THE FIELD EQUATIONS OF GRAVITATION 249effect. The orbital contributions of earlier days have turned out to be nearly
entirely quenched. The quantum mechanical theory of ferromagnetism, given by
Heisenberg in 1928 [Hlb] provided the basis for a refined treatment of the cor-
responding gyromagnetic effects [H2]. Experimentally, the g value for ferromag-
netic materials has been found to lie close to 2 (except for Fe 7 S 8 ) with deviations
<10 per cent [S2]. The first experimental indications for g ~ 2 were published
in 1915 by Barnett (then at Ohio State University). In his earlier-mentioned paper
on the Barnett effect [B2], he concluded that 'the magnitude ... is within the
experimental error equal to twice the ... value computed,' the latter value being
g = 1. However, further measurements done by him in 1917 gave g ~ 1, 'but
the experimental errors. .. are such that great importance cannot, in my opinion,
be attached to the discrepancies [with his earlier results]' [B4]. In the period
1918-20, three independent measurements of the EdH effect were reported. In
chronological order, these came from Princeton [S3], the ETH in Zurich [B5],
and Uppsala [Al]. The answers found were g « 1.96, 1.88, and 1.87, respec-
tively. From that time on, the 'gyromagnetic anomaly' (as it was often called) was
firmly established. Inevitably this led to fairly widespread speculations about
'planetary motions of [positively-charged] constituents of nuclei' [B6]. The first
one to suspect a connection between the anomalous Zeeman effect and this new
gyromagnetic anomaly was Alfred Lande [LI] in 1921, the same year Heisenberg
expressed the opinion in a letter to Pauli that g = 2 could occur only in ferro-
magnetic bodies [H3].
Since de Haas was from Leiden, where the spin was discovered, it was only
natural that I would ask Uhlenbeck whether the EdH effect had played any role
in the discovery of the electron spin by him and Goudsmit (knowing that the effect
is not mentioned in their paper). Uhlenbeck replied that he knew of the effect
because he was in Leiden but that this subject was not in the center of attention
at that time. 'Had Ehrenfest thought it pertinent, he would surely have mentioned
it to us.' Thus the EdH effect served to confirm rather than stimulate subsequent
theoretical developments.
As to Einstein, his interest in gyromagnetism continued after de Haas's depar-
ture. In 1916 he published another paper on the EdH effect. It contains the design
of a new experimental arrangement* for determining g [E37]. He also remained
interested in the activities at the Reichsanstalt. In 1916 he was appointed member
of its Kuratorium (board of governors) and played an active role in the planning
and design of its experimental projects [K6].
Let us now return to our main topic, Einstein's final formulation of his theory
of general relativity.
"The idea was to flip the remanent magnetization of a premagnetized iron cylinder. This method
has the advantage that the cylinder is exposed to a magnetic field for such a brief time (=slO~^3 s)
that irritating side effects are largely eliminated.