an air pressure in the direction C. Such pressure, acting upon
the keel-surface, presses the tail back towards its first position in
which the aeroplane is upon its course B.

What I have described is continually going on during flight, but in
a well-designed aeroplane such stabilizing movements are, most of the
time, so slight as to be imperceptible to the pilot.

If an aeroplane was not stabilized in this way, it would not only be
continually trying to leave its course, but it would also possess a
dangerous tendency to "nose away" from the direction of the side gusts.
In such case the gust shown in the above illustration would turn the
aeroplane round the opposite way a very considerable distance; and the
right wing, being on the outside of the turn, would travel with greater
velocity than the left wing. Increased velocity means increased lift;
and so, the right wing lifting, the aeroplane would turn over sideways
very quickly.

LONGITUDINAL STABILITY.--Flat surfaces are longitudinally stable owing
to the fact that with decreasing angles of incidence the centre line of
pressure (C.P.) moves forward.

The C.P. is a line taken across the surface, transverse to the direction
of motion, and about which all the air forces may be said to balance, or
through which they may be said to act.

Imagine A to be a flat surface, attitude vertical, travelling through
the air in the direction of motion M. Its C.P. is then obviously along
the exact centre line of the surface as illustrated.

In B, C, and D the surfaces are shown with angles of incidence
decreasing to nothing, and you will note that the C.P. moves forward
with the decreasing angle.

Now, should some gust or eddy tend to make the surface decrease the
angle, i.e., dive, then the C.P. moves forward and pushes the front of
the surface up. Should the surface tend to assume too large an angle,
then the reverse happens--the C.P. moves back and pushes the rear of the
surface up.

Flat surfaces are, then, theoretically stable longitudinally. They are
not, however, used, on account of their poor lift-drift ratio.

As already explained, cambered surfaces are used, and these are
longitudinally unstable at those angles of incidence producing a
reasonable lift-drift ratio, i.e., at angles below: about 12 degrees.

A is a cambered surface, attitude approximately vertical, moving through
the air in the direction M. Obviously the C. P. coincides with the
transverse centre