CHAPTER X

.

CONTENTS.

Gothic details—Pillars—Windows—Circular windows—Bays—Vaults—Buttresses— Pinnacles—Spires—Decoration—Construction—Furniture of churches— Domestic architecture.

ALTHOUGH in the preceding pages, in describing the principal churches of France, mention has been made of the various changes of detail which took place from the time of the introduction of the pointed style till its abandonment in favour of the revived classical, still it seems necessary to recapitulate the leading changes that were introduced. This will be most fitly done before we leave the subject of French architecture, that being on the whole the most complete and harmonious of all the pointed styles, as well as the earliest.

PILLARS.

Of these details, the first that arrests the attention of the inquirer is the form of the pillars or piers used in the Middle Ages, inasmuch as it is the feature that bears the most immediate resemblance to the typical forms of preceding styles. Indeed, the earlier pillars in the round-arched style were virtually rude imitations of Roman originals, made so thick and heavy as to bear without apparent stress the whole weight of the arches they supported, and of the superincumbent wall. This increase of the weight laid upon the pillars, and consequently in their strength and heaviness, was the great change introduced into the art of building in the early round Gothic style. With the same requirements the classic architects either must have thickened their pillars immensely, or coupled them in some way. Indeed the Romans, in such buildings as the Colosseum, placed the pillars in front and a pier behind, which last was the virtual support of the wall. The Gothic architects improved on this by adding a pillar, or rather a half pillar, on each side, to receive the pier arches, and carrying up those behind and in front to support the springing of the vault or roof, instead of the useless entablature of the Romans.

By this means the pier became in plan what is represented in figs. 1 and 2 in the diagram (Woodcut No. 648). Sometimes it was varied, as represented in fig. 3, where the angle-shafts were only used to lighten the apparent heaviness of the central mass; in other examples both these modes are combined, as in fig. 4, which not only constructively, but artistically, is one of the most beautiful combinations which the square forms are capable of, combining great strength with great lightness of appearance, and variety of light and shade.

These four forms may be said to be typical in the South, where the style was derived so directly from the Roman square pier combined with an attached circular pillar.

In the North the Normans, and generally speaking, all the Frankish tribes, used the circular pillar in preference to the square pier, and consequently the variations were as shown in figs. 5, 6, 7, and 8, which, though forming beautiful combinations, wanted the accentuation produced by the contrast between the square and round forms.

[Illustration: 648. Diagram of Plans of Pillars.]

The architects after a time seemed to have felt this, and tried to remedy it by introducing ogee forms and sharp edges, with deep undercut shadows, thus applying to the pillars those forms which had been invented for the mouldings of the ribs of the vaults, and for the tracery of the windows. The expedient was perfectly successful at first, and, so long as it was practised in moderation, gave rise to some of the most beautiful forms of pillars to be found in any style. It proved, however, too tempting an opportunity for the indulgence of every sort of quirk and quibble; and after passing through the shapes shown in figs. 9 and 10, where the meaning of all the parts is still sufficiently manifest, it became as complicated as fig. 11, and sometimes even more cut up, so that all meaning and beauty was lost. It became moreover very expensive and difficult to execute, so that in later times the architects reverted, either to circular pillars, or to such a form as that shown in fig. 12, which was introduced in the 16th century. The change may have been partly introduced from motives of economy, and also to some extent from a desire to imitate the flutings of classical pillars; but from whatever motive it arose, it is singularly unmeaning and inartistic; and as the capital was at the same time omitted, the whole pillar took an appearance of cold poverty entirely at variance with the true spirit of Gothic art. This last change showed, perhaps more clearly than those introduced into any other feature, how entirely the art had died away before the classical styles superseded it.

WINDOWS.

Before painted glass came into use, very small apertures sufficed to admit the required quantity of light into the churches. These openings retained their circular-arched heads long after the pointed form pervaded the vaults and pier arches, because the architects still thought them the most beautiful; they moreover occupied so small a portion of the wall spaces that their lines neither came in contact nor interfered with the constructive lines of the building itself; but when it was required to enlarge them for the purpose of receiving large pictures, the retention of the circular form was no longer practicable.

[Illustration: 649. Window, St. Martin, Paris. (From ‘Paris Archéologique.’)]

[Illustration: 650. Window of Nave of Cathedral at Chartres.]

[Illustration: 651. Window in Choir of Cathedral at Chartres.]

The Woodcut No. 622, showing the side elevation of Notre Dame at Paris, illustrates well three stages of this process as practised in the 12th and 13th centuries. It exhibits first the large undivided window without mullions, the glass being supported by strong iron bars; next, that with one mullion and a circular rose in the head; and lastly, in the lower storey, a complete traceried window. The transition from the old small window to the first of these is easily explained, and the Woodcut No. 649, representing one of the windows in St. Martin at Paris, will explain the transition from the first to the second. Instead of one large undivided opening, it was often thought more expedient to introduce two lancets side by side; but as these never filled, nor could fill, the space of one bay so as to follow its principal lines, it became usual to introduce a circular window of greater or less size between their heads. This, with the rude construction of the age, presented certain difficulties which were obviated by carrying the masonry of the vault through the wall so as to form a discharging arch. When once this was done it required only a glance from an experienced builder to see that if the discharging arch were strong enough, the whole of the wall between the buttresses might be removed without endangering the safety of the building. This was accordingly soon done. The pier between the two lancets became attenuated into a mullion, the circle lost its independence, and was grouped with them under the discharging arch, which was carried down each side in boldly splayed jambs, and the whole became in fact a traceried window.

