Development and Character of Gothic Architecture/Chapter 2
CHAPTER II
GOTHIC CONSTRUCTION IN FRANCE
By France in this chapter, and indeed throughout this book, is meant the France of the twelfth and thirteenth centuries—that is, the Royal Domain of the Capetian Dynasty and portions of a few contiguous provinces, chiefly Champagne, Burgundy, Picardy, Orleanais, and Berry. To this region the early Gothic movement was confined. Indeed, its earliest manifestations were circumscribed by even narrower limits, those, namely, of the Ile-de-France—that is, of the region of which the larger part is now included in the departments of the Seine and the Oise.
Though many of the works of this early art have perished, much yet remains, and the beginning and course of development of the new style may, by careful examination and comparison of the characteristics of existing buildings, be made out with substantial correctness, though no other sources of information exist. For such scanty written records of building as have been preserved are wholly devoid of information respecting principles and methods of construction. We are compelled, therefore, to rely upon independent study of the buildings themselves.
We need not stop to consider the earlier innovations in buildings in which the round arch alone continues to be employed—as in the apse of St. Martin des Champs at Paris, and in the Collegiate Church of Poissy, where the vaulting compartments are merely separated by transverse ribs,—but rather begin our investigation of the growth of Gothic in France, with those monuments in which the pointed arch first appears as a constructive device in vaulting.
The new principles of construction are first distinctly and skilfully exemplified in the Abbey Church of St. Denis, which dates from 1137 to 1141. And the origin of Gothic, so far as existing monuments exhibit it, is, on the Continent, now commonly traced to that building only. But the Abbey Church of Morienval, near Crépy-en- Valois, appears to anticipate, though in a halting manner, some of the principles that are carried out so remarkably in St. Denis. This church, a construction of the eleventh century, has an apse which dates from the end of the eleventh or the beginning of the twelfth century, and a rudimentary apsidal aisle whose vaults have diagonal ribs, pointed archivolts, [1] and even rudely pointed transverse arches, though of these last arches one has no rib, while the other has a rudely adjusted and very heavy round-arched one. The diagram 2 [2] (Fig. 11) illustrates the form and construction of one compartment of these vaults, the one marked a on the reduced plan of the apse given at A. It will be seen that the narrow archivolt a, in the plan B, whose elevation is at a', is both stilted and pointed, in order to bring its crown up to nearly the same level as that of the wider spanned round arch b, whose elevation is at b'; and that the transverse arch c, situated at b in the plan A, is more acutely pointed for the same reason, while the transverse arch d assumes the form of an irregular ellipse. It is interesting, as showing that constructive exigencies alone here brought about the use of the pointed arch, to notice further, that the ridges of the longitudinal cells, g and h, having to pass through the intersection of the diagonals, and curving with the form of the apse, bring the crowns of these transverse arches to one side, rather than over the centres of their bases, and thus give unequal curvature to their sides. This would hardly have been tolerated had it been merely a preference for the form of the pointed arch which determined its use. These awkward arches are so plainly the result of a groping struggle with the difficulties of vaulting a curved oblong space, that they seem to show beyond question that the pointed arch was not introduced as an admired form, but that it came unsought
FIG. 11.
in the course of constructive experiment. The interest of this vault lies chiefly in what it exhibits of experiment in the application of new principles as yet but feebly apprehended. This groping procedure is shown again in the forms given to the plans of the diagonal ribs which are curved instead of being straight, in order, in some measure, to avoid the extreme one-sided position of the ridges of the longitudinal cells, which must pass through their intersection, and also to avoid, in some measure, the excessive inequality of the sides of the transverse arches. By this means, too, the inequality of magnitude in the transverse cells, e and f, is in some degree lessened. The perspective elevation of this vault is shown in Fig. 12. [3]
FIG. 12.
Incomplete and awkward as is the system adopted at Morienval, we have here a form of vault construction such as had been before unknown; a form which, though imperfectly carried out, already contains some of the most characteristic structural features, and involves some of the most important principles of the Gothic vault. The apse of Morienval, therefore, rather than that of St. Denis, must be regarded as the first step known to us on the way to the Gothic style. The full value and far-reaching consequences of what was here rudely accomplished were, indeed, not yet recognised; but everything was sure in time to follow on such a beginning.
Between Morienval and St. Denis no intermediate steps in the line of progress can now be traced. Yet it would seem that such must have existed; for not only does St. Denis show a consistent advance in the direction that had been indicated at Morienval, but the mastery of the new principles, the skill of execution, and the comparative lightness and elegance of the work, are such, that it is hard to believe so much could have been accomplished at a single stride.
Of the work wrought under Suger at St. Denis the greater part has been destroyed; but the aisles of the choir and apse and the apsidal chapels remain in excellent condition. The construction is on a much larger scale than that of Morienval, and in place of the rudimentary apsidal aisle we have here amply developed double aisles to both apse and choir—foreshadowing the vast and magnificent aisles of Paris, Chartres, and Amiens,—whose vaults show no sign of hesitation and little executive imperfection. They are furnished with a full system of sustaining ribs, of which the transverse and longitudinal ones are pointed. The diagonals are round arches, and project vigorously, effectually strengthening the groins. Their intersection is far above the crowns of the enclosing arches, and the cells are thus necessarily much domed. [4] Here, for the first time among existing monuments, does the rib system wholly determine the forms and constitute the strength of the vaults. The architect of St. Denis was a master of the principles involved in his scheme. His complex plan is carried out in all its details with astonishing sureness. [5] In the curved compartments of the apsidal aisle, for instance, the difficulty which in Morienval led to the one-sided form of the transverse arches, is overcome by the application of the new principle that the forms and arrangements of the ribs shall determine those of cells. Thus an equal-sided pointed arch spans the end of the compartment, and the 
FIG. 13. ridge of the cell is made to meet its crown. Moreover, the unequal dimensions of the transverse cells, which the vault of Morienval exhibits, is largely avoided by disposing the opposite branches of each diagonal rib, so that in plan they meet at an angle, thus bringing the intersection of the diagonals in the centre of the compartment (Fig. 13). The difficulties of vaulting a structure of such plan by any of the older methods of vaulting were very considerable, and when accomplished, the result was necessarily very unsightly. For the varying dimensions of the spans to be arched over inevitably occasioned either most awkward differences of level in the springing of the arches, or else equally awkward differences in the elevations of their crowns. But by the use of the pointed arch, which readily adjusts itself to any span without a change in height, such a plan is easily vaulted, and with beautiful rather than unsightly effect.
The original arcade piers of the choir and sanctuary [6] do not exist, the piers having been reconstructed, together with the high vaults, in the thirteenth century; but the shafts which divide the aisles remain unchanged. They are monoliths of slender proportions, though heavy enough for their work, which is considerable. They taper a little and have a slight entasis—a lingering characteristic of ancient art which appears also in several other early Gothic buildings. Their capitals are considerably advanced towards Gothic form by spreading to receive, upon an ample abacus surface, the threefold load of transverse, longitudinal, and diagonal ribs. It is to be regretted that we have now no means of ascertaining what were the forms of the choir vaults and what was their system of supports. The bases of the piers of the sanctuary are all that remain of this system, which was, however, there can be little doubt, substantially like that of the nearly contemporaneous structures presently to be noticed.
In Morienval and St. Denis we have, then, the Gothic principle of vaulting in its inception and in its earliest complete character, as displayed in ground -story vaults of small dimensions. The next examples show the nature of the early vaults and vaulting systems on a larger scale.
It is impossible to be precise in chronological sequence, but the Cathedrals of Senlis and of Noyon must, it would seem, have followed very soon after Suger's work at St. Denis. They are beyond doubt nearly contemporaneous buildings, and M. Vitet has shown [7] that the early portions of the existing Cathedral of Noyon must have been begun as early as 1150. The eastern portions only of these buildings belong to this early date; the naves were not completed till later in the century. In Senlis all to the east of the transept, and in Noyon both choir and transept, illustrate the progress that had been made by the middle of the twelfth century. Both of these churches have apsidal aisles, apsidal chapels, and vaulted triforium galleries. The choir of Senlis is of the unchanged primitive construction, with a few minor exceptions, up to the level of the clerestory string; but its present high vaults and clerestory are the incongruous and ill-proportioned work of a much later epoch. The arrangement and the forms of the piers indicate that the original vaults were sexpartite, and, after those of St. Denis, they must have been among ~the very earliest vaults of considerable scale constructed on Gothic principles. Hardly any vaults of this form remain in the Ile-de-France of a date earlier than those of the choir of the Cathedral of Paris, which were completed about 1180. The earliest progress of sexpartite vaulting, in France proper, cannot therefore now be studied; but there are sexpartite vaults in the Norman churches of Caen which afford instructive illustration of the beginnings of such constructions. Of these the vaults of the Abbaye-aux-Hommes are the earliest, dating from the early part of the twelfth century. [8] They are beyond doubt among the earliest, if they be not the very earliest, sexpartite vaults that were anywhere built. But they are not Gothic vaults, notwithstanding their transverse and diagonal ribs, because the rib system is neither complete nor independent. It is not complete because there are no longitudinal or wall ribs; and it is not independent, though it is not incorporated with the shell of the vault, because the forms of the arches do not determine the forms of the vaults, but are themselves determined by these forms. The mind of the builder was preoccupied with the traditional methods of vaulting, based upon the intersecting principle, and the ribs of his vaults were accordingly made to follow the old forms; hence the transverse ribs are round arched, and the diagonals are elliptical—forms opposed to Gothic principles as exerting the maximum rather than the minimum of thrust. The lateral cells, however, take a new form, which is a step in the direction of Gothic, since it is necessitated by the positions and the curves of the intermediate and diagonal ribs, to which these cells have to accommodate themselves. In covering the triangular spaces enclosed by these ribs and the clerestory wall, it was impossible to avoid those twisted surfaces which have often been mistakenly regarded as among the defects, though they really are among the essential characteristics, of Gothic vaults. In this instance, however, the twisted surface is less pronounced than it would otherwise be, because the longitudinal arches are made to assume an upright elliptical form—which however, adds to the awkwardness of the whole effect. [9] But with all their defects these vaults mark an early step of much importance; and the sexpartite form was a novelty which stimulated the artists of the neighbourhood of Paris to devise improvements, by means of which the true sexpartite Gothic vault was produced.
FIG. 14.
The motives which, in the Abbaye-aux-Hommes, led to the adoption of the sexpartite vault are not clear, but the forms of the piers which differ alternately in having respectively a single engaged shaft, and a broad pilaster in addition to such a shaft, probably suggested it. This arrangement of the piers, which the builders of the vault found existing in the originally unvaulted edifice, was beyond doubt derived from the Lombard churches, though it differs in one important particular from the Lombard example. The Church of San Michele of Pavia, for instance, exhibits an alternate arrangement of piers, but one which, if designed for vaulting, was clearly intended for quadripartite vaults over square compartments, each embracing two bays of the aisles; and hence the intermediate piers, having no function in connection with the high vaults, do not rise above the triforium. The arrangement will be understood by reference to the plan and elevation (Figs. 14 and 15) of one bay of this church. [10] The existing quadripartite vaults do not belong to the original construction, and it is quite possible, indeed perhaps probable, that no vaults were originally intended, but that the main piers carried at first mere transverse arches, such as may still be seen in the
FIG. 15.
basilica of San Miniato at Florence, and such as were constructed also in the Church of Notre-Dame du Pre at Mans. [11] In these cases the intermediate trusses of the timber roofs rest on the walls without any supporting shafts, which are, of course, not required. But in the Abbaye-aux-Hommes (Fig. 16) the shafts of the intermediate piers are carried up to the top of the wall, [12] and hence, when the vaulting was undertaken, they may naturally have suggested the intermediate arch of the sexpartite vault. However this may be, the piers, though not originally designed or adapted for
FIG. 16.
vaulting, are now, having been shortened and slightly modified for the purpose, [13] fairly well suited to such vaults as they carry. The two sides of the pilaster and the engaged shaft of the main pier carry respectively the main transverse rib and the two diagonals, while the intermediate shaft carries the intermediate transverse rib; and, there being no wall ribs, no other supports are required.
The principle of the sexpartite vault, with an appropriate vaulting system, thus, it would appear, accidentally developed and rudely embodied, in the Abbaye-aux-Hommes, became fruitful immediately in the Ile-de-France; where all chance work and groping experiment were being rapidly superseded by constructive foresight and consistency. Here in the Cathedral of Senlis there was nothing fortuitous or incongruous. The vaults and vaulting system were simultaneously conceived, and were, in all respects, parts of one whole. This is fully indicated by the piers, though not a stone of the primitive vaults remains in place. As the rib skeleton now determined the forms of the vaults, so with equal strictness did it determine the forms of the piers. These piers of Senlis are alternately massive and slender; the main pier being in section (Fig. 17) primarily a square intersected unsymmetrically by a rectangle. Against the four projecting faces of this pier rise, respectively, four engaged columns, a, b, c, and d, as shown in the section. Of these a supports the main transverse rib of the high vault, b and c support the archivolts of the ground-story, and d supports the transverse rib of the aisle vault. On each of the angles of the square central mass are worked round shafts, f, g, i, and j: f and g to carry the diagonal ribs of the high vault, and i and j to carry the diagonal ribs of the aisle vaults. While in the re-ntering angles on the choir side are placed the shafts e and h, which carry the longitudinal ribs. The only capitals on the ground-story level are those of the archivolt columns and the aisle vaulting shafts. The five members of the great vaulting group ascend without interruption to the springing of the high vaults; and the whole pier is built up of coursed masonry, admirably cut and closely jointed. The intermediate pier consists, on the ground-story, of a plain round column, from whose capital rise three slender vaulting shafts to support the intermediate transverse rib and the two clerestory ribs, the two ground-story archivolts and the transverse and diagonal ribs of the aisle vaults.
