n we begin to
examine the order of the leaves, we find it varied carefully. Let A
stand for scarlet bud, _b_ for blue leaf, _c_ for two blue leaves on one
stalk, _s_ for a stalk without a leaf, and R, for the large red leaf.
Then, counting from the ground, the order begins as follows:

_b_, _b_, A; _b_, _s_, _b_, A; _b_, _b_, A; _b_, _b_, A; and we think we
shall have two _b_'s and an A all the way, when suddenly it becomes _b_,
A; _b_, R; _b_, A; _b_, A; _b_, A; and we think we are going to have
_b_, A continued; but no: here it becomes _b_, _s_; _b_, _s_; _b_, A;
_b_, _s_; _b_, _s_; _c_, _s_; _b_, _s_; _b_, _s_; and we think we are
surely going to have _b_, _s_ continued, but behold it runs away to the
end with a quick _b_, _b_, A; _b_, _b_, _b_, _b_![69] Very often,
however, the designer is satisfied with _one_ surprise, but I never saw
a good illuminated border without one at least; and no series of any
kind was ever introduced by a great composer in a painting without a
snap somewhere. There is a pretty one in Turner's drawing of Rome with
the large balustrade for a foreground in the Hakewell's Italy series:
the single baluster struck out of the line, and showing the street below
through the gap, simply makes the whole composition right, when
otherwise it would have been stiff and absurd.

224. If you look back to Fig. 48 you will see, in the arrangement of the
battlements, a simple instance of the use of such variation. The whole
top of the tower, though actually three sides of a square, strikes the
eye as a continuous series of five masses. The first two, on the left,
somewhat square and blank, then the next two higher and richer, the
tiles being seen on their slopes. Both these groups being couples, there
is enough monotony in the series to make a change pleasant; and the last
battlement, therefore, is a little higher than the first two,--a little
lower than the second two,--and different in shape from either. Hide it
with your finger, and see how ugly and formal the other four battlements
look.

225. There are in this figure several other simple illustrations of the
laws we have been tracing. Thus the whole shape of the walls' mass being
square, it is well, still for the sake of contrast, to oppose it not
only by the element of curvature, in the ring, and lines of the roof
below, but by that of sharpness; hence the pleasure which the eye takes
in the projecting point of the roof. Also, because the walls are t