[Illustration: 652. Window at Rheims.]

[Illustration: 653. Window at St. Ouen.]

In the cathedral at Chartres we have examples of the two extremes of these transitional windows. In the windows of the aisles of the nave (Woodcut No. 650) the circle is small and insignificant, and only serves to join together the two lancets. In the clerestory (Woodcut No. 651), which is somewhat later, the circle is all important and quite overpowers the lower part. Here it is in fact a circular window, supported by a rectilinear substructure. In both these instances the discharging arch still retains its circular form, and the tracery is still imperfect, inasmuch as all the openings are only holes of various forms cut into a flat surface, whereas to make it perfect, it is necessary that the lines of two contiguous openings should blend together, being separated by a straight or curved moulded mullion, and not merely pierced as they are in this instance. This may perhaps be better illustrated by one of the windows of the side-aisles at Rheims, where the pointed Gothic window has become complete in all its essential parts. Even here it will be observed how awkwardly the circle fits into the spherical triangle of the upper part of the window. Indeed, there is an insuperable awkwardness in the small triangles necessarily left in fitting circles into the spaces above the lancets, and beneath the pointed head of the openings. When four or five lights were used instead of two, this defect became more apparent; and even in the example from St. Ouen (Woodcut No. 653). one of the most beautiful in France, the architect has not been able to obviate the discordance between the conflicting lines of the circle and spherical triangle. At last, after two centuries of earnest trial, the builders of those days found themselves constrained to abandon entirely these beautiful constructive geometric forms, for tracery of a more manageable nature, and in place of the circle they invented first a flowing tracery, of which the window at Chartres (Woodcut No. 654) is an exquisite example; and then having shaken off the trammels of constructive form, launched at once into all the vagaries of the flamboyant style. In this style stone tracery was made to look bent and twisted, as willow wands. Its forms, it must be confessed, were always graceful, but constructively weak, and frequently extravagant, showing a complete contrast to the contemporary perpendicular style followed in England. That failed from the stiffness of its forms; this from the fantastic pliancy with which so rigid a material as stone was used. Greatness or grandeur was as impossible in flamboyant tracery, as grace and beauty were with the perpendicular style; still for domestic edifices, and for the smaller churches erected in the 16th century, it must be confessed the flamboyant style has a charm it is impossible to resist. It is so graceful and so fantastically brilliant, that it captivates in spite of our soberer reason, lending as it does an elegance to every edifice where it is found, and finding its parallel alone among the graceful fancies of the Saracenic architects of the best age.

[Illustration: 654. Window at Chartres.]

CIRCULAR WINDOWS.

By far the most brilliant examples of this class in France are to be found among the great circular windows with which the west ends and transepts of the cathedrals were adorned. There is, I believe, no instance in France of the great straight-mullioned windows of which our architects were so fond, and even where the east end terminates squarely, as at Laon, it has a great rose window. There can be little doubt that the circle, so long as it was wholly adhered to, was the noblest form architecturally, both externally and internally; but when the triforium below it was pierced, and the lower angles outside the circle were filled with tracery, making it into something like our great windows, the result was a confusion of the two modes, in which the advantages of neither were preserved.

Of the earlier circular windows, one of the finest is that in the western front at Chartres (Woodcut No. 655), of imperfect tracery, like the greater part of that cathedral, but of great size and majesty. Its diameter is 39 ft. across the openings, and 44 ft. 6 in. across to the outer mouldings of the circle. Those of the transepts are smaller, being only 33 ft. across the opening, but show a considerable advance in the art of tracery, which by the time they were executed was becoming far better understood.

[Illustration: 655. West Window, Chartres.]

[Illustration: 656. Transept Window, Chartres.]

[Illustration: 657. West Window, Rheims.]

[Illustration: 658. West Window, Evreux.]

If space admitted, it would be easy to select examples to trace the progress of the invention between these early efforts and the almost perfect window that adorns the centre of the west front at Rheims (Woodcut No. 657); and again from this to that at Evreux (Woodcut No. 658). In the latter instance, the geometric forms have given way to the lace-work of flowing tracery, of which this is a pleasing example. It is further remarkable in respect that all the parts of the tracery or mullions are of the same thickness, whereas it is usual in flowing or flamboyant tracery to introduce a considerable degree of subordination into the parts, dividing them into greater or smaller ribs, thus avoiding confusion and giving to the whole a constructive appearance which it otherwise would not possess. This is very apparent in such a window as that which adorns the west front of St. Ouen, at Rouen, where the parts are distinctly subordinated to one another, and have consequently that strength and character which it is so difficult to impart. It also exemplifies what was before alluded to, viz., the mode in which the lower external angles of the circle were filled up, and also, in a far more pleasing manner than usual, the mode in which the pierced triforium is made to form part of the decoration. Owing to the strong transom bar here employed, there is strength enough to support the superstructure; but as too often is the case, when this is subdued and kept under, there is a confusion between the circular and upright parts, which is not pleasing. It is then neither a circular nor an upright window, but an indeterminate compound of two pleasing members, in which both suffer materially by juxtaposition.

