string courses of the walls form the capitals of the
shafts of the traceries; and the abaci of the vaulting shafts of the
apse.

[Illustration: Fig. XXVIII.]

Sec. XXV. We have hitherto spoken of capitals of circular shafts only:
those of square piers are more frequently formed by the cornice only;
otherwise they are like those of circular piers, without the difficulty
of reconciling the base of the bell with its head.

Sec. XXVI. When two or more shafts are grouped together, their capitals
are usually treated as separate, until they come into actual contact. If
there be any awkwardness in the junction, it is concealed by the
decoration, and one abacus serves, in most cases, for all. The double
group, Fig. XXVII., is the simplest possible type of the arrangement. In
the richer Northern Gothic groups of eighteen or twenty shafts cluster
together, and sometimes the smaller shafts crouch under the capitals of
the larger, and hide their heads in the crannies, with small nominal
abaci of their own, while the larger shafts carry the serviceable abacus
of the whole pier, as in the nave of Rouen. There is, however, evident
sacrifice of sound principle in this system, the smaller abaci being of
no use. They are the exact contrary of the rude early abacus at Milan,
given in Plate XVII. There one poor abacus stretched itself out to do
all the work: here there are idle abaci getting up into corners and
doing none.

Sec. XXVII. Finally, we have considered the capital hitherto entirely as
an expansion of the bearing power of the shaft, supposing the shaft
composed of a single stone. But, evidently, the capital has a function,
if possible, yet more important, when the shaft is composed of small
masonry. It enables all that masonry to act together, and to receive the
pressure from above collectively and with a single strength. And thus,
considered merely as a large stone set on the top of the shaft, it is a
feature of the highest architectural importance, irrespective of its
expansion, which indeed is, in some very noble capitals, exceedingly
small. And thus every large stone set at any important point to
reassemble the force of smaller masonry and prepare it for the
sustaining of weight, is a capital or "head" stone (the true meaning of
the word) whether it project or not. Thus at 6, in Plate IV., the stones
which support the thrust of the brickwork are capitals, which have no
projection at all; and the large stones in the window