n order that it may
continue the curve of the cusp. This modification is not to be given to
it in practice without considerable nicety of workmanship; and some
curious results took place in Venice from this difficulty.

At _l_ (Plate III.) is the shape of the Venetian side stone, with its
cusp detached from the arch. Nothing can possibly be better or more
graceful, or have the weight better disposed in order to cause it to nod
forwards against the keystone, as above explained, Ch. X. Sec. II., where
I developed the whole system of the arch from three pieces, in order that
the reader might now clearly see the use of the weight of the cusp.

Now a Venetian Gothic palace has usually at least three stories; with
perhaps ten or twelve windows in each story, and this on two or three of
its sides, requiring altogether some hundred to a hundred and fifty side
pieces.

I have no doubt, from observation of the way the windows are set
together, that the side pieces were carved in pairs, like hooks, of
which the keystones were to be the eyes; that these side pieces were
ordered by the architect in the gross, and were used by him sometimes
for wider, sometimes for narrower windows; bevelling the two ends as
required, fitting in keystones as he best could, and now and then
varying the arrangement by turning the side pieces _upside down_.

There were various conveniences in this way of working, one of the
principal being that the side pieces with their cusps were always cut to
their complete form, and that no part of the cusp was carried out into
the keystone, which followed the curve of the outer arch itself. The
ornaments of the cusp might thus be worked without any troublesome
reference to the rest of the arch.

Sec. XIX. Now let us take a pair of side pieces, made to order, like that
at _l_, and see what we can make of them. We will try to fit them first
with a keystone which continues the curve of the outer arch, as at _m_.
This the reader assuredly thinks an ugly arch. There are a great many of
them in Venice, the ugliest things there, and the Venetian builders
quickly began to feel them so. What could they do to better them? The
arch at _m_ has a central piece of the form _r_. Substitute for it a
piece of the form _s_, and we have the arch at _n_.

Sec. XX. This arch at _n_ is not so strong as that at _m_; but, built of
good marble, and with its pieces of proper thickness, it is quite strong
enough for all practical purpo