132 
POPULAR SCIENCE REVIEW. 
inches of air, is incapable of heating the residuum of air within 
the globe, and unless the radiant heat is very great, the glass 
is not appreciably warmed. I placed one candle about 12 
inches from B, and found its light (or heat) capable of 
producing slow black repulsion. Eleven other candles were 
then lighted on the same side of the globe, and at the same 
distance from it; they produced quick rotation. Ten of them 
were then suddenly blown out ; the vanes soon became stationary, 
and reverse motion set in, the black discs moving against the 
light of the two candles. Under other conditions, these two 
candles would have caused them to be repelled with a velocity 
equal to If revolution per minute. With three lighted candles, 
and then two of them blown out, the same effect was produced, 
though not with the same intensity. By raising or lowering 
the flame of a benzine lamp, rotation in either direction can be 
produced ; the degree of reverse motion, I find, always depend- 
ing on the degree of diminution of light. These effects cannot 
be produced with radiometer A. 
Into a large glass jar, filled with a strong solution of alum, 
I placed A, the globe being surrounded with nearly three 
inches of the solution. When the solution was of the same 
temperature as the air of the room, I placed a lighted candle 
one inch from the jar, another exactly opposite, one on either 
side, and so increased the number to nine, but no movement of 
the discs was caused, except when the sixth candle was lighted. 
Then the discs moved through an angle of 1 35°, and again 
became stationary. The reason of this will presently be 
explained. When the tenth candle was lighted, the black 
discs were repelled with a velocity of 2J complete revolutions 
per minute ; with the eleventh candle, 3^ ; twelfth, 4 ; 
thirteenth, 4J ; fourteenth, 5 ; fifteenth, ; sixteenth, 8^. 
By gently stirring the top of the solution with a glass rod the 
movement was increased to lOf revolutions per minute. I 
find the action of light which has passed through alum solution 
much greater when the solution is not perfectly still. When 
the sixth candle was lighted an accidental shake given to the 
table produced a little motion in the liquid, and the discs 
moved as mentioned above. If the solution be stirred too 
violently, the colder liquid from the bottom is forced upwards, , 
and rotation ceases. This can also be effected by allowing a 
little of the solution, chilled in a pipette, to run down the side 
of the globe. By the same means, a little of the solution when 
heated will produce repulsion of either disc, as in the experiment 
with a warm finger, already mentioned under section I. 
I tried on B the effect of light after it had passed through 
different thicknesses of alum solution. The repulsion of the 
black was always much decreased, varying with the thickness 
