ECCLESIASTICAL
261
ECCLESIASTICAL
are found at Lyons (St-llartin d'Ainay), at Lesterps,
Civray, and Carcassonne (St-Nazaire) (Dehio and v.
Bezold, op. cit., PI. 122. figs. 3-6). An improvement
on this design, in view of the illumination of the nave,
consists in giving to the vaulting of the aisles the form
of a "rampant" arch, as at Silvacanne, and from this
it was but a step to the arrangement by which the sec-
tion took the form of a simple quadrant, as at Partlie-
nay-le-^'ieux, Preuilly, and Fontfroide. This method
of quadrant vaulting, as ^'iollet-le-Duc and others
have observed, provides a kind of continuous internal
"flying buttress", though it is by no means certain
that the idea of the flj'ing buttress in the Gothic archi-
tecture of Xorthern France was actually suggested by
these Southern buildings (Viollet-le-Duc, Diet., I,
173). In point of stability, the hall-churches of the
eleventh century leave nothing to be desired. Their
great defect is want of light (Viollet-le-Duc, Diet., I,
176). And this defect almost equally affects a class
of buildings which may be described as two-storied
hall-churches, and which are found principally, if not
exclusively, in Au\-ergne and its neighbourhood.
These are furnished, like a few of the Roman basilicas
and certain Byzantine churches, with a gallery, which
is not a mere triforium contrived in the thickness of the
walls, but a chamber of equal dimensions with the
aisle. This arrangement not only affords additional
space, but also, bj- reason of the greater height of the
edifice, might seem to facilitate the provision of a more
liberal supply of light, unimpeded by neighbouring
buildings. This last-mentioned advantage is, how-
ever, almost entirely negatived by the circumstance
that, in this class of buildings, each bay of the gallery
is subdivided by means of coupled or grouped arches,
so that the additional obstructions offered to the pas-
sage of the light almost entirely counterbalance the
possible gain through additional fenestration. We
say "the possible gain" because, in fact, the galleries
of these churches are Ijut sparingly provided with
windows. In these churches (which to the English
reader should be of special interest by reason of their
affinity in point of construction to the Westminster
cathedral) the aisle is usually cross-vaulted, while the
gallery has a quadrant vault abutting in the wall of
the nave just below the springing of the transverse
arches. The most noteworthy examples are found at
Clermont-Ferrand (Notre-Dame du Port"i, I.s.soire (St-
Paul), and Conques. To the same family belongs,
moreover, the great church of St-vSernin at Toulouse,
already mentioned, which is distinguished from those
previously named by having a double aisle. At Ne-
vers the church of St-Etienne resembles those at Cler-
mont, Issoire, and Conques, except that it is provided
with a range of upper windows which break through
the barrel- vaulting, somewhat after the fashion which
afterwards became so common in Italy in churches of
the Renaissance period.
The inherent shortcomings of the barrel-vault, es- pecially when used as a roof for the nave of an aisled church, have been .sufficiently illustrated. These dis- advantages, so far as structural stability and fenes- tration are concerned, might indeed be overcome by adopting the system of a succession of transverse bar- rel-vaults, such as are seen in the unique instance of the church of St-Philibert at Tournus. Such a con- struction is, however, " ponderous and inelegant, and never came into general use " (Moore, Gothic .\rchitec- ture, 42). The .system of cros.s-vaulting, which has now to be considered, may be regarded as a combina- tion of longitudinal with transverse barrel- vaulting, inasmuch as it may be described as consisting of a cen- tral barrel which is penetrated or intersected by a series of transverse vaults, corrcsjionding of course to the successive bays or compartments of the nave. The advantages of cross-vaulting are threefold. In the first place the total amount of the outward lateral thrust is very greatly diminished, since one-half of it is
now replaced by longitudinal thrusts, which, being op-
posed in pairs, neutralize one another. Secondly, all
that is left of the lateral thrust, as well as the longitu-
dinal thrusts, and the whole of the vertical pressure,
instead of being distributed throughout the whole
length of the building, is now collected and delivered
at definite points, namely the summits of the columns
or pillars. Thirdly and lastly, a perfectly developed
system of cross-vaulting makes it possible so to
heighten the clerestory windows that their archivolts
shall reach the utmost interior height of the buikling,
and so to broaden them that their width between re-
veals may approximate very closely to the interval be-
tween column and column below. By these improve-
ments (as ultimately realized in the perfected Gothic
of the thirteenth century) the somewhat rudimentary
design of the ancient Roman basilica may be said to
have reached the highest development of which it is
capable. The gradual development of cross-vaulting,
it is to be observed, did not take place in those dis-
tricts of Southern and Central France which had al-
ready become the home of the barrel-vault and to a
less degree of the cupola, but first in Lombardy, then
in Germany, and finally in Northern France and in
England. In these countries the evolution of the
Romanesque timber-roofed basQican church had —
with local variations of course — reached a far more
advanced stage than was ever attained in those regions
in which the adoption of barrel-vaulting at a relatively
early date had in a manner put a check on architec-
tural progress. And it is noteworthy that in Lom-
bardy and Germany, when cross-vaulting was first
adopted, its development was far less complete than
in Northern France, and that in like manner the ad-
vance towards perfection was both less rapid and less
complete in Normandy than in Picardy and the Ile-de-
France. These two districts were the last to adopt the
system, but it was here that it was, within the brief
space of less than fifty years (1170-1220), brought to
its final perfection. The reason may probably have
been, as Dehio and von Bezold suggest, that the archi-
tects of the Ile-de-France, in the days of Philip .\ugus-
tus and St. Louis, were less trammelled than those of
Normandy by the traditions of a school. The com-
parative lack of important architectural monuments
of an earlier date left them, say these writers, a more
open field for their inventive enterprise (op. cit., I,
■IIS).
The simplest form of cross- vaulting is of course that which is formed by the intersection of two cylindrical barrel-vaults of equal span. And this, without the use of ribbed groining, was the method mostly adopted by the Roman builders in their civic edifices. In the case of a pillared or columned church, however, this method had its disadvantages. In particular, having regard to the dimensions of the aisle and its vaulting, the builders of Northern Europe had all but univer- sally adopted the plan of so spacing tlie colunms and pillars which flank the nave that the intervals between them should be one-half the width of the church. Now the only means by which an equal height could be given to vaults of unequal span was the use of the pointed arch; and so it came about that the pointed arch was adopted, not primarily for aesthetic reasons, but rather for constructive purposes. And the same is to be said of the use of ribbed groining. The medie- val builders, who, as has been said above, possessed neither a tenacious mortar nor the command of an abundant supply of rough labour, and who therefore could not — even had they wished it — have adopted the niassi\-e concrete masonry of the Romans, were driven by the very necessities of the case to aim at lightness in the construction of their vaults, and at the same time to depend for stability not on the cohesion of the materials, but on the reduction of thrusts to a minimum, and on their skilful transmission to points where they could be effectively resisted. It was, then,
