f a tree cannot be
so treated.

The awning of the classic hypaethral hall or court was often reproduced
in Roman arabesques. Sometimes we find it in a classical tomb, painted
over the ceiling, and recalling its original use. This was revived in
the Cinque-cento Renaissance; and again in Adams' "Eighteenth Century
Decorations," it became an accepted pattern, called "the shell,"
losing its original motive, and descending to fill up the panels of
tea-caddies and surround keyholes. When thus reduced to the appearance
of a little ruff, it needs some thought to recognize it, and give it
credit for its first motive.

[Illustration: Fig. 16.]

It is amusing to find how a form which it seems impossible to reduce
to a pattern, will yet fall into one by a judicious arrangement of
light and shadow, and by repetition. There is a little frieze in one
of the Indian cases on the staircase in the British Museum, which is
extremely pretty and effective. It consists of a repetition of little
balconies with recesses and pillars and figures in pairs. I give it as
illustrating the way conventional patterns grow. This balcony pattern
is of the sixth century, A.D.

[Illustration: Fig. 17.
Indian Balcony Pattern, from steps of tope of Jamal-Zartri,
Afghanistan. British Museum.]

The ancient palmated pattern called Chrysoclavus, from the beginning
of our era to the thirteenth century was partly a nail-headed design,
and had become a Christian symbol. It was, probably, originally the
primitive spot pattern; afterwards promoted to being an ornament of
discs in colour or metal: this was Assyrian, Etruscan, and
Mycenaean.[128] (Pl. 70.)

Among the conventional patterns which have apparently no hidden
meaning, but which clearly show their descent, are the Chinese and
Japanese wicker and lattice-work designs. The beauty of these is
wonderful.

Semper shows that wicker (including bamboo work) was the foundation of
all Chinese civilized life, for constructing houses, bridges,
utensils, and for decoration. He gives this wicker-work origin to the
universal key pattern, which may, however, have a double source--the
wave, and the wicker-work.

We find the Key pattern in a tomb at Essiout, in Egypt, painted
perhaps about 1600 B.C., in company with some other very old
friends,[129] the Tuscan border, the Egg and Tongue, and the Bead, the
Daisy, and the Wave. (Pl. 17, No. 2.) We meet it everywhere in ancient
and modern decoration.