ad established the
magnetism of flame, he repeated the experiments and augmented the
results. He passed from flames to gases, examining and revealing their
magnetic and diamagnetic powers; and then he suddenly rose from his
bubbles of oxygen and nitrogen to the atmospheric envelope of the
earth itself, and its relations to the great question of terrestrial
magnetism. The rapidity with which these ever-augmenting thoughts
assumed the form of experiments is unparalleled. His power in this
respect is often best illustrated by his minor investigations, and,
perhaps, by none more strikingly than by his paper 'On the Diamagnetic
Condition of Flame and Gases,' published as a letter to Mr. Richard
Taylor, in the 'Philosophical Magazine' for December, 1847. After
verifying, varying, and expanding the results of Bancalari, he submitted
to examination heated air-currents, produced by platinum spirals placed
in the magnetic field, and raised to incandescence by electricity. He
then examined the magnetic deportment of gases generally. Almost all
of these gases are invisible; but he must, nevertheless, track them in
their unseen courses. He could not effect this by mingling smoke with
his gases, for the action of his magnet upon the smoke would have
troubled his conclusions. He, therefore, 'caught' his gases in tubes,
carried them out of the magnetic field, and made them reveal themselves
at a distance from the magnet.

Immersing one gas in another, he determined their differential action;
results of the utmost beauty being thus arrived at. Perhaps the
most important are those obtained with atmospheric air and its two
constituents. Oxygen, in various media, was strongly attracted by the
magnet; in coal-gas, for example, it was powerfully magnetic, whereas
nitrogen was diamagnetic. Some of the effects obtained with oxygen
in coal-gas were strikingly beautiful. When the fumes of chloride of
ammonium (a diamagnetic substance) were mingled with the oxygen, the
cloud of chloride behaved in a most singular manner,--'The attraction
of iron filings,' says Faraday, 'to a magnetic pole is not more striking
than the appearance presented by the oxygen under these circumstances.'

On observing this deportment the question immediately occurs to
him,--Can we not separate the oxygen of the atmosphere from its nitrogen
by magnetic analysis? It is the perpetual occurrence of such questions
that marks the great experimenter. The attempt to analyze atm