DOC.
480
MARCH
1918 495
to
be
taken
over
the
system
and
the
gravitational field
belonging
to
the
system.
This
state of affairs
can
be
expressed
in
the
following way.
As
far
as
its
gravitational
influence
at
a great
distance
is
concerned, any
(quasi-static) system can
be
replaced
by a
point
mass.
The
gravitational
mass
of
this
point
mass
is
given by
J(‘Xt
+
tt)dV,
i.e., by
the total
energy (more
precisely,
total
“rest
energy”)
of
the
system, exactly
as
the inertial
mass
of
the
system.
Formally,
the
possibility
of this
physically
important interpretation is based
on
the
fact
that
the
same
quantities,
Evo
+
tvo,
which enter in
the
conservation
law
(22),
can, by
virtue
of
the
field
equations
(23),
be written
as a
“divergence”
(i.e.,
in
the
form
Ed/dxp
()
)
of
certain
expressions
constructed
out
of
the
guv's
&
their
first
derivatives.
From
(22)
it
can
be concluded
that the
same
integral
j(E44
+
t44)dV
also deter-
mines the
system’s
inertial
mass.
Without
the introduction and
interpretation
of
t,
it
is
impossible
to
say
that
the
inertial
and
gravitational
mass
of
a
system
coincide.-
I
hope
that this
anything
but
complete explanation
will
allow
you
to
guess
what
I
mean.
In
the
first
place, though,
I
hope
that
you
abandon
your
view
that
I
had
regarded
an
identity,
that
is, an
equation
that
does not
impose any
conditions
on
the
quantities
occurring
in
it,
as
the
energy
law.
On
your
comparison
of
the
old
theory
with
the
new
one,
I comment
that
in
the
case
of
the
old
theory,
the
guv's
are
not
determined
adequately by
the
equations
Knv
=
0.
Since all
(up,
er)
=
0
must
be
satisfied,
of
course.[4] A
substitution
by
Kuv
=
0
could
only
take
place
if
specific
boundary
conditions
were
imposed.
Thus this
problem
also leads to
the
fundamentally important boundary
conditions which
previously
had not been
given
sufficient notice.
In
great
respect, yours truly,
A. Einstein.
Previous Page Next Page