DOC. 282
SEPTEMBER
1911
203
that
the time of travel
from
one
plane to
the
other determines the
magnitude
of the
energy
quantum,
one
obtains
for
the electrons
moving
in the
direction of the electric
force,
where the thermal
and
electric
energies
add
up[4]
h=
RT
+
sXH;
T
and for
electrons
moving
in the
opposite
direction
h
=
RT

eXH.
T"
For
the
average
velocity
of
the
electrons
moving
in the
direction of
the
electric
force

1
(X
X
sX2H
v
=

2
r'
T"
V
'
T
/
h
Wait! Some evil
spirits are
at work
here:
if
one
tidies
up
the
equation,
X2
appears
in it
again,[5]
which
already
follows from the
dimensions,
after
all.
Does
my previously
used
hypothesis
nX
=
const.
then
still work
reasonably
well
here?
F2
Conductivity
=

X
(nX).
In accordance
with the
earlier
argument,
it does not
seem
too
improbable to
me
that
X
·
T
=
const.
for
pure metals.[6]
And what about the thermal
conductivity?
With
regard to it,
one
might
want to
substitute
h=
RT',
h^
=
RT";
the
flow
of
energy
in
time
1
t t
1
n
2
= 12nR2h{T'2

T"2)
=
12nR2h2TLT
=
12nR2h2TX()
is,
if
I
h
h
tt
2
\r_'2
r
/
calculated
and
reasoned
correctly,
even
without numerical
computation, dealing
a
death
R2
( A
T\
blow
to
the
whole
musing
=
nX

T
,
in accordance,
as
I
see
in
retrospect,
with
h
\
X
)
the
WiedemannFranz
law.[7]
I
don't
want to
do the
numerical
computation, not just
because
I
am
lazy,
but
also
because
the
whole
letter
would have
to
go
in
the
waste
basket
if
the
computation
were
to
go wrong.
LampaEhrenhaft.[8]
Dimensional
arguments:
natural
units
of
measure
of
two
kinds.
Beside
Ģ,
the
quantum
of
action.

I
must
now
send off
the
letter;
for that
reason
I will not
elaborate
on
these
questions
any
further. Your
colleague
Lampa
discusses
Ehrenhaft's
experiments
in
Das
Wissen
für Alle.[9]
What
do
you
think of
it?