FIG. 17.
The arches of the great arcade are pointed and of one order. The triforium arches are also pointed, and are of two orders—the one not concentric with the other, an unusual arrangement, but one which occurs also in the apse of St. Denis. These triforium arches are carried by monolithic shafts, compactly grouped with the pier, and are not divided. Figures 18 and 19, an elevation and a section of one double bay, will afford a clearer idea than words can do of the character of this structure. And the perspective view (Fig. 20) of the opposite bay [14] will convey some impression of the beauty and the degree of Gothic expression that were reached in this interesting monument.
FIG. 18.
FIG. 19.
requirements. The principle fortuitously foreshadowed in the Abbaye-aux-Hommes is here developed with both scientific and artistic precision.
FIG. 20.
The same constructive logic is exhibited in other portions of this interior. The vaults of the aisles, of the apsidal chapels, and of the triforium gallery show no defects of principle; and they are admirable in workmanship and remain in perfect condition. They are so closely similar to the aisle vaults of St. Denis that no further description is necessary.
It may here be remarked that we have reached the time of greatest perfection in masonry. After 1130 the walls, piers, and vaults of the twelfth century are unrivalled for fineness of facing and jointing. They are, in this respect, in striking contrast to those of the more hasty constructions of the thirteenth century. And nowhere do we find skill in the manipulation of carefully selected material more admirably exhibited than in this building.
A noticeable characteristic of the choir of Senlis is that it exhibits a nearer approach to Gothic principles and expression within than without. The interior is indeed frankly Gothic, though still massive, while what remains of the exterior is, in its broad walls and round-arched openings, almost strictly Romanesque. In this and in other early monuments of France, we see the style in process of formation according to the vital law of all organic development—the functions to be exercised calling into being and giving appropriate form to the requisite organs. We do not find the change from Romanesque to Gothic beginning in a mere arbitrary transformation of external forms and details—external forms and details are, in France, the last things to change. The growth of the Gothic principle begins at the very heart of the fabric and gradually works outward till every part is reached. Operating first imperfectly in the diminutive vaults of Morienval, perfecting the vault forms in St. Denis, and now from the high vaults of Senlis creating for itself an appropriate, though not a final system of internal supports, it moves on, as we shall see, in this creative fashion, till the full development is accomplished.
What was the precise mode of buttressing the high vaults we have now no means of knowing. It is possible that flying buttresses may have sprung over the aisle roofs; but there is hardly an instance of an external flying buttress of this early epoch, and it appears more probable that the triforium vaults formed the only abutments to the piers, which are almost heavy enough to bear the thrusts independently. As may be seen in Fig. 19, the capitals, [15] from which the high vaults sprang, were situated below the level of the clerestory string, so that these vaults may have been sufficiently abutted by those of the triforium.
The choir of the Cathedral of Noyon has many points of resemblance to that of Senlis. It is on a somewhat larger scale, is lighter in the proportions of its parts, and, in some respects, betokens a freer exercise of the inventive talents of that great body of secular builders which was now taking the lead in architectural constructions, and finding scope for its genius in these communal cathedrals that were beginning to rise, in quick succession, in the newly-chartered towns.
Noyon had been one of the first cities to organise a commune; and it had done so under the fortunate circumstance of its bishop having taken the initiative in the work, so that from the first there was harmony between the ecclesiastical and civil authorities, [16] which is curiously imaged in the church. [17] The vaults of this choir happily remain in good preservation. Unlike the original vaults of Senlis, they are quadripartite in oblong compartments; and hence we have here a uniform, rather than an alternately varied, series of vaulting shafts and piers. The transverse ribs alone are pointed, and the round-arched longitudinal ribs are so much stilted as to bring their crowns up nearly to the level of the crowns of the diagonal ribs. There is, therefore, hardly any doming in these vaults. Three vaulting shafts, 
FIG. 21 resting on the capitals of the ground-story piers, sustain the transverse and diagonal ribs respectively. The piers of the ground-story are, in the choir proper (with exception of two massive clustered ones, which were designed to support towers against the east side of the transept), plain round columns with a single engaged shaft (as in the section, Fig. 21). In the sanctuary the columns are more slender and have no engaged shaft; they are monolithic, and have a slight entasis, like those of the aisles of St. Denis, and their capitals are of distinctly Gothic form in their functional adaptation to the new structural conditions.
The ground-story arcades are round arched except in the sanctuary, where they are pointed. The triforium openings are coupled pointed arches divided by groups of three detached monolithic shafts. Over the triforium is an arcaded gallery in the thickness of the wall, composed of diminutive trefoiled round arches, and the clerestory openings are round arched and undivided.
Unhappily the buttress system has been reconstructed, so that the original forms are uncertain. It can, however, hardly be doubted that flying buttresses sprang over the aisle roofs in true Gothic fashion, for the clerestory rises so high above the vaulted gallery that the vaults of this gallery could hardly have formed effectual abutments to the high vaults; while the walls and piers, being less massive than those of Senlis, could hardly have sustained the vault pressures without reinforcement On the whole, the Gothic principle is far advanced in this choir of Noyon, though its interior is by no means so pure and harmonious in design as that of Senlis. The nave was constructed at a somewhat later, though still at an early epoch; and in it a fuller apprehension of the freedom afforded by the new principles of construction is indicated in the more slender proportions of piers and shafts, and in the increased magnitude of openings. In these respects, indeed, the nave of Noyon is hardly surpassed by any other monument of the twelfth century. Like Senlis, this nave was designed to carry, and it probably did at first carry, sexpartite vaults;[18] its piers differ from those of Senlis in their lighter proportions and in minor details only. The ground-story arcade has pointed arches of one order, and the triforium openings consist of a pointed arch in each bay spanning a sub-order of two pointed arches. These are divided by a shaft of unprecedented slenderness, and the tympanum is pierced with a trefoil, an early step in the direction of plate tracery. There can be little question that flying buttresses substantially like those now existing—which probably date from the time of the reconstruction of the vaults in the thirteenth century—abutted against the main piers; and thus the Gothic constructive system was complete in its principal features here at Noyon soon after the middle of the twelfth century.
Though there are many minor puzzles presented by the respective peculiarities of these early buildings which, in the absence of clear documentary records, often prevent an exact determination of their chronological relationships, yet there can be little question that such constructions as the choir of St. Germain des Prés at Paris, the nave of St. Stephen's at Beauvais, the Churches of St. Leu d'Esserent, and St. Martin at Laon, besides many others, were nearly contemporaneous with Senlis and Noyon. Each of these buildings presents characteristics peculiar to itself, which show a degree of individual independence in the builders that is no less striking than their common allegiance to the leading idea which was, day by day, gaining distinctness, and was rapidly transforming the art of building. Some of these peculiarities we shall have occasion to refer to farther on, while we now pass to the consideration of the structural forms exhibited by some of the larger buildings of the second half of the twelfth, and the beginning of the thirteenth centuries—the vast Cathedrals of Paris, Laon, Chartres, Bourges, Reims, and others,—in which the highest perfections of the system were reached, and in which the astonishing rapidity of the Gothic development is shown.
But first it may be noticed in passing that the two modes of vaulting—the quadripartite and the sexpartite—go along together from the first, though in the earlier monuments the sexpartite form is the most common. Even in vaults so early as those of the Abbaye-aux-Dames of Caen, where a curious kind of sexpartite system [19] prevails, the vault of the westernmost bay is quadripartite on an oblong plan. In the choir of Noyon all the vaults are, as we have seen, quadripartite in oblong compartments, and they are the same in the choir of St. Germain des Pres. The idea that the sexpartite vault was developed first and was gradually superseded by the quadripartite form does not seem correct. [20]
Of the greater cathedrals the one in which the Gothic principles are first distinctly and systematically manifest is that of Paris. And this wonderful monument, notwithstanding all that it has suffered from violence and so-called restoration, exists to-day in almost complete constructive integrity. [21] Here is a vast central aisle so admirably roofed with stone that the construction has lasted intact [22] for seven hundred years, and will probably, if not injured by violence, last for centuries to come. These vaults are of the sexpartite form, and those of the choir, being about contemporaneous with those which originally covered the nave of Noyon, doubtless in the main show us what these were. The transverse ribs are pointed, the longitudinal ribs are pointed (here they differ from the vaults of the nave of Noyon, whose longitudinal ribs appear to have been round arched), and the diagonal ribs are semicircular. The intersection of the diagonal ribs is at a higher level than the crowns of the transverse ribs, which, in turn, are higher than the crowns of the longitudinal ribs. The vaults have thus a distinctly domed form which, at this period, was almost universal. All these ribs are independent arches which determine the forms of, and actually sustain, the vault shells. In vaults of this form the lateral cells are, as I have already remarked, necessarily oblique to the axis of the nave, and their surfaces assume forms which are difficult to define. Indeed, more or less obliquity and irregularity of surface is a constant and necessary characteristic of true Gothic vaults, even of those which are quadripartite. Gothic vaults are never simple intersecting pointed vaults. The new constructive principles did not admit of such forms. Gothic vault-forms do not admit of description in geometric terms. They vary according to the spans, the altitudes, the curves, and the points of springing of the arches that compose the rib system, and it is by the forms and relations of these arches only that such vaults can be described. In the vaults of Paris the filling-in consists of successive courses of arched masonry reaching from rib to rib over each triangular space of the plan. The beds of these successive courses are not parallel one with another, but incline variously according as the mason found necessary or convenient in developing the twisted concave surfaces required by the varying spans and positions of the ribs. In early vaults, like these of Paris, the courses usually have a considerable rise near the springing, from the longitudinal rib toward the diagonal; and they become gradually more level as they approach the crown of the vault, where they are more nearly parallel. But perfectly parallel they hardly ever can be, since each course is properly a portion of a surface which is concaved in all directions. The masonry of these vaults, especially in the choir, is perfectly faced and closely jointed.
The vaulting shafts are slender, and rise from the great capitals of cylindrical columns which constitute the piers of the ground-story. The flying buttresses were originally double—that is to say, the piers which divide the double aisles were carried up above the roof, where each one received the head of a flying buttress which sprang from the outer buttress, over the outer aisle, and gave foothold to another flying buttress which spanned the inner aisle and abutted against the great piers. [23] The principle of equilibrium maintained by opposing thrusts is here completely developed. The inert principle is wholly abandoned. The maximum of internal space for circulation and for prospect is attained by attenuation of supports, and if the maximum of size in external openings is not reached, it is not because any obstacles stand in the way, but only because the idea of having the largest possible openings has not yet presented itself to the minds of the builders.
A somewhat detailed consideration of the leading constructive characteristics of this building, and a comparison of them with those of other kindred buildings, will enlarge our understanding of the Gothic principles which are peculiar to France. These principles are substantially carried out in all of the great churches which were erected in this region between 1160 and 1220; but they were carried out more completely and consistently in some of them than in others. Hardly one of them unites all the perfections of which the entire group affords illustration; and the great variety of forms, under which the same leading idea struggles for embodiment, gives striking evidence of the active spirit of invention which animated this remarkable building movement.
One marked peculiarity of the Cathedral of Paris is that its piers are not adapted in form and magnitude to its sexpartite vaults. In this respect the adjustment of piers to vaults is the precise reverse of that which we have seen in the nave of Noyon, where the piers, fashioned for sexpartite vaults, are now covered with those of the four-celled form. It would seem that in Paris quadripartite vaults must have been intended when the plan was laid out, and that, for some now unknown reason, the sexpartite form was adopted after the building had been carried up to the springing of the vaults. For up to this level the construction of the piers is perfectly adapted for quadripartite vaulting—all the ground-story columns being of equal magnitude, and each of them carrying a group of three vaulting shafts, of which one group is precisely like another. The incongruity thus presented in both Paris and Noyon between the forms of the vaults and the adjustment of their supports is a serious defect in each of these otherwise admirable structures as they now exist; a defect which so contradicts the logic of the Gothic system as to leave little doubt that it was in each case the result of changes made in the original project—the changes having been wrought at Noyon probably at a later date than the original construction of the building, and at Paris after the construction had reached the springing of the vaults. [24] But in the Cathedral of Paris, though the same general incongruity exists in both choir and nave, there is a marked difference in the forms, the arrangements, and the adjustments of the respective vaulting systems. In the choir, the original design of 1163, [25] the vaulting shafts rise without break from the capitals of the ground-story to the springing of the vaults, varying in magnitude according to their respective loads. They are built up in courses—the central one being engaged against a projecting pilaster, and the lateral ones against the face of the pier. In the main group—that which carries the main transverse rib and two diagonals—the capitals are all on the same level. The central capital is set even with the wall, while the lateral ones are set diagonally, in the direction of the diagonal ribs. [26] These lateral capitals carry, besides the diagonal ribs, each a small shaft which rises to support the longitudinal rib whose springing is at a higher level—an arrangement of great significance as we shall presently see. In the intermediate group the arrangement is different. Here the central shaft only has its capital at the level of the springing of the great ribs, the side shafts rising unbroken to the higher points of springing of the longitudinal ribs, where they receive their capitals.
