DOC.
382
SEPTEMBER
1917 375
where
21
behaves
like
a
vector under Lorentz
transformations
and div
means
the
four-dimensional
divergence,
in
the
sense
of the
special theory
of
relativity.
In
(1),
(2)
(3)
(4)
The
minus
sign on
the
left in
(1)
and
(3)
seems
to
me
necessary
so
that the
T44
calculated
from
your equations
gives
the
energy
density
with
the
correct
sign.
(911
should have
a
sign allowing
the
variation
principle
to be
written
as 6
f
(0*
+
9Jl)dT.)[7]
It
appears
to
me
that
Droste also has
the
same
sign
error
in
equation
(17)
of his
dissertation,[8]
and
I
am
going
to
investigate
the
matter
properly,
together
with him.
But
because it
is
entirely possible
that
I
am
completely
wrong,
I
don’t
want to
go
into
further
detail
about the
sign question yet.
The
matter does
not
affect
the
results
derived in
the
following.
From
the
expression
in
(3)
for
-0*
and from
equations
(29)
and
(32)
in
your
Annalen
paper,[9]
it
is
evident
that
0* consists
of
nothing
but
terms
of
the
form
F(gab)-gcedgfh9.
Hence
one
has
X-7777 =
20
dg&
p,f
(5)
and the last
[of
your]
expressions
(20)
in the Berl.
Ber.[10] yields
for
the
diagonal
sum
of
the
small
t’s
£:
=
«*
(6)
Because
Ei
=
e,
(1)
produces
ECS
+
O
=
diva.
(7)