wire round a magnetic needle he bends its upper portion to the west:
the north pole of the needle immediately swerves to the east: he
bends his loop to the east, and the north pole moves to the west.
Suspending a common bar magnet in a vertical position, he causes it
to spin round its own axis. Its pole being connected with one end
of a galvanometer wire, and its equator with the other end,
electricity rushes round the galvanometer from the rotating magnet.
He remarks upon the 'singular independence' of the magnetism and the
body of the magnet which carries it. The steel behaves as if it
were isolated from its own magnetism.

And then his thoughts suddenly widen, and he asks himself whether
the rotating earth does not generate induced currents as it turns
round its axis from west to east. In his experiment with the
twirling magnet the galvanometer wire remained at rest; one portion
of the circuit was in motion relatively to another portion. But in
the case of the twirling planet the galvanometer wire would
necessarily be carried along with the earth; there would be no
relative motion. What must be the consequence? Take the case of a
telegraph wire with its two terminal plates dipped into the earth,
and suppose the wire to lie in the magnetic meridian. The ground
underneath the wire is influenced like the wire itself by the
earth's rotation; if a current from south to north be generated in
the wire, a similar current from south to north would be generated
in the earth under the wire; these currents would run against the
same terminal plate, and thus neutralise each other.

This inference appears inevitable, but his profound vision perceived
its possible invalidity. He saw that it was at least possible that
the difference of conducting power between the earth and the wire