Figure 22 will illustrate these features. In this figure A exhibits the plan of the group of abaci of the capitals of the main vaulting shafts, and the sections of the ribs which they support; B is the plan of the abacus of the intermediate capital, with the section of the intermediate rib, and the sections of the side shafts; C is a perspective view of the main group, and D a perspective view of the intermediate group. It will thus be seen that here in the choir the main and the intermediate groups of vaulting shafts differ according to their respective functional exigencies.
FIG. 22.
But in the nave, which was probably completed by 1196,[27] the vaulting system exhibits no such differences. Here the vaulting shafts (which are of unprecedented slenderness, and are not engaged with the pier, but are detached, as in the section, Fig. 23) are all of the same magnitude, notwithstanding the unequal weights which they have to carry. 
FIG. 23. They are not, as in the choir, built up in courses of small stones, but are each composed of several, in most cases five, lengths of strong cliquart. Each group has all of its capitals (A, B, Fig. 24) at the springing of the great ribs. This arrangement is perfectly adapted to quadripartite vaulting, but it is ill adapted to the six-celled vaults which are reared upon these capitals. For while in the main group (A, Fig. 24), the abaci are well covered by the three great ribs, and the bases of the small shafts which carry the longitudinal ribs, the lateral abaci of the intermediate group B, in the same figure, have the larger portions (e, in the plan D) of their surfaces unoccupied, since there are no diagonal ribs to be supported here. This is manifestly illogical as well as unpleasing to the eye. The only arch in this pier which springs from this level being the intermediate transverse rib, the central shaft which carries it
FIG. 24.
is the only one that requires a capital here. The lateral shafts ought to continue unbroken to the higher point from which the longitudinal ribs spring, as they do in the intermediate system of the choir.
In the Church of Mantes, a construction of the same epoch which bears much resemblance to Paris, the vaults are also sexpartite; and they are here prepared for by piers which are alternately massive and slender—the massive ones having all their vaulting members built up from the pavement—as at Senlis and Noyon. But the vault supports, unlike those of Senlis and Noyon, consist, as in Paris, of three shafts both in the main and in the intermediate piers—the longitudinal ribs not being independently provided for below the springing of the vaults, but rising from shafts carried by the same capitals which support the diagonal ribs.
Sexpartite vaults occur again in the Cathedral of Laon, a building nearly contemporaneous with the nave of Paris, and here we meet with still another arrangement of supports (Fig. 25). The ground-story piers are, as in Paris, simple round columns whose capitals support the vaulting shafts;
FIG. 25.
but instead of three such shafts to each pier, an arrangement which affords no provision for the alternately varying number of ribs to be carried, there are five shafts to the main piers and three to the intermediate ones, thus giving each rib in the vault its own independent support. In other words, the vaulting system of Laon is the same as that of the choir of Paris, except that the shafts which carry its longitudinal ribs, instead of resting, at the main piers, on the capitals from which the diagonal ribs spring, descend with the other vaulting shafts to the ground-story capitals. From this [28] level, therefore, the system is as logical as the systems of Senlis and Noyon. But the want of adequate support in the lower piers for the heavy main groups is a defect which mars this otherwise magnificent interior. [29] It should be added that the vaulting shafts of Laon vary in magnitude in conformity with the weights they have to support the central shaft which carries the main transverse rib being the largest, those which carry the diagonal ribs smaller, and those of the longitudinal ribs the smallest.
The Cathedral of Bourges, constructed mainly during the first quarter of the thirteenth century, [30] has also sexpartite vaults, and the disposition of its piers and shafts is peculiar though almost entirely logical. These piers are, in effect, gigantic round columns from the pavement to the springing of the vaults. As they rise through the arcade spandrels they leave something less than a quarter of their diameters projecting, and in conformity with the construction of the vault they are alternately massive and slender. Engaged in them are slender, coursed, vaulting shafts—not closely grouped according to the usual arrangement, but widely separated upon the great cylindrical surface. The adjustment of these vaulting shafts to the vaults is substantially the same as at Laon—the main shafts alone having capitals where the great ribs spring, while the supports of the longitudinal ribs rise unbroken to the higher level of the springing of these ribs. The one illogical feature of the vaulting system of Bourges is that its lesser shafts do not vary in magnitude according to their functions. With this exception there is hardly any Gothic building which exhibits greater constructive propriety; for a complete structural continuity is maintained, in both main and intermediate piers, from the pavement to the crowns of the vaulting arches.
Another mode of sexpartite vault support occurs in the Cathedral of Sens (second half of the twelfth century) and in Notre-Dame of Dijon, built about 1225. In these instances the main pier has three vaulting shafts, and the intermediate pier but one—the longitudinal ribs being carried by shafts which rise from a ledge at the clerestory level.
Of these various modes of adjustment of sexpartite vaults to their supports the most logical are perhaps on the whole the earliest. In these every member in the vault has, except in the intermediate piers, its own support from the pavement; and these supports are graduated in size in conformity with the weights with which they are charged. In Bourges the continuity of support from the pavement in all of the piers makes the system of its construction more logical than that of Senlis or of Noyon, except for the defect of the equal magnitude of the vaulting shafts.
These examples are enough to show how great are the minor differences exhibited by these early Gothic buildings. No two buildings ever show precisely the same arrangements of structural parts; yet every one of them exhibits the clear apprehension by their builders of the governing principles of the new style. The differences are largely due to local and individual differences of genius. Each locality developed to some extent its own natural modes and predilections, which modified the central influence that went forth from Paris and its neighbourhood, and thus produced more or less mixed forms of art. Sens and Dijon, for instance, show the united influences of Burgundy and the Ile-de-France, Bourges is a creation of the school of Poitou modified by the central school, while the Cathedral of Reims is a product of the school of Champagne, with a large infusion of influence from that of the Ile-de-France. [31]
We have thus far considered the dispositions and adjustments of vaulting systems to vaults, for the most part in buildings of the twelfth century and of the first quarter of the thirteenth. We have now to examine these adjustments in the more fully developed structures of the thirteenth century; and it may be well to begin with the early developments of the piers on the ground-story level—developments which constitute one of the most interesting branches of the subject, and afford important illustration of some of the fundamental principles of the Gothic style.
In the transitional buildings with sexpartite vaults the main piers, made up of square members and engaged vaulting shafts, like those of Senlis and Noyon, could hardly be improved as regards their functional adjustment and expression. But they were so massive as to take up a great deal of room, and hence were more or less inconvenient. It was probably to avoid this inconvenience that plain round columns were employed for the ground-stories in Paris and Laon; but these round columns were soon felt to be unsatisfactory, as affording no independent supports for the various members of the superstructure. Such columns did not partake of the new principles that now characterised every other constructive member of the building. Attempts to improve them were therefore made, and a new and strictly functional form was soon devised, a very early, perhaps the first, example of which may be studied in the nave of the Cathedral of Paris.
FIG. 26.The first step in the change appears to have been connected with a new adjustment to its load of the form of the abacus of the great capital of the round column,—an adjustment rendered necessary by the employment of two arch orders in the great arcade instead of one. In the choir of Paris the arches of the great arcade are of one order on the choir side, and of two orders on the side of the aisle, as in the plan (Fig. 26). The transverse rib, a, of the aisle vault is so wide that the diagonals, b and c, which are also rather wide, leave little of the abacus surface unoccupied on the aisle side; and the bases of the vaulting shafts, d, e, f, on the choir side are so spread out that the square abacus which carries this compound load fits it sufficiently well. But in the nave (Fig. 27), where the pier arches are of two orders on both sides, and where the vaulting shafts and the ribs of the aisle vault are smaller and more compactly grouped, the square abacus is not so well fitted to its load. Large portions of its surface, a, b, c, and d, are left unoccupied, notwithstanding that its corners are cut away in order to diminish this useless surface. But with this measure the builders appear not to have been satisfied; and in order to give the lower pier a more functional correspondence with
FIG. 27.
the superstructure, they before long did away with the device of terminating the pier by a capital at this level, and starting afresh to erect upon it the high vault supports. As a final result the true Gothic pier was produced in which all the vaulting members receive continuous support from the pavement, substantially as in the main piers of Senlis and Noyon, but superior to these by exhibiting the utmost compactness consistent with the complete discharge of their function.
The first modification in the nave of Paris occurs in the sixth pier counting from the transept. Here a smaller column, to augment the support of the vaulting shafts, is incorporated with the great round column, and corresponding additions are made to the great base and to the capital. Larger portions are cut off from the corners of the abacus, whose plan, thus modified, is shown in Fig. 28, while a section of the pier is shown in Fig. 29. Piers of similar section had occurred earlier in the choir of Noyon, and even in some
FIG.28.
Norman buildings, as in the transept of Ely in England; but in these cases they are not, as they plainly are in Paris, links in a chain of progress. This was, however, but a partial, and perhaps on the whole even a doubtful, improvement. It provided an independent support for the vaulting shafts, but left the 
FIG. 29. archivolts and the ribs of the aisle vaults without such supports. It had, moreover, an awkward appearance, and the abacus of the capital was still ill adjusted to its load. It was next seen that if the vaulting shafts were to have separate support in the lower pier, the other members of the superstructure ought to be supported in like manner. Accordingly in the seventh and westernmost pier this idea was carried out, and an almost perfect Gothic pier was constructed—a pier which furnished the type that was henceforth employed with many variations of proportion and detail, and which attained its 
FIG. 30.highest perfection in the nave of the Cathedral of Amiens. The section of this pier is shown in Fig. 30, its abacus surface with the plan of the imposed load, in Fig. 31, and a perspective view taken from the opposite triforium, in order to show as much as can be seen of the upper surface of the abacus and of the form of its load, is given in Fig. 32. It will be seen on the plan that the portion of the abacus which covers the capital of the great round column is now circular; that the abaci of the lesser capitals of the engaged shafts are square in agreement with the sections of the sub-archivolts and of the transverse rib of the aisle vault, which they respectively support; and that over the engaged column, which falls under the vaulting shafts, a portion of the great
FIG. 31.
abacus projects in a segmental curve, forming a band to this column, which is not provided with a capital—partly, perhaps, because it is really the lower member of the group of vaulting shafts which have their capitals at the springing of the vaults, and partly also because a capital here would present a much greater abacus surface than the bases of the
FIG. 32.
vaulting shafts would cover. The unsymmetrical plan of the impost (Fig. 31), where one of the lateral vaulting shafts advances forward of the other, is caused by the thickening of the arcade spandrel on that side in order to reinforce the great piers of the western tower. The adjustment of this compound abacus to its load could hardly be improved; there is even less unoccupied space here than in the square abacus of the choir already noticed. The form of the compound capital, as shown in the perspective view, is also admirable, the lesser heights of the smaller members being well proportioned to their lesser diameters, and these again harmonising well with the central mass. The entire group is one of the most beautiful, as it is also one of the grandest, in the whole range of Gothic design.
FIG. 33.
Another mode of reinforcing the lower pier is that which occurs in the nave of Laon (section, Fig. 33), as an exception to the plain round columns which form the prevailing support. In this case five detached monolithic shafts are grouped with the great cylinder, four of them being placed so as to support the angles of the abacus, and the fifth sustaining the central member of the group of vaulting shafts. This is certainly, in many respects, a fine pier, but it is not so compact, nor so consonant with Gothic principles, as that of Paris. It was, therefore, not so well worthy of adoption, and in fact it was not employed elsewhere.
The lower piers (A, Fig. 34) of the choir of Soissons are interesting as being designed on the same general principle as the sixth pier of Paris (B, in the same figure), while exhibiting a marked improvement upon it. Here the engaged column is more slender than at Paris, and instead of an independent abacus to its capital, the octagonal abacus of the great capital is made to project so as to cover it. The engaged column being a part of the original design and not an experimental interpolation, as at Paris, the whole lower pier is so adjusted to the superstructure as to bring this smaller column fairly under the vaulting system to be sustained. Another improvement is that of the diagonal positions given to the plinths of the lateral vaulting shafts in correspondence with the positions of the vaulting capitals. These plinths are thus parallel also with the sides of the great abacus, where it projects to cover the capital of the lower vaulting column; the adjustment of the great abacus to its load is, moreover, as close as possible, and the whole design is both compact and elegant.
FIG. 34.
We may now examine the vaults and vaulting systems of the more advanced Gothic of the first half of the thirteenth century, in which the continuity of members, from the pavement upwards, becomes an unvarying principle. Not that every individual member of the superstructure now has an independent support from the foundation, but there is at least one independent shaft on the ground-story level sustaining each main group above.
From the beginning of the thirteenth century the quadripartite vault, which was now greatly improved in form, came into general, though not invariable,[32] use; and in connection with it the form of the lower pier, which had been developed in the west bay of Paris, was generally adopted.
The improvements in the vault consisted in replacing the round longitudinal arch—such as that of the choir of Noyon and of St. Germain des Prés—by a pointed arch (such as had been employed in the sexpartite vaults of Paris, and other ontemporaneous buildings); and in making the crowns of all the arches reach more nearly to the same level, thus doing away with the excessive doming that had characterised the earlier quadripartite vaults.
