Page images
PDF
EPUB

small pinnacles, and a very beautiful battlement. The whole front is panelled, and the buttresses, which have a very bold projection, are ornamented with various tiers of nichework of excellent composition and most delicate execution. We may here incidentally notice that the east front is fine, but mixed with early English, which style extends to the transepts, while the nave and aisles are decorated, terminating with perpendicular, and finished with the west façade above noticed.

In concluding this portion, we cannot withhold naming the most elaborate work on the subject of this section, published from drawings made by C. Wickes, in S vols. fol. 1853– 59. Its chief drawback is that the illustrations are pictorial and not geometric, which might have been obviated by a plan and section to each. Our sketch of the varieties of towers and spires will be found filled up, in Rev. G. A. Poole's History of Ecclesiastical Architecture.

In Ireland, the Dominican Abbey, commonly called the Black Abbey, at Kilkenny, had a tower placed on the south of the altar in a most singular way. At the Franciscan Church, the tower was placed at the east end of the nave, with a chancel at the end; the tower was much narrower than the nave, but exactly the width of the lofty arch supporting it, so that now the roof has gone, the construction appears extremely bold and hazardous. This building was one of a numerous class. Except the round towers. which ceased to be built when the English went to Ireland, and the low Cistercian towers, the Irish churches up to that period were almost towerless. In a few instances other towers could be named, as the fine massive one of the Trinitarian Friary, at Adare; but suddenly, in the 15th century, it became the practice to build to the Franciscan and Dominican structures these lofty and slender additions. The nave was shut out from the choir by two transverse walls placed close together and pierced each with a narrow arch; above them rose the slender tower, standing as it were on the apex of the gables, instead of spreading over the width of the nave. They were finished with a peculiar battlemented parapet. There is no instance of two western towers to the medieval churches in Ireland; and a mediæval spire is not known to exist in that country.

In Scotland, the spires are chiefly of the middle pointed period, but not erected until about the middle of the 15th century. Short octagonal stone spires form a very common termination to towers of late date; they generally carry small pedimental headed 1 ghts either on all or on the cardinal faces, and are for the most part plain, though, as at Corstorphine, at Aberdeen, and at Crail, in Fifeshire, they are banded by two or three embattled strings or coronæ into stages. Sometimes, as at the two former places, there are small pinnacles at the angles; while at Corstorphine, and St. Andrew's at Aberdeen, a lumpish semi-pyramidal abutment on the angles is extremely suggestive of the brouch.

The construction of the tower and spire is of such importance as to require much attention. A tower built for the reception of bells intended to be rung, should have a solid foundation, not merely four arches nearly as wide as the tower itself, leaving four piers not much bigger than the thickness of the wall which they support. Bells require a tower to themselves, for it is known that they will spoil the best clock ever fixed. In Sir C. Wren's towers, and others built by his imitators, the substance of the walls was concentrated at the angles, leaving a moderate sized arch on each side, and only the same internal area as would exist in the case of four straight walls. This is sound construction, and is well displayed in the tower of Antwerp Cathedral. Such an arrangement also admits of a staircase being carried up in the substance of the wall, without diminishing strength, besides, a desirable object in some large towers, doing away with the necessity for buttresses. The tower, if thus carried up its whole height, will be more fit to support an octangular or circular spire or lantern. The mean internal area should be half the external area, and then, if well built and of good materials, the tower will safely bear as many bells as can be hung on one level.

There should be an offset to support the ringing floor and the bell floor, so that no timber be run into the wall to act as battering rams. Neither should a bell be hung on cross beams resting on the walls, but always in a trussed cage. As regards sound, one level of bells is considered better than two tiers. It is wonderful that some of the early brick or stone cones or pyramids (shown in fig. 1211.) have stood, for they were evidently built in level but gathering courses, even in the 11th century, around a light frame of timber, which was either removed or left to decay. As soon as the principle of diminution upward was acknowledged, two systems of construction presented themselves; the first is direct carriage of the upper storey from the basement floor; the other is a false-bearing; the weight being, in either case, thrown as much as possible upon the angles, even to the extent upon each floor of an opening in the centre of each side, which is the weakest part of a blank tower. In the first case there are two varieties, one being the pyramidal roof square on plan; the other being the pyramidal roof octagonal on plan. The latter, whether completed externally as a broach or otherwise, requires to be carried as low down the tower for support as possible; and in some cases, as at St. Léonard, in France, the octagon is more judiciously placed with four angles over the centres of the sides of the

tower, than with four faces over the corners of the tower, which then require to be loaded by pinnacles. These are set diagonally more advantageously than when square with the tower, because they thus have a larger base. The greater height given in the middle of the 12th century to the spire rendered such precaution inevitable; and at the same time it became evident that if the spire were to be no longer square on plan, it must not seem to rise abruptly out of a square.

