of the ornament be, it must clearly have
relief of some kind, and must present projecting surfaces separated by
incisions. But it is a very material question whether the contour,
hitherto broadly considered as that of the entire bell, shall be that of
the _outside_ of the projecting and relieved ornaments, or of the
_bottoms of the incisions_ which divide them; whether, that is to say,
we shall first cut out the bell of our capital quite smooth, and then
cut farther into it, with incisions, which shall leave ornamental forms
in relief, or whether, in originally cutting the contour of the bell, we
shall leave projecting bits of stone, which we may afterwards work into
the relieved ornament.

Sec. XXXVII. Now, look back to Fig. V., p. 65. Clearly, if to ornament the
already hollowed profile, _b_, we cut deep incisions into it, we shall
so far weaken it at the top, that it will nearly lose all its supporting
power. Clearly, also, if to ornament the already bulging profile _c_ we
were to leave projecting pieces of stone outside of it, we should nearly
destroy all its relation to the original sloping line X, and produce an
unseemly and ponderous mass, hardly recognizable as a cornice profile.
It is evident, on the other hand, that we can afford to cut into this
profile without fear of destroying its strength, and that we can afford
to leave projections outside of the other, without fear of destroying
its lightness. Such is, accordingly, the natural disposition of the
sculpture, and the two great families of capitals are therefore
distinguished, not merely by their concave and convex contours, but by
the ornamentation being left outside the bell of the one, and cut into
the bell of the other; so that, in either case, the ornamental portions
will fall _between the dotted lines_ at _e_, Fig. V., and the pointed
oval, or vesica piscis, which is traced by them, may be called the Limit
of ornamentation.

Sec. XXXVIII. Several distinctions in the quantity and style of the
ornament must instantly follow from this great distinction in its
position. First, in its quantity. For, observe: since in the Doric
profile, _c_ of Fig. V., the contour itself is to be composed of the
surface of the ornamentation, this ornamentation must be close and
united enough to form, or at least suggest, a continuous surface; it
must, therefore, be rich in quantity and close in aggregation; otherwise
it will destroy the massy character of the profile it a