instead of straight ones, in the ornaments of the buttresses and pinnacles. In this style the architects seem to have had an aversion to flat surfaces, as well as to right angles among their mouldings. They were fond of dividing the thickness, and increasing the apparent intricacy, by giving to each half a different ramification; making for instance two sets of mullions and tracery in one opening, one before the other, and totally without correspondence.[[8]] They divided the mouldings into separate parts, and placed those of their bases at different heights, one set of vertical mouldings passing between the bases of other vertical mouldings, and the bases of these again, are interrupted by the high plinths of the former bases, as if each penetrated the solid stone, and reappeared again where that did not cover it; many of these fancies are evidently taken from basket work.
The remaining fragment of the church of St. John the Baptist at Soissons, belongs to this style, and the new tower and spire at Chartres form a most beautiful specimen. The shrine work that surrounds the choir at Chartres is also an exquisite miniature example, which I shall mention more particularly hereafter. At about two leagues from Chalons, at a small village called L’Épine, is a little church of the fourteenth century. One tower only has been completed, and crowned with an elegant spire; but had the front been finished, it would offer perhaps the most beautiful specimen of Gothic external composition in the world. The arch of the doorway is large; even more so than the usual proportion in French churches, and its ornaments reach to the top of the rose window over it. The spire is short, with little flying buttresses at its base. It rises from an octagonal turret placed on the tower. Many of the parts themselves may be thought clumsy, but they are beautifully disposed, and every little defect vanishes in the perfection of the whole. Inside also it is an elegant building, if you except the white wash and yellow wash with which it is at present variegated. The front, including the two first arches of the nave, appears to be somewhat posterior to the rest of the church.
If we compare these examples with the buildings in our own country, we shall find the first nearly to correspond with the earliest specimens of what has been called the early English. The eastern parts of the cathedral at Canterbury form the best example I can cite to you; Salisbury Cathedral, and the transept at York, both agree with it in some particulars, while in others they approach to the second French style. Of this, after making some allowance for national differences, Westminster Abbey will furnish you a pretty good idea, or the eastern end of the cathedral at Lincoln. The nave at York would also belong to this style, excepting the vaulting and the west window. Of the third style good examples are rather deficient in England as well as in France; and perhaps it might be considered only as a variety of the second, yet it has a distinct and peculiar character. In our own buildings it is marked by a more complicated arrangement of the ribs on the vaulting; and in general it may be observed, that the English architects paid more attention to the enrichment of this part than the French. After this the two nations held a different course, and I can produce you no parallel to my fourth French style; nor have I met in France with any building like the choir at York, King’s College Chapel at Cambridge, or that of Henry the Seventh at Westminster.
I hope this general view of the subject will enable you to comprehend more easily my accounts of particular buildings, but in order to explain myself more fully, I shall request your attention to a very curious building of early date, which has been the subject of much controversy in France.
The church of St. Germain des Prés claims to be the oldest in Paris.[[9]] The first edifice was begun by Childebert in 557, and finished in 558, a degree of expedition which does not announce much magnificence; yet we are told that it was in the shape of a cross, and that the fabric was sustained by large marble columns, the ceiling was gilt, the walls painted on a gold ground, the pavement composed of rich mosaic, and the roof externally covered with gold. This description is by Gislemer, a monk of the abbey, who lived at the end of the ninth century, after the church had been twice burnt by the Normans; and perhaps it rather gives us the author’s opinion of what a church ought to be, than what this once was. The building however does not appear to have been totally destroyed, since Morard, who became abbot in 990, perceiving that the repairs since its ruin by the Normans, had been hastily and slightly executed, determined to pull it down entirely and rebuild it; and he is said to have had the satisfaction of completing it, nearly as it exists at present, before his death, which happened in 1014. We learn from the inscription which was formerly legible on his tomb, that he added to the church a tower, containing a bell (signum): this addition may seem to throw some doubt on the extent of the works executed by him. A dedication took place in 1163, but we cannot suppose the building stood complete and useless for all that period. The old cloister was taken down in 1227, and another begun and finished in the course of the same year by Eudes, the abbot. A new refectory was commenced in 1236, and in 1244 the great chapel of the Virgin was undertaken. These were executed from the designs of Pierre de Montereau, and are cited as proofs of his exquisite taste and skill. The Chapter House, and a beautiful chamber which adjoined it, were constructed about the same time, and the dormitory over them in 1273; but all these parts have been destroyed during the revolution. A new cloister was erected in 1555, but in 1579 the church is described as being much out of repair, and though some restorations and alterations were made in 1592, yet in 1644 it was in a most dilapidated and dangerous condition. The nave was covered with the fragments of the ceiling, and in parts with the tiles of the roof; the pavement was so sunk, that it was necessary to descend to it by steps; and the vaulting of the transept threatened to fall in. The whole of these deficiencies were repaired in the course of two years, the vaulting of the transept was renewed, and the nave for the first time vaulted with stone. The pillars were ornamented with composite capitals, some of the windows enlarged, a new doorway opened to the south, and an alteration made in the disposition of the choir, which seems to have been the only part of the fabric which had been kept in sufficient repair. As it now stands the church is not a very large one. The inside is low and gloomy;[[10]] in the nave and part of the choir, the piers consist of four half columns attached to a square pillar, the vaulting of the nave is slightly pointed, but the known recent date of this part renders its form of little consequence, nor is that of the choir of much more historical value. The piers of the chevet are cylindrical. All the arching is round, except that of the chevet, where the French and Whittington say it was pointed from necessity; but this is not very evident: the openings are smaller, but this is not the only way of carrying the arches to the same height. This may be done in the first place by making a Gothic arch formed from two centres, with a larger radius than the semicircular arches, (a pointed arch with the same radius would not rise so high) or with an arch from two centres, and the same radius on a base somewhat more elevated, or lastly, by a semicircular arch on a much more elevated base. The following diagram will explain this better than words.