Octagonal steeples, with octagonal spires not built through, but resting upon them, seem to be considered now as dangerous experiments in construction. Yet one at Guebwiller, in France, is a central steeple of four stages, including the pendentives. At Schelestadt is another of the same kind. This plan does not seem to have been in favour after the commencement of the 13th century.

When the French architects determined to trust their octagonal spires to the upper storeys of their steeples, they seem to have been careless about allowing the pendentives to approach points of weakness. The student will gather a good lesson on this point from the section of the steeple at the Abbaye de la Trinité, at Vendôme, given in Viollet le Duc's Dictionnaire. In the steeple of the cathedral at Chartres, the pendentives of the octagon sit upon the four pinnacles, which are thus each obliged to take a part of the weight of the spire; the other part being thrown upon the four faces of the octagonal drum, which are weighted by heavy gables. At the bottom the spire is 31 in. thick, and at top 113 in. in a length of 156 ft. 8 in., built of hard Berchère stone. The roofs of the pinnacles are 19 in. thick. It is to be noticed that the danger of a fall, which was so imminent as to cause the destruction of the steeple at St. Denis, is attributed in great part to the increase of weight given to it during a course of restoration, by using the stone of St. Pierre instead of that of Vergelé. Some French spires have a very curious effect, due to the presence of a simulated hip in the centre of their sides for the whole or part of the height: but still more extraordinary were the slits in that of St. Denis, and the slit with two transoms in that of St. Nicaise, at Reims,

The spire of the church at Langrune, near the sea-coast, north of Caen, in Normandy, has at its base in the interior, a sort of buttress of thin stone resting on the thicker walls of the tower, which runs up for a great height to each of the angles and sides of the spire. They are pierced so as to afford a free passage all round at the base of the spire; and may have been provided to assist in strengthening it on account of its exposed position. It has been drawn by Rev. J. L. Petit in his Architectural Studies.

It will be found that the stone spires of the 12th century were high in regard to the rest of the steeple. The proportions at St. Denis were 38 to 35; those at Chartres are 60 to 42; but in time these proportions were altered so much that the spires of St. Nicaise at Reims (end of 13th century), and those of the front of the cathedral in that city, are scarcely half the height of the tower instead of equal or superior to it. Murphy, in his account of the Batalha, remarks that no settled proportion seems to have been observed in the dimensions in general; they varied from four times the width of the base to eight times.

As regards the jointing of the stones of which spires are composed, their security seems to be wholly the result of an accurate working of the beds and vertical joints, and the adhesion of naturally good and properly applied mortar. In modern work it is questionable whether such aids as dowelling and cramping should be altogether dispensed with. Iron must not be used, for reasons given in an earlier portion of this work. One method used at present to steady and tie in the spire, is that of the insertion of an intermediate stage or floor of timber framing. Sir C. Wren, when rebuilding the upper portion of the (former) spire of Chichester Cathedral which had been forced out of the upright, placed two intermediate stages connected with a pendent beam of timber about 80 feet in length attached to the finial stone; each stage was about 3 inches less in diameter than the spire at their levels; these restored the spire if it departed from the upright. A similar pendulum, with two stages, to act in like manner, has been introduced by Gibbs in his Spire of St. Martin's in the Fields, London. Iron rods have of later years been used to effect this purpose.

When the beds of the stones are horizontal, one course of binders secured with dovetailed dowels will perhaps be enough in the height; but when the beds are inclined, two or three of these courses in its height would be an effectual means of preventing its spread. It has been considered that a spire is stronger when the beds are set at right angles to the face, but if not well set, water gets in, and sudden frosts do much injury. It is probable, however, that a large number of steeples would, were examination possible, be found to have been well chained with timber or with metal The former material appears to have been employed in the church at Châteauneuf (Sâone et Loire).

