But again, in order that the cut-off may operate in the same manner when backing as when going ahead, this eccentric must be symmetrically situated with respect to both C and G; and since L O M bisects and is perpendicular to G C, it follows that if the cut-off eccentric be fixed on the shaft, its center must be located at H, the intersection of C T with L M. This would require the edge of the cut-off valve at the given instant to be at Q, perpendicularly over H; and the travel over the main valve would be equal to twice C H, the virtual lever arm of the eccentric, the actual traverse in the valve chest being twice O H, the real eccentricity.

This being clearly excessive, let us next see what will occur if the lever arm, CH, be reduced as in the diagram to CK.  The edge of the cut-off valve will then be at N; it instantly begins to close the port.  CN, but not so rapidly as the main valve opens the port, AB.

The former motion increases in rapidity, while the latter decreases; therefore at some point they will become equal in velocity, and the openings of the two ports will be the same; and the question is, Will this maximum effective port area give a sufficient supply of steam?

This diagram is the same as the one actually used in the engine under consideration, in which it was required to follow a minimum distance of 5 inches in the stroke of 22.  Under these conditions it is found that the actual port opening for that point of cutting off is three-fifths of that allowed when following full stroke, whereas the speed of the piston at the time when this maximum opening occurs is less than half its greatest speed.

This, it would seem, is ample; but we now find the eccentric, K, no longer in the right position for backing; when the engine is reversed it ought to be at, P, the angle, POL, being equal to the angle, KOL.  By leaving it free, therefore, to move upon the shaft, by the means above described, through the angle, KOP, the desired object is accomplished.  The real eccentricity is now reduced in the proportion of OK to OH, while the lengths of the cut-off valves, and what is equally important, their travel over the back of the main valve, are reduced in the proportion of CK to CH, in this instance nearly one-half; a gain quite sufficient to warrant the adoption of the expedient.

The third, and perhaps the most notable, peculiarity is the manner of suspending and operating the main link.  As before stated, this link is used only for reversing, and is therefore always in “full gear” in one direction or the other; and the striking feature of the arrangement here used is that, whether going ahead or backing, there is no slipping of the link upon the link block.

The link itself is of the simplest form, being merely a curved flat bar, L, in which are two holes, A and B (Fig. 7), by which the link is hung upon the pins, which project from the sides of the eccentric rods at their upper ends.