420
DOC.
26
THE PROBLEM OF SPECIFIC HEATS
entropy
does
not
depend (the
mode of thermal interaction between K and the
environment).
Further,
one
could
surmise
that the heat absorbed
by
K
when
the latter
is
irradiated
is
not
exactly
equal
to
the radiation incident
on
K,
so
that the fluctuations of the heat
taken
up
by
K
are
not
equal
to
the fluctuations of
the
radiations
in
the
given wavelength
range
that
are
incident
on
the
surface
f.
Such
a
conception
does
not
necessarily
amount
to
an
actual
violation of the
energy law,
because
it
is.
possible
to
assume
that the
hypothesized
difference between the two
quantities
of
energy
is
going
to
accumulate.
Of
course,
one
then
faces
the
task of
picturing
the
mechanism
of
such
an accumulation, just
as, analogously,
one
is
faced with
the task of
picturing
the immense disorderliness
in
the
spatial
distribution of the radiation
energy.
If
we
reject
this
accumulation
hypothesis as
well,
then
we
must resolve to
abandon the
energy
law in its
present form,
and conceive
it
as
a
law
that
can
claim statistical
validity only,
in
analogy
with
the
conclusions from
the
second law
of
thermodynamics.18
Who
would have
the
audacity
to
give
a categorical
answer
to
these
questions?
I
only
intended
to show
here
how
fundamental
and
deep–
rooted the
difficulties
are
in
which
the
radiation
formula enmeshes
us even
if
we
view
it
as a purely
empirical given.
§
3.
The
Quantum
Hypothesis
and
the General Character
of
the
Related
Experiments
The
positive
results
produced
by
the
investigations
described in the last
section
can
be
summarized
as
follows:
When
a
body
absorbs
or
emits
thermal
energy
by
a
quasi–
periodical
mechanism,
the
statistical
properties
of the mechanism
are
such
as
they
would
be
if the
energy
were
propagated
in whole
quanta
of
the
magnitude
hv.
Though
we
have
little
insight
into
the
details
of the mechanism
by
which nature
produces
this
property
of
these
processes,
we
must
expect
all
the
same
that the
disappearance
of
such
an
energy
of
a
periodic
character
is accompanied
by
the
generation
of
packets
of
energy
in
the form
of discrete
quanta
of
magnitude hv,
and
second,
that
energy
in
discrete
quanta
of
magnitude
hv must
be
available,
so
that
energy
of
a
periodic
character in the
frequency
region
v may
be
produced.
In
particular,
if
a
radiation
in
the
frequency range
Av
is
capable
of
producing a
certain
type
of
effect,
e.g.,
a
certain
photochemical
reaction,
at
18
In addition
to
what
has
been
said in
the
text,
let
me
point
out
that the formula for the
energy
fluctuations
e2
can
also
be
applied
to
a
radiation-filled
space
that
is
bounded
by light-scattering,
nonabsorbing
walls and
that
can exchange
radiation of the
frequency range
dv
with
some body.
Of
course,
one
would
again
arrive
at
a
fluctuation formula of
similar
construction.
In this
case
the
accumulation
hypothesis
is inconceivable,
so
that
only
the
choice between
the
hv-structure and
the
abandoning
of the
strict
validity
of the
energy principle seems
to remain.