ould tend to prove.

[Illustration: EXPERIMENTS WITH DOUBLE-BARRELLED GUNS.]

Fig. 1 represents a single barrel fitted with sights and firmly
attached to a heavy block of beech. This was placed on an ordinary
rifle rest, being fastened thereto by a pin at the corner, A, the
block and barrel being free to revolve upon the pin as a center.
Several shots were fired both with the pin in position and with it
removed, the barrel being carefully pointed at the target each time.
No practical difference in the accuracy of fire was discernible under
either condition. When the pin was holding the corner of the block,
the recoil caused the barrel to move from right to left in a circular
path; but when the pin was removed, so that the block was not attached
to the rest in any way, the recoil took place in a line with the axis
of the bore. It will be observed that the conditions which are present
when a double barreled gun is fired in the ordinary way from the
shoulder were in some respects much exaggerated in the apparatus, for
the pin was a distance of 3 in. laterally from the axis of the barrel,
whereas the center of resistance of the stock of a gun against the
shoulder would ordinarily be about one-sixth of this distance from the
axis of the barrel. This experiment would apparently tend to prove
that the recoil does not appreciably affect the path of the
projectile, as it would seem that the latter must clear the muzzle
before any considerable movement of the barrel takes place.

With a view to obtain a further confirmation of the result of this
experiment, it was repeated in a different form by a number of shots
being fired from a "cross-eyed" rifle,[1] in which the sights were
fixed in the center of the rib. Very accurate shooting was obtained
with this arm.

[Footnote 1: A cross-eyed rifle is one made with a crooked stock
for the purpose of shooting from the right shoulder, aim being
taken with the left eye.]

A second theory, often broached, in order to account for the
divergence of the charge, is that the barrel which is not being fired,
by its _vis inertia_ in some way causes the shot to diverge. In order
to test this, Mr. Phillips took a single rifle and secured it near the
muzzle to a heavy block of metal, when the accuracy of the shooting
was in no way impaired.

So far the experiments were of a negative character, and the next step
was made with a view to discover the actual cause of the divergence
refe