DOC. 38
THEORY
OF SPECIFIC HEAT 391
heats decrease with
temperature,
but that
they
all
approach zero
at
small absolute
temperatures
(small
T/ßv),
as
Einstein's
formula
implies.
[28]
This
assertion is
generally
confirmed
by
the tables in
Landolt and
Börnstein
1905,
pp.
385-406.
[29] Nernst
1911c,
pp.
308-309, indicates that
discrepancies
do
occur
between
Einstein's
pre-
dictions and
experimental
data,
especially
at
lower absolute
temperatures (small T/ßv)
where
Einstein's
predicted
curve
is too low. For Ein-
stein's
response,
see
Einstein
1911f.
[30]
The work in
question may
be Roscoe
et
al.
1898,
a copy
of
which
is
in
Einstein's
personal
library.
[31]
The data for Ybeob
in
the
table
are
taken
from
Landolt and
Börnstein
1905,
p.
620.
[32]
The data in the table below
are
derived
from data in
Landolt and
Börnstein
1905, p.
384,
which in
turn
are
derived from
Weber,
H.
F. 1875. In the table below the
temperatures
given
in
Landolt and
Börnstein
1905, p. 384,
have
been converted
from Celsius to absolute
and the
specific
heats
multiplied by 12,
the
atomic
weight
of
carbon,
to obtain the atomic
specific
heat,
c.
[33]
In the
third column
of
the
table, 9,6638
should be 0,6638.
[34]
This
prediction
is
modified in Einstein
1907d
(Doc. 42).