[Illustration: 659. West Window, St. Ouen. (From Pugin.)]

I believe it is safe to assert, that out of at least a hundred first-class examples of these circular windows, which still exist in France, no two are alike. On the contrary, they present the most striking dissimilarity of design. There is no feature on which the French architects bestowed more pains, or in which they were more successful. They are, indeed, the _chefs-d’œuvre_ of their decorative abilities, and the most pleasing individual features of their greater churches. At the same time, they completely refute the idea that the pointed form is at all necessary for the production of beauty in decorative apertures.

BAYS.

It may be useful here to recapitulate what has been said of the subdivision of churches into bays, or, as the French call them, _travées_. The two typical arrangements of these are shown in Woodcuts Nos. 616 and 617, as existing before the introduction of the pointed forms. In the first a great gallery runs over the whole of the side aisle, introduced partly as a constructive expedient to serve the purpose for which flying buttresses were afterwards employed, partly as enabling the architect to obtain the required elevation without extraordinarily tall pillars or wide pier-spaces, both which were beyond the constructive powers of the earlier builders. These galleries were also useful as adding to the accommodation of the church, as people were able thence to see the ceremonies performed below, and to hear the mass and music as well as from the floor of the church. These advantages were counterbalanced by the greater dignity and architectural beauty of the second arrangement (Woodcut No. 617) where the whole height was divided into that of the side-aisles and of a clerestory, separated from one another by a triforium gallery, which represented in fact the depth of the wooden roof requisite to cover the side-aisles. When once this simple and beautiful arrangement was adopted, it continued with very little variation throughout the Middle Ages.[358] The proportions generally used were to make the aisles half the height of the nave. In other words, the string-course below the triforium divided the height into two equal parts; the space above that was divided into three, of which two were allotted to the clerestory, and one to the triforium.[359] It is true there is perhaps no single instance in which the proportions here given are exactly preserved, but they sufficiently represent the general division of the parts, from which the architects only deviated slightly, sometimes on one side, sometimes on the other, according to their taste or caprice. The only really important change afterwards introduced was that of glazing the triforium gallery also, by adopting a flat roof, or one nearly so, over the side-aisles, as the nave in the church of St. Ouen at Rouen, or by covering each bay by a pyramidal roof not seen from the interior, as is shown in the Woodcuts Nos. 621 and 641; the whole walls of the church, with the slight exception of the spandrils of the great pier-arches, having thus become walls of glass, the mass of the vault being supported only by the deep and bold constructive lines of which the framework of the glazed surfaces consists.

In England, we have not, as far as I am aware, any instance of a glazed triforium, but it is one of the most fascinating features in the later styles of the French architects, and where it retains its coloured glass, which is indispensable, produces the most fairy-like effect. It is however, questionable whether the deep shadow and constructive propriety of the English practice is not on the whole more satisfactory. In a structure of glass and iron nothing could be more appropriate than the French practice; but in a building of stone and wood more solidity is required to produce an effect which shall be permanently pleasing.

VAULTS.

It has already been explained how essential a part of a Gothic church the vault was, and how completely it was the governing power that gave form to the art. We have also seen the various steps by which the architects arrived at the intersecting vault, which became the typical form in the best age. In France especially the stone vault was retained throughout as a really essential feature, for though the English were so successful in the art of constructing ornamental wooden roofs, the practice never prevailed in France.

[Illustration: 660. Diagram of Vaulting.]

In the best age the arrangement of the French vaults was extremely simple. The aisles were generally built in square compartments, the vaults of which were first circumscribed, each by four equal arches (Woodcut No. 660), of which A A were transverse ribs or _arcs doubleaux_ as the French called them, and were used, as we have seen, in the old tunnel-vaults. These arches, as springing from the main points of support, were the principal strengtheners of the vault, and served as permanent centres for the superstructure. B was called the _formeret_, and was a rib built into the wall, of the same form as the transverse ribs, and so called because, being the first constructed, it gave the form to the vault. Lastly, there were two more ribs springing from angle to angle, and intersecting one another at C. These were called _ogives_, from the Latin word _augere_, to strengthen,[360] the chief object of their employment however being to serve as centering. In Roman vaulting similar ribs were employed, but the spaces between were subsequently filled in flush with concrete. In Renaissance and in modern work (such as in cellar or dock-vaults, for instance), when built in brick, stone voussoirs are used for the groins, because the brickwork used there would be liable to be crushed or fall out; here also the stone is flush with the brickwork, but the Mediæval architects recognised the value of the rib, not only as a permanent centre, but as suggesting the appearance as well as the reality of strength.

