line will again be on the lateral face; and when again bent to the
south, the line will be on the original convex surface. Now, instead
of bending the sapling, let us suppose that the cells along its
northern surface from the base to the tip were to grow much more
rapidly than on the three other sides, the whole shoot would then
necessarily be bowed to the south; and let the longitudinal growing
surface creep round the shoot, deserting by slow degrees the northern
side and encroaching on the western side, and so round by the south,
by the east, again to the north. In this case the shoot would remain
always bowed with the painted line appearing on the several above
specified surfaces, and with the point of the shoot successively
directed to each point of the compass. In fact, we should have the
exact kind of movement performed by the revolving shoots of twining
plants. {9}

It must not be supposed that the revolving movement is as regular as
that given in the above illustration; in very many cases the tip
describes an ellipse, even a very narrow ellipse. To recur once
again to our illustration, if we suppose only the northern and
southern surfaces of the sapling alternately to grow rapidly, the
summit would describe a simple arc; if the growth first travelled a
very little to the western face, and during the return a very little
to the eastern face, a narrow ellipse would be described; and the
sapling would be straight as it passed to and fro through the
intermediate space; and a complete straightening of the shoot may
often be observed in revolving plants. The movement is frequently
such that three of the sides of the shoot seem to be growing in due
order more rapidly than the remaining side; so that a semi-circle
instead of a circle is described, the shoot becoming straight and
upright during half of its course.