156 THE STONES OF VENICE
most part found in the one, and severity of ornament, with stern discipline of the fancy, in the other.
Of these two families of capitals, both occur in the Byzantine period, but the concave group is the longest-lived, and extends itself into the Gothic times. In the account which I gave of them in the first volume, they were illustrated by giving two portions of a simple curve, that of a salvia leaf.1 We must now investigate their characters more in detail; and these may be best generally represented by considering both families as formed upon the types of flowers,-the one upon that of the water-lily, the other upon that of the convolvulus. There was no intention in the Byzantine architects to imitate either one or other of these flowers; but, as I have already so often repeated, all beautiful works of art must either intentionally imitate or accidentally resemble natural forms;2 and the direct comparison with the natural forms which these capitals most resemble, is the likeliest mode of fixing their distinctions in the reader’s mind.
The one then, the convex family, is modelled according to the commonest shapes of that great group of flowers which form rounded cups, like that of the water-lily, the leaves springing horizontally from the stalk, and closing together upwards. The rose is of this family, but her cup is filled with the luxuriance of her leaves; the crocus, campanula, ranunculus, anemone, and almost all the loveliest children of the field, are formed upon the same type.
The other family resembles the convolvulus, trumpetflower, and such others, in which the lower part of the bell is slender, and the lip curves outward at the top. There are fewer flowers constructed on this than on the convex model; but in the organization of trees and of clusters of herbage it is seen continually. Of course, both of these conditions are modified, when applied to capitals, by the enormously greater thickness of the stalk or shaft, but in other respects the parallelism is close and accurate; and the reader had better
1 [See Vol. IX., Plate vii., p. 268.]
2 [See, e.g., Vol. IX., pp. 70, 253, 293, 409.]
[Version 0.04: March 2008]