The roof of the nave was composed of precisely the same parts, only that, being twice as wide as each compartment was broad, the length of the transverse ribs and of the intersecting ogives was greater in proportion to the formerets than in the aisles. Another addition, and certainly an improvement, was the introduction of ridge-ribs (D D), marking the point of the vault. These could not of course be used with circular arches, where there was no centre line for them to mark; and it probably was from this cause that the French seldom adopted them, having been accustomed to vaults not requiring them. Another reason was that all their earlier vaults were more or less domical, or in other words the point C was higher than the points A or B, though this is more apparent in hexapartite vaults, or where one compartment of the nave-vaults takes in two of the aisles, than in quadripartite, like those now under consideration. Still all French vaults have this peculiarity more or less, and consequently the longitudinal ridge-rib, where used, has an up and down broken appearance, which is extremely disagreeable, and must in a great measure have prevented its adoption. There is, however, at least one exception to this rule in France, in the abbey church of Souvigny, represented in the Woodcut No. 661, where this rib is used with so pleasing an effect that one is surprised it was not in more general favour.

[Illustration: 661. Abbey Church, Souvigny. (From ‘L’Ancien Bourbonnais.’)]

These are the only features usually employed by French architects: but we do sometimes find tiercerons, or secondary ogives, used to strengthen as well as to ornament the plain faces of the vaults, one or two on each face, as at E E (in Woodcut No. 660); small ribs or _liernes_, F F, from _lier_, to bind, were also occasionally used to connect all these at the centre, where they formed star patterns, and other complicated but beautiful ornaments of the vault. These last, however, are rare and exceptional in French vaulting, though they were treated by the English architects with such success that we wonder they were not more generally adopted in France. The most probable explanation appears to be that the French architects depended more on colour than on relief for the effect of their vaults, while in England colour was sparingly used, its place being supplied by constructive carving. Whatever may have been the comparative merits of the two methods when first used, the English vaults have a great advantage now, inasmuch as the carving remains, while the paintings of the others have perished, and we have no means left of judging of their original effect.

One of the most beautiful features of French vaulting, almost entirely unknown in this country, is the great polygonal vault of the semi-dome of the chevet, which as an architectural object few will be disinclined to admit is, with its walls of painted glass and its light constructive roof, a far more beautiful thing than the plain semi-dome of the basilican apse, notwithstanding its mosaics. Still, as the French used it, they never quite surmounted the difficulties of its construction; and in their excessive desire to do away with all solid wall, and to get the greatest possible surface for painted glass, they often distorted these vaults in a very unpleasing manner.

The chevet of Pontigny (Woodcut No. 643) presents a good example of the early form of vault, which owing to the small size of the windows and general sobriety of the composition, avoids the defects above alluded to. Of the later examples there are few, except that of Souvigny, represented in Woodcut No. 661, where the difficulty has been entirely conquered by constructing the spandrils with pierced tracery, so that the vault virtually springs from nearly the same height as the arch of the windows, and a very slight improvement would have made this not only constructively, but artistically perfect. This is a solitary specimen, and one which, though among the most beautiful suggestions of Gothic art, has found no admirers, or at least no imitators.

Notwithstanding this difficulty of construction, these pierced semi-domes are not only the best specimens of French vaulting, but are among the most beautiful inventions of the Middle Ages, and form a finer termination to the cathedral vista than either the great windows of the English, or the wonderful rose windows of the French cathedrals.

BUTTRESSES.

The employment of buttresses was a constructive expedient that followed almost indispensably on the use of vaults for the roofing of churches. It was necessary either to employ enormously thick walls to resist the thrust, or to support them by some more scientific arrangement of the materials. The theory of the buttress will be easily understood from the diagram (Woodcut No. 662), representing seven blocks or masses of masonry, disposed first so as to form a continuous wall, but which evidently affords very little resistance to a thrust or push tending to overturn it from within. The left-hand arrangement is, from the additional breadth of base in the direction of the thrust, much less liable to fall outwards, provided the distance of the blocks from one another is not too great, and the mass of the vault does not press heavily on the intermediate space. This last difficulty was so much felt by the earlier French architects that, as we have seen, in the South of France especially, they used the roof of the side-aisle as a continuous buttress to resist the thrust of their tunnel-vaults. It was surmounted also by the introduction of intersecting vaults, inasmuch as by this expedient all the thrusts were collected together at a point over each pier, and a resisting mass applied on that one point was sufficient to give all the stability required. This, and the desire of raising the lights as high as possible into the roof, were the principal causes that brought this form of vaulting into general use; still it has not yet been shown that the continuous vault is not artistically the more beautiful of the two forms, if not constructively so also.

[Illustration: 662. Diagram of Buttresses.]

There was yet another difficulty to be mastered, which was that the principal vault to be abutted was that over the nave or central part of the church, and buttresses of the requisite depth would have filled up the side-aisles entirely. The difficulty first presented itself in the building of the basilica of Maxentius (Woodcut No. 203), and was there got over in something like the manner practically adopted in the Middle Ages, except that the arch was there carried inside, whereas the Gothic architects threw the abutting arch across on the outside and above the roof.

