opens C., closes A. and opens B. If these mechanical movements are
completed it must be plain that so long as the air (or steam) pressure
is continued in the forked pipe the piston will automatically cut off
its supply and open its escape at each alternate end, and move back and
forth. Any boy can see how a backward and forward movement may be made
to give motion to a crank. All other details in an engine are questions
of convenience in construction, and not questions of principle or manner
of action.

Of older readers, I might request the supposition that, in Fig. 2, only
the valves A. and B. were automatically and invariably opened and closed
by the action of the piston-rod of Fig. 3, and that C. and D. were
controlled solely by the governor, before mentioned, which we will
suppose to be located at E. Then the escape of the steam ahead of the
piston must always come at the same time with reference to the stroke,
but the supply will depend upon the requirements of each individual
stroke, and the work it has to do, and afford to the piston a greater or
less push, as the emergencies of that particular instant may require.
This arrangement would be one of regularity of movement and of economy
in the use of steam. That which is needed is supplied, and no more. This
is the principle and the object of the Corliss cut-off, and of all
others similar to it in purpose. Their principle is that _only the
escape is automatically controlled by the movements of the
piston-rod_, occurring always at the same time with reference to the
stroke, while _the supply is under control of the movement of the
governor_, and regulated according to the emergencies of the
movement. The governor, in any of its forms, as ordinarily applied,
performs only half of this function. It regulates the general supply of
steam to the cylinder, but the supply-valve continues to be opened,
always to full width, and always at the same moment with reference to