d upward by the centrifugal action of the screw will,
by impinging on the overhanging stern, press the vessel forward in the
water, just in the same way as is done by the wind when acting on an
oblique sail. I believe, the two revolving vanes without any twist or
obliquity on them at all, would propel a vessel if set well forward in the
dead wood or beneath the bottom, merely by the ascent of the water up the
inclined plane of the vessel's run; and, at all events, a screw so placed
would, in my judgment, aid materially in propelling the vessel when her
progress was resisted by head winds.

594. _Q._--But you said there are some kinds of screws which profess to
accomplish this?

[Illustration: Fig. 49. THE EARL OF DUNDONALD'S PROPELLER.]

_A._--There are screws which profess to counteract the centrifugal velocity
given to the water by imparting to it an equal centripetal force, the
consequence of which will be, that the water projected backward by the
screw, instead of taking the form of the frustum of a cone, with its small
end next the screw, will take the form of a cylinder. One of these forms of
screw is that patented by the Earl of Dundonald in 1843, and which is
represented in fig. 49. Another is the form of screw already represented in
fig. 48, and which was patented by Mr. Hodgson in 1844. Mr. Hodgson bends
the arms of his propellers backward, not into the form of a triangle, but
into the form of a parabola, to the end that the impact of the screw on the
particles of the water may cause them to converge to a focus, as the rays
of light would do in a parabolic reflector. But this particular
configuration is not important, seeing that the same convergence which is
given to the particles of the water, with a screw of uniform pitch bent
back into the form of a parabola, will be given with a screw bent back into
the form of a triangle, if the pitch be suitably varied between the centre
and the circumference.

595. _Q._--Then the pitch may be varied in two ways?

_A._--Yes: a screw may have a pitch increasing in the direction of the
length, as would happen in the case of a spiral stair, if every successive
step in the ascent was thicker than the one below it; or it may increase
from the centre to the circumference, as would happen in the case of a
spiral stair, if every step were thinner at the centre of the lower than at
its outer wall. When the pitch of a screw increases in the direction of its
length, the leadin