some cases. Upon the impact of a shell with armor-plate the
physical phenomena occur instantaneously and the resultant forces are
so great that it is impossible to mechanically record their action. A
study of the stresses in the shell can, however, be made on a
theoretical basis.

In the first place, if the projectile were twenty calibres in length
and of a material offering less resistance to torsional stress than
steel and rotated at the high velocities indicated we would find that
upon impact the torsion would be plainly evident as per the following:

Assume a projectile A of length twenty calibres, about to penetrate an
armor-plate B of thickness sufficient to prevent complete penetration
by the shell in question.

[Illustration: Fig. 27]

The tendency of the impact is to stop the rotation of the projectile,
owing to the friction between the surfaces in contact, but owing to the
length of the projectile the point receives this retarding influence
before it can be transmitted throughout the body of the shell to its
base. The consequent result is that the head will finally come to a
stop while the base is still rotating, however slightly that may be.

Theoretically considering the projectile to be composed of a series of
discs a line drawn parallel to the major axis, while at rest, would be
represented by the line _cd_. Upon impact, however, the rotative force
tends to create a twisting couple with the result that each disc will
tend to slide on its preceding disc, so that by the time these twisting
couples have been transmitted to the base of the shell the original
line _cd_ will have taken some such position as _de_.

The objection to the present method of forging shells is as a result,
the grain or fibre of the metal lies parallel with the major axis of
the forging, the forging process causing an elongation of the ingot and
the metal grain following the direction of elongation. Consequently any
flaws occurring in the material will extend parallel to the grain or
major axis. If a flaw remains undiscovered in a finished projectile--as
is sometimes the case--the projectile is not only weakened thereby, but
the element of weakness lies in such a direction that the compression
forces and counterforces produce very much the same results as would a
wedge driven into a niche, i.e. the separation of adjacent material.
The author is in possession of a shell in which a longitudinal flaw was
revealed in the ogive by the