IIEAT AND NOT LIGHT A MOTIVE POWER. 
135 
rate of about one complete revolution in two minutes. This 
was tried with my radiometer B. With A the velocity 
would be greater. According to the theory which has been 
suggested to me, the repulsion or attraction of the black discs 
is caused by the direct action of heat, and it seems to be able 
to account for all the phenomena I have as yet observed. The 
way in which it affects the discs is somewhat similar to that 
shown in the “ Trevelyan rocker.” Let us suppose a radiometer 
with good reflecting and absorbing discs to be at rest, a candle 
to be then placed at a distance of 6 inches from it ; the 
light or heat from the flame, having crossed 5 inches of 
common air, is incapable of heating the residuum of air within 
the globe. It is arrested and absorbed by a black surface, 
which becomes warm, and heats the adjacent particles of air. 
They expand, and give the disc a push into the cooler air 
behind, which the reflecting, non-radiating surface was in- 
capable of warming. The ordinary black repelling motion 
would be thus produced ; the greater the heat the more the 
expansion, and the quicker the repulsion. Remove the source 
of heat, and chill the globe. The friction soon overcomes the 
momentum, and the discs are at rest ; the air inside is warm ; 
the black absorbs a little of the heat close to its surface ; causes 
contraction of the air. A space is formed into which the disc 
is pushed by the hot air behind it — the backward or “ reverse 
motion ” would thus be caused. 
This reverse motion can never be very great, as the (motive) 
heat is limited in quantity, and it is expended by gradual 
conversion into mechanical action, which has friction to over- 
come, and also is lost by radiation from the outside surface of 
the globe. It is evident that unless the reflecting surfaces are 
good, this backward motion could not be obtained, as both 
sides of the vanes would cool the adjacent air by absorption, and 
the difference between the two actions woidd not be sufficient 
to cause the rotation. This is why I can get little, if any, 
reverse motion in my radiometer A. 
When heat is applied to the globe direct, as is done by 
touching it with the finger, &c., the action is somewhat dif- 
ferent. The glass becomes warm, and it warms the air within, 
producing expansion, which repels whichever side of a vane is 
nearest to the warmest part of the globe. If this happens to 
be the white side, no rotation can be produced, because the 
continued application of heat repels the white, and the opposite 
black disc (on the next approaching vane) absorbs the heat, and 
has, as in the first case, a strong tendency to move in the 
opposite direction ; but should it happen that a black disc is 
nearest to the warm glass, rotation immediately ensues. The 
black being repelled by the expanded air, and taking with it 
