ppeared sufficient to
the Venetians for all ordinary purposes; and they used it in common
windows to the latest Gothic periods, but yet with certain differences
which at once show the lateness of the work. In the first place, the
rose, which at first was flat and quatrefoiled, becomes, after some
experiments, a round ball dividing into three leaves, closely resembling
our English ball flower, and probably derived from it; and, in other
cases, forming a bold projecting bud in various degrees of contraction
or expansion. In the second place, the extremities of the angle leaves
are wrought into rich flowing lobes, and bent back so as to lap against
their own breasts; showing lateness of date in exact proportion to the
looseness of curvature. Fig. 3 represents the general aspect of these
later capitals, which may be conveniently called the rose capitals of
Venice; two are seen on service, in Plate VIII. Vol. I., showing
comparatively early date by the experimental form of the six-foiled
rose. But for elaborate edifices this form was not sufficiently rich;
and there was felt to be something awkward in the junction of the leaves
at the bottom. Therefore, four other shorter leaves were added at the
sides, as in fig. 13, Plate II., and as generally represented in Plate
X. Vol. II. fig. 1. This was a good and noble step, taken very early in
the thirteenth century; and all the best Venetian capitals were
thenceforth of this form. Those which followed, and rested in the common
rose type, were languid and unfortunate: I do not know a single good
example of them after the first half of the thirteenth century.

But the form reached in fig. 13 was quickly felt to be of great value
and power. One would have thought it might have been taken straight from
the Corinthian type; but it is clearly the work of men who were making
experiments for themselves. For instance, in the central capital of Fig.
XXXI. Vol. II., there is a trial condition of it, with the intermediate
leaf set behind those at the angles (the reader had better take a
magnifying glass to this woodcut; it will show the character of the
capitals better). Two other experimental forms occur in the Casa Cicogna
(Vol. II.), and supply one of the evidences which fix the date of that
palace. But the form soon was determined as in fig. 13, and then means
were sought of recommending it by farther decoration.

The leaves which are used in fig. 13, it will be observed, have lost
the C