DOC. 182
OCTOBER
1909 137
Now I would like to tell
you a
few
things
about
your
experimental arrangement,
as
used
by
Mr.
Müller.[6]
What the
theory
yields
for
the
discharge
tube
is,
as we
know,
the
following.
If the
quantity
of
electricity
J
passes
through
the
tube
during
time
r,
then J
is
subjected
to
fluctuations. If
we
take the
deviation
of J
from
its
mean
value
J0
to be
j,
then the
theory
yields
j2.
However,
rather than
measuring
j
directly,
one
measures
the
fluctuations
tt
of the
potential,
using
the
following
experimental
setup[7]
Ate
ground
Brons.
resist.
w
capacity
of
lead + electrometer
+
Bronson
electr.
=
c
What
is
the
relationship
between the
mean
square
fluctuation of
the
potential (tt2)
and
j2? I
found the formula
72
_
2c
~2
J

ttt
w
Of
course,
j2
is
also
proportional to
the
time
r.
To
be
sure,
in this
case
tt
is not
the
potential
fluctuation
at
the
inversion
points
but
at arbitrarily
(randomly)
chosen
times.
But this
should
not matter too much.
However,
one
does
not
measure
tt
directly
either; instead,
one
measures
the
slow
deflection of the needle of electrometer.
This does not
cause a
noticeable
error
if
the
transient
time
of
the
electrometer
is
small
compared
with
c .w.
If
this
is
not
the
case,
the
observations
can
nevertheless
be
evaluated
absolutely
if the constants
of the
electrometer
have
been determined.
But it
certainly
would be
preferable to
satisfy
the
above
condition.
Finally,
I
would also like
to
ask
that
you
not
address
me so
formally
but that
you
treat
me as a
comrade instead.
With
best
regards,
your
A.
Einstein
I
haven't
written
for such
a
long
time because the
lecturing,[8]
to
which
I
am
not
accustomed,
has
totally
worn me
out.
My
address
is
Moussonstr.
12.
The
c.w
in Mr.
Müller's
setup is
small
compared
with
the transient time of
his
electrometer. The relation
given
in the
letter
is
therefore
of
no
value for his
work.[9]