There is one most important characteristic of French-Gothic vaults that often passes unnoticed, the real significance of which has not, I believe, as yet been explained by those writers who have noticed it. It is that (alluded to above, page 39, in connection with the vaults of the Abbaye-aux-Hommes at Caen, and of Notre Dame at Paris) of the twisted surfaces caused by stilting the longitudinal arches, so that their springing begins at a much higher level than that of the main arches.
A frequent misunderstanding of the Gothic vault has arisen from supposing that, by taking advantage of the properties of the pointed arch, all its ribs were made to spring from the same level and reach the same height, however they might vary in span. [33] It is indeed true that the use of the pointed arch made this possible, but it is equally true that in strictly Gothic vaults the pointed arch was never so used; in such vaults the longitudinal rib was always stilted. This fact is noticed by Willis, [34] who merely remarks in relation to it that "it is a very universal arrangement of clerestory vaults, and is productive of great beauty and convenience, but it leads to some difficulty in the form and arrangement of the vaulting surface." Other writers have supposed that this arrangement was intended to provide for largeness of clerestory openings. Thus Sir Gilbert Scott [35] says, "The side arches were sometimes stilted, not from any necessity, but merely to afford greater space for clerestory windows." But that it was not adopted because it was productive of beauty or convenience, nor to afford greater space for clerestory windows, a just consideration of the structural exigencies involved would show beyond question, even if it were not proved by the fact that the same peculiarity is frequent long before the clerestory opening is developed so as to fill the whole space beneath the longitudinal rib. In fact, the opening occupies but a comparatively small portion of this space in all early Gothic buildings, as may be seen in Paris, in Mantes, in Laon, in St. Leu d'Esserent, in the Collegiate Church of St. Frambourg at Senlis, and in many others. Fig. 35, a perspective view of one bay of the clerestory of St. Leu d'Esserent, will illustrate this point. Here the springing, a, of the longitudinal rib will be seen to be above the springing, b, of the main ribs by almost half the vertical height of the vault. It will be seen, too, that the intrados of the flying buttress, visible through the window, meets the pier at the same level. It is well known that the thrusts of the great vault ribs are not confined to their points of springing, but that there is a tendency in the arches, when firmly abutted at these points, to rise at their haunches, in consequence of which they require to be reinforced in these parts. Now the method here employed by which the line, a, b, is made to rise vertically to the level, a, brings the triangular vault surface, b, a, c, into a plane, which is inclined to the pier in the direction of the thrust of the diagonal rib; and as the diagonal rib of the next adjoining compartment, with the corresponding portion of vault surface, is inclined to the same pier in the opposite direction, the obliquity of each pressure is neutralised; and as the haunch of the transverse rib is reinforced by a solid filling-in, a perfect concentration of thrusts upon the pier is secured—the greatest power of these thrusts falling where the flying buttress is brought to bear. [36] The horizontal section (Fig. 36), taken at the level, a (Fig. 35), will more fully explain the form of this portion of the vault and the manner in which the pressures are gathered upon the pier.
FIG. 35.
Here a, b, and c are the great ribs whose thrusts, in the direction of the arrows, are concentrated upon the pier, d,and are counterbraced by the flying buttress, e. In other words, the section through the vaulting conoid at half its vertical height gives the triangle, a, b, c, in A, and not the square, a, b, c, d, in B in the same figure.
FIG. 36.
No single feature could be chosen which would exhibit more clearly the essential principles of Gothic construction. It exhibits, in fact, its governing characteristic, upon which, more than upon anything else, every other characteristic depends. And in view of this hardly any one can fail to see the error that even so learned an authority as Sir Gilbert Scott commits when he remarks that the stilting of the clerestory arches did not arise from any necessity, but was effected merely to afford greater space for clerestory windows. Without this concentration of thrusts, as far up as they extend, that compactness of the pier, which is so essential to the Gothic system, could not exist.
How this form of the clerestory was afterwards taken advantage of for larger openings we shall see when we come to consider modes of enclosure. For the present we must confine our attention to the forms and adjustments of the vaults, the vaulting supports, and the general framework of the buildings of the early thirteenth century in France. Meantime, however, it may be remarked that the ribs of the vaults of these, as of the earlier buildings, consist only of such as have a constructive office—namely, the transverse, the diagonal, and the longitudinal ribs. Ridge ribs and surface ribs, which sometimes appear later, do not occur at this epoch. Of the constructive ribs none are ever wanting, nor are independent supports for them ever wanting in the piers. Throughout the building there is a structural reason for every member that meets the eye, though the degree of perfection with which minor structural exigencies are met continues to vary.
In the nave of St. Leu d'Esserent, whose vaults we have just considered, the lower piers and the vaulting shafts are wrought substantially on the model of those of the westernmost pier of Paris; but they differ in having a complete capital over the engaged column which sustains the vaulting shafts, as well as in having the central portion of the great abacus in the form of a square set diagonally to the axis of the nave, instead of a circle, and also in having the longitudinal rib shaft rest upon the clerestory ledge instead of resting upon the same capital which carries the diagonal rib. This nave is one of the very earliest of the constructions of the thirteenth century, and its design, in many points, resembles rather the work of the latter part of the twelfth century—to which epoch it might be supposed to belong were it not for the forms of the lower piers, the character of its capitals, and the forms of the clerestory openings which shall be noticed farther on.
The nave of the Cathedral of Chartres followed quickly after that of St. Leu, which it closely resembles in main features, though the design is carried out on a much grander scale. Here the vaults, which in St. Leu are constructed on very low arches, are more acutely pointed; even the diagonal ribs having the pointed form. The stilting of the longitudinal rib is still more marked in these vaults; its
FIG. 37.
springing being more than half the vertical height of the vault above the springing of the great ribs, and the principle of concentrated thrusts upon the pier is therefore more distinctly emphasised (Fig. 37). The lower piers are alternately round columns with engaged octagonal shafts, and octagonal columns with engaged round shafts. The main vaulting column has no capital, but is simply banded by the abacus moulding, as in the westernmost pier of Paris. The great abacus is in form like that of St. Leu, except that over the main vaulting shaft it assumes the curve of the band just mentioned. Five vaulting shafts, instead of three, as at St. Leu, rise from this abacus, the shaft of the longitudinal rib descending with the rest to this level. These shafts are in three magnitudes corresponding with the magnitudes of ribs, which they respectively carry. The great compound pier capitals are admirably proportioned; and the arrangements and proportions throughout—those of the vaults, the piers, the ground-story arcade, the triforium, and the clerestory make up a whole which is perhaps the most harmonious that had been devised up to its date, and one that was hardly ever surpassed, notwithstanding its comparatively massive construction.
In the Cathedral of Reims, the lower portions of which, with exception of the westernmost bays, are contemporaneous with Chartres,[37] the structural system is again substantially the same, though the proportions and the general style of the details differ considerably. The narrow arches of the vaults are stilted in the same manner, the vaulting shafts vary functionally in magnitude, and they all descend to the capitals of the lower piers, which piers are of the type of Paris, St. Leu, and Chartres. The great compound capitals are, however, not so well composed as those of Chartres, since the smaller members are of equal height with the central ones. As in Paris and Chartres, the main vaulting shaft has no capital, though it has much the appearance of possessing one in consequence of the carved ornament with which it is banded.
We now come to the building in which the greatest perfections of the Gothic system are realised—the nave of the Cathedral of Amiens, which was begun in the year 1220. Not only is this nave the grandest in scale of any in France—being in height forty-two metres from the pavement to the crown of the vault, and in width nearly fifteen metres from centre to centre of its piers, but its design may justly be considered as the crowning glory of Gothic art, and the grand summing-up of the principles and constructive forms that had been gradually taking shape since the beginning of the twelfth century. 
FIG. 38. At first glance it may appear that the longitudinal rib is not so much stilted as are those of the buildings before noticed; but this is only because the capitals of their shafts are not situated at the true impost level. The arch is really stilted above these capitals, so that the vault surfaces continue to rise vertically for a considerable height above the springing of the great ribs, whose pressures fall directly upon the pier (Fig. 38), as in the previous examples.
The main vaulting shaft is now, for the first time in a quadripartite vaulting system, perfectly continuous from the pavement; that is to say, it is not sustained by a separate member in the lower pier, as at Paris, Chartres, and Reims; and it is not only continuous, but it also has the same diameter throughout. The shafts of the diagonal ribs rest upon the great pier capital as before, and those of the longitudinal ribs are brought down to the triforium ledge, uniting the clerestory and triforium into one composition in a manner which finds earlier illustration in the choirs of St. Germain des Prés and St. Remi of Reims, and further development in St. Denis and Sees. The abaci of the capitals are everywhere adjusted in form to the sections of the members which they carry; and those of the vaulting shafts are set in the directions of the ribs which they sustain.
I have remarked above that no Gothic building unites all the perfections of which the entire group affords illustration; but of the nave of Amiens it may be said that a more admirable carrying out of Gothic principles can hardly be imagined.
The nave of St. Denis resembles in some of its features, though not in its proportions, that of Amiens, to which it is not long subsequent in date. 
FIG. 39. Its vaults, however, are in form more like the earlier ones of which those of the Church of St. Leu d'Esserent are an instance. Their arches are of low sweep, and do not give to the vaults so much soaring expression as was common in contemporaneous constructions. The lower piers are largely a return to the Romanesque type—consisting of square members with engaged shafts, as in the section, Fig. 39. The three principal vaulting shafts rise without interruption from the pavement, while those of the longitudinal arches rest, as at Amiens, upon the triforium ledge. In the choir still another vaulting system is employed. Here the lower piers are plain round columns, against which the group of principal vaulting shafts rise, cutting through their capitals, and to these capitals the shafts of the longitudinal ribs descend, so that above the ground-story there are five shafts against each pier, as at Chartres and Reims.
The enormous, though ill-proportioned and yet magnificent choir of Beauvais presents, as it now exists, no new features in its vaulting system. The existing sexpartite vaults are not of the original design, but were probably constructed towards the end of the thirteenth century when, after serious ruptures had taken place in consequence of faulty construction, intermediate piers had to be introduced, and the whole design to be largely remodelled.[38]
FIG. 40
Minor differences of adjustment in the internal vaulting systems of the French churches of this time are no less frequent than in those of the earlier times; but the general principles which govern them are everywhere the same; and, as we have now considered the main types of construction, we need not investigate them further, but may pass to the consideration of the forms of external support which complete the anatomy of the Gothic structure.
One of the earliest remaining examples of a Gothic buttress system is that of the Church of St. Martin of Laon (Fig. 40). The pier buttress, a, which is incorporated with that portion of the internal pier upon which the vault thrusts are gathered, is a plain, square-edged mass of masonry reinforced by a flying buttress, b, which springs from the great outer buttress, c, the one inert member of the structure. The flying buttress is a square-edged rampant arch loaded with masonry, whose upper surface is brought into a right line which slopes a little less steeply than the chord of the arc, and is covered by a plain flat coping. The very massive lower buttress, c, is adjusted to the flying buttress, b, by a simple set-off, d, which penetrates the roof of the aisle, and is carried on a substructure of masonry over the transverse arch of the aisle, abutting against the pier at the springing of the vaults. These vaults are thus effectually propped both at their haunches and at their springing, but the whole construction is somewhat needlessly heavy.
Flying buttresses of lighter construction occur in the apse of St. Leu d'Esserent (Fig. 41), which dates from the last quarter of the twelfth century, or perhaps a little earlier. The pier buttress does not, in this system, rise above the abutting arch—the semicircular wall of the apse presenting an unbroken surface above this level, while below this point it is of the same form as that of St. Martin of Laon. The lower or outer buttress has three set-offs and rises to a considerable height above the roof of the aisle before the arch, whose intrados is set even with its inner face, springs. The back of the flying buttress is no longer a continuous sloping right line extending to the outer face of the system, but it is met by a level surface in which the outer portion of the buttress terminates, and the whole is covered, as at St. Martin, by a flat coping. In this case no portion of a lower prop is visible above the roof, though there probably is one beneath it.
Some improvements upon these forms are shown in the buttresses (Fig. 42) of the nave of the same building, which date from the end of the twelfth, or the beginning of the thirteenth century. The buttresses of the apse, by the number and depth of their set-offs, indicate that the builder considered their efficacy to depend upon a far -projecting foundation from which, by an inclined outline, somewhat of an oblique action, from the ground upwards, should be secured; but it was soon seen that the same result might be reached on a less projecting foundation by giving more weight to the upper part of the buttress. [39] Accordingly the outer faces of these buttresses of the nave rise more vertically and have a more equal volume at different levels. A further improvement which they exhibit is that of the gabled form
FIG. 41.
in which their tops terminate—a form which is better adapted for protection, as more readily shedding water, and one which is also more pleasing to the eye. The pier buttress is, like that of the apse, stopped at the level of the coping of the flying buttress, the wall above being entirely unbroken throughout its length, which is the case also in some other undeveloped Gothic buildings, as in the apse of St. Remi of Reims, and, I believe, also in that of St. Germain des Fres of Paris. This pier buttress is not, as in the preceding examples, in the form of a continuous pilaster-like member, but is based upon a substructure (a, Fig. 42), which rises through the roof and is carried above the transverse arch of the aisle vault.
FIG. 42.
Many of these early flying buttresses were ill adjusted from want of accurate knowledge where they should abut. It required repeated experiment to teach the precise points upon which they should be brought to bear. This flying buttress of St. Leu effectually meets the higher pressures exerted by the vault haunches, but those at the springing were not securely braced. It was apparently thought that the substructure a, including that portion of it which is beneath the aisle roof, would form a sufficient abutment to these lower pressures, but this did not prove to be the case. 
