AMPÈRE'S
MOLECULAR CURRENTS
149
tive involvement
in
the
problem
was a
short
paper,
submitted
in
February 1916,
in
which
he
described
a new
and
simple experiment
to
show the existence of
Ampère's
currents.
It
appears
he
successfully
conducted the
experiment
himself.[18]
In
the
meantime,
De
Haas continued
to
refine the
experimental setup.
In
Septem-
ber
1915
he submitted
a
paper
in which
a
modified
experiment
was
proposed
that al-
lowed
better control
of the
experimental
sources
of
error.
One of its features
was a
new
method
to
eliminate disturbances due
to
the
magnetic
field of the earth.
Al-
though
he
did
not set out to
make
any
precision
experiments,
he
did achieve satisfac-
tory qualitative agreement
between
experiment
and
theory. Together
with his
wife,
De
Haas also
published a
paper containing
a new
theoretical discussion of
Maxwell's
original
experiment.[19]
Their conclusion
was
that the
effect would
be
hardly percep-
tible.
IV
Einstein and
De
Haas
were
not
the
only
ones
who
performed
experiments
on
Ampère's
currents.
Beginning
in
1909
the
American
physicist
Samuel Barnett had
tried
to
measure
the
magnetization
of
an
iron bar that is
suddenly
set
into rotational
motion. His
most recent results,
published
in
1915, corresponded
to
a
gyromagnetic
factor
g
twice
as
large
as
the
one
found
by
Einstein and
De Haas.[20] In
the fall of
1915
Einstein and De Haas
published
a
short
note (Einstein
and
De
Haas 1915d
[Doc. 23])
acknowledging
Barnett's
work,
without
commenting
on
the
discrepancy
between their
respective experimental
results.
They only
concluded that the
two
ex-
periments complemented
each other "in
a
most
satisfactory
manner."[21]
As
time
passed,
however,
evidence accumulated from other Einstein-De Haas-like
experiments
that
g
had in fact
a
value of
approximately
two
rather than
one.[22]
Ein-
stein took the results
seriously
and
urged
new
experiments
to
decide the
matter.[23]
At the third
Solvay Congress
in 1921 De
Haas reviewed the
experimental
results. His
own
most
recent
work
pointed to
a
value of
g
still close
to
one,
and he
was
reluctant
to accept
other
values.[24]
Eventually,
however,
evidence became
overwhelming
in
favor
of
a
value of
two,
and the
conclusion
was
inevitably
drawn that the
experiment
of Einstein and
De
Haas had been
flawed.[25]
[18]See
Einstein
to
Arnold
Sommerfeld,
8 February 1916.
[19]De
Haas and
De Haas
1915.
[20]See
Barnett
1915b,
1915d.
[21]"in
erfreulicher Weise." Einstein and
de
Haas 1915d.
[22]See
Stewart
1918, Beck 1919,
Arvidsson
1920.
[23]See, e.g.,
his
comments in
Einstein
to
Paul
Ehrenfest, 22
March
1919.
[24]See
De Haas
1923.
[25]See
Galison
1987, pp.
56-65,
for
a
discussion of the later
history
of determinations
of
the
gyromagnetic
factor.