DOCS.
59,
60
MARCH
1915
71
could be
taken
instead;
because
if
the
mass
is
critical
at
the
developmental
stage,
then
for
binary
stars
the
mass
of
the
components
is
much
more
critical
than the
overall
mass.
5)
You should
provide
me
with
a
bit
more
precise
information
on
how
the
radii
of
the
eclipses
are
estimated.[6]
6)
Is
the
absence
of
the
effect
really
proven
for
types
that
yield
smaller
masses
using
the
binary-star
method?
Why
do
you pass
over
this
important
question
so
quickly?[7]
7)
Did
you keep yourself open
to
the
hypothesis
that
for
stars
with
only one
visible
component,
both
components
have
masses
in
the
same
order of
magnitude?
It
is
a
shame that
your descriptions
are
not
detailed
enough
to
be
able to
es-
timate the
uncertainty
attached
to
your
estimates.
Thus
a
nonspecialist
cannot
get
a
notion
of
the
reliability of
your
calculations. A
much
more
in-depth
pre-
sentation
would
be
desirable.
The worst in
this
regard
is
the
specification
of
the
mean
densities.
In
the
formula in
the fourth
column
of
the
proofs, “V0.1”
is inadvertently
indicated[8]
instead
of
“3V0.1.”
With best
regards, yours,
Einstein.
60. To
Tullio Levi-Civita
[Berlin,]
5
March
1915
Highly
esteemed
Colleague,[1]
By examining my paper
so
carefully, you
are
doing
me a
great favor.[2]
You
can
imagine
how
rarely
someone
delves
independently
and
critically
into this
subject.
I also cannot
help admiring
the
uncommon sureness
with which
you
make
use
of
a
language
that
is
foreign
to
you.
When
I
saw
that
you
are directing
your
attack
against
the
theory’s
most
important
proof,
which
I
had
won
by
the
sweat of
my
brow,
I
was
not
a
little
alarmed, especially
since
I
know
that
you
have
a
much better command
of
these
mathematical
matters
than
I. Nevertheless,
upon thorough
consideration
I
do
believe I
can
uphold my
proof.[3]
I
start
with the
second
part
of
your
letter
in which
you
intend
to
show with
an
example
that the
result
of
§14,
that
Euv/-g
is
a
tensor,
is not correct.
For
this