DOC.
42
233
However,
one
has
to
take into consideration
that
the
frequency
v0
(for
the
co-moving
observer)
is
unknown,
so
that the
above
relation is
not
acces-
sible
to
direct
experimental
verification.
But,
it
may
be assumed
that
v0
is also
equal
to
the
frequency
emitted
or
absorbed
by
the
same
ion while
at
rest,
and
this for the
following
reason.
From
the fact that
one
and
the
same
line
spectrum
is
formed under
very
different conditions,
we
conclude that the
frequency
v0
does not
depend
on
interactions
between
moving
ions
and
the
stationary
gas,
but is
a
characteristic of the ion
only;
from
this
one
directly concludes
with
the
help of
the
principle
of
relativity
that
v0 must
equal
the
frequency
of radiation emitted
or
absorbed
by an
ion
at rest.
The equation
v
-
1
v
2
V
thus
gives
directly
the
second
order effect
sought.
The
numerical values
presented
by
Mr.
Stark for the effect
are more
than
ten
times
larger
than those resulting
from
the formula
presented.
It
seems
likely
to
me
that
reliable results with
regard to
this
problem
can
be
expected
only
after
it
has
been
possible
to
obtain
(nonluminous?)
canal
rays
in
a
completely
gas-free
space.
[4]
[5]
Bern,
March
1907.
(Received
on
17 March
1907)
Doc. 42
CORRECTION
TO
MY
PAPER: "PLANCK'S
THEORY OF RADIATION,
ETC."
by
A.
Einstein
[Annalen
der
Physik
22
(1907):
180-190]
In the above-cited
paper,
which
was
published
in this year's
January
issue,
I wrote:
"According
to
Drude's
investigations, these
proper
fre-
quencies
are
to be
attributed
to
the
ponderable
atoms
(atom
ions)
themselves.
The
most
obvious conclusion
seems
therefore
to be to
consider