consequently the mixture would be so rarefied that on the
compression stroke the pressure would possibly be extremely low and
totally inadequate to produce efficient working. Moreover, working at
such a high vacuum as this would not only prevent us obtaining a normal
explosion in the cylinder, but would upset the working of the exhaust
valve. The latter being held down on its seat during the suction stroke
by means of a spiral spring would be lifted off its seat by suction (the
partial vacuum in the cylinder), and any burnt gases which happened to
be hanging about in the exhaust port or pipe would be drawn into the
cylinder again, and tend to damp the ensuing explosion. Too early
closing of the exhaust should be avoided almost as rigorously as too
late. The latter will affect the working in a similar way to the exhaust
being lifted on the charging stroke by suction; on the other hand, if it
closes too soon, the entire volume of burnt gases will not have been
swept out of the cylinder, and the effect will again be to damp the
following explosion.

The gas valve opens just after the crank is above the back centre and
closes just before the front centre is reached, that is, opening a
little after the air valve and closing a shade before it, thus every
particle of gas is used in the cylinder, due to a draught of air being
drawn in after the gas valve has been closed.

The settings of the valve being of primary importance, no matter what
size engine we are dealing with, and being also the most confusing
matter for anyone unacquainted with gas engines to grasp, it will not be
out of place to suggest a simple method of checking these settings.

Let us begin by pulling the fly-wheel round backwards until we feel the
piston is on the compression stroke, then from this point--the crank
being about 45 deg. above the front centre--pull the wheel round until the
crank is in the position for the exhaust opening (see fig. 28). In this
position there should be but the _slightest_ play in the exhaust lever,
showing that the valve is _just_ on point of opening; and by keeping
one's hand on the lever whilst the fly-wheel is pulled round _very
slowly_ (it is a good plan to get some one else to do the pulling
round), it is possible to ascertain the precise point at which the valve
opens. Next pull round till the crank is in the position for the air
valve opening, and observe that it is set correctly. Then go on to a
trifle above the back