s, so that the socketed
arrangement is not seen: it is shown as it would appear in a
longitudinal section. The balconies are not let into the circular
shafts, but fitted to their circular curves, so as to grasp them, and
riveted with metal; and the bars of stone which form the tops of the
balconies are of great strength and depth, the small trefoiled arches
being cut out of them as in Fig. III., so as hardly to diminish their
binding power. In the lighter independent balconies they are often cut
deeper; but in all cases the bar of stone is nearly independent of the
small shafts placed beneath it, and would stand firm though these were
removed, as at _a_, Fig. II., supported either by the main shafts of
the traceries, or by its own small pilasters with semi-shafts at their
sides, of the plan _d_, Fig. II., in a continuous balcony, and _e_ at
the angle of one.

There is one more very curious circumstance illustrative of the Venetian
desire to obtain horizontal pressure. In all the Gothic staircases with
which I am acquainted, out of Venice, in which vertical shafts are used
to support an inclined line, those shafts are connected by arches rising
each above the other, with a little bracket above the capitals, on the
side where it is necessary to raise the arch; or else, though less
gracefully, with a longer curve to the lowest side of the arch.

But the Venetians seem to have had a morbid horror of arches which were
not _on a level_. They could not endure the appearance of the roof of
one arch bearing against the side of another; and rather than introduce
the idea of obliquity into bearing curves, they abandoned the arch
principle altogether; so that even in their richest Gothic staircases,
where trefoiled arches, exquisitely decorated, are used on the landings,
they ran the shafts on the sloping stair simply into the bar of stone
above them, and used the excessively ugly and valueless arrangement of
Fig. II., rather than sacrifice the sacred horizontality of their arch
system.

[Illustration: Fig. IV.]

It will be noted, in Plate XI., that the form and character of the
tracery bars themselves are independent of the position or projection of
the cusps on their flat sides. In this respect, also, Venetian traceries
are peculiar, the example 22 of the Porta della Carta being the only one
in the plate which is subordinated according to the Northern system. In
every other case the form of the aperture is determined, either