RomanArchitectureAgreat deal of conjecture has been expended on the question as to the genesis ofthe Roman basilica. For present purposes it may be sufficient to observe thatthe addition of aisles to the nave was so manifest a convenience that it mightnot improbably have been thought of, even had models not been at hand in thecivic buildings of the Empire. The most suitable example that can be chosen astypical of the Roman basilica of the age of Constantine is the church of S. Maria Maggiore.
And this, not merely because, in spite of certain modernalterations, it has kept in the main its original features, but also because itdeparts, to a lesser extent than any other extant example, from the classicalideal. The lateral colonnade is immediately surmounted by a horizontalentablature, with architrave, frieze, and cornice all complete. The monolithiccolumns, with their capitals, are, moreover, homogenous, and have been cut fortheir position, instead of being like those of so many early Christian churches,the more or less incongruous and heterogeneous spoils of older and non-Christianedifices. Of this church, in its original form, no one however decidedly histastes may incline to some more highly developed system or style of architecturewill call in question the stately and majestic beauty. The general effect isthat of a vast perspective of lines of noble columns, carrying the eye forwardto the altar, which, with its civory or canopy, forms so conspicuous an object,standing, framed, as it mere, within the arch of the terminal apse, which formsits immediate and appropriate background.
S. Maria Maggiore is considerablysmaller than were any of the other three chief basilicas of Rome (St Peter’s,St. Paul’s, and the Lateran). Each of these, in addition to a nave of greaterlength and breadth, was furnished (as may still be seen in the restored StPaul’s) with a double aisle. This, however, was an advantage which was notunattended with a serious drawback from a purely esthetic point of view. For agreat space of blank wall intervening between the top of the lateral colonnadeand the clerestory windows was of necessity required in order to give support tothe penthouse roof of the double aisle.
And it is curious, to say the least,that it should not have occurred to the builders of those three basilicas toutilize a portion of the space thus enclosed, and at the same time to lightenthe burden of the wall above the colonnade, by constructing a gallery above theinner aisle. It is true, of course, that such a gallery is found in the churchof S. Agnese, where the low-level of the floor relatively to the surface of theground outside may have suggested this method of construction; but whereas, inthe East, the provision of a gallery (used as a gynaeceum) was usual from veryearly times, it never became otherwise than exceptional in the West. Taking Eastand West together, we find among early and medieval basilican churches examplesof all the combinations that are possible in the arrangement of aisles andgalleries.
They are the single aisle without gallery, which is, of course, thecommonest type of all; the double aisle without gallery, as in the three greatRoman basilicas; the single aisle with gallery, as in S. Agnese; the doubleaisle with single gallery, as in St. Demetrius at Thessalonica; and finally, asa crowning example, though of a later period, the double aisle surmounted by adouble gallery, as in the Duomo at Pisa. These, however, are modifications inthe general design of the building. Others, not less important, though they areless obviously striking, concern the details of the construction. Of these thefirst was the substitution of the arch for the horizontal entablature, and thesecond that of the pillar of masonry for the monolithic column.
The formerchange, which had already come into operation in the first basilica of St. Paulwithout the Walls, was so obviously in the nature of an improvement in point ofstability that it is no matter for surprise that it should have been almost. universally adopted. Colonnaded and arcaded basilicas, as we may call them, forthe most part older than the eleventh century, are to be found in the mostwidely distant regions, from Syria to Spain, and from Sicily to Saxony; and thelack of examples in Southern France is probably due to the destructive invasionof the Saracens and Northmen and to the building of new churches of a differenttype, in the eleventh and succeeding centuries, on the ruins of the old. Thechange from column to pillar, though in many cases it was no doubt necessitatedby lack of suitable materials — for the supply of ready-made monoliths frompagan buildings was not inexhaustible — proved, in fact, the germ of futuredevelopment; for from the plain square support to the recessed pillar, and fromthis again to the grouped shafts of the Gothic cathedrals of later times, theprogress can be quite plainly traced.
Mention should here be made of a class ofbasilican churches, in which as in S. Miniato, outside Florence, and in S. Zenone, Verona, pillars or grouped shafts alternate, at fixed intervals, withsimple columns, and serve the purpose of affording support to transverse archesspanning the whole width of the nave; a first step, it may be observed, tocontinuous vaulting. ROMANESQUE TYPES Something must now be said of the veryimportant alterations which the eastern end of the basilican church underwent inthe process of development from the Roman to what may conveniently be groupedtogether under the designation of “Romanesque” types. When, instudying the ground-plan of a Roman basilica, we pass from the nave and aislesto what lies beyond them, only two forms of design present themselves.