The spire, built cir. 1315, of St. Aldate's Church, Oxford, had to be taken down in 1865. The tower is about 56 feet high; the spire, about the same height to the weathercock, was for 10 feet down from it of solid stone, similarly to that shown in fig. 1213. The cause of its failure was that a 14-inch iron bar coupled at the angles and inserted in

the first course of stone 7 inches thick at the base of the spire, had rusted, in some place, entirely through, bursting the stone inside and out. The angle pinnacles alone sustained the spire for many years.

Nearly all the spires of Normandy are said to have been executed in thin slabs of stone; they are all about 7 inches thick at the bottom, and about 4 inches thick at the top, and are almost all executed in the Creuilly stone. In Caen, especially, that stone was employed in the steeples, though it had to be brought about 12 or 14 miles. The joints are (probably) set at right angles to the face of the stone. The spire at Batalha is about 7 inches thick, independent of the carved work, though almost a fourth part of its superficies is perforated its stones are said to be keyed together by means of dovetailed pieces of pine wood (Murphy). The slender stone ribs of the octagonal spire of Freiburg Cathedral are girded together at intervals of about 15 feet by means of double horizontal ribs or bands of limestone; in the middle of each of these bands an iron cramp is inserted, so that one half of the thickness of the metal is fixed in the under course of the stone-work, and the other half in the upper course, in order to prevent all thrust. The space between the rib and the horizontal bands is filled up with perforated tracery, so that the appearance of great lightness, united with great boldness, is imparted to the whole. Plate XI. of Moller's work shows a careful representation of the joints, explaining in what manner the stones are connected together, both in the principal members and the ornamental parts. The spires of Strasburg and Constance Cathedrals, and that of St. Stephen's Church at Vienna, present other examples of open work spires. The thickness of the decorated spire to the staircase in the north tower of the west front of Peterborough Cathedral, is about 11 inches at 2 feet above the wall of the tower, where the octagon commences, and is about 10 feet diameter (shown in Robson, Masons' Guide). The methods adopted of strengthening Salisbury spire and tower, are related by Price in his work published in 1750, who states that it is 400 feet high from the pavement to the extreme top, but to the top of the capstone or ball only 387 feet as previously noticed. It is only 9 inches thick at the bottom, diminishing to 7 inches.

The

The outline of a tower in elevation should be a parabolic curve, for strength as well as appearance, as it will not then present a top-heavy appearance, The difficulty in designing a tower and spire in the Roman or Italian style is to prevent a telescopic effect; and in the mediæval style the appearance of an extinguisher is too often obtained. entasis to the spire, and due diminution of the tower (though the former is usually held not to have existed, some spires being formed of two and even three lines at different angles), are desirable both for appearance and strength. They are common features in Essex and Middlesex, and the absence of them may be noticed by any one going from Essex into Suffolk, the round towers in which county have the entasis, but not those of later date. The tower of All Saints' Church, Colchester, possesses it, and diminishes from 21 feet to 19 feet, having internally an offset at each floor and at the roof, so that no timbers run into the walls.

A mathematical method of setting out the entasis for a spire was furnished by Mr. Thomas Turner, of Hampstead, to the Builder for 1848, through the late Professor Cockerell, R.A. But as he states that the ordinates may be obtained very nearly true by taking a thin lath and bending it to the extent required, we do not consider it necessary here to do more than to refer to the paper. In the reconstruction of the spire to St. Stephen's Church, at Vienna, an iron framework was introduced to support the light stone ribs, until near to the summit, which was made wholly of iron.

The iron spires at Rouen, Bruxelles, and Auxerre, are the only three we have noted.

[blocks in formation]

The introduction into this work of the investigation of the principles of proportion, as propounded by the late E. Cresy, renders it necessary that some preliminary details should be considered, before the student passes on to those pages. These details will consist of the result of the use of numbers, as given by the late Mr. Gwilt and appended to the previous editions, and of the enquiry by modern investigators into the use of the triangle and of the square during the medieval period. The subject is interesting, and a verv enticing

one, and we regret that our limited space will not allow us to do more than merely enter upon it. We would warn the student that should he feel inclined to devote any time to this subject himself, he must not be content with the measurements he may usually find in publications, but must found his theories on those taken by himself to be in any degree

certain of his deductions.