[Illustration: 663. Flying Buttress of St. Ouen. (From Batissier, ‘Histoire de l’Art.’)]

Several of the previous woodcuts[361] show the system of flying buttresses in various stages of advancement. The view of one of those of the choir of St. Ouen (No. 663) exhibits the system in its greatest degree of development. Here there are two vertical and two flying buttresses, forming a system of great lightness, but at the same time of immense constructive strength, and when used sparingly and with elegance, as in this instance, constituting an object of great beauty. The abuse of this expedient, as in the cathedral at Cologne and elsewhere, went very far to mar the proper effect.

The cathedral at Chartres presents a singular but very beautiful instance of an earlier form of flying buttress: there the immense span of the central vault put the architects on their mettle to provide a sufficient abutment, and they did it by building what was literally an open wall across the aisle (see Woodcut No. 628), strongly arched, and the arches connected by short strong pillars radiating with the voussoirs of the arch. Nothing could well be stronger and more scientific than this, but the absence of perpendicularity in the pillars was unpleasing to the eye then as now, and the contrivance was never repeated.

[Illustration: 664. Flying Buttress at Amiens. (From Chapuy.)]

A far more pleasing form was that adopted afterwards at Amiens (Woodcut No. 664) and elsewhere, where a series of small traceried arches stand on the lower flying buttress, and support the upper, which is straight-lined. Even here, however, the difficulty is not quite got over; the unequal height of these connecting arches, and the awkward angle which the lower supports make with the curvilinear form on which they rest, deprive them of that constructive propriety which alone secures a perfectly satisfactory result in architecture. The problem indeed is one which the French never thoroughly solved, though they bestowed immense pains upon it. Brilliant as the effect sometimes is of the immense mass of pinnacles and flying buttresses, they are seldom so put together as to leave an entirely satisfactory result on the mind of the spectator. Taken all in all, perhaps the most pleasing example is that of Rheims (Woodcut No. 629)—those on each side of the nave especially—where two bold simple arches transmit the pressure from a bold exquisitely pinnacled buttress to the sides of the clerestory, and in such a manner as to leave no doubt whatever either as to their purpose or their sufficiency to accomplish their object.

Notwithstanding the beauty which the French attained in their flying buttresses, it is still a question whether they did not carry this feature too far. It must be confessed that there is a tendency in the abuse of the system to confuse the outlines and to injure the true architectural effect of the exterior. Internally it no doubt enabled them to lighten their piers and increase the size of their windows to an unlimited extent, and to judge fairly we must balance between the gain to the interior, and the external disadvantages. This we shall be better able to do when considering the next constructive expedient, which was that of the introduction of pinnacles.

PINNACLES.

The use of pinnacles, considered independently of their ornamental purposes, is evident enough. It is obvious that a wall or pillar which has to resist the thrust of a vault or any other power exerted laterally, depends for its stability on its thickness, its solidity, and generally on its lateral strength. A material consideration, as affecting this solidity, is that of weight. The most frequent use of pinnacles by the French was to surmount the piers from which the flying buttresses sprang. To these piers weight and solidity were thus imparted, rendering them a sufficiently steady abutment to the flying arches, which in their turn abutted the central vaults.

It must be understood that these expedients of buttresses and pinnacles were only employed to support the central roof of the nave. The vaults of the aisles were so narrow as not to require any elaborate system of abutments for their support—the ordinary thickness of the walls would have sufficed for that purpose; but they also had the advantage of the use of the supports designed for the larger vaults.

As a general rule the English architects never hesitated to weight their walls so as to apply the resistance directly on the point required, and not only adorned the roofs of their churches with pinnacles, but raised towers and lanterns on the intersections on all occasions. The French, on the other hand, always preferred placing these objects, not _on_ their churches, but rather grouped around them, and springing from the ground. This, it is true, enabled them to indulge in height and lightness internally to an extent unknown in England. This extravagance proved prejudicial to the true effect even of the interior, while externally the system was very destructive of grace and harmony. A French cathedral is generally solid and simple, as high as the parapet of the side-aisles, but above this base the forest of pinnacles and buttresses that spring from it entirely obscure the clerestory, and confuse its lines. Above this again the great mass and simple form of the high steep roof, unbroken by pinnacles or other ornaments, contrasts unpleasingly with the lightness and confused lines immediately below it. This inconsistency tends to mar the beauty of French cathedrals, and even of their churches, though in the smaller buildings the effect is less glaring owing to the smallness of the parts.

SPIRES.

An easy transition leads from pinnacles to spires, the latter being but the perfect development of the former, and each requiring the assistance of the other in producing a thoroughly harmonious effect. Still their uses were widely different, for the spire never was a constructive expedient, or useful in any way. Indeed, of all architectural features, it is the one perhaps to which it is least easy to apply any utilitarian rule.

Towers were originally introduced in Christian edifices partly as bell-towers, partly as symbols of power, and sometimes perhaps as fortifications, to which may be added the general purpose of ornamenting the edifices to which they were attached, and giving to them that dignity which elevation always conveys.

