122
DOC. 26
THEORY OF TETRODE AND SACKUR
[p. 2]

E
s
=
(2a)
E
denotes here the
mean
energy
of
the
system
under consideration. We make this
formula the basis
of
all
our subsequent
considerations. We
prefer
it
over
formula
(2)
becausefrom
a
thermodynamic point
of
viewentropy
is,
in contrast
to
the free
energy, a completely
determined function
of
a
("macro") state,
up
to
an
additive
constant.
The
precise meaning
of
equation (2a)
is
as
follows. The basic
assumption
(later
to be
modified)
is that all molecular
processes (including
chemical
ones) can
be
viewed
as
movements of the smallest
particles governed by
laws
of
classical
mechanics.
Macroscopically,
i.e.,
phenomenologically
in
terms
of
the usual
thermodynamic parameters
(volume,
pressure,
etc.),
the
state
of
a system
is
determined when
we
know:
1)
the
temperature
0,
2)
the
energy
E
as a
function
of
the molecular variables
q1...pn.
Every
not
purely
thermal
change
of state in the
system (e.g., a change
in
volume)
is
accounted for
by a change
of the function
E.
In order to
express this, one
views
E
as
not
only dependent upon
q1...qn
but also
dependent upon
certain
parameters
ax;
and
every
not
purely
thermal
change
of
state in the
system corresponds
to
a
change
in the
values of the
parameters
ax.
If
S1
and
S2
denote the
entropy
values
of
two states of
the
physical system
considered,
thenas
statistical mechanics
tells usthe
transition
of the
system
from the first into the second state involves
a change
in
entropy,
S2

S1,
which is
equal
to
the
change on
the
righthand
side
of
(2a).
From what has been
said,
one sees
that
equation (2a)
remains valid if
an
additive
constant
is included
on
the
righthand
side, or more
precisely, a quantity
that
depends
neither
upon
0
nor on
the other
thermodynamic parameters
of
state
ax.
We shall
use
this feature
in the
following.
Equation
(2a) produces
the
entropy
difference of
two
states when the
energy
function,
in the
sense
of molecular
theory,
is known. It therefore
answers
also in
a
very
distinct
manner
the
question
of
the constant of
integration
of
entropy[5]
whose
[p.
3] meaning
has
been
especially clearly
stressed
by
Nernst and answered
by
him
through
the known Nernst theorem
(in
a
distinct
manner).
Equation
(2a)
has the
advantage over
Nernst's theorem
in
the
regard
that its
statement is not related to the
zero point
of absolute
temperature,
and
therefore
it
can
be tested without
extrapolating experimental
results down to absolute
zero.
But
on
the other
hand,
there
seems
to be,
on
first
sight,
a
grave
and almost
devastating
drawback in the
use
of
this
equation:
a) Equation
(2a)
presupposes
that
we
have
a complete system
of
molecular
mechanicswhich is not the
case.
Meanwhile,
the
equation can
have
only
true