Should the engine be standing still in the position shown by Fig. 6, with the crank on the dead center, when ignition occurred, there would be no more tendency to turn the crank to the right than to the left, providing of course, that there was no effect from the momentum of a revolving fly wheel. If ignition occurred with the crank inclined ever so little toward the right, the pressure of the piston would force the crank downwards in a right handed direction. If the crank were inclined to the left, the tendency would be for left handed rotation. If the ignition system were arranged so that the spark occurred when the crank was inclined towards the right every time that the piston came up on the compression stroke, we should have continuous rotation in a right hand direction. By shifting the sequence of the spark so that it occurred with the crank on the left we would cause the engine to stop and reverse to left handed rotation. This is exactly the method used in reversing two stroke motors in practice, the change in the ignition being accomplished by advancing or retarding the mechanism that dispatches the spark (“Timer” or “Commutator”). [Illustration: Fig. F-9. Cross Section of Fairbanks-Morse Three Port—Two Stroke Cycle Engine, with Parts Named. ] This is an advantage not possessed by the four stroke cycle engine of the ordinary type, as the cams and valve mechanism require reversal as well as a reversal of the ignition system. This relation between the valve action and rotation in a four stroke cycle engine may be illustrated by the following example. Consider the piston at the end of the compression stroke in an engine designed for right hand rotation. After ignition, under the proper conditions, the piston would descend turning the crank to the right until it reached the bottom of the stroke, at which point the exhaust valve would open and relieve the pressure in the cylinder. Let us now consider an attempt at reversing the engine by causing the spark to occur before the piston reached the end of the compression stroke with the crank still inclined toward the left. In this case the piston would force the crank down in a left hand direction until it reached the end of the stroke. The exhaust valve would not open to relieve the pressure, as the exhaust cam would be moving away from the valve rod instead of toward it. Should the crank swing a little past the dead center, because of its momentum, the inlet valve would be opened instead of the exhaust, and the contents of the cylinder would shoot through the intake pipe and carburetor. This would bring matters to a close as far as rotation was concerned.