In thegreat majority of instances the terminal apse opens immediately on the nave,with the necessary result, so far as internal arrangements are concerned, thatthe choir, as we should call it, was an enclosure, quite unconnected with thearchitecture of the building, protruding forwards into the body of the church,as may still be seen in the church of S. Clemente in Rome. In the four greaterbasilicas, however, as well as in a few other instances, a transept wasinterposed between the nave and the apse, affording adequate space for the choirin its central portion, while its arms (which did not project beyond the aisles)served the purpose implied in the terms senatorium and matroneum. Now it isnoteworthy that the transept of a Roman basilica is, architecturally speaking,simply an oblong hall, crossing the nave at its upper extremity, and formingwith it a T-shaped cross, or crux immissa, but having no organic structuralrelation with it. But it was only necessary to equalize the breadth of transeptand nave, so that their crossing became a perfect square, in order to give tothis crossing a definite structural character, by strengthening the pieces atthe four angles of the crossing, and making them the basis of a more or lessconspicuous tower.
And this was one of the most characteristic innovation orimprovements introduced by the Romanesque builders of Northern Europe. In fact,however, before this stage of development was reached, the older basilicandesign had undergone another modification. For the simple apse, openingimmediately to the transept, church builders of all parts of Europe had alreadyin the eighth century substituted a projecting chancel, forming a fourth limb ofthe cross, which now definitively assumed the form of the crux commissa, bycontrast with the crux immissa of the Roman basilica. The earliest example of aperfectly quadrate crossing, with a somewhat rudimentary tower, appears to havebeen the minster of Fulda, built about A.
D. 800. It was quickly followed by St. Gall (830), Hersfeld (831), and Werden (875); but nearly two centuries were toelapse before the cruciform arrangement, even in the case of more importantchurches, can be said to have gained general acceptance (Dehio and v.
Bezold,Die kirchliche Baukunst des Abendlandes, I, 161). The differences which havealready been mentioned were, however, by no means the only ones whichdistinguished the Romanesque from the Roman transept. The transept of aRomanesque church, especially of those which were attached to monasteries, wasusually provided with one or more apses, projecting from the east side of itsnorthern and southern arms; and from this it appears, plainly enough, that thepurpose, or at least a principal purpose, of the medieval transept, was to makeprovision for subsidiary altars and chapels. A pair of transept apses,projecting eastwards, already makes its appearance at Hersfeld and Werden. AtBernay, Boscherville (St- Georges), and Cerisy-la-Forêt(St-Vigor), each arm of the transept has two eastern apses, correspondingrespectively to the aisle and to the projecting arm.
The same arrangement isfound also at Tarragona. At La Charité,a priory dependent on Cluny, each arm had three apses, so that there were sevenin all, immediately contiguous to one another, and varying in depth from thecentral to the northern and southern members of the system. The plan of Clunyitself was that of a cross with two transverse beams. Of the western transepteach arm had two apses; of the eastern each had three, two projecting eastwardsand one terminal. Saint-Benoît-sur-Loirehad likewise a double transept, furnished on the same principle with sixsubsidiary apses.
Among English cathedrals — it may here be mentioned — bothCanterbury and Norwich have a single chapel projecting from each arm of theirrespective transepts; and at E1y the “Galilee” porch, which has theform of a western transept, opens eastwards into two apsidal chapels, contiguouson either side to the main walls of the cathedral. Far more important in theirbearing on the later history of architecture than these developments of thetransept were certain changes which gradually took place in connection with thechancel. It is not unusual in Romanesque churches, to find the chancel flanked,like the nave, with aisles, terminating in apsidal or square-ended chapels. Butin more considerable edifices especially in France, the aisle is often carriedround as an ambulatory behind the chancel apse; and when this is the case, theambulatory most commonly opens into a series of radiating chapels.
These are, inthe earliest examples, entirely separate from one another, being sometimes twoor four, but more usually three or five, in number. In later examples the numberof chapels increases to seven or even nine; and they are then contiguous,forming a complete corona or chevet. The first beginnings of this system go backto so early a date as the fifth century. De Rossi has argued, apparently on goodgrounds, that some early Roman, Italian, and African basilicas were furnishedwith an ambulatory round the apse. This form of design, however, was soonabandoned in Italy, and in the Romanesque pre-Gothic period it cannot be said tohave been usual anywhere except in France, where it proved a seed rich with thepromise of future developments. The earliest instance of its adoption there wasalmost certainly the ancient church of St-Martin of Tours, as rebuilt by BishopPerpetuus in A.
D. 470. This edifice, as Quicherat has shown, had a semicircularambulatory at the back of the altar, in which, a few years later, was placed thetomb of Perpetuus himself. From Tours the type seems to have passed to Clermont-Ferrand(Sts.