The plan on which the earlier Christian churches were constructed, wrote Mr. Gwilt, was that of a cross: he omitted to notice, however, the Italian basilican plan and the domical Greek plan; but he justly observes that (in western Europe) after the 10th century it would perhaps be difficult to find a cathedral deviating from a cruciform plan. At

Fig. 1219.

the beginning of the 9th century, in an inauguration (of a church) sermon, the preacher observes, "In dextro cornu altaris quæ in modum crucis constructa est;" and again, “ In medio ecclesiæ quæ est instar crucis constructa.” (Acta SS. Benedict.) Round churches, as at Aix la Chapelle, in Germany, Rieux and Merinville, in France, with Little Maplestead. Cambridge, Northampton, and the Temple Church, London, in England, are not enough in number to affect the rule. It was in the 13th century that the termination of the choir was changed from a circular to a polygonal form. The general ordonnance of the plan was, however, not changed, and seems almost to have sprung from the laws and proportions upon which surfaces and solid bodies are dependent. The square and its diagonal, the cube and its sides, appear, at least the latter or the side of the former (fig. 1219.), to furnish the unit on which the system is based. Hence the numbers 3, 5, and 7, become the governing numbers of the different parts of the building. The unit in the Latin cross, placed at the intersection of the nave, gives the development of a perfect cube, according to the rules of descriptive geometry. Here are found the number 3, in the arms of the cross and the centre square; the number 5, in the whole number of squares, omitting the central one; and the number 7, counting them in each direction. The foot, however, of the cross was, in time, lengthened to repetitions of five and six, and even more times. In monumental churches, formed on such a system, there necessarily arises an unity of a geometrical nature; and the geometrical principles emanating therefrom guided not only their principal, but their secondary, detail. Even before the 13th century there seems to have been some relation between the number of bays into which the nave was longitudinally divided, and the exterior and interior divisions whereof the apsis consisted; but after the introduction of the pointed style, this relation became so intimate, that from the number of sides of the apsis the number of bays in the nave may be always predicated, where the work has been carried out as it was originally designed. From the examination of many, indeed most, of the churches in Flanders, this circumstance had been long known to us; but for its first publicity, the antiquary is indebted, we believe, to M. Ramée, in 1843.

The connection of the bays of the nave with the terminating polygon of the choir was such, that the polygon is inscribed in a circle, whose diameter is the measuring unit of the nave, and generally of the transepts, and forms always the side of the square intercepted

Fig. 1220.

by them. It is most frequently octagonal (fig. 1220.), and generally formed by three sides of the octagon. When this is used, the governing number will be found to be 8, or some multiple of it. Thus, in the Abbaye aux Hommes, at Caen (this, however, is previous to the 13th century), the termination of the choir is by a double octagon, and the number of bays in the nave is eight. The same occurs at St. Stephen's, at Vienna; in the Church of St. Catherine, at Oppenheim; at Lichfield Cathedral; at Tewkesbury Abbey, and at almost every example that is known. It may be well here to observe, that English cathedrals, partly from their great deficiency in symmetry, on account of their not having been finished on their original plans, do not afford that elucidation of the theory that is found in those on the Continent. twenty-four instances of them we have sixteen in which the terminations are square instead of polygonal; when polygonal, the rule seems to have been always followed. It must be noted, however, that in contradistinction to the rest of Europe, England kept steadily, as a rule, to a square east end; and though at Canterbury and Tewkesbury, and a few other noted examples, the circular form appears, yet often, as at Peterborough and Westminster, the curved apse was capped with a rectilinear addition, protesting, as it were, against the foreign element.

In

An eastern termination of the choir in three bays may be produced from the octagon, by omitting the sides in the direction of the length of the building, as in fig. 1221. In fig. 1222. the three sides will be found to be those of a hexagon; and in this case the number 6

1

governs the other parts.

