being licentious, though in early work
often very lovely), it is literally impossible, without consulting the
context of the building, to say whether the cusps have been added for
the sake of beauty or of strength; nor, though in mediaeval architecture
they were, I believe, assuredly first employed in mere love of their
picturesque form, am I absolutely certain that their earliest invention
was not a structural effort. For the earliest cusps with which I am
acquainted are those used in the vaults of the great galleries of the
Serapeum, discovered in 1850 by M. Maniette at Memphis, and described by
Colonel Hamilton in a paper read in February last before the Royal
Society of Literature.[72] The roofs of its galleries were admirably
shown in Colonel Hamilton's drawings made to scale upon the spot, and
their profile is a cusped round arch, perfectly pure and simple; but
whether thrown into this form for the sake of strength or of grace, I am
unable to say.

Sec. XCVI. It is evident, however, that the structural advantage of the
cusp is available only in the case of arches on a comparatively small
scale. If the arch becomes very large, the projections under the flanks
must become too ponderous to be secure; the suspended weight of stone
would be liable to break off, and such arches are therefore never
constructed with heavy cusps, but rendered secure by general mass of
masonry; and what additional _appearance_ of support may be thought
necessary (sometimes a considerable degree of _actual_ support) is given
by means of tracery.

[Illustration: Fig. XVIII.]

Sec. XCVII. Of what I stated in the second chapter of the "Seven Lamps"
respecting the nature of tracery, I need repeat here only this much,
that it began in the use of penetrations through the stone-work of
windows or walls, cut into forms which looked like stars when seen from
within, and like leaves when seen from without: the name foil or feuille
being universally applied to the separate lobes of their extremities,
and the pleasure received from them being the same as that which we feel
in the triple, quadruple, or other radiated leaves of vegetation, joined
with the perception of a severely geometrical order and symmetry. A few
of the most common forms are represented, unconfused by exterior
mouldings, in Fig. XVIII., and the best traceries are nothing more than
close clusters of such forms, with mouldings following their outlines.

Sec. XCVIII. The term "folia