its the admission of high-pressure steam to both high and
low-pressure cylinders for starting a train, or moving it up heavy
grades.

REVERSING GEARS.

[Illustration: FIGS. 30, 31, 32.--Showing how a reversing gear alters
the position of the slide-valve.]

The engines of a locomotive or steamship must be reversible--that is,
when steam is admitted to the cylinders, the engineer must be able to
so direct it through the steam-ways that the cranks may turn in the
desired direction. The commonest form of reversing device (invented by
George Stephenson) is known as Stephenson's Link Gear. In Fig. 30 we
have a diagrammatic presentment of this gear. E^1 and E^2 are two
eccentrics set square with the crank at opposite ends of a diameter.
Their rods are connected to the ends of a link, L, which can be raised
and lowered by means of levers (not shown). B is a block which can
partly revolve on a pin projecting from the valve rod, working through
a guide, G. In Fig. 31 the link is half raised, or in "mid-gear," as
drivers say. Eccentric E^1 has pushed the lower end of the link fully
back; E^2 has pulled it fully forward; and since any movement of the
one eccentric is counterbalanced by the opposite movement of the other,
rotation of the eccentrics would not cause the valve to move at all, and
no steam could be admitted to the cylinder.

Let us suppose that Fig. 30 denotes one cylinder, crank, rods, etc., of
a locomotive. The crank has come to rest at its half-stroke; the
reversing lever is at the mid-gear notch. If the engineer desires to
turn his cranks in an anti-clockwise direction, he _raises_ the link,
which brings the rod of E^1 into line with the valve rod and presses
the block _backwards_ till the right-hand port is uncovered (Fig. 31).
If steam be now admitted, the piston will be pushed towards the left,
and the engine will continue to run in an anti-clockwise direction. If,
on the other hand, he wants to run the engine the other way, he would
_drop_ the link, bringing the rod of E^2 into line with the valve rod,
and drawing V _forward_ to uncover the rear port (Fig. 32). In either
case the eccentric working the end of the link remote from B has no
effect, since it merely causes that end to describe arcs of circles of
which B is the centre.

"LINKING UP."

If the link is only partly lowered or raised from the central position
it still causes the engine to run accordingly, but the movement of the
valve is decrea