ould easily be
devoted to the subject, and even then much valuable information would
have to be omitted from lack of space. It is, therefore, impossible,
in only a section of a chapter, to do more than indicate in the
briefest manner a few salient features concerning these problems. The
suspension of weights from the lightest possible gas compartment must
be based on the ordinary principles of calculating the distribution
loads as in ships and other structures. In the non-rigid, the envelope
being made of flexible fabric has, in itself, no rigidity whatsoever,
and its shape must be maintained by the internal pressure kept slightly
in excess of the pressure outside. Fabric is capable of resisting
tension, but is naturally not able to resist compression. If the car
was rigged beneath the centre of the envelope with vertical suspensions
it would tend to produce compression in the underside of the envelope,
owing to the load not being fully distributed. This would cause, in
practice, the centre portion of the envelope to sag downwards, while
the ends would have a tendency to rise. The principle which has been
found to be most satisfactory is to fix the points of suspension
distributed over the greatest length of envelope possible proportional
to the lift of gas at each section thus formed. From these points the
wires are led to the car. If the car is placed close to the envelope
it will be seen that the suspensions of necessity lie at a very flat
angle and exert a serious longitudinal compression. This must be
resisted by a high internal pressure, which demands a stouter fabric
for the envelope and, therefore, increased weight. It follows that the
tendency of the envelope to deform is decreased as the distance of the
car from the gas compartment is increased.

One method of overcoming this difficulty is found by using the
Astra-Torres design. As will be seen from the diagram of the North Sea
airship, the loads are excellently distributed by the several fans of
internal rigging, while external head resistance is reduced to a
minimum, as the car can be slung close underneath the envelope.
Moreover, the direct longitudinal compression due to the rigging is
applied to a point considerably above the axis of the ship. In a large
non-rigid many of these difficulties can be overcome by distributing
the weight into separate cars along the envelope itself.

We have seen that as an airship rises the gas contained in the en