FIG. 43. The piers subsequently yielded at the springing, and it was found necessary to add a second arch beneath the first. Experience, in fact, showed that the pressures of a vault cannot be concentrated upon any single point, but only upon a line which extends over a considerable portion of the pier from the springing point upwards.
In the buttress system of the nave of Noyon (Fig. 43), which dates from the time of the reconstruction of the vaults early in the thirteenth century, the flying buttress assumes an improved form, in being both narrower and deeper, thus covering at once a greater vertical and a diminished lateral extent upon the pier,—a form more in accordance with the exigencies of the vault pressures. The intrados of the flying buttress, which in St. Leu is on a level with the impost of the longitudinal vault rib, is here in Noyon considerably below this level, while its upper part reaches as high as that of St. Leu; and instead of a shallow clerestory buttress terminating where the arch abuts, there is a vigorously salient one reaching to the top of the wall. The flying buttress is thus brought to bear upon a line (which is already considerably fortified by a pier buttress) rather than upon a point. Just what the form of the structure may be under the roof I am not aware; but as this nave has a high vaulted triforium gallery, there is doubtless an abutment of some kind carried over its vaults to meet the pier at the springing of the high vaults. It may be added that this buttress system has proved effectual, the vaults having stood without yielding for more than six hundred years. As a minor improvement the back of the flying as well as the top of the upright buttress assumes the gabled form, and a small finial marks the first attempt to render pleasing by ornament this latter important functional member.
FIG. 44.In the choir of Soissons, a monument but little subsequent in date to the time of the reconstruction of the buttress system of Noyon, still further improvements were made in the form of the flying buttress. Here two arches, one above the other, were established, and resistance to the thrusts of the vault was thus distributed vertically over a yet greater portion of the pier. The top of the inner half of the outer buttress is here carried up above the back of the flying buttress, helping by its weight to neutralise the vault pressures, and preparing the way for the pinnacle which was soon after introduced. The pier buttress assumes under each arch the form of an engaged shaft with base and capital. Shafts in this place had occurred earlier in the buttress systems of St. Remi of Reims and the choir of St. Germain des Prés, and they now became very frequent features.
The flying buttresses of Soissons were quickly followed by a beautiful variation on the same principle in those of the Cathedral of Chartres, where the superposed arches are united by an open shafted arcade. Henceforth the employment of two arches in the buttress system became practically constant; and perhaps the grandest illustration of the type is that which is afforded in the nave of Amiens (Fig. 44). In double-aisled buildings a double system of flying buttresses was introduced, one system over each aisle. In these cases the dividing piers of the aisles rise through the aisle roof, above which they receive the heads of the outer system of arches and give foothold to the inner one. The flying buttresses of the apse of Reims (Fig. 45) are of this form.
FIG. 45.
The evolution and adjustment of the pinnacle, which is so conspicuous a feature in this last example, was rapid after the device of weighting the top of the buttress had been introduced. At Chartres, where the superincumbent weight terminates in a truncated oblong pyramid in place of the gabled coping of Soissons, we get one of the intermediate steps of this development; but at Chartres this weight is still placed over the inner portion only of the buttress; it was presently seen, however, that it would be more effectual if set farther out upon it. Accordingly at Amiens (Fig. 44)[40] this weight is set flush with the outer face of the buttress, in the form of an upright square mass of masonry crowned by a steep pyramid, and the Gothic pinnacle stands forth in essential completeness. But the inventive faculties of the Gothic artists were fertile in variations upon this feature in which the constructive and decorative functions are so equally joined; and among the grandest results of their inventive activity are the pinnacles of Reims, which date from about 1240, and combine in one magnificent design the forms both of Soissons and Amiens. In this design the inner portion of the buttress, capped with a gable, rises far above the solid part of the outer portion, receiving the thrust of the upper abutting arch; while over the outer portion is an open-shafted canopy of elegant design, surmounted by a massive octagonal pyramid and by four lesser pyramids covering the angles of the square base on which they rest.
Thus the forms of these external supports, no less than those of the interior, were gradually developed as the mechanical exigencies involved were more and more perfectly apprehended. But, as with everything else in the Gothic system, a fine artistic spirit was always equally and simultaneously operative, which made beauty of form as imperative as constructive fitness; and hence these hard-working members are also among the most ornamental features of the Gothic edifice, so much so that their important constructive office has sometimes been largely lost sight of.[41] But in French Gothic, after 1160 at the latest, the stability of the structure is always absolutely dependent upon this member.
From the vaults and their internal and external supports, which together constitute the real structure of a French Gothic building, we may now pass to the development of the Gothic modes of enclosure. In early Gothic buildings massive walls filled the spaces between the piers much as they had done in Romanesque constructions. The openings remained small and were often even round arched, as at Senlis and Noyon. The nave of the Cathedral of Paris
FIG. 46.
affords a good illustration at once of the early forms of wall and opening, and of the changes that were quickly introduced as the Gothic idea began to take more distinct form in the minds of the builders. Of the two bays of the clerestory of that building shown in Fig. 46, [42] the one on the spectator's right retains its original form. It is the bay next the transept—the great pier, c, being one of the four piers of the crossing. In this bay the clerestory window is a simple pointed arched opening above the level of the springing of the vaults, and although larger than such openings had been in Romanesque design, it nevertheless is simply an opening in a wall, the area of the solid still 
FIG. 46. bis.being greater than that of the void. Beneath the clerestory is a circular opening, filled with a peculiar and beautiful form of tracery, occupying the space between the vaults of the triforium gallery and the timber roof which covers them. The whole design is one that exhibits a good deal of massive wall space, and an eye not quick to recognise leading structural features might not readily perceive that this is really a building whose stability resides not in its walls, but in its framework.
Early in the thirteenth century the original vaults of this nave, which had been completed in the preceding century, were damaged by fire and had to be repaired. [43] It would appear, indeed, that their lateral cells were wholly reconstructed and somewhat changed in form; for the longitudinal arches of the original cells which remain in place fall considerably below the present vault surfaces, [44] as may be seen in Fig. 46. Contemporaneously with this repairing and remodelling of these nave vaults great changes were making in other parts of the building—chiefly in the clerestory—in conformity with developments that had elsewhere taken place. These developments consisted chiefly in the enlargement of apertures, and in the dividing of them by mullions and simple forms of tracery. The apertures of the nave of St. Leu d'Esserent (Fig. 47) show the first step of this development in the grouping of two or more openings under one enclosing arch. Grouped openings had, indeed, been employed much earlier. Twin round-arched windows occur in the clerestory of Noyon; and in the triforium of
FIG. 47.
St. Germer (Oise) coupled round arches surmounted by a circular opening occur. The same arrangement is partly carried out at a still earlier date in the triforium of the Norman Church of Cerisy-la-Foret (Manche), where the circle over the arches is not pierced through the tympanum, but is formed by a moulding on its face. In the triforium of the nave of Noyon are coupled pointed arches with a pierced trefoil in the tympanum above. But already a new and far-reaching development of these germ forms had begun whose progress was most rapid. In the clerestory openings of St. Leu we have coupled pointed arches surmounted by a circle, within which is a thinner plane of masonry pierced with a six-foiled opening. Internally the longitudinal rib forms an encompassing pointed arch to the group, and externally there is a similar arch in the wall which projects beyond the plane of the grouped openings, throwing the design into two orders, as in Fig. 47. Windows like these of St Leu, though more enriched in having moulded archivolts and shafted jambs, occur also in the clerestory of the choir of Soissons. In these examples we have the radical forms of a great variety of subsequent Gothic apertures.
As I have before remarked, constructive development was the moving cause of change in every portion of the building, and the enlargement of the apertures was rather due to the nature of the construction than a result of endeavour to produce a beautiful and imposing effect, though such endeavour doubtless became also operative as constructive development went on.
The apertures of the clerestories of St. Leu and of Soissons were followed almost immediately by those of the apsidal chapels of the Cathedral of Reims, which date from about 1215, and though designed on the same general scheme, present an entirely new character (Fig. 48). For here, instead of solid spandrels between the main openings, we have lesser openings following the outlines of the larger ones; and thus instead of grouped openings we have rather a single opening divided by slender bars of stone. Moreover, these bars are no longer left, like the former pierced masonry, with plain flat surfaces, but are lightened and enriched by mouldings which, on the jambs and on the mullion, assume the forms of slender shafts with capitals and bases. In short, what may be called the pierced plane of masonry of St. Leu is here converted into Gothic tracery.
The great change referred to above, which was wrought soon after its first completion in the clerestory of Paris, consisted chiefly in substituting for the old ones openings like those of the apsidal chapels of Reims. But these new openings of Paris mark one further step in the development of tracery. The tracery of Reims, as will be seen in the figure, is made up of many small pieces of stone jointed as in ordinary arch
FIG. 48.
construction; but that of Paris is composed of fewer, and consequently larger pieces,—a method which rendered it possible to make the bars very much more slender, and yet to secure equal strength. One of these new clerestory windows is seen in the left bay in Fig. 46. To accomplish the change the pretty circular opening, seen in the unaltered bay to the right, had to be sacrificed; a string-course was inserted far below the springing of the vaults, down to which level the splays of the new openings were brought. The tracery is more simple than that of Reims, the sexfoil being omitted from the circle, and it is also more slender. The form of the window head is changed from that of the original window into a more acutely pointed arch, which is nearly concentric with the remodelled vault above it. [45] Both it and the vault disagree strikingly with the old longitudinal rib which remains undisturbed, affording an instance of the frankness with which traces of what previously existed are allowed to remain in mediæval work when occasion gives rise to changes or corrections. Such traces add much to the historic value of these monuments, and not unpleasantly correspond with what we frequently find in the works of old designers in the sister arts of painting and engraving, where corrections are made with little attempt at concealment, as in the case of the well-known horse's hoof in Dürer's engraving of the Knight and Death. [46]
In the case of these clerestory windows of Paris it may, I think, be questioned whether they were improvements to the building. There was a severe and simple beauty in the older design, with its unique rose in the upper triforium, that is not wholly compensated for by the tall new windows which do not fit very well into the old work. The windows, moreover, though their tracery, in its mode of construction and in its lightness, is an advance upon the tracery of Reims, are, in their relation to the building, not so distinctly Gothic; for the windows of Reims completely fill the spaces between the piers of the building, while those of Paris still leave a large wall space remaining.
Very soon after it had been recognised in the apse of Reims that the openings might safely be made equal to the entire spaces between the piers, the now growing art of painting on glass led to the universal practice of making them so. Vast and resplendent colour-designs in glass, softening the light and affording a grateful warmth of effect, thus became the leading mode of enrichment of the interior by colour.
The tracery by which these great openings were divided was constructively necessary to support the expanses of enclosing glass against the force of winds; and the greater the area of the opening the larger was the number of tracery bars required to afford this support. Hence the extensively subdivided tracery of the vast apertures of fully developed Gothic buildings—like the Cathedral of Amiens—which, by its very nature, afforded a beautiful mode of enrichment and adornment of the windows.
The enlargement of the clerestory opening, to the extent of doing away with the wall entirely beneath the vault rib, resulted in an important simplification of the construction, of which we have, I think, the first instance in the clerestory of the nave of Amiens, where the longitudinal rib and the archivolt of the opening, as has been pointed out by M. Viollet-le-Duc, become one and the same member, while the longitudinal rib shaft becomes a member of the window jamb.
As before noticed, the clerestory and triforium in the nave of Amiens are united into one grand composition by the descent of the longitudinal rib shafts and the shaft of the central mullion to the level of the triforium string. These three members divide each bay of the triforium into two bays, in each of which is a pointed arch encompassing a sub-order of three pointed arched openings. The same arrangement is repeated, with minor variations, in the nave of St. Denis. In this case each of the three mullions of the clerestory window descends through the triforium, dividing it into four bays—an arrangement which gives to the triforium a somewhat unpleasant effect by making its main divisions to consist of a series of upright rectangles.
The triforium passage in France, when it is not a vaulted gallery, is always enclosed by a thin screen of masonry, hiding the open space between the aisle vaults and the timber roof over them. Nothing but masonry is, therefore, to be seen in the interior of a French church, in consequence of which it has a strikingly consistent and monumental effect that is in some degree lost when the timbers of the triforium are exposed to view.
In the aisles, as in the clerestory and triforium, the openings, in the fully developed style, reach from pier to pier—a thin wall for enclosure, carried up to the window sills, being all the wall that is visible from the interior.
Thus were the walls of the former style practically suppressed, their place being taken by screens of painted glass, sustained by a slender framework of stone wrought into beautiful geometrical designs.
We have now examined the leading structural developments of French Gothic buildings as far as concerns their longitudinal bays, and it only remains to examine those of the eastern and western terminations, and also of transept ends, and of towers and spires.
The traditional semicircular apse, greatly enlarged and varied in the developed style by a polygonal form, is the characteristic eastern termination of French churches, though in exceptional cases this form gives place to the square, as at Laon Cathedral, where the original round apse was replaced at an early period by a square end.
A more beautiful termination than the round or polygonal apse, as designed by the mediæval architects of France, it is hard to conceive. No part of the Gothic edifice does more honour to these builders. The low Romanesque apse—vaulted as in older times with a plain semi-dome—presented no constructive difficulties, and produced no very imposing effect. But the soaring French chevet, with its many-celled vault, its divided stories, and its encircling aisles, taxed the utmost inventive and executive skill, and deserves universal admiration.