From the tower the spire arose first as a wooden roof, and as height was one of the great objects to be attained in building the tower, it was natural to eke this out by giving the roof an exaggerated elevation beyond what was actually required as a mere protection from the weather. When once the idea was conceived of rendering it an ornamental feature, the architects were not long in carrying it out. The first and most obvious step was that of cutting off the angles, making it an octagon, and carrying up the angles of the tower by pinnacles, with a view to softening the transition between the perpendicular and sloping part, and reducing it again to harmony.

One of the earliest examples in which this transition is successfully accomplished is in the old spire at Chartres (Woodcut No. 627); the change from the square to the octagon, and from the tower to the pyramid, being managed with great felicity. The western spires of St. Stephen’s abbey at Caen (Woodcut No. 613), though added in the age of pointed Gothic to towers of an earlier age, are also pleasing specimens. But perhaps one of the very best in France, for its size and age, is that of St. Pierre at Caen (Woodcut No. 665), uniting in itself all the properties of a good design without either poverty or extravagance. The little lantern of Ste. Marie de l’Épine (Woodcut No. 644), though small, is as graceful an object as can well be designed; and the new spire at Chartres (Woodcut No. 627), as before remarked, is, except as regards the defects inherent in its age, one of the most beautiful in Europe.

This feature is nevertheless, it must be confessed, rarer in France than might be expected. This is perhaps owing to many spires having been of wood, to their having been allowed to decay, and to their removal; while in other instances it is certain that the design of erecting them has been abandoned in consequence of the tower, when finished, having been found insufficient to bear their weight.

[Illustration: 665. St. Pierre, Caen. (From Chapuy.)]

The ruined church of St. John at Soissons has two, which are still of great beauty. At Bayeux are two others, not very beautiful in themselves, but which group pleasingly with a central lantern of the Renaissance age.[362] And at Coutances there are two others of the best age (Woodcut No. 634), which combined with a central octagonal lantern make one of the most beautiful groups of towers in France. Here the pitch of the roof is very low, and altogether the external design of the building is much more in accordance with the canons of art prevalent on this side of the Channel than with those which found favour in France.

Of the earlier French lanterns, this at Coutances is perhaps the best specimen to be found: of the latter class there is none finer than that of St. Ouen (Woodcut No. 666); and had the western towers been completed in the same character, in accordance with the original design, the towers of this church would probably be unrivalled. Even alone the lantern is a very noble architectural feature, and appropriate to its position, though some of the details mark the lateness of the age in which it was erected.

[Illustration: 666. Lantern, St. Ouen, Rouen. (From a print by Chapuy.)]

Notwithstanding the beauty of these examples, it must be confessed that the French architects were not so happy in their designs of spires and lanterns as they were in many other features.

[Illustration: 667. Corbel. (From Didron, ‘Annales Archéologiques.’)]

[Illustration: 668. Capitals from Rheims.]

It would be in vain to attempt to enumerate all the smaller decorative features that crowd every part of the Gothic churches of France, many of which indeed belong more to the department of the sculptor than to that of the architect, though the two are so intimately interwoven that it is impossible to draw the line between them. It is, however, to the extreme care bestowed on these details and their extraordinary elaboration that the Gothic churches of the best age owe at least half their effect. There are many churches in Italy of the Gothic and Renaissance ages, larger and grander in their proportions than some of the best French examples, but they fail to produce a similar effect because these details are all—if the expression may be used—machine-made. The same forms and ornaments are repeated throughout, and too frequently borrowed from some other place without any evidence of thought or fitness in their application, and consequently call up no responsive feeling in the mind of the spectator. On this side of the Alps, in the best age, every moulding, every detail, exhibits an amount of thought combined with novelty, and is always so appropriate to the place or use to which it is applied, that it never fails to produce the most pleasing effect, and to heighten to a great extent the beauty of the building in which it is found. The corbel for instance represented in Woodcut No. 667 is as much a niche for the statue as a bracket to support the ends of the ribs of the vaults, and is one of the thousand instances which are met with everywhere in Gothic art of that happy mixture of the arts of the mason, the carver, and the sculptor, which, when successfully combined, produce a true artistic effect. These combinations are so numerous and so varied that it would be hopeless to attempt to classify them, or even to attempt to illustrate the varieties found in any single cathedral.[363]

The same may be said of the capitals of the pillars, which in all the best buildings vary with every shaft, and appear to have been executed after the architect had finished his labours, by artists of a very high class. In the best age, in France at least, as in the examples from Rheims, shown in Woodcut No. 668, they would appear to have retained a reminiscence of a Roman Corinthian order, but to have used it with a freedom entirely their own.

CONSTRUCTION.

