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 decreased. When running at high speed the engineer "links up" his reversing gear, causing his valves to cut off early in the stroke, and the steam to work more expansively than it could with the lever at _full_, or _end_, gear; so that this device not only renders an engine reversible, but also gives the engineer an absolute command over the expansion ratio of the steam admitted to the cylinder, and furnishes a method of cutting off the steam altogether. In Figs. 30, 31, 32, the valve has no lap and the eccentrics are set square. In actual practice the valve faces would have "lap" and the eccentric "lead" to correspond; but for the sake of simplicity neither is shown. OTHER GEARS. In the Gooch gear for reversing locomotives the link does not shift, but the valve rod and its block is raised or lowered. The Allan gear is so arranged that when the link is raised the block is lowered, and _vice versa_. These are really only modifications of Stephenson's principle--namely, the employment of _two_ eccentrics set at equal angles to and on opposite sides of the crank. There are three other forms of link-reversing gear, and nearly a dozen types of _radial_ reversing devices; but as we have already described the three most commonly used on locomotives and ships, there is no need to give particulars of these. Before the introduction of Stephenson's gear a single eccentric was used for each cylinder, and to reverse the engine this eccentric had to be loose on the axle. "A lever and gear worked by a treadle on the footplate controlled the position of the eccentrics. When starting the engine, the driver put the eccentrics out of gear by the treadle; then, by means of a lever he raised the small-ends[5] of the eccentric rods, and, noting the position of the cranks, or, if more convenient, the balance weight in the wheels, he, by means of another handle, moved the valves to open the necessary ports to steam and worked them by hand until the engine was moving; then, with the treadle, he threw the eccentrics over to engage the studs, at the same time dropping the small-ends of the rods to engage pins upon the valve spindles, so that they continued to keep up the movement of the valve."[6] One would imagine that in modern shunting yards such a device would somewhat delay operations! PISTON VALVES.