Examples of this arrangement are, the minster at Freiburg-im

Breisgau; the cathedral at Cologne, where the apsis is do

Fig. 1221.

decagonal, and there are six bays in
the nave; and the abbey at West-
minster, where the eastern end is
hexagonal, and there are found

twelve bays in the nave. In re-
spect of a nonagonal termination,

Fig. 1222.

the most extraordinary instance of a coincidence with the above-mentioned rules occurs in the duomo of Milan, commenced at the end of the 14th century. Its apsis is formed by three sides of a nonagon, and the bays in the nave are nine in number. One third of the arc contained under the side of an equilateral triangle seems to be the governing dimension. The number 3, submultiple of 9, pervades the structure. There are three bays in the choir, and the like number in the transepts The vault of the nave is subtended by an equilateral triangle. The lower principal windows are each designed in three bays The plan of the columns in the nave in each quarter contains three principal subdivisions, and, in a transverse section of the nave, the voids are just one-third of the solids. These are curious points, and much more worthy of investigation than many of the unimportant details which now-a-days so much occupy the attention of archaeologists. If the stem of the plant is right, the leaves and fruit will be sure to grow into their proper forms.

Figs. 1223. and 1224. show the decagonal terminations of an apsis. In the first, a side of the polygon faces the east; in the second, the angle of the polygon is on the axis of the

Fig. 1223.

church. The last case is of rare
occurrence. Examples of it are,
however, found in the church at
Morienval, and in the choir of
the dom-kirche of Naumburg.
The first case is illustrated by
a variety of examples—such are

Fig. 1224.

the cathedrals at Reims, Rouen, Paris, Magdeburg, and Ulm, with the churches of Ste. Elizabeth at Marburg, that at St. Quentin, &c., and, in this country, the cathedral at Peterborough; all of which have either five or ten bays in the nave. The dodecagon, as a termination, is subject to the same observations as the hexagon: indeed they were anticipated by the mention of the cathedral at Cologne. Under the figure of the heptagon must be classed the magnificent cathedral of Amiens, wherein seven chapels radiate round the choir end, and there are as many bays in the nave (fig. 237.). The choir at Beauvais is terminated by a double heptagon; and, had the church been completed, it would doubtless have had seven or fourteen bays in the nave. At Chartres, the choir is also terminated by a double heptagon, and the nave contains seven bays. In the duomo at Florence, the eastern termination is octagonal, and there are four bays in the nave; this is an example of the expiring Gothic style in Italy.

On an examination of the principal churches on the Continent, in and after the 13th century, it would appear that the practice of regulating the details was dependent on the number of sides in the apsis, or of bays in the nave. Thus, if the choir is terminated by three bays, formed on an octagonal plan, we find 3, or a multiple of it, is carried into the subdivision of the windows. So, if the number 5 is the dominant of the apsis, that number will be found transferred to the divisions of the windows; and in like manner the remainder is produced. There are two or three other matters affecting the monuments of art erected in and after the 13th century. The aisles are usually half the width of the nave, though instances occur where the width is equal. Mary churches have two apsides— such are the cathedrals at Nevers, and at St. Cyr; and in Germany, St. Sebald at Nuremberg; the dom-kirche at Mayence; the abbey church at Laach; the cathedrals of Bamberg, Worms, and others. So far Mr. Gwilt.

"It remains to observe," writes Professor Cockerell, in the Archæological Journal, 1845 "upon the mysterious numbers employed by Wykeham in the plans of his chapels at Winchester and Oxford, which are divided longitudinally by 7, and transversely by 4, equal parts. In the first, the chapel consists of 6 of these parts, and the ante-chapel of 1; in the second, the chapel consists of 5, and the ante-chapel of 2; the width being equal to 4, corresponding with the entire figure of the vesica piscis."

The recurrence of the number 7, "a number of perfection," is constant; accordingly we find it employed in the following remarkable instances, sometimes in the nave, and sometimes in the choir. In the cathedrals of York, Westminster, Exeter, Bristol, Durham, Lichfield, Paris, Amiens, Chartres, and Evreux; in the churches of Romsey, Waltham, Buildwas, St. Alban's (Norman portion), and Castle Acre; and in St. George's Chapel, at Windsor, Roslyn Chapel, and many others. See also the notice on page 1011.

« PreviousContinue »