chemistry under Boussingault, undertook experiments with the
view of determining why the baryta lost its power of absorbing oxygen.

They found that it was owing to molecular and physical changes caused
in it by impurities in the air used and by the high temperature
employed for decomposing the dioxide. They discovered that by heating
the dioxide in a partial vacuum the temperature necessary to drive off
its oxygen was much reduced. They also found that by supplying the air
to the baryta under a moderate pressure, its absorption of oxygen was
greatly assisted. Under these conditions, and by carefully purifying
the air before use, they found that it became possible to use the
baryta an indefinite number of times. Thus the process became
practically, as it was theoretically, continuous.

After securing patent protection for their process, Messrs. Brin
erected a small producer in Paris, and successfully worked it for
nearly three years without finding a renewal of the original charge of
baryta once necessary. This producer was exhibited at the Inventions
Exhibition in London, in 1885. Subsequently an English company was
formed, and in the autumn of last year Brin's Oxygen Company began
operations in Horseferry Road, Westminster, where a large and complete
demonstration plant was erected, and the work commenced of developing
the production and application of oxygen in the industrial world.

[Illustration: APPARATUS FOR MAKING OXYGEN.]

We give herewith details of the plant now working at Westminster. It
is exceedingly simple. On the left of the side elevation and plan are
shown the retorts, on the right is an arrangement of pumps for
alternately supplying air under pressure and exhausting the oxygen
from the retorts. As is shown in the plan, two sets of apparatus are
worked side by side at Westminster, the seventy-two retorts shown in
the drawings being divided into two systems of thirty-six. Each system
is fed by the two pumps on the corresponding side of the boiler. Each
set of retorts consists of six rows of six retorts each, one row above
the other. They are heated by a small Wilson's producer, so that the
attendant can easily regulate the supply of heat and obtain complete
control over the temperature of the retorts. The retorts, A, are made
of wrought iron and are about 10 ft long and 8 in. diameter.
Experience, however, goes to prove that there is a limit to the
diameter of the retorts beyond which the results be