Vitalis and Agricola), and thence, many centuries later, to Orléans(St-Aignan, 1029). Meanwhile, in 997, the church of St. Martin had been rebuilt,and in the foundations of this edifice, which can still be traced, we find whatis probably the earliest example of a chevet or corona of radiating chapels. Itserved, in its turn, in the course of the following century, as the model, inthis respect, of Notre-Dame de la Couture at Le Mans (c.
1000), St-Remi at Reims(c. 1010), St-Savin at Saint Savin (1020-30), the cathedral at Vannes (c. 1030),St-Hilaire at Poitiers (1049), and the abbey church at Cluny, as rebuilt in1089. Shortly before 1100 the church of St. Martin was once more rebuilt, on ascale of greater splendour; and once more the new building became the model forother churches, chief among which were those of St-Sernin at Toulouse (1096), ofSantiago at Compostela (c.
1105), and of the cathedral at Chartres (1112). ROMANESQUE VAULTING The history of ecclesiastical architecture in Western Europeduring the relatively short period which alone deserves to be regarded as one ofmore or less continuous and steady advance, and which extends, roughly speaking,from 1000 to 1300, may be described as the history of successive and progressiveattempts to solve the problem, how best to cover with stone vaulting a basilicanor quasi-basilican church, that is to say, a building of which the leadingfeature is a nave flanked with aisles and lighted with clerestory windows (Dehioand v. Bezold, op. cit. I, 296; Bond, op.
cit. , 6). It was the conditions ofthis problem, and the failure, more or less complete, of all previous attemptsto solve it satisfactorily, and by no means a mere aesthetic striving afterbeauty of architectural form, which led step by step to the development of theGothic architecture of the thirteenth century in its unsurpassed andunsurpassable perfection. The advantages of a vaulted, as compared with atimber, roof are so obvious that we are not surprised to find, dating from thetenth century or at latest from the beginning of the eleventh, examples ofbasilican churches with vaulted aisles.
Indeed these first attempts atcontinuous vaulting would probably have been made much earlier, but for theinvasions of Saracens and Northmen, which delayed till that period the firstbeginnings of a steady development in ecclesiastical architecture, but which bytheir wholesale destruction of pre-existing buildings may be said to haveprepared the way for that same development. The vaulting of the nave, however,in the case of any church of considerable size, was a very different matter; andit was not until the eleventh century was well advanced that the problem wasseriously faced. And when at last it was definitely taken in hand, this was doneunder pressure of dire necessity. Everyone who is at all conversant withmedieval chronicles, or with the history of the cathedrals of Western Europe,must be aware how extremely frequent were the disasters caused byconflagrations, and it was natural enough that the church-builders of the laterMiddle Ages should aim at making their buildings, at least relatively,fire-proof. The simplest form which the vaulting of a rectangular chamber cantake is, of course, the cylindrical barrel-vault; and this is, in fact, the formwhich was adopted in many of the earliest examples of vaulted roofs, especiallyin the south of France; a form, too, which was extensively used in Italy duringthe age of the Renaissance.
But, though simplest alike in conception and inconstruction, the cylindrical barrel-vault is in fact the least satisfactorythat could be devised for its purpose; and the objections which militate againstits employment are equally valid against that of the barrel-vault whose crosssection forms a pointed arch. Of these objections the chief is that thehorizontal thrust of a barrel-vault is evenly distributed throughout its entirelength. Theoretically, then, this thrust requires to be met, not by a series ofbuttresses, but by a continuous wall of sufficient thickness to resist theoutward pressure at any and every point along the line. Moreover, the higher thewall, the greater is the thickness needed, assuming of course that the wallstands free, like the clerestory wall of an aisled church.
Much, too, willdepend on the cohesiveness of the vaulting itself; and as the Romanesquechurch-builders were either unacquainted with, or unable to use, the methods bywhich the Romans and the Byzantines respectively contrived to give an almostrigid solidity to their masonry, it is no matter for surprise that in two largeclasses of instances they should have been content to sacrifice either theclerestory or the aisles to the advantages of a vaulted roof and to theexigencies of stability. Of aisleless churches indeed, we must forbear here tospeak. But of an important group of buildings which German writers havedesignated Hallenkirchen (hall- churches) a word must be said, as theyunquestionably played a part in preparing the way for the final solution of theproblem of vaulting. The most rudimentary form of hall-church is that in whichthe nave and aisles are roofed with three parallel barrel-vaults, those of theaisles springing from the same level as those of the nave.
Examples are found atLyons (St-Martin d’Ainay), at Lesterps, