the
arches carrying the tower had a most marked effect on the
architectural expression of the interior. At Tewkesbury, for instance,
while the lower piers are designed in the usual way toward the north
and south sides (viz., as portions of a pier of nearly square
proportion standing under the angle of the tower), in the east and
west direction the tower piers run out into great solid masses of
wall, in order to insure a sufficient abutment for the tower arches.
On the north and south sides the solid transept walls were available
immediately on the other side of the low arch of the side aisle, but
on the east and west sides there were only the nave and choir arcades
to take the thrust of the north and south tower arches, and so the
Normans took care to interpose a massive piece of wall between, in
order that the thrust of the tower arches might be neutralized before
it could operate against the less solid arcaded portions of the walls.
This expedient, this great mass of wall introduced solely for
constructive reasons, adds greatly to the grandeur of the interior
architectural effect. The true constructive and architectural
perception of the Normans in this treatment of the lower piers is
illustrated by the curious contrast presented at Salisbury. There the
tower piers are rather small, the style is later, and the massive
building of the Normans had given way to a more graceful but less
monumental manner of building. Still the abutment of the tower arches
was probably sufficient for the weight of the tower as at first built;
but when the lofty spire was put on the top of this, its vertical
weight, pressing upon the tower arches and increasing their horizontal
thrust, actually thrust the nave and choir arcades out of the
perpendicular toward the west and east respectively, and there they
are leaning at a very perceptible angle away from the center of the
church--the architectural expression, in a very significant form, of
the neglect of balance of mass in construction.

But while the buttress in Gothic architecture has been in process of
development, what has the vault been doing? We left it (Fig. 92) in
the condition of a round wagon vault, intersected by another similar
vault at right angles. By that method of treatment we got rid of the
continuous thrust on the walls. But there were many difficulties to be
faced in the construction of vaulting after this first step had been
taken, difficulties which arose chiefly from