It has been shown that the exigencies of a Gothic cathedral were a stone roof, a glass wall, and as great an amount of space on the floor, as little encumbered with pillars and points of support, as could be obtained. The two first of these points have been sufficiently insisted upon in the preceding pages; the last, however, demands a few more remarks, as the success achieved by the masons in the Middle Ages in this respect was one of their chief merits, though it was but a mechanical merit after all, and one in which they hardly surpassed their masters the Romans. The basilica of Maxentius, for instance, covers a space of 68,000 sq. ft., or about the average size of a French cathedral, and the points of support, or in other words the piers and walls, occupy only 6900 sq. ft., or between a 9th and a 10th part of the whole area. If we turn to the great cathedral of St. Peter’s at Rome, we find the points of support occupying more than one-fourth of the whole area, though built on the model, and almost a copy, of the Roman basilica. At St. Mary’s at Florence they occupy one-fifth; and in St. Paul’s, London, and the Pantheon at Paris, the walls and pillars occupy in the first rather more, in the other rather less, than one-sixth. If from these we turn to some of the Mediæval examples, we find for instance at

The whole area. Solid. Ratio.

Bourges 61,591 11,908 0·181, or between 1-5th and 1-6th.

Chartres 68,261 8,888 0·130, or 1-8th.

Paris 64,108 7,852 0·122, or between 1-8th and 1-9th.

St. Ouen 47,107 4,637 0·090, or between 1-10th and 1-11th.

The figures, however, at Bourges include a heavy and extended porch not belonging to the original design, which if omitted would reduce the fractional proportion considerably; and if the unbuilt towers of St. Ouen were excluded, the proportion of the points of support to the area would be less than one-twelfth.

Our best English examples show a proportion of rather less than one-tenth, and though they have not the great height and wide-spreading vaults of the French cathedrals, their spires and pinnacles externally perhaps more than counterbalance this. Taken altogether it may generally be stated that one-tenth is about the proportion in the best Gothic churches of the best age. When we find it exceed this, it is obvious that the lightness of the walls and pillars has been carried to excess, and even in St. Ouen, if there is an error, it is on this side. There can be no question that to produce a satisfactory effect a church requires solidity, and apparent as well as real strength; for, without affecting the extreme massiveness of Egyptian art, with its wonderful expression of power and durability, there is an opposite extreme far more prejudicial to true architectural effect in parading, as it were, mechanical contrivances of construction, so as to gain the utmost utilitarian effect with the least possible expenditure of means. This the Egyptians utterly despised and rejected, and heaped mass on mass, even at the expense of any convenience or use for which the building might have been designed. The French architects, on the other hand, made it their study to dispense with every ton of stone they could possibly lay aside. This system they undoubtedly carried too far, for without looking at such extreme examples as the choir of Beauvais or St. Ouen, everywhere in France we find a degree of airy lightness and tenuity of parts destructive of many of the most important conditions of architectural excellence.

FURNITURE OF CHURCHES.

Little less thought and expense were probably bestowed upon what we may call the furnishing of Gothic churches than upon the fabrics themselves. Though the objects included in this denomination were altogether of a lower class of art, they were still essential parts of the whole design, and we cannot fairly judge of the buildings themselves without at least endeavouring to supply their minor arrangements.

It is not easy to do this in France, nor indeed in any part of Europe, as no one church or chapel displays at the present day all the wealth and ornament which once belonged to it.

There is scarcely a single church in France with its original altar, the most sacred and therefore generally the most richly adorned part of the whole. These have either been plundered by the Huguenots, rebuilt in the execrable taste of the age of Louis XIV., or destroyed during the Revolution.

The cathedrals of Amiens and Rouen are among the few which retain their original stalls; and the enclosure of the choir at Chartres is one of the most elaborate pieces of ornamental sculpture to be found. That at Alby has been before alluded to, and fragments of this feature still exist in many cathedrals.

[Illustration: 669. Rood-Screen from the Madeleine at Troyes. (From Arnaud, ‘Voyage dans l’Aube.’)]

The Rood-screens, or _Jubés_, which almost all French churches once possessed, are rarer than even the other parts of these enclosures. A good example of them is found in the church of the Madeleine at Troyes (Woodcut No. 669), which gives a favourable idea of the richness of decoration that was sometimes lavished on these parts. Though late in age, and aiming at the false mode of construction which was prevalent at the time of its execution, it displays so much elegance as to disarm criticism. It makes us too regret the loss of the rood-screens of St. Ouen’s (of which we can alone judge from drawings) and of the larger cathedrals; though of these we are able to form some idea by following out the design of the lateral screens, of which they formed a part.

If to these we add the altars of the minor chapels, with the screens that divided them from the nave, the tombs of wealthy prelates and nobles, the organ galleries, with their spiral stairs and richly-carved instrument cases, and all the numberless treasures of art accumulated by wealth and piety, we may form some idea of what a Mediæval cathedral really was, though scarcely one now exists in any part of Europe in an entire state.

DOMESTIC ARCHITECTURE.

It is probable that specimens remain sufficient to elucidate in an archæological point of view the progress of domestic architecture in France, and thereby to illustrate the early manners and customs of the people; but these remains are much less magnificent and are less perfectly preserved than the churches and cathedrals, and have consequently received comparatively little attention.

Had any of the royal palaces been preserved to our day, or even any of the greater municipal buildings, the case might have been different. The former have, however, perished, without an exception; and as regards the latter, France seems always to have presented a remarkable contrast to the neighbouring country of Flanders.

[Illustration: 670. Hôtel de Ville de St. Antonin.]

