398
DOC.
25
SOLVAY
DISCUSSION REMARKS
be
sure)
that
a
transition from
A
to B in
a way
that
is
purely
static from
the
viewpoint
of
molecular
mechanics
is
always possible
in
principle.
In his
lecture,
Nernst
only briefly
mentioned
the
problem
of
heat conduction
(on p. 231).
See Einstein
1911g
(Doc. 21),
§4,
and Einstein 1914
(Doc. 26), p.
341,
for
a
more
detailed
account
of Einstein's
contemporary
thoughts
on
this
problem.
Nos.
[191]-192 (Nernst
et al.
1914,
p.
244;
Nernst
et al.
1912, p. 303)
VIII. SOMMERFELD
14)
The considerable thermal
conductivity
of insulators
can
neither be
explained by
the conventional
mechanical
theory
nor
with
the
auxiliary
concept
of
energy quanta.
According
to
both
conceptions,
during
the
time
of half
an oscillation,
the
oscillation
energy
bound
to
an
atom
should
not
propagate
farther than
up
to
the
immediately
neighboring atom,
and
consecutive
energy
transfers of
this kind should
be
conceived
as
mutually
independent
processes.
But
on
the
basis
of these
assumptions one
arrives at
values
for thermal
conductivity
that
are
far
too small.
It
appears,
accordingly,
that
at low
temperatures
thermal motion does
not
possess
the character of
complete
disorder.
In
the
introductory
section
of
his
lecture,
Sommerfeld introduced his version of the
quantum
hypothesis,
which
he considered to be
compatible
with classical
electrodynamics (see Sommerfeld
1914, p.
294)
in the form
of
the
principle
that
in
"every purely
molecular
process," p. 254,
the
quantity
of
action
h
2ic
is
exchanged
where
t
is
the
duration of the
process,
H the
Lagrangian,
and
h
Planck's
constant
(for a
historical
study
of
Sommerfeld's
work, see, e.g.,
Hermann
1971, pp. 103-123).
In
the
context
of
his
talk,
Sommerfeld
restricted
the
notion
of
a purely
molecular
process
to
the
interaction between
an
electron and
an
atom
(Sommerfeld 1914, p.254),
but
demonstrated
the
relativistic invariance
of
the action
integral
for
the
case
of
a
single mass point,
which is taken up
by
Einstein in
his
comment.
The
function
L
-
U mentioned
by
Einstein
is
the
Lagrangian
written in
terms
of the
kinetic
energy
L
and
potential energy
U.
No.
197
(Sommerfeld
et al.
1914, p. 301; Sommerfeld
et al. 1912,
p. 373)
15)
It
seems
to
me
that Sommerfeld's
interpretation
of the
physical
meaning
of
Planck's
constant
h
involves
a difficulty
in
that the function L
-
U
for
a freely
movable
particle
can
hardly
be set
equal
to
zero,
so
that the
existence
of
a freely
movable
mass point decomposes, so
to
speak,
into
quanta
of
action,
which
it
does in
a manner
that
depends
on
the
state
(of
velocity)
of the coordinate
system.
The
following
comments
by
Einstein refer
to
Sommerfeld's
analysis
of
X
rays generated by
the
impact
of
electrons
on an
obstacle. In
a
letter
to Besso,
Einstein had earlier stressed his view
that
Sommerfeld had
postulated
his
hypothesis on
the
role of collision
times in this
process
without
any
theory
(see
Einstein
to