362 
ME. W. CEOOKES ON EEPULSION RESULTING EEOM RADIATION. 
A comparison of these figures with those usually given in text-books to represent the 
distribution of heat in the spectrum will be a sufficient proof that the mechanical action 
of radiation is as much a function of the luminous rays as it is of the dark heat-rays. 
194. In the intervals of spectrum work I tried many other experiments with the 
apparatus. I was anxious to get the exact time of the oscillations of the beam in a 
vacuum, and tried many ways of starting the initial impulse. 
A candle held close to the screen, and the shutter momentarily opened and closed, 
sent the index with some violence to the extreme limit of the scale. It then slowly 
came back to zero and there stopped. Magnesium wire used in the same way produced 
the same effect. There was no oscillation, as there would have been if the impulse had 
been given by a material blow. The movement of the beam, as shown by the spot of 
light, seemed as if it were held in check by a force acting the whole time of its move- 
ment, and not only for the time the light acted. The impression conveyed was that the 
beam was swinging in a viscous fluid, and the more perfect the vacuum the greater 
appeared to be the viscosity (107). Thinking that the heat-rays from the candle might 
be absorbed by the black pith and so raise its temperature, I interposed screens of water, 
alum, and thick plates of glass, so as to cut off the ultra-red rays. Still the apparent 
resistance to free oscillation continued. 
I then, without interfering with the vacuum, and without letting radiation fall on the 
pith surface, gave the apparatus a sudden twist round, so as to cause the beam to knock 
against the side of the tube. This set it swinging through a large arc, and the oscilla- 
tions kept up with perfect freedom for several minutes, declining in amplitude at each 
oscillation till the beam ultimately came to zero. This perfectly free movement is in 
strong contrast to the constraint under which the beam moves when the initial blow is 
given by a ray of light instead of by a mechanical push. 
The same effect was noticed during the experiments with the spectrum. A ray in the 
blue, falling on the pith, would drive the index twenty divisions along the scale. It 
would then gently come back to zero, where it would stop, occupying the same time to 
come back as it did to go forward. If, however, after the action of the light had entirely 
ceased, I gave the tube a slight jerk, so as to cause the beam to swing through the same 
arc, the index on returning to zero would pass perhaps 15 degrees the other side, and 
would thus oscillate for some time from one side to the other of zero, taking many 
minutes to come to rest. 
195. This phenomenon enables me to advance a step towards an explanation of the 
mechanical action of radiation, although I fear I shall have to make some assumptions 
which are scarcely yet proved. 
A ray of light falls on a white surface and is reflected back again ; it does no work 
there (170, 171); but if the ray falls on a black surface it is absorbed and quenched. 
What becomes of it ? It seems to me probable that the ray becomes converted into 
thermometric heat, and that its energy is in whole or in great part used up in raising 
the temperature of the dark body. But having thus become warm, the powerful 
