'The views and opinions of Brigadier General John Jacob, CB. Aide-decamp to the Queen; Aide-de-camp to the Governor General of India, &c. &c. &c.; late commanding in chief the British forces in Persia; at present commanding the Sind Irregular Horse, and Political Superintendent of the North-West frontier of Sind.' [‎233] (264/481)

233
observed in my practice. There may be some such ten
dency to deviation from such cause, but it must be of small
amount, and is not preceptible.
The real facts of the matter seem to be these: consider
ing the person of the rifleman as a cylindrical column, the
recoil of a gun fired from the right shoulder is a force acting
at a distance from the axis of that column, and therefore
tending not only to drive it back, but to run it round on
its axis. The muzzle of the gun, therefore, must receive a
certain degree of motion to the right hand when fired.
It was found here, in many thousands of trials, that
on a calm day the balls of the guns which recoiled with
the greatest force, fired from the right shoulder, always
deviated to the right hand; this deviation amounting to
about twenty feet at 800 yards.
That the cause of this is correctly assigned above is thus
proved. The same gun fired from the left shoulder threw
the ball to the left. With the guns recoiling most, such
as the 8-gauge rifles, used with a four-ounce ball and
three-drachm charge, the deviation was very great; with
the 32-gauge rifles, of which the recoil was slight, this
regular deviation did not occur at all. These effects were
observed always to occur, without any respect whatever to
the amount of twist given to the grooves of the rifle. In
practice, a perfect remedy was found for this deviation, or
for any other similar errors, caused by the action of wind
or anything else. When taking aim, leaning the gun
over a little to the left perfectly counteracted the deviation
to the right; and by practice the habit is formed of
instinctively adjusting the amount of the lateral inclination
thus given to the sight, so as correctly to compensate the
error.
Thus the matter stands at present. The limit of the
31