It is now hardly possible to trace the earliest development of the Gothic chevet. That of St. Denis was destroyed in the thirteenth century, and certainly not one of an earlier date exists complete. The apse of Senlis has lost its original vault, and is therefore serviceable as an example only so far up as the clerestory string. But the apse of the Cathedral of Noyon remains complete and in excellent condition. Like all early Gothic apses it retains the semicircular plan, which had been constant in Romanesque design. Its high vault (plan, Fig. 49) [47] is in five cells, whose ribs intersect upon the first transverse rib of the choir, and are abutted by the diagonal ribs of the first rectangular compartment. In order to effect the abutment this compartment is made tripartite—that is, its diagonal ribs, instead of each consisting of two branches abutting in the centre of the compartment, and thus, in plan,
FIG. 49.
following the diagonals of the rectangle, consist of but one branch which runs to the centre of the eastern transverse rib. These branches are abutted, however, by the first two ribs respectively of the apse compartment; and hence the two compartments together may be considered as forming one octopartite vault.
In elevation each bay of the apse is like a single bay of the straight part of the edifice, except that it is narrower and is on a curved plan; its vault cells also are substantially like the lateral cells of the rectangular compartments, except that those portions which are comprised between the ribs and the piers up as high as the springing of the narrow arches which span the ends of the cells, become merely vertical walls instead of having the conoidal form. In fact, the vaulting system and the forms of the lower portions of the vaults are in the apse very much as they are in the intermediate system in sexpartite vaulting.
FIG. 50.
The apse of the Cathedral of Paris is the next one of importance, and it is among the most admirable and majestic apses in France. Its high vault is, like that of Noyon, in five cells, and its ribs intersect, in the same manner, on the transverse rib of the adjoining rectangular compartment, being abutted as before by the ribs of this compartment (Fig. 50).[48] But beyond this there is a difference, which is worthy of notice, between these two constructions. In the choir of Noyon, the vaults being quadripartite, the ribs of the compartment adjoining the apse would not naturally have furnished an abutment for those of the apse; for the diagonals of a quadripartite vault intersect in the middle of the compartment. Therefore in order to secure abutment this compartment was made tripartite as we have seen, uniformity being sacrificed, in a truly Gothic spirit, to constructive exigency. But in Paris, and also at Bourges, the vaults of the choir being sexpartite, the plan is so arranged that a half of a double compartment shall adjoin the apse. The transverse rib upon which the apse ribs intersect thus corresponds to the intermediate rib of a sexpartite compartment, and hence this rectangular compartment is naturally tripartite, and forms a natural abutment to the compartment of the apse. For a sexpartite system no better arrangement could be devised.
Even for quadripartite vaulting the arrangement of Noyon, though logical and effectual, is not the best. The marked disparity in form and in the size of the respective cells, which it occasions between the easternmost compartment and the other compartments of the choir, was a defect which the builders were not slow to observe and to correct.
A better arrangement was discovered at Chartres (Fig. 51), and was afterwards more perfectly carried out at Amiens. At Chartres the centre of the arc of the apse is, as at Paris set eastward of the line spanned by the easternmost transverse rib; but instead of lengthening the apse ribs (as they are lengthened at Paris) to meet this transverse rib, they are shortened so as to meet a little eastward of the true centre, while additional ribs, a, are introduced, springing from the easternmost piers of the choir, whence they also converge upon the same point. These additional ribs thus furnish effectual abutment to the others, and the vault of the apse is rendered independent of the vaults of the choir. The choir vaults are now uniformly quadripartite; the awkward expedient adopted at Noyon being no longer necessary. At Chartres, too, the semicircular or segmental form is replaced by that of a polygon. The introduction of the two additional ribs in the vault gives eight instead of five cells to the apse, and the plan thus becomes a polygon of seven sides exclusive of the long side marked by the transverse rib which separates the apse from the choir.
FIG. 51.
The apse of Amiens (Fig. 52)[49] is perhaps the grandest in which this polygonal plan is carried out. And its plan, though perfectly simple, is most ingenious and perfect. The centre of the arc is in this case set to the eastward still farther than at Chartres, and upon this centre the ribs are made to converge. These ribs are now consequently all of the same length; and the whole effect is harmonious and admirable in the highest degree.
Among the finest earlier Gothic apses is that of St. Remi of Reims, which was constructed toward the end of the twelfth century. Below the clerestory it closely resembles the apse of Paris; its lower piers, its vaulting system, and the forms of ground-story and triforium arcades (like Paris it has a vaulted triforium gallery) being almost identical in design. It is, however, in advance of Paris as regards attenuation of supports and general lightness of construction. Like the Cathedral of Noyon it has a second triforium; and it is noticeable that this second triforium is united with the clerestory by shafts which embrace both stories, as at Amiens and St. Denis. The openings of the clerestory and the outer openings of the vaulted gallery are three in each bay; they are closely grouped, and entirely fill the spaces between the piers. This is not a common form of opening in France, and it is not strictly in accordance with the Gothic system; for although the openings quite fill the spaces between the piers (the dividing members having scarcely more volume than mullions), they do not quite fill the space beneath the vault rib, since their solid tympanums necessarily occupy a good deal of it.
FIG. 52.
There is hardly a construction in France of more interest than this portion of St. Remi as showing how the Gothic style is but an evolution from the Romanesque. Externally the plain round apse, the projecting apsidal chapels, and the general quietness of effect are strikingly reminiscent of the old monastic style; while yet the bold flying buttresses and the large openings, nearly suppressing the walls, bespeak a structure on strictly Gothic principles. This apse is another striking illustration, too, of the fact that the development of Gothic construction was from within outward, the internal arrangements necessitating those of the exterior, and the exterior assuming the Gothic expression first, as here shown, in the large constructive members, and finally in every detail. Fig. 53 and 54, portions of the interior and exterior, respectively, of this apse will illustrate these remarks.
FIG. 53.
The vaulting of the apsidal aisles presented difficulties which had been very embarrassing to the constructors previous to the employment of the pointed arch and the adoption of a system of supporting ribs. These difficulties grew out of the peculiar forms of the spaces to be vaulted—forms which presented great disparity between the spans of the pier archivolts and those of the wall arches. The intersection of a cone (which was necessarily the form of the transverse vault) with a barrel vault on a segmental plan
FIG. 54.
gave groins whose plans were wavy lines, and which were so salient in parts as to possess little strength. Many unsightly results were produced in the efforts of the Romanesque builders to construct such vaults at once securely and conveniently. [50] To illustrate these experiments would require more space than it is worth while to give to them in this connection. It is enough here to say that the difficulties which they attempted unsuccessfully to solve with the round arch were wholly removed, as we have seen in St. Denis, by the use of the pointed arch and of groin ribs. 
FIG. 55 In addition to what has already been shown concerning the vaults of St. Denis one further illustration of the flexibility of the Gothic system, in the construction of such vaults, is afforded by the Cathedral of Paris.
The form of each vaulting compartment of an apsidal aisle is, of course, trapezoidal; and where two such compartments adjoin each other concentrically, as they do in the double apsidal aisles of Paris (Fig. 55), there results an awkward difference in the lengths of their sides. The simple arrangement employed at St. Denis and elsewhere might still, of course, be employed; but the disparity in the spans of the arches and in the magnitudes of the cells would, on this arrangement, be very great. To obviate this the architect of the apse of Paris adopted another method whereby all the longitudinal arches are rendered of nearly equal span. This result is obtained by dividing the longest side of the compartment A into two parts by the insertion of a pier, a; and the longest side of the compartment B into three parts by the insertion of two piers, b and c. There are no intersections of the arches, and consequently no common keys in these compartments, each line in the plan representing a complete arch. [51]
In regard to chapels the Gothic apse in France differs considerably in different buildings according as the ecclesiastical, or the communal, influences predominated in the foundation. In the one case chapels are general and pronounced, while in the other they are either wholly wanting or are but slightly developed. The Cathedrals of Senlis, Noyon, and Chartres are instances of the first, and those of Paris and Bourges of the second class. In the more developed Gothic, as at Amiens and Beauvais, apsidal chapels become general, and add great richness to both the interior and exterior effects of the building.
FIG. 56.
The combination of apse, apsidal aisle, and apsidal chapels called for a wonderful degree of structural ingenuity, and led to some of the most charming interior effects. And among such combinations hardly any are more admirable than those which were constructed towards the close of the twelfth century, of which St. Leu d'Esserent (Fig. 56) affords a typical example.
The plans of nearly all large French churches include transepts. Bourges, among cathedrals of the first magnitude, is exceptional in having none. In some large buildings, as in Paris, the transept is of slight projection. In others, as at Noyon and Laon, it is more extended. In some cases it is without aisles, as at Paris and Noyon. In others it has aisles, as at Chartres, Amiens, and Reims. More rarely, as at Sens, there is an eastern, but no western aisle. And in some transepts, as in those of Laon and Sens, there are chapels opening out of the eastern aisle. The transept usually terminates in a square end; but at Noyon. both ends are apsidal, while at Soissons one end is round and the other is square. The round transept end is formed by a continuation of the side walls and buttresses, with their horizontal divisions. But in the usual square termination an appropriate façade is given, which is often of a very imposing character. This facade corresponds, in its main divisions, with the interior, being in three stories, exclusive of the gable. Where there are aisles the vertical divisions are marked by buttresses which divide the facade into three bays. The flying buttresses appear over the roofs of the aisles, and the whole structure is largely expressive of the form of the interior. On the ground-story of the central compartment there is a great portal, while the aisle ends usually have windows instead of doors. At the triforium level is a range of arched openings, while in the division above is usually a vast wheel window surmounted by a gable. The façade of the south transept of Amiens answers to this description completely. But every great church exhibits more or less conspicuous variations from every other in this, as in its other main parts. In some cases the transept portals are so vast and so richly adorned as to almost equal, as at Paris, and sometimes even to surpass, as at Chartres, those of the main façade. The transepts of Chartres are provided with vast and unique porches which, with their respective portals, are among the grandest architectural productions of the Middle Ages.
Of the majestic aspect of the great west end of a French Gothic cathedral too much in praise can hardly be said, notwithstanding that in it the constructive principles which distinguish the style, and which most excite our wonder and admiration, are least manifest. These façades are sometimes criticised on the ground that they disguise the true character of the structure behind them. It is, perhaps, true that an entirely satisfactory design for a western façade was hardly ever realised in a large Gothic church, though at Paris, Amiens, and Reims we have west fronts of magnificent, and for the most part appropriate, character.
It may be said in behalf of these designs that it is not an imperative principle that a façade should wholly express the structure of the building of which it forms the front. The façade rarely can do this in any architecture. But it may be admitted as a principle that unnecessary concealment of internal arrangements is an architectural offence; and hence those horizontal arcades which connect the towers in the grandest of Gothic fronts, hiding the forms of the gables behind them, may seem at first not easily defensible. It should be remembered, however, that the gable over a Gothic nave is not the true roof, and that the form of the vault is not wholly incongruous with the horizontal arcade. To the eye at least this arcade harmonises well with the total composition—a composition which is determined largely by the great lateral towers; and to the eye, if not in reality, this arcade performs the function of steadying the towers. As for the towers themselves, it would be hard to conceive any more appropriate termination for the aisles. And yet from the front view they quite mask the whole of that wonderful mechanism of flying buttresses which reveals so much of the distinctive character of Gothic art. But it is always easy to get a view which commands the whole structural system, and in such a view we are impressed with the majesty and appropriateness of the mighty towered western front. In fact, criticise it as we may, it is hard to see what better could be done. Without the towers the front would be wanting in that special emphasis and dignity which are appropriate to a great communal edifice. In the smaller and less important churches a treatment like that of transept ends was sometimes adopted. The small Church of Vaux-sous-Laon (Fig. 57) has a façade of this form. The whole structural character of the building is perhaps as fully expressed in this façade[52] as it could be; and where there is, as in this case, a transept and a central tower, the whole together makes a charming and impressive composition.
FIG. 57.
The practice of terminating the western extremities of the aisles of large churches by towers was established in the Romanesque period; and an instance of the simple character of the resulting façade is afforded by the Abbaye-aux-Hommes (Fig. 58). These towers are flanked by plain pilaster buttresses of two orders, which rise without set-offs to nearly the level where the towers disengage themselves from the central compartment of the façade. At this level they are banded by a string-course which continues across the entire front, and above it they are carried up three stories higher.[53] The façade is divided into three stories marked by plain string-courses. In the ground-story three round arched portals of modest dimensions, and each of three orders, open into the nave and aisles respectively; while in the central compartment are three round arched openings in each of the upper stories, and in each tower compartment is one such opening in each of the same stories. A low gable over the central compartment, with a single small arched opening, completes a design which is simple even to baldness.
FIG. 58.
It will be instructive to follow some of the changes which were wrought by the Gothic builders upon this simple scheme, and to see what different character they finally give it, retaining to the last these leading component elements. The development was less rapid here than in the other parts of the building; and, indeed, it was not till near the end of the twelfth century that the distinctly Gothic impress began to predominate over the Romanesque characteristics. Still the germs of the new style appear in the façade of St. Denis—in the larger dimensions of its recessed portals, in the presence of the pointed arch in some of its openings, in the large wheel window of the upper compartment of the central bay, and in the general character of its sculptured enrichments.
