that the cut off valve will engage with them earlier or later as desired.  This device was designed specially as an inexpensive method of changing the common slide valve into an automatic cut off.  The cut off would not be as quick as in other cases we have cited, depending here upon the movement of the lower valve alone, and that, too, is in its slowest movement; whereas in the other cases, the edges approaching each other, by the differing movement of the valves the cut off is very rapid, provided the distance to travel is not long.  In this device considerable noise must result by the cut off valve striking the cut off blocks, and a considerable amount of leakage is likely to occur past this valve.

But there is one great objection in the valve gears thus far cited, that the travel of the expansion valve upon the main valve is variable.  I have in mind the case of a Kendall & Roberts engine, which had been run for a long time at no better economy than would be obtained from a plain slide valve engine, and when it was attempted to get an earlier cut off by separating the two cut off valves, they had worn so much in their old place on the valve that shoulders were found sufficient to cause a disagreeable noise and a leaky valve.  This is very apt to occur, not only where the valve is run for a long time on one seat, but in cases of variation of the travel of the expansion valve.  The result is that a change will bring about a leaky valve, something that every engineer abhors.

The construction of the Buckeye engine, which is also of this type, is such that the travel of the valve on the back of the main valve is always the same, no matter what the cut off may be.  Then this engine makes use of our second proposition as a means of effecting the cut off, viz., by advancing the eccentric.  You will readily observe that anything that will cause the cut off valve to reach a certain point earlier in the stroke will bring about an earlier cut off as it hastens everything all around.  This is the plan pursued in the Buckeye, in which the governor, of the shaft type, turns the eccentric forward or back according as the load demands.  Then, in addition, the valve is balanced partially, the attempt not being made to produce an absolutely balanced valve, on the ground that there should be friction enough to keep the surfaces bright and to prevent leakage.  The most perfect valve will, of course, be entirely balanced under all conditions of pressure so as to move with perfect ease.  With the riding cut off valve in connection with the plain slide valve, this is not accomplished, and it does not matter whether it is partially unbalanced to prevent leakage or not, the fact that it is not entirely balanced prevents it reaching the ideal valve.

[Illustration:  Fig. 5]