NOVEMBER 22, 1906| 
NATURE 
QI 
unless these conditions are fulfilled the disturbing forces 
due to heating of the surface, convection currents and radio- 
meter effects may easily have a large moment either way 
round c. But these disturbing forces take time to develop, 
as Nichols and Hull showed, while the tangential push 
of the light acts instantly. Always when the beam is first 
directed on to B the motion in the first second or two 
is away from aB. 
It has been urged that this experiment is not con- 
clusive in that the lampblack is granular, and the force 
observed may be due to normal pressure against the sides 
of the grains. But if the back surface of the disc is 
blackened, so that the surface is much smoother, the action 
is as great. 
Another form of the experiment which we have lately 
ae 
Fic. 4. 
made is perhaps better. A horizontal disc of mica, about 
2 inches in diameter, is suspended in the case by a quartz 
fibre (Fig. 4). The dise is blackened on its under face. Ifa 
beam of light aB is incident at 45° at B, it tends to push the 
dise one way round. The gas action due to heating may 
possibly, and sometimes does, act against this push. But 
if an equal beam cp is sent from the other side instead 
of aB, the heating, and therefore the gas action, is the 
same, while the tangential push is in the opposite direc- 
tion, and the deflection now is always less in the direction 
of the arrow than it was before, and the difference gives 
twice the effect due to the tangential push of either. 
Another experiment, rather different in kind, even more 
clearly shows that light carries a stream of momentum. 
Fic. 5. 
Two glass prisms Bp (Fig. 5) were fixed at the end of a 
torsion arm and suspended by a fibre from c. A beam of 
light AB was directed horizontally so as to pass through 
the two prisms and emerge parallel to its original direction 
along pe. Always the torsion arm turned as indicated by 
the arrow, just as a pipe would tend to turn if it were 
bent as the beam of light is bent and carried a stream of 
water—a stream of forward momentum. 
I will not now dwell on the interesting modification of 
the third law of motion which we must make to reconcile 
with it these experiments on light. It is enough to say that 
Wwe must admit the luminiferous medium into momentum 
transactions just as long ago we admitted it into trans- 
actions with energy. 
NO. 1934, VOL. 75 | 
Let us now see how this way of regarding a beam of 
light leads us to expect a modification of the pressure 
when the receiving or the emitting surface is moving. 
First, let us suppose that the receiving surface is moving 
Let a (Fig. 6a) be the source. 
towards the source. 
@) 
Let 
B ¢ 
Fic. 6. 
B be the receiving surface, moving towards a with 
velocity u. If B were at rest at cC it would receive in one 
second the radiation and the momentum in length cb=v, 
the velocity of light. But when a given wave starts from 
D, let the surface start from B, and let them meet at the 
end of a second. Then B has evidently absorbed the 
momentum in length sp=u+wu, and it has received more 
than it would have done if at rest in the ratio u+w: uv. 
The pressure, therefore, is increased, and by the fraction 
u/u. It is easy to see from Fig. 6b that if B is moving 
away from the source it receives less momentum, has less 
pressure than if it were at rest, and the decrease is again 
by the fraction u/u. We may call this the ‘‘ Doppler 
reception effect,’’ ‘‘ Doppler ’’ since he was the first to 
point out the effect of motion on radiation. 
If the source is moving there is a nearly equal effect 
upon it. The pressure is increased if it advances and is 
decreased if it retreats, but the effect arises in a different 
way. It is now due to alteration of wave-length. The 
source crowds up and shortens the waves it sends forward, 
putting into them more energy and more momentum, and 
so suffering an increase in pressure, while it draws away 
from and lengthens the waves it sends backward, putting 
into them less energy and momentum, and so suffering a 
decrease in pressure. The alteration of pitch produced in 
sound by motion of the source is familiar to all. 
We can easily deduce the alteration in pressure if we 
make the reasonable assumption that the amplitude, the 
height or depth of the waves sent out from the source, 
depends on its temperature alone, and not on its motion. 
A Cc B 
Fic. 7. 
Let us imagine, by way of illustration, that the source 
moves with half the velocity of light, so that a wave 
which would be acs (Fig. 7) is packed into half the, space 
a’c’n’. With waves of the same height, the energy in a 
