nding current of mingled steam and water. It is also
necessary that this uptake should be practically direct, and it should
not be composed of frequent enlargements and contractions. Take, for
instance, a boiler well known in Europe, copied and sold here under
another name. It is made up of inclined tubes secured by pairs into
boxes at the ends, which boxes are made to communicate with each other
by return bends opposite the ends of the tubes. These boxes and return
bends form an irregular uptake, whereby the steam is expected to rise to
a reservoir above. You will notice (Fig. 8) that the upward current of
steam and water in the return bend meets and directly antagonizes the
upward current in the adjoining tube. Only one result can follow. If
their velocities are equal, the momentum of both will be neutralized and
all circulation stopped, or, if one be stronger, it will cause a back
flow in the other by the amount of difference in force, with practically
the same result.

[Illustration: 4880 Horse-power Installation of Babcock & Wilcox Boilers
at the Open Hearth Plant of the Cambria Steel Co., Johnstown, Pa. This
Company Operates a Total of 52,000 Horse Power of Babcock & Wilcox
Boilers]

[Illustration: Fig. 9]

In a well-known boiler, many of which were sold, but of which none are
now made and a very few are still in use, the inventor claimed that the
return bends and small openings against the tubes were for the purpose
of "restricting the circulation" and no doubt they performed well that
office; but excepting for the smallness of the openings they were not as
efficient for that purpose as the arrangement shown in Fig. 8.

[Illustration: Fig. 10]

Another form of boiler, first invented by Clarke or Crawford, and lately
revived, has the uptake made of boxes into which a number, generally
from two to four tubes, are expanded, the boxes being connected together
by nipples (Fig. 9). It is a well-known fact that where a fluid flows
through a conduit which enlarges and then contracts, the velocity is
lost to a greater or less extent at the enlargements, and has to be
gotten up again at the contractions each time, with a corresponding loss
of head. The same thing occurs in the construction shown in Fig. 9. The
enlargements and contractions quite destroy the head and practically
overcome the tendency of the water to circulate.

A horizontal tube stopped at one end, as shown in Fig. 10, can have no
proper circulation