To judge by the eye, the arches of St. Germain des Prés lie nearly in the middle between the second and third, i. e. between b and c. The base is considerably elevated by a perpendicular line continued above the capital, and the radius of the curve is smaller than that of the semicircle of the arches in the square part. As the architects have in some degree availed themselves of this elevated base, it is evident that they might, by doing it a little more, have preserved the semicircular form, and they must have been conscious that they had it in their power to do so. There is no gallery along the nave, but we find one round the choir, with square-headed openings. It has been much disputed whether this was, or was not the original form. M. Du Fourny contends, that as the first ceiling was of wood, and probably flat, it was highly natural that they should make these openings square-headed; but I think he is wrong. On the two towers at the entrance of the choir, we see openings, the lower parts of which are exactly similar to those abovementioned, and they are divided in the same manner by a little pillar; but these are arched above. It is probable that the arches have been removed in order to make room for the windows of the clerestory above, which in fact come down to them, but of which the lower part is filled up. All the windows are round-headed, (except those of the little chapels) without tracery or division of any sort, ornamented with a billetted moulding externally, and entirely plain within. Those of the chapels are pointed, but with the same ornament, and equally without tracery. There are some Saxon (or Norman) arcades below these windows; but there can be no doubt that these chapels in their present form (exclusively of the vaulting) are somewhat posterior to the church. The vaulting of the aisles is circular, and remarkably arched on the ridge, so as to present nearly a succession of portions of domes. The capitals, as usual in the Norman architecture, are very various, some resemble baskets, others are formed of a collection of figures of animals: some bear a resemblance to the Corinthian, but the masses are smaller in proportion to the size of the capital, and the relief less strongly marked. In this I think the artist judged rightly, and that the looseness of the Corinthian foliage would have been inconsistent with the massiveness of such a pillar. Here are also some decidedly composite capitals under the vaulting, but these were probably introduced in the repairs of 1644. Whittington says, that the proportions of the columns of the choir approach nearly to those of the Corinthian order. The shafts of the latter have full eight diameters in height: those of the former about four and a half.
The western tower is entirely incrusted in a wall of modern masonry except at the top, where we may observe a story of what we call Saxon architecture, with circular headed windows divided by little columns. In the other two towers, which flank the clerestory of the choir, the arches are also semicircular; but the openings are separated by piers, not by columns, and the workmanship of both, though somewhat differing, is more rude than that of the western tower. Judging by the little portion still exhibited, I should conclude this the latest of the three. Yet, as the masonry of these two ruder towers forms an essential part of the edifice, and the aisles are continued through their lower story, without exhibiting any difference of style in that part, we can hardly suppose them prior to the rest, more especially as the arches of the recesses, corresponding with the gallery of the choir, are surmounted with pointed arches. I cannot attribute this form to any alteration, because these arches do not correspond in style with any other restoration of the building. I must therefore be content to attribute the body of the church to Morard, excluding the vaulting, and to doubt about all the rest.
The western portal of this church exhibits the pointed arch. It is at present ornamented with shafts set in reveals, but some of them are restorations, and occupy the place either of statues, or of columns with statues attached to them: above is a series of ten small figures, whose faces have been broken by the Iconoclasts of the eighteenth century. The lower figures have been adduced as proofs of the antiquity of the tower, because they are supposed (two of them at least) to represent the family of Childebert, but the conclusion certainly does not follow from the premises, and I have no doubt that this portal, ancient as it is, was posterior to the body of the church. To make a theory for the chronology of this edifice from the dates we find in books, compared with the evidence of the architecture, we may suppose the bulk of the western tower to have been built in the eighth century, but nothing of this work remains exposed to view: the body of the church, northern tower, and lower part of the southern tower by Morard, between 990 and 1014: the upper part of the southern tower very shortly after. The upper part of the western tower followed. The western portal was certainly posterior to 1028: the reasons for fixing on this date are derived from the cathedral at Chartres. Bulliart does not give any representation of the arch of this doorway, and Whittington’s whole theory seems to indicate that he supposed it semicircular. I shall resume this subject in my observations on Chartres.