228
DOCS.
231,
232
JULY
1916
with
the
complementary
eq.
kC(d)i
=
(d)
L
k
The
symbols
A,
B,
Tik
have
meanings analogous
to
those for
A,
B,
Tik,
which
appear
in
my
eqs.
(ch. VII);[13]
thus,
for
ex.,
Tik
=
Tik
+
Tuik,
where
Tik
relates to
the
electrical
f[ield]
and where
Tuik,
is
due
to
actual
matter.
The
10
eqs.
arising
from
(r) will
be able to read
(352,
353,
Ch.
VII):
(1+Sim)irnl
=
lm)k(g)¿Cgim
k
1

^
'
Okrn^ik
• • •
(a')
k
If
to
these
10
eqs.
(a')
the
compl[ementary] eq.
(d)
is
added,
the
11
eqs. are
obtained.
*(*)*££*“
(ik,
Im)
=
^2(gkm^ik
çydirn^kk
• • •
(a")
k
l
k
¿
kC(g)i
=

(d")
Z
k
But
from
the
10
eqs.
(a")
the
11th
(d") can
be
easily
deduced.
Thus,
it
is
demon
strated that
(r)
and
(d)
provide
the
physical meaning
of
your
ten
eqs.
The
result
of this
is
that
tAu
(341,
Ch.
VII) agrees
without
modification
with
this
case.
Please

gr[eatest] re[spect].
232. To
Théophile
de
Donder
[Berlin,]
8 July 1916
Esteemed
Colleague,
I
hasten
to
answer
your
questions.[1]
Beforehand, however,
a
general
remark.
It
is,
in
my
opinion,
of
disadvantage
not to
look
at
the
indices of
a
tensor,
whether
they
have
a
covariant
or
contravariant transformation
character;
you
obscure
the
general
idea
unnecessarily
through
this
shortcoming
in
the
designation.
When
you
therefore
write,
e.g.,
my
field
equations
with
vanishing energy
scalar for
matter,
you
must
not write
k\/ZIgJ2gkl(ik, lm)
=
Y,9kmTik
kl k