DOC.
1
MANUSCRIPT
ON
SPECIAL RELATIVITY 27
If,
in
conjunction
with such
reflections,
one
considers in how
simple
a
fashion
the
theory
of
relativity
interprets Fizeau's
result
(
),
one can
hardly
avoid
the
impression that,
given
the
present
state
of
physical
experience,
this
theory is
to
be
viewed
as
the
most
natural
one.[49]
In
case I,[50]
the
velocity
of
light
would
depend
on
the
state
of motion
of
the
light
[p. 20a]
source.[51]
Light rays
of
most
diverse
propagation
velocities could be
present
at
one
and the
same
location
at
the
same
time. The
physical properties
of the
light
would
not
be
determined, then,
by
the
frequency
alone,
since the
wavelength
and the
frequency
would be
independent
of each
other;
the
light originating
from
stars
that
are moving
relative
to
us
would have
to
be
physically distinguishable
from the
light
originating
from
light
sources
at rest.
Experience
has
yielded nothing
of the
sort.
The
most
convincing argument against
this
point
of
view,
which had been advocated
by
Ritz,
has been
put
forward
by
the Dutch
astronomer Pexider.[52]
The individual
stars
of stellar
systems (double stars)
must
have
sent
us
light
of different velocities
during
different
epochs
of their orbits.
Thus,
the time of
propagation
of the
light
from double
stars to
us
would be
different for different
epochs.
The
temporal sequence
of the
epochs
as
traced
by
us
with the
help
of
the
Doppler principle
would be different from
that in
reality;
a
simple
calculation shows
that, indeed,
if the
underlying hypothesis
were
borne
out
by
the
facts,
the indicated
influence would
have
to
be
so
considerable
that it would have been
absolutely impossible
for the
astronomers to
miss
it.
The
untenability
of this
conception
can
surely
be viewed
as
being definitively proved.
In
case
II,[53]
the motion
of
the
surrounding
matter
would have
an
influence
on
the
propagation velocity
of
light
in
stationary
media. The influence of the motion
of
the
matter traversed
by light
on
the result of Fizeau's
experiment
would be
attributable
to
the fact that the
Earth
is
moving
relative
to
the medium traversed
by
the
light.
Then
one
would have
to
assume
that Fizeau's
experiment
would
turn out
differently
if
it
were
conducted
on a
sufficiently
small celestial
body.
Even
though
this
conception cannot
be viewed
as
definitively disproved,
still its
extraordinary
complexity already
makes
its correctness
improbable.
It is
hardly plausible
that the
velocity
of
light
in
a
medium would
depend
on
the
state
of motion of the
surrounding
bodies,
but would do
so
to
a
noticeable
degree only
if these bodies
possessed
a
very
large mass
relative
to
the
body
traversed
by
the
light.
Also,
it has
not
proved possible
thus far
to
base
a
serviceable
theory
on
this
assumption.
On the other
hand,
since the Maxwell-Lorentz
theory requires
the
principle
of the
constancy
of
the
(vacuum)
velocity
of
light,
there
is
every
reason
to
stick with
it.
It
turns out
that the result of the Fizeau
experiment can
be derived
quantitatively
from
(a)
and
(b)
without
introducing
the
more
special assumptions
of
the
Maxwell-Lorentz
theory.
Laue
was
the first
to
draw attention
to
this
important circumstance,
to
which
we
will
return
in
§11.[54]