A more distinct approach to the Gothic type is shown in the façade of the Cathedral of Senlis (Fig. 5 9), which dates from the end of the twelfth century. Although in main features it is almost the same as that of the Abbaye-aux-Hommes, it nevertheless shows a new spirit—a spirit that bespeaks the vigorous activity of the Gothic genius. Here are the same plain square-edged tower buttresses dividing the front vertically into three bays. The central bay is divided horizontally into three stories by simply moulded string-courses, the upper one of which breaks around both towers and buttresses. The whole width of the central bay in the ground-story is occupied by a recessed portal in five orders with pointed arches, which is probably the earliest of the unparalleled series of portals of distinctly Gothic type which so distinguish the architecture of France in the twelfth and thirteenth centuries. Over this portal is a great pointed arched opening of four orders, which must originally have been divided by one or more shafts and a pierced tympanum, but whose present dividing members are incongruous interpolations of a later time. In the third story is a small circular opening of three orders—also filled with tracery of a later date,—and on either side of it is a pointed arched niche [54] of two orders, enshrining a statue. There is a smaller pointed doorway of three orders in the ground-story of each tower bay, the stilted and pierced tympanums of which are of curious design. Above the doorway in the south tower is a small pointed arched aperture of two orders, while the corresponding opening in the north tower is smaller and round arched. Above each of these openings there is an obtusely pointed blind arcade
FIG. 59.
of two arches on slender shafts, and over these again are small circular openings like that on the same level in the central bay. The tower buttresses here, unlike those of the Abbaye-aux-Hommes, continue up to the cornice of the first story above the roof of the nave. In this story the space between the buttresses is occupied by coupled pointed arched openings of two orders. The north tower was not completed above this level; but that on the south is surmounted by a spire of early thirteenth century construction, and of almost unrivalled beauty.
Unlike the main body of the building, the Gothic façade is largely the result of a mere modification and enrichment of Romanesque forms rather than a growth from new constructive principles. The façade with its towers is, for the most part, simply a storied edifice, which may, indeed, as at Reims, by the width of its openings, the slenderness of their dividing members, and the general upward impulse of its lines, attain a high degree of Gothic expression, but it is not a structure into which any dynamic principle largely enters. This being so, the criticism which early Gothic façades, like the Cathedral of Paris, sometimes receive, as wanting in Gothic character, is not well founded. The façade of Paris (Fig. 60) is one of the most Gothic, as it is also one of the grandest of structures of this kind. It is certainly the most remarkable façade that had been erected up to its time; and for dignity, sobriety, and extreme beauty of details it has hardly been surpassed. The general scheme of design is still the same as in the Abbaye-aux-Hommes; but such is the treatment of its elements that the Gothic spirit is manifest in every part. Its large divisions are admirably proportioned and beautifully subdivided. The Romanesque characteristics have completely disappeared from the apertures, arcades, and even from the moulding profiles, while in Senlis the Romanesque profiles are largely retained. Three majestic portals on the ground-story, above them a great central rose flanked on each side by twin pointed apertures and a circle embraced by an encompassing arch, a magnificent shafted arcade, sheltering twenty-eight sculptured statues, dividing these stories, an elegant, though gigantic, open arcade over all, and the two towers, each pierced by coupled pointed openings, rising one story higher, make up a most impressive architectural composition.
The magnificence, without extravagance, of the still richer façade of Amiens (Fig. 61) is, in some respects, beyond all praise. In addition to the leading features of 
FIG. 60
FIG. 6l.
The west fronts of Senlis, Paris, Amiens, and Reims sufficiently illustrate the development and the characteristics of the French-Gothic façade. Its typical form, as exhibited in the Cathedral of Amiens, is truly a marvel of architectural design. With the given elements and conditions it is hard to see how a more consistent and beautiful product could have been created. The arch, the shaft, the buttress, and the dividing string are employed in it not only with mechanical propriety, but with the most subtle artistic feeling. Indeed, by the simple adjustment of proportions the Gothic builders, in these structures, wrought wonders.
Nor ought we to convey the impression that structural developments were altogether wanting in this part of the building, though it is true that here such developments were fewer, and were hardly ever apparent externally.
M. Viollet-le-Duc has shown, [56] for instance, that in vast buttresses like those of the façade of Paris the tendency to settlement of the mortar beds is greater at the inner part, which is the more heavily weighted, than at the outer face. This inequality of settlement would tend to cause rupture were the buttress constructed of mere horizontal beds of masonry all of the same material. But if the weight falling on the lower portions of the inner part were diminished by a system of props carried up obliquely at different levels from the outer face, the tendency to rupture would be neutralised. Accordingly it has been found that the buttresses of the façade of Paris have an internal structure consisting of a superposed series of inclined masses of harder masonry which conduct a great part of the weight from the inner portion to the outer face of the buttress. To secure the outer face against yielding to the outward push of these props the bonding of the masonry is reinforced, at the levels where they find foothold, by cramps of iron. The buttress thus becomes really an organic structure, exhibiting something of the dynamic principle which reigns throughout the rest of the building.
Of external features none is more striking, and after the flying buttress, none shows more of the Gothic spirit, than the stone spire with which, in the design, if not in the executed work, the tower was crowned. It is a feature, too, 
FIG. 62 which, more emphatically perhaps than any other, marks the communal spirit and influence. The spire formed the governing feature in any general view of the mediaæval town, and was a sign of municipal power and prosperity. It was natural, therefore, that the spire should call forth the special enthusiasm and effort of the lay builders.
Before the twelfth century nothing like a true spire had been built. The French towers of the eleventh century, when roofed with stone, had these roofs constructed in the form of a low square pyramid, like those which still exist over the towers that flank the apse of Morienval (Fig. 62), and date from about the middle of the eleventh century.
In Normandy a more pointed pyramid frequently took the place of the low one, as at Basly and Rosel (Calvados), but it was still on a square base.
In the Ile-de-France, however, the true octagon spire, surmounting the square tower, with pinnacles occupying the angles of the square, occurs early in the twelfth century, as in the small churches of St. Vaast de Longmont, Chamant near Senlis, and others. Of these Chamant (Fig. 63) is especially interesting as exhibiting features which were afterwards magnificently amplified in the unique spire of the Cathedral of Senlis. These features are the acutely gabled dormers, [57] with pierced tympanums, one on each of the four faces of the octagon that are even with the tower walls, and, above these dormers, the openings pierced through each face of the octagon. 
FIG. 63. There are probably few spires of earlier date; and from such simple types the progress was surprisingly rapid—each new construction showing some innovation, and generally some improvement, upon what had been accomplished before. There were difficulties to be overcome of no slight magnitude. To manage the transition from the square plan of the substructure to the octagonal plan of the spire, so as to secure both stability and beauty, was by no means an easy matter when there were no precedents to guide the constructor.
If we regard these early spires as experiments, which indeed they were, in a new form of structure, we may well wonder at their cleverness; but if we judge them by the subsequent achievements in the building of spires, we appreciate the points in which they fail.
The adjustment of the octagon to the square, as exhibited in this spire of Chamant, was but partially successful. The transition is too abrupt—the upper story of the tower is not happily connected with its substructure and superstructure by any continuity of members; but a great improvement was very soon made, and the typical Gothic spire was brought into existence almost at one stride, in the Cathedral of Chartres.
The south tower and spire of this cathedral (Fig. 64) were constructed between 1140 and 1160. 
FIG. 64. Here the octagonal pyramid, instead of rising directly from the square tower, is raised upon an octagonal drum, and Four pinnacles built up vertically against the diagonal walls of this drum are based upon the angle buttresses of the tower, continuing their lines and crowned by pyramidal summits at the level where the great pyramid begins. Pointed arched openings in the four cardinal faces of the vertical octagon are surmounted by acute gables which rise far above the angle pyramids and abut the inclined walls of the spire. Coursed half-round mouldings strengthen and adorn the spire angles, [58] and a similar moulding is wrought upon the middle of each face. The transition from the square to the octagon is thus accomplished in the happiest manner; and the continuity of lines, all of which have a perfectly structural function and expression, is complete from the ground to the apex.
FIG. 65.
In spires of this form the diagonal walls of the octagon are sustained by squinches in the re-entering angles of the tower, and these, with their superstructure, bind and weight its walls, and thus help to consolidate the fabric; while the oblique pressures of the spire are reduced to a minimum by thinness of masonry, and by the weight of the abutting dormers.
Of early Gothic spires on a large scale few now exist. In many of the cathedrals and larger churches the progress of the works was arrested before these features were reached, as at Paris; and in others spires were constructed and subsequently destroyed, as at Laon. But such constructions as we have noticed are enough to show that the imaginative and mechanical resources of the Gothic artists were largely displayed in them as well as in other parts of the fabric.
In the general external aspect of the French-Gothic church, except in that which is presented by the front, the structural system is everywhere plainly expressed. We see at a glance that this is not an architecture of walls and trussed timber roofs, but that it is an architecture of vaults maintained by piers and buttresses. So marked is this appearance that M. Renan, in his excellent 
FIG. 66.
The general proportions of the exterior are sometimes criticised. But it is often forgotten that hardly any of the great early churches were completed according to the original designs; and that not one of them has come down to us without having undergone considerable and often very damaging alterations. The churches which were the most nearly completed, and have suffered least from alterations, are generally remarkable for justness and harmony of proportions.
As illustration of the general external aspect two examples may be taken—one an early and the other a later structure. The first is the Abbey Church of St. Leu d'Esserent (Fig. 66), and the second is the Cathedral of Reims (Fig. 67). Both exhibit parts which belong to different periods of construction; but these parts are, for the most part, of admirable character, and they assist in forming exceedingly grand and harmonious wholes.
In the view of St. Leu we have the apse, the apsidal chapels, the eastern towers, and the choir, all of which date probably from about 1170. [60] We have also the nave—which must have been completed before the close of the twelfth century, or very early in the thirteenth—and one of the western towers, which is a remnant of an earlier edifice. With exception of the western tower, which is ill adjusted to the reconstructed nave, the total composition is conspicuously fine in its groupings and just in its proportions. It is, in fact, one of the best remaining examples of the simple 
FIG. 67.
The general view of Reims presents a striking consistency and harmony of parts which is equalled by that of few other French cathedrals, notwithstanding that the total composition comprises portions that were wrought at successive periods from 1212 to the early part of the fifteenth century. The earlier and later portions, with exception of the gable of the transept, are hardly seen in the general view. The visible portions of the choir, the transept, and the nave were, for the most part, all built in the thirteenth century, after 1240, about which time the work was recommenced after a delay that followed the construction of the lower portions of the east end. [61]
These two examples afford an interesting comparison—the one showing the unadorned condition of external features which is characteristic of the time when structural exigencies were first being successfully met, and the other showing the richness of the full development—when the edifice stands forth clothed in a vast wealth of appropriate adornment. Each condition has its own proper charm; but there is something in the restraint, and even in the severity, of the early structure which holds our admiration almost more than the magnificence of the later one. [62]
We have now, perhaps, examined enough of these structural forms to have gained a fair understanding of them, and of their animating principles. They were the result of an unparalleled impulse which had its source in the social improvement which marked the eleventh century. The numbers of churches that were erected in the Ile-de-France during the first half of the twelfth century is astonishing. No other part of Europe can show anything like it. [63] And this activity continued and gathered force for at least three-quarters of a century after 1150.
We have examined but a few (though we have examined some of the most important) of the buildings of this remarkable epoch; but in nearly all the rest we should find substantially the same progress and the same leading characteristics. The movement was general throughout the region where it arose, though there was a promptness of development in some localities that was not shown in others. The spontaneous character of the movement is conspicuous also as well as its general prevalence. All the elements of the new style were products of the soil on which it grew, or of the immediately adjacent districts. Of imitation of foreign developments there was none; there was not anything elsewhere at this time to imitate. After the beginning of the twelfth century every traditional element of building was here subjected to a process of re-creation which was not superficial but radical.
It will be seen that the foregoing summary of the structural characteristics of French Gothic agrees with that which was given of Gothic construction in the preceding chapter. For, in fact, any correct definition of Gothic must be derived from analysis of the buildings of the Ile-de-France, as will, I think, plainly appear after we have examined the contemporaneous architecture of other countries.
A constant characteristic of this French Gothic is that structural and artistic principles go hand and hand in it. It does not appear that two independent processes were carried on in the minds of the designers—the one mechanical, and the other artistic. Artistic qualities are so wrought into the construction, have so little independent existence, that we feel the two principles to be inseparable in this art. The builders were artists; and they invariably wrought with a steady regard for and sense of proportion and harmony, not less than of mechanical propriety. I would especially emphasise this, lest from this somewhat lengthened examination of its structural growth it should be in any measure inferred that Gothic architecture was such a growth merely.
We have in this chapter followed with some fulness the development of French Gothic from its inception to its maturity. We have found the first distinct indication of the formation of a new style in the apsidal vaults of Morienval, succeeded by vaults constructed on thoroughly Gothic principles in the Abbey Church of St. Denis. We have seen the development extended in the Cathedrals of Senlis and Noyon. We have found in the Cathedral of Paris every functional member complete, and the Gothic system therefore fully developed, though more or less that is unessential, preventing the perfect expression of the Gothic spirit, still clings to it. We have next seen these lingering remains of Romanesque art—chiefly the extensive and massive wall enclosures—gradually eliminated, and the Gothic spirit more thoroughly pervading every part of the building in the early portions of Reims, in the remodelled portions of Paris, and finally in the nave of Amiens, where the transformation from the old style to the new is for the first time wholly accomplished.
