198 DOC.
198
MARCH
1916
Counter-example (ideal case)
gravity
Here
the
mixed
crystal
certainly
exists: Its
entropy
does not
disappear.[3]
Now
you
will
say
that this
case
does not exist in
reality,
since
the
potential
energies
of
conceivable
cases
never are
exactly
identical.
Therefore,
the
following
case as a
true
analogy
to
the
real
one:
At
high
temperatures
the distribution
is
practically
uniform
despite
gravity.
We
cool it down
now
to
absolute
zero.
As
a
result two
cases are
possible,
namely,
gravity
1)
Transition
from
one
chamber to
another
is
so rare
that the
case
where transi-
tion from
one
chamber
to
another
during cooling
can
take
place
occurs
very
rarely.
2)
Transition
from
one
chamber
to
another
is
so
frequent
that
it
occurs
very
frequently
during
the
cooling period.
In
case
(1)
the
mixed
crystal
still exists
at
T
=
0
and does
not
satisfy
Nernst’s
theorem.
In
case
(2)
the
particle is
certainly
at
the bottom
in
the
first chamber
at
T
=
0.
The
mixed
crystal as
such
does not
exist
at
T
=
0. Nernst’s
theorem
applies.
In
any
case,
it
cannot
be asserted
that
Nernst’s theorem
applies
to
a
mixed
crystal
at
T
=
0.
For in
your opinion
it
simply
does not exist at
T
=
0. According
to
your view,
case
(1)
is
not
present
in
nature.
But
I
hold
another
view.
As
I
do not
see
why
mixed
crystals
should not be-
have
like
chemical
mixtures,
about
which
thermodynamically
realizable
cooling
is
known to be
possible
without the
creation
of
a
chemical
or
statistical
equilibrium
affecting
the
structure.
Do
you
not
agree
with this?
With best
regards, yours,
Einstein.