ernally only. But there is not the smallest excuse for carving the
soffit, and not the outside; for, in that case, we approach the building
under the idea of its being perfectly plain; we do not look for the
soffit decoration, and, of course, do not see it: or, if we do, it is
merely to regret that it should not be in a better place. In the
Renaissance architects, it may, perhaps, for once, be considered a
merit, that they put their bad decoration systematically in the places
where we should least expect it, and can seldomest see it:--Approaching
the Scuola di San Rocco, you probably will regret the extreme plainness
and barrenness of the window traceries; but, if you will go very close
to the wall beneath the windows, you may, on sunny days, discover a
quantity of panel decorations which the ingenious architect has
concealed under the soffits.

The custom of decorating the arch soffit with panelling is a Roman
application of the Greek roof ornament, which, whatever its intrinsic
merit (compare Chap. XXIX. Sec. IV.), may rationally be applied to waggon
vaults, as of St. Peter's, and to arch soffits under which one walks.
But the Renaissance architects had not wit enough to reflect that people
usually do not walk through windows.

Sec. IX. So far, then, of the Southern archivolt: In Fig. LXIX., above,
it will be remembered that _c_ represents the simplest form of the
Northern. In the farther development of this, which we have next to
consider, the voussoirs, in consequence of their own negligence or
over-confidence, sustain a total and irrecoverable defeat. That
archivolt is in its earliest conditions perfectly pure and
undecorated,--the simplest and rudest of Gothic forms. Necessarily, when
it falls on the pier, and meets that of the opposite arch, the entire
section of masonry is in the shape of a cross, and is carried by the
crosslet shaft, which we above stated to be distinctive of Northern
design. I am more at a loss to account for the sudden and fixed
development of this type of archivolt than for any other architectural
transition with which I am acquainted. But there it is, pure and firmly
established, as early as the building of St. Michele of Pavia; and we
have thenceforward only to observe what comes of it.

Sec. X. We find it first, as I said, perfectly barren; cornice and
architrave altogether ignored, the existence of such things practically
denied, and a plain, deep-cut recess with a single mighty shadow