two
of them it is composed merely of three concentric lines, parallel with
the sides of the triangle; in the third, it is a wreath of beautiful
design, which I have drawn of larger size in fig. 2, Plate V., that the
reader may see how completely the surface is left undestroyed by the
delicate incisions of the chisel, and may compare the method of working
with that employed on the white stones, two of which are given in that
plate, figs. 4 and 5. The keystone, of which we have not yet spoken, is
the only white stone worked with the light incision; its design not
being capable of the kind of workmanship given to the floral ornaments,
and requiring either to be carved in complete relief, or left as we see
it. It is given at fig. 1 of Plate IV. The sun and moon on each side of
the cross are, as we shall see in the fifth Chapter, constantly employed
on the keystones of Byzantine arches.

Sec. XXIX. We must not pass without notice the grey and green pieces of
marble inserted at the flanks of the arch. For, observe, there was a
difficulty in getting the forms of the triangle into anything like
reconciliation at this point, and a mediaeval artist always delights in a
difficulty: instead of concealing it, he boasts of it; and just as we
saw above that he directed the eye to the difficulty of filling the
expanded sides of the upper band by elongating his triangles, so here,
having to put in a piece of stone of awkward shape, he makes that very
stone the most conspicuous in the whole arch, on both sides, by using in
one case a dark, cold grey; in the other a vigorous green, opposed to
the warm red and purple and white of the stones above and beside it. The
green and white piece on the right is of a marble, as far as I know,
exceedingly rare. I at first thought the white fragments were inlaid, so
sharply are they defined upon their ground. They are indeed inlaid, but
I believe it is by nature; and that the stone is a calcareous breccia of
great mineralogical interest. The white spots are of singular value in
giving piquancy to the whole range of more delicate transitional hues
above. The effect of the whole is, however, generally injured by the
loss of the three large triangles above. I have no doubt they were
purple, like those which remain, and that the whole arch was thus one
zone of white, relieved on a purple ground, encircled by the scarlet
cornices of brick, and the whole chord of color contrasted by the two
precious fragmen