of the bulb for ten seconds. 
The rotation stopped, and I could not start it again for some minutes, although I dropped 
ether on the bulb several times in the interval. 
When the radiometer had once more acquired the temperature of the air, I dropped 
ether on the bulb, not in cethe ntre, but so that the ether wetted only half of the bulb. 
The arm which was nearest to the part most wetted by the ether rushed towards that 
part and remained, as it were, fixed opposite to it, refusing to move away, although I 
tried to equalize the temperature by dropping ether on the other parts of the bulb, and 
to drive it round by bringing a candle near. Not until the candle came within 6 inches 
of the bulb did the arms begin to rotate, which they then did with a rush, as if 
suddenly relieved from a state of tension. 
170. These results appear at first sight anomalous. I think, however, they admit of 
an explanation which is in keeping with the facts, if I may make one supposition. 
The great difference between a lampblacked and a white surface is only an optical 
one. Pith reflects a considerable amount of light, and lampblack absorbs a large 
quantity of light, but it is unsafe to carry the analogy into the ultra-red region of 
the spectrum. We know of many white powders, optically identical, which in their 
thermic relations are as wide apart as pith and lampblack ( e . cj. powdered alum and 
mdccclsxvi. 3 c 