No town in France proper seems to have possessed in the Middle Ages prior to the end of the 15th century either a municipality or a town-hall of any note. When necessary to discuss communal business it was the custom to meet in the open air, or occasionally in the churches or cloisters. There is one notable exception to this in the town-hall of St. Antonin, in the department of Tarn and Garonne, which is a remarkable edifice of the 12th century, and though partially restored retains still the principal features of its early design (Woodcut 670). The ground storey, used as a market, consists of a series of pointed arches, the one on the left being a passage-way through. On the first floor is a fine room, lighted by three windows, each subdivided by three shafts. The two piers separating the windows (and which on the inner wall support segmental arches carrying the wall above) are decorated with sculpture representing Adam and Eve and Moses. The second storey, which rises into the roof, is lighted by three double windows. Of later examples at the end of the 15th and commencement of the 16th centuries there exist still the town-hall of Compiègne, a beautiful example, with central tower; and at Saumur, St. Quentin, Orleans, Bruges, and Beaugency a series of small but interesting buildings, some flamboyant and others showing early Renaissance influence.

In a work like the present, which is barely sufficient in extent to admit of all the great typical examples of architectural art being enumerated, much less described, it is evident that to domestic art a very subordinate position must be assigned. Perhaps it ought to be omitted altogether. There are, however, so many beauties in even the most insignificant productions of the great ages, that it may be expedient at least to direct attention to the subject, and the three examples here given may serve to illustrate the forms of the art at the three great epochs of the French Gothic style.

[Illustration: 671. House at Cluny. (From Gailhabaud.)]

The first (Woodcut No. 671) is from a house at Cluny, and exhibits the round-arched arcade with its alternate single and coupled columns, which arrangement was usual at that period, and of which examples are found all over the South of France and as far north at least as Auxerre.

The second (Woodcut No. 672) represents a house at Yrieix, and shows the pointed Gothic style in its period of greatest development; and although the openings are of larger extent than would be convenient in this climate, they are not more so than would be suitable, while they give, in the South of France, great lightness and elegance to the façade. The third example is from the portal of the Ducal Palace at Nancy (Woodcut No. 673), and is an instance of the form the style took when on the verge of the Renaissance. It is not without elegance, though somewhat strange and unmeaning, and, except as regards the balconies, the parts generally seem designed solely for ornament without any constructive or utilitarian motive.

[Illustration: 672. House at Yrieix. (From Gailhabaud.)]

One of the most extensive as well as one of the best specimens of French domestic architecture is the house of Jacques Cœur, at Bourges, now used as the town-hall. It was built by the wealthy but ill-used banker of Charles VII., and every part of it shows evidence of careful design and elaborate execution; it was erected too at an age before the style had become entirely debased, and as a private residence situated in a town, and therefore without any attempt at fortification, is the best that France now possesses.

The château of Meilhan (Cher) is nearly a repetition of the same design, but at least a hundred years more modern.

Rouen possesses several examples of domestic architecture of a late date; so does Paris—and among others, the celebrated Hôtel de Cluny. Few of the great towns are however without fragments of some sort, but hardly any are of sufficient importance to deserve separate notice or illustration.

France is not so rich as either Germany or England in specimens of castellated architecture. This does not apparently arise from the fact of no castles having been built during the Middle Ages, but rather from their having been pulled down to make way for more convenient dwellings after the accession of Francis I., and even before his time, when they had ceased to be of any use. Still the châteaux of Pierrefonds and Coucy are in their own class as fine as anything to be found elsewhere. The circular keep of the latter castle is perhaps unique, both from its form and dimensions; but being entirely gutted inside its architectural features are gone, and it is now difficult to understand how it was originally arranged, and by what means it was lighted and rendered habitable.[364]

[Illustration: 673. Portal of the Ducal Palace at Nancy. (From Dusomerard.)]

Tancarville still retains some of the original features of its fortifications, as do also the castles of Falaise and Gaillard.

The keeps of Vincennes and Loches are still remarkable for their height, though they hardly retain any features which can be called strictly architectural. In the South, the fortified towns of Carcassonne and Aigues Mortes, and in the North, Fougères, retain as much of their walls and defences as almost any place in Europe. The former in particular, both from its situation and the extent of its remains, gives a singularly favourable and impressive idea of the grave majesty of an ancient fortalice. But for alterations and desecrations of all sorts, the palace of the popes at Avignon would be one of the most remarkable castles in Europe: even now its extent and the massiveness of its walls and towers are most imposing.

These are all either ruins or fragments; but the castle of Mont St. Michel, in Normandy, retains nearly all the features of a Mediæval fortress in sufficient perfection to admit of its being restored, in imagination at least. The outer walls still remain, encircling the village, which nestles under the protection of the castle. The church crowns the whole, and around it are grouped the halls of the knights, the kitchens and offices, and all the appurtenances of the establishment, intermingled with fortifications and defensive precautions that must have made the place nearly impregnable against such engines of war as existed when it was erected, even irrespective of its sea-girt position.

## BOOK IV.

BELGIUM AND HOLLAND.

##