e to
draw the leaves of their natural size, they would be so small that their
forms would be invisible in the darkness; and were he to draw them so
large as that their shape might be seen, they would look like laurel
instead of olive. So he arranges them in small clusters of five each,
nearly of the shape which the Byzantines give to the petals of the lily,
but elongated so as to give the idea of leafage upon a spray; and these
clusters,--his object always, be it remembered, being _decoration_ not
less than _representation_,--he arranges symmetrically on each side of
his branches, laying the whole on a dark ground most truly suggestive of
the heavy rounded mass of the tree, which, in its turn, is relieved
against the gold of the cupola. Lastly, comes the question respecting
the fruit. The whole power and honor of the olive is in its fruit; and,
unless that be represented, nothing is represented. But if the berries
were colored black or green, they would be totally invisible; if of
any other color, utterly unnatural, and violence would be done to the
whole conception. There is but one conceivable means of showing them,
namely to represent them as golden. For the idea of golden fruit of
various kinds was already familiar to the mind, as in the apples of the
Hesperides, without any violence to the distinctive conception of the
fruit itself.[51] So the mosaicist introduced small round golden berries
into the dark ground between each leaf, and his work was done.

[Illustration: Plate IV.
Mosaics of Olive-tree and Flowers.]

Sec. XVII. On the opposite plate, the uppermost figure on the left is a
tolerably faithful representation of the general effect of one of these
decorative olive-trees; the figure on the right is the head of the tree
alone, showing the leaf clusters, berries, and _interlacing_ of the
boughs as they leave the stem. Each bough is connected with a separate
line of fibre in the trunk, and the junctions of the arms and stem are
indicated, down to the very root of the tree, with a truth in structure
which may well put to shame the tree anatomy of modern times.

Sec. XVIII. The white branching figures upon the serpentine band below are
two of the clusters of flowers which form the foreground of a mosaic in
the atrium. I have printed the whole plate in blue, because that color
approaches more nearly than black to the distant effect of the mosaics,
of which the darker portions are generally compose