SEPTEMBER II, 1913] 
on a measured interval fixed on a rigid block of 
matter is independent of the aspect of that block with 
respect to any motion of the earth through space. 
A definite and most interesting result : but it may be, 
and often is, interpreted loosely and too widely. — 
It is interpreted too widely, as I think, when Prof. 
_.Einstein goes on to assume that no_ non-relative 
motion of matter can be ever observed even when 
light is brought into consideration. The relation of 
light to matter is very curious. The wave front of a 
progressive wave simulates many of the properties 
of matter. It has energy, it has momentum, it exerts 
force, it sustains reaction. It has been described as 
a portion of the mass of a radiating body—which gives 
it a curiously and unexpectedly corpuscular “ feel.” 
But it has a definite velocity. Its velocity in space 
relative to the zther is an absolute constant indepen- 
dent of the motion of the source. This would not be 
true for corpuscular light. 
Hence I hold that here is something with which 
our own motion may theoretically be compared; and 
I predict that our motion through the zther will some 
day be detected by help of this very fact—by com- 
paring our speed with that of light: though the old 
astronomical aberration, which seemed to make the 
comparison easy, failed to do so quite simply, because 
it is complicated by the necessity of observing the 
position of a distant source, in relation to which the 
earth is moving. If the source and observer are 
moving together there is no possibility of observing 
aberration. Nevertheless, I maintain that when 
matter is moving near a beam of light we may be able 
to detect the motion. For the velocity of light in 
space is no function of the velocity of the source, nor 
of matter near it; it is quite unaffected by source or 
receiver. Once launched it travels in its own way. 
If we are travelling to meet it, it will be arriving at 
us more quickly; if we travel away from it, it will 
reach us with some lag. And observation of the 
acceleration or retardation is made by aid of Jupiter’s 
satellites. We have there the dial of a clock, to or 
from which we advance or recede periodically. It 
gains while we approach it, it loses while we recede 
from it, it keeps right time when we are stationary 
or only moving across the line of sight. 
But then, of course, it does not matter whether 
Jupiter is standing still and we are moving, or vice 
versa: it is a case of relative motion of matter again. 
So it is if we observe a Doppler effect from the right- 
and left-hand limbs of the rotating sun. True, and 
if we are to permit no relative motion of matter we 
must use a terrestrial source, clamped to the 
earth as our receiver is. And now we shall observe 
nothing. 
But not because there is nothing to observe. Lag 
must really occur if we are running away from the 
light, even though the source is running after us at 
the same pace; unless we make the assumption—true 
only for corpuscular light—that the velocity of light 
is not an absolute thing, but is dependent on the speed 
of the source. With corpuscular light there is nothing 
to observe; with wave light there is something, but 
we cannot observe it. 
But if the whole solar system is moving through 
the zther I see no reason why the relative zther drift 
should not be observed by a differential residual effect 
in connection with Jupiter’s satellites or the right and 
left limbs of the sun. The effect must be too small 
_to observe without extreme precision, but theoretically 
it ought to be there. Inasmuch, however, as relative 
' motion of matter with respect to the observer is in- 
volved in these effects, it may be held that the detection 
of a uniform drift of the solar system in this way is 
_ not contrary to the principle of relativity. It is con- 
trary to some statements of that principle; and the 
NO. 2289, VOL. 92| 
NATURE 4! 
cogency of those statements breaks down, I think, 
whenever they include the velocity of light; because 
there we really have something absolute (in the only 
sense in which the term can have a physical mean- 
ing) with which we can compare our own motions, 
when we have learnt how. 
But in ordinary astronomical translation—trans- 
lation as of the earth in its  orbit—all our 
instruments, all our standards, the whole contents of 
our laboratory, are moving at the same rate in the 
same direction; under those conditions we cannot 
expect to observe anything. Clerk Maxwell went so 
far as to say that if every particle of matter simul- 
taneously received a graduated blow so as to produce 
a given constant acceleration all in the same direc- 
tion, we should be unaware of the fact. He did not 
then know all that we know about radiation. But 
apart from that, and limiting ourselves to compara- 
tively slow changes of velocity, our standards will 
inevitably share whatever change occurs. So far as 
observation goes, everything will be practically as if 
no change had occurred at all; though that may not 
be the truth. All that experiment establishes is that 
there have so far always been compensations; so that 
the attempt to observe motion through the ether is 
being given up as hopeless. 
Surely, however, the minute and curious compensa- 
tions cannot be accidental, they must be necessary? 
Yes, they are necessary; and I want to say why. 
Suppose the case were one of measuring thermal 
expansion; and suppose everything had the same 
temperature and the same expansibility ; our standards 
would contract or expand with everything else, and 
we could observe nothing; but expansion would occur 
nevertheless. That is obvious, but the following 
assertion is not so obvious. If everything in the 
Universe had the same temperature, no matter what 
that temperature was, nothing would be visible at all; 
the external world, so far as vision went, would not 
appear to exist. Visibility depends on radiation, on 
differential radiation. We must have differences to 
appeal to our senses, they are not constructed for 
uniformity. 
It is the extreme omnipresence and uniformity and 
universal agency of the zther of space that makes 
it so difficult to observe. To observe anything you 
must have differences. If all actions at a distance 
are conducted at the same rate through the ether, 
the travel of none of them can be observed. Find 
something not conveyed by the ether, and there is 
a chance. But then every physical action is trans- 
mitted by the wther, and in every case by means 
of its transverse or radiation-like activity. 
Except perhaps Gravitation. That may give us 4 
clue some day, but at present we have not been 
able to detect its speed of transmission at all. No 
plan has been devised for measuring it. Nothing 
short of the creation or destruction of matter seems 
likely to serve: creation or destruction of the gravita- 
tional unit, whether it be an atom or an electron, 
or whatever it is. Most likely the unit of weight 
is an electron, just as the unit of mass is. 
The so-called non-Newtonian Mechanics, with mass 
and shape a function of velocity, is an immediate 
consequence of the electrical theory of matter. The 
dependence of inertia and shape on speed is a genuine 
discovery, and, I believe, a physical fact. The Prin- 
ciple of Relativity would reduce it to a conventional 
fiction. It would seek to replace this real change in 
matter by imaginary changes in time. But surely 
we must admit that Space and Time are essentially 
unchangeable: they are not at the disposal even of 
mathematicians; though it is true that Pope Gregory, 
or a Daylight-saving Bill, can play with our units, 
can turn the 3rd of October in any one year into the 
