ospheric
air by magnetic force proved a failure, like the previous attempt to
influence crystallization by the magnet. The enormous comparative power
of the force of crystallization I have already assigned as a reason for
the incompetence of the magnet to determine molecular arrangement; in
the present instance the magnetic analysis is opposed by the force of
diffusion, which is also very strong comparatively. The same remark
applies to, and is illustrated by, another experiment subsequently
executed by Faraday. Water is diamagnetic, sulphate of iron is strongly
magnetic. He enclosed 'a dilute solution of sulphate of iron in a tube,
and placed the lower end of the tube between the poles of a powerful
horseshoe magnet for days together,' but he could produce 'no
concentration of the solution in the part near the magnet.' Here also
the diffusibility of the salt was too powerful for the force brought
against it.

The experiment last referred to is recorded in a paper presented to
the Royal Society on the 2nd August, 1850, in which he pursues the
investigation of the magnetism of gases. Newton's observations on
soap-bubbles were often referred to by Faraday. His delight in a
soap-bubble was like that of a boy, and he often introduced them into
his lectures, causing them, when filled with air, to float on invisible
seas of carbonic acid, and otherwise employing them as a means of
illustration. He now finds them exceedingly useful in his experiments
on the magnetic condition of gases. A bubble of air in a magnetic field
occupied by air was unaffected, save through the feeble repulsion of its
envelope. A bubble of nitrogen, on the contrary, was repelled from the
magnetic axis with a force far surpassing that of a bubble of air.
The deportment of oxygen in air 'was very impressive, the bubble being
pulled inward or towards the axial line, sharply and suddenly, as if the
oxygen were highly magnetic.'

He next labours to establish the true magnetic zero, a problem not so
easy as might at first sight be imagined. For the action of the magnet
upon any gas, while surrounded by air or any other gas, can only be
differential; and if the experiment were made in vacuo, the action of
the envelope, in this case necessarily of a certain thickness, would
trouble the result. While dealing with this subject, Faraday makes
some noteworthy observations regarding space. In reference to the
Torricellian vacuum, he says, 'Perhaps it is hardly ne