lines of gravitating, luminiferous, or electric forces.'

Pure space he assumes to be the true magnetic zero, but he pushes
his inquiries to ascertain whether among material substances there
may not be some which resemble space. If you follow his experiments,
you will soon emerge into the light of his results. A torsion-beam
was suspended by a skein of cocoon silk; at one end of the beam was
fixed a cross-piece 1 1/2 inch long. Tubes of exceedingly thin glass,
filled with various gases, and hermetically sealed, were suspended
in pairs from the two ends of the cross-piece. The position of the
rotating torsion-head was such that the two tubes were at opposite
sides of, and equidistant from, the magnetic axis, that is to say
from the line joining the two closely approximated polar points of
an electro-magnet. His object was to compare the magnetic action of
the gases in the two tubes. When one tube was filled with oxygen,
and the other with nitrogen, on the supervention of the magnetic
force, the oxygen was pulled towards the axis, the nitrogen being
pushed out. By turning the torsion-head they could be restored to
their primitive position of equidistance, where it is evident the
action of the glass envelopes was annulled. The amount of torsion
necessary to re-establish equidistance expressed the magnetic
difference of the substances compared.

And then he compared oxygen with oxygen at different pressures.
One of his tubes contained the gas at the pressure of 30 inches of
mercury, another at a pressure of 15 inches of mercury, a third at a
pressure of 10 inches, while a fourth was exhausted as far as a good
air-pump renders exhaustion possible. 'When the first of these was
compared with the other three, the effect was most striking.'
It was drawn towards the axis when the magnet was excited, the tube