In all these processes the action of the gas is impeded by the bulky
presence of its fellow constituent of air, nitrogen. We may say, for
instance, in homely phrase, that whenever a fire burns there are four
volumes of nitrogen tending to extinguish it for every volume of
oxygen supporting its combustion, and to the same degree the nitrogen
interferes with all other processes of atmospheric oxidation, of which
most metallurgical operations may be given as instances. If, then, it
has become possible to remove this diluent gas simply and cheaply in
order to give the oxygen free play in its various applications, we are
doubtless on the eve of a revolution among some of the most extensive
and familiar of the world's industries.

A series of chemical reactions has long been known by means of which
oxygen could be separated out of air in the laboratory, and at various
times processes based on these reactions have been patented for the
production of oxygen on a large scale. Until recently, however, none
of these methods gave sufficiently satisfactory results. The simplest
and perhaps the best of them was based on the fact first noticed by
Boussingault, that when baryta (BaO) is heated to low redness in a
current of air, it takes up oxygen and becomes barium dioxide
(BaO_{2}), and that this dioxide at a higher temperature is
reconverted into free oxygen and baryta, the latter being ready for
use again. For many years it was assumed, however, by chemists that
this ideally simple reaction was inapplicable on a commercial scale,
owing to the gradual loss of power to absorb oxygen which was always
found to take place in the baryta after a certain number of
operations. About eight years ago Messrs. A. & L. Brin, who had
studied chemistry under Boussingault, undertook experiments with the
view of determining why the baryta lost its power of absorbing oxygen.