started at Toledo and Ancor, Ohio.

At Muscle Shoals a mushroom city of 20,000 sprang up on an Alabama
cotton field in six months. The raw material, air, was as abundant there
as anywhere and the power, water, could be obtained from the Government
hydro-electric plant on the Tennessee River, but this was not available
during the war, so steam was employed instead. The heat of the coal was
used to cool the air down to the liquefying point. The principle of this
process is simple. Everybody knows that heat expands and cold contracts,
but not everybody has realized the converse of this rule, that expansion
cools and compression heats. If air is forced into smaller space, as in
a tire pump, it heats up and if allowed to expand to ordinary pressure
it cools off again. But if the air while compressed is cooled and then
allowed to expand it must get still colder and the process can go on
till it becomes cold enough to congeal. That is, by expanding a great
deal of air, a little of it can be reduced to the liquefying point. At
Muscle Shoals the plant for liquefying air, in order to get the nitrogen
out of it, consisted of two dozen towers each capable of producing 1765
cubic feet of pure nitrogen per hour. The air was drawn in through two
pipes, a yard across, and passed through scrubbing towers to remove
impurities. The air was then compressed to 600 pounds per square inch.
Nine tenths of the air was permitted to expand to 50 pounds and this
expansion cooled down the other tenth, still under high pressure, to the
liquefying point. Rectifying towers 24 feet high were stacked with trays
of liquid air from which the nitrogen was continually bubbling off since
its boiling point is twelve degrees centigrade lower than that of
oxygen. Pure nitrogen gas collected at the top of the tower and the
residual liquid air, now about half oxygen, was allowed to escape at the
bottom.

The nitrogen was then run through pipes into the lime-nitrogen ovens.
There were 1536 of these about four feet square and each holding 1600
pounds of pulverized calcium carbide. This is at first heated by an
electrical current to start the reaction which afterwards produces
enough heat to keep it going. As the stream of nitrogen gas passes over
the finely divided carbide it is absorbed to form calcium cyanamid as
described on a previous page. This product is cooled, powdered and wet
to destroy any quicklime or carbide left unchanged. Then it is charged
into aut