nconsciously kicks the rudder-controls, first to one side, then
to the other. The needle begins to swing around, and the compass is thus
rendered useless for the time being. For the next minute or two, until
it is safe to leave the clouds, the pilot must now keep the machine
straight by instinct, and trust to his sense of direction.

A similar mishap often happens when flying through cloud. Pilots have
been known to declare that all compasses are liable to swing of their
own accord when in clouds, though the real explanation is probably that
they themselves have disturbed the needle unduly by a continuous
pressure on each side of the rudder-bar in turn, thus causing an
oscillation of the rudder and a consequent zigzagged line of flight. The
trouble is more serious than it would seem to the layman, as when the
compass is out of action, and no other guides are available, one tends
to drift round in a large circle, like a man lost in the jungle. Should
the craft be driven by a rotary engine, the torque, or outward wash
from the propeller, may make a machine edge more and more to the left,
unless the pilot is careful to allow for this tendency.

Such a drift to the left has taken us well to the north of a straight
line between Toutpres and Passementerie, as we discover on leaving the
clouds for a second or two, so as to correct the error with the aid of
landmarks. But the compass has again settled down to good behaviour, and
we are able to get a true course before we climb back to the sheltering
whiteness.

A flight inside the clouds is far from pleasant. We are hemmed in by a
drifting formlessness that looks like thin steam, but, unlike steam,
imparts a sensation of coldness and clamminess. The eye cannot penetrate
farther than about a yard beyond the wing tips. Nothing is to be seen
but the aeroplane, nothing is to be heard but the droning hum of the
engine, which seems louder than ever amid the isolation.

I am bored, cold, and uncomfortable. Time drags along lamely; five
minutes masquerade as half an hour, and only by repeated glances at the
watch do I convince myself that we cannot yet have reached the next
objective. I study the map for no particular reason except that it is
something to do. Then I decide that the Lewis gun ought to be fired as a
test whether the working parts are still in good order. I hold the
spade-grip, swing round the circular mounting until the gun points to
the side, and loose five rounds in