In France, then, the Gothic style is found germinating by the beginning of the twelfth century in Morienval, has accomplished its structural transition by 1163 in the design for the choir of Paris, and has reached its fullest distinctive perfection by 1220 in the design for the nave of Amiens. [64]
In the next succeeding chapters we shall see how far and how early a similar novel system of building existed elsewhere in Europe.
- ↑ Three of these archivolts are pointed, and one is round, there being four arches in all to the apse arcade. The Report of the "Congres Archéologique de France" for the year 1877 thus alludes to this apse: "Nous voudrions aussi dégager de l'époque ogival certains églises où, malgre l'ogive et même la nervure, on trouve dans les moulures, dans les colonnes ou dans quelques dispositions générales, des formes qui rappellent par trop encore soit le XIe siécle, soit les premiers rudiments de la transition. Le chœur de l'église de Morienval mérite à cet égard d'être cité tout le premier: la nervure et l'ogive avaient eté fort peu pratiquées lorsqu'il fut construit, et peut-être, ou plutôt selon toute probabilité, c'est là que les habitants du Valois virent pour la première fois ces germes fécunds d'un nouvel art de bâtir."
- ↑ I am indebted for this diagram to Mr. G. F. Newton, who kindly took the pains to go for me from Paris to Morienval, and secure the data which my own notes had not fully included.
- ↑ This figure is copied from a drawing by M. Boesviswald, architect of the French Government, and published with his permission.
- ↑ More or less doming is a constant characteristic of French-Gothic vaults. The courses of masonry, from rib to rib, are all distinctly arched, and the ridges of the cells are consequently arched also, even in Amiens Cathedral.
- ↑ See Viollet-le-Duc, Dict. s.v. Trait, p. 201, et seq.; and Voute, pp. 503-505.
- ↑ The sanctuary is the semicircular part of the choir.
- ↑ Notre-Dame de Noyon. Par I. Vitet. Paris, 1845.
- ↑ See Église Ste. Trinité et I'Eglise St. Étienne à Caen. Par V. Ruprich Robert. Caen, 1884.
- ↑ In plain intersecting vaults the courses of masonry are, of course, all parallel with the axes, and hence the surfaces of the lateral cells are everywhere at right angles with the clerestory wall upon which they trace segmental curves. But when the lateral cell is divided, by an intermediate rib, into two smaller cells, these cells must have oblique axes; and their surfaces are necessarily also oblique. Moreover, the shell having to fit itself on to the triangle formed by the diagonal and intermediate ribs (which are portions of arches of different curvature) and the clerestory wall, has, of necessity, to assume an irregular shape. The efforts to keep the courses of masonry, as nearly as might be, parallel with the axes of the oblique cells may have been the cause of the elliptical forms of the wall arches.
- ↑ Figures 14, 15, and 16 are copied from M. V. Ruprich-Robert's Architecture Normande.
- ↑ See ViolIet-le-Duc, s.v. Travée, p. 242.
- ↑ A similar alternate arrangement of piers with intermediate shafts carried up to the top of the wall was not unfrequent in Norman buildings, both in Normandy and in England, as in Jumieges and Norwich. Constructively this is useless alike" for vaulting (as vaults would have to spring from a lower level) and for timber roofs (which require no shafts whatever). It bespeaks a want of constructive logic in the Norman builders which, as we shall see in the following chapter, frequently appears also in the Anglo-Norman pointed architecture.
- ↑ To adapt the main pier more perfectly to the vaulting the pilaster was cut away at a level about two metres below the springing of the vaults, and short lateral round shafts were here introduced to support the diagonal ribs; a corbel on each side being interposed.
- ↑ These illustrations include only what remains of the original construction. The clerestory piers and vaulting are, therefore, wanting, and the wall is left incomplete in the section because subsequent alterations have obliterated the old design in the parts omitted.
- ↑ The original capitals remain in place in the easternmost piers; the additional height to the springing of the present vaults being reached by building up on them.
- ↑ A. Thierry, Lettres sur l'Hist. de France, p. 223, et seq.
- ↑ The ecclesiastical elements of design appear in the pronounced apsidal chapels and the apsidal terminations of the transepts. It may be, too, that the round arches which mingle so curiously in this building with the pointed arches are concessions to ecclesiastical traditional preference.
- ↑ The existing vaults date from the thirteenth century (L. Vitet, Monographie de l'Eglise Notre-Dame de Noyon; and Viollet-le-Duc, s.v. Cathédrale) and are quadripartite. The original transverse ribs remain, and are alternately massive and slender in conformity with the requirements of sexpartite vaults.
- ↑ The vaults of the Abbaye-aux-Dames are really quadripartite over square divisions embracing two bays each, with a dividing transverse arch carrying a vertical wall up to the crown of the vault. This might almost seem like an experiment tending toward the true sexpartite form, were it not pretty well ascertained that they were of later construction than the true sexpartite vaults of the neighbouring Church of the Abbaye-aux-Hommes. The same form of vault occurs in several other Norman churches, as that of the Priory of St. Gabriel (Calvados), and that of Ouestreham, both of which are illustrated in M. Ruprich-Robert's Architecture Normande.
- ↑ Viollet-le-Duc, s.v. Construction, p. 103, refers to sexpartite vaults as constructed "suivant la méthode des premiers constructeurs Gothiques."
- ↑ Notre-Dame of Paris not only remains structurally in comparative perfection, but also retains a large part of its original sculpture. Within the building everything above the ground-story is the genuine untouched work of the twelfth and thirteenth centuries, and of the exterior the sculptures of the tympanums, the archivolts, and large portions of the jambs of the great portals remain as originally executed.
- ↑ The vaults of the choir are absolutely perfect; those of the nave have had to undergo some slight repairs since their partial reconstruction early in the thirteenth century.
- ↑ The present flying buttresses, consisting of arches which clear both aisles at a single span, are alterations of the thirteenth century.
- ↑ M. Viollet-le-Duc, s.v. Construction, remarks that the ill adjustment of the piers to the vaults in the Cathedral of Paris had in his early studies of the building greatly puzzled him; but that in the course of later investigations he perceived that the necessities of the sexpartite vault system were really provided for by the monolithic shafts which surround every alternate one of the piers which divide the aisles—those so reinforced being opposite the great piers which carry the main transverse ribs and the diagonals of the vaults. This is hardly a justification of the construction as it exists. For this mode of reinforcement does not accommodate itself to the extra weight as well as thrust that falls on the main piers. And it is a mode that does not appeal to the eye at all. Moreover, it exists in the nave only. The dividing piers of the choir aisles are all alike.
- ↑ Notre-Dame et ses premiers architectes, etc Par M. C. Bauchal. Paris, 1882.
- ↑ This is the usual arrangement of the vaulting capitals so long as the ribs retain a square section; but it shows the logic of the Gothic system to find that when the rib section is reduced to more or less of a wedge-shape—as it is in simple form at St. Denis, Senlis, and some other early buildings, and in more complex form at Amiens and elsewhere—then the central capital is, except in the earliest buildings, set diagonally, and the lateral ones are set even.
- ↑ See note 1, p. 55.
- ↑ The Cathedral of Laon was begun almost immediately after the establishment of the Commune in 1191. See Viollet-le-Duc, s.v. Cathedrale, p. 305.
- ↑ In two of the piers, on each side of the nave, the great cylindrical columns are each reinforced by four lesser shafts supporting the four angles of the abacus. This and many other variations of structure which occur here and in many other of these great buildings show an ever-ready disposition to experiment as new ideas were suggested.
- ↑ The existing edifice was commenced during the first years of the thirteenth century, but the vaults were not reached before 1230. Viollet-le-Duc, s.v. Cathédrale, p. 294; and s.v. Architecture, p. 235.
- ↑ The Carte des Monuments Historiques de France, prepared by the Commission of Historic Monuments for the French Government, is a valuable aid to the student in respect to the various schools and their mutual influences during the twelfth century.
- ↑ The sexpartite vaults of Bourges date from the latter part of the first half of the thirteenth century. Those of the choir of Beauvais are still later. But though thus sometimes still employed, the sexpartite form had now become as unusual as, nearly a century before, the quadripartite form had been.
- ↑ See, for instance, Ferguson's History of Architecture, vol. i. p. 517.
- ↑ In his essay on the Construction of the Vaults of the Middle Ages, published in the Transactions of the Royal Institute of British Architects for the Year 1842.
- ↑ Lectures on the Rise and Development of Medieval Architecture, vol. i. p. 63. London, 1879.
- ↑ This was the intention, but actually the flying buttresses are not brought to bear precisely on the points of greatest thrust in St. Leu d'Esserent, as will be seen farther on.
- ↑ The Cathedral of Reims was begun in 1212. Viollet-le-Duc, s~v. Cathedrale, P. 315-
- ↑ See Viollet-le-Duc, s.v. Cathcdrale, p. 336.
- ↑ See Viollet-le-Duc, s.v. Centre-Fort, p. 297.
- ↑ The upper portions of the buttresses of Amiens have been remodelled in the Flamboyant style. The pinnacle given in the illustration is taken from Viollet-le-Duc, s.v. Cathedrale, Fig. 20, p. 329.
- ↑ An eminent English architect recently remarked to me that, in his opinion, the flying buttress was not really necessary to the stability of a vaulted building, citing, in support of this opinion, the nave of Salisbury Cathedral, where the external flying buttress does not occur.
- ↑ This being a perspective view, looking upwards from the opposite triforium, all the forms appear a little foreshortened.
- ↑ Viollet-le-Duc, s.v. Cathédrale, p. 292.
- ↑ They do so in some, though not in all, of the bays. Indeed, great and very puzzling variations are exhibited in the different bays of this clerestory. For instance, in the first five bays on the north side of the nave, counting from the choir, the original longitudinal ribs are surmounted by other arches, in each of which the extrados is more acutely pointed than the intrados, which follows the form of the old arch, thus giving a more pointed form to the vault cell than it had originally. But the sixth, seventh, and eighth bays have their old arches raised by stilting to the new level, and thus these cells have the same form that they had originally. In the sixth and seventh bays the outline of the window head is not concentric with the archivolt, but is rendered more pointed by a singular filling-in, as shown in Fig. 46 bis.
- ↑ Many more particulars concerning the changes that were made in this building at this time are given by Viollet-le-Duc, s.v. Construction, and elsewhere; but those noticed above are, for the most part, not referred to by him.
- ↑ Motives of economy doubtless had much to do with the preservation of old work when changes were made in these buildings.
- ↑ This figure is copied from Vitet's Monograph on the Cathedral of Noyon.
- ↑ The arc of the plan of the apse is in this case more than half a circle, its centre being somewhat eastward of the point of intersection of the ribs. Two of these ribs are consequently longer than the others.
- ↑ Fig. 52 is copied from Viollet-le-Duc.
- ↑ See Viollet-le-Duc, s.v. Voute, p. 489, et seq.
- ↑ For full particulars regarding the mode, at once simple and ingenious, of filling in these triangular spaces, see Viollet-le-Duc, s.v. Voute, p. 512.
- ↑ I have seen this church from a distance only; and I am not sure whether this façade is at the eastern or the western end.
- ↑ The spires which now crown these towers do not, of course, belong to the original construction. The original roofs were replaced by them in the thirteenth century.
- ↑ The niche is not a Gothic feature, but as yet the Gothic façade was but beginning to develop.
- ↑ In later forms of Gothic this artifice is carried to excess, as at Rouen and Cologne.
- ↑ S.v. Construction, p. 158, et seq.
- ↑ Gabled dormers, though of a more primitive form, occur in the Norman pyramidal roof of St. Contest Calvados). If they occur elsewhere in Norman design they are certainly rare.
- ↑ Similar mouldings had been employed in some Norman constructions, as in the Tower of St. Loup (Calvados), figured in M. Kuprich-Robert's Architecture Nonnande.
- ↑ "Discours sur l'État des Beaux-Arts" in the Hist. Littéraire de la France au Quatorziéme Siécle. Par Le Clerc et Renan. Paris, 1865.
- ↑ I judge of the date by the style of the design alone. It is, on the one hand, so closely similar in character to the style of the eastern portions of the neighbouring Churches of Senlis and Noyon as to warrant the belief that the construction was not much subsequent to them in date; while, on the other hand, the style of the capitals of the interior is such as to indicate that they must have been wrought subsequently to those of the choir of Paris, which was begun in 1163.
- ↑ For further particulars concerning the dates of the different portions of this cathedral, see Viollet-le-Duc, s.v. Cathedrale, p. 315, et seq.
- ↑ Although St. Leu is a church of moderate dimensions, while Reims is a cathedral of the first magnitude, the comparison still affords a good illustration of the primitive and the mature characteristics respectively. Among the larger buildings of the time of St. Leu hardly any retain so much of their original aspect.
- ↑ A glance at the Carte des Monuments Historiques de France, indiquant les Écoles d'Art du territoire Français pendant la première moitié du XIIe Siccle, published by the French Government, shows this district thickly studded with churches, while in the neighbouring provinces they were more or less sparsely scattered.
- ↑ I take the dates of the beginnings of Paris and Amiens as marking the steps of progress indicated in the text, because the architects had unquestionably matured their schemes, as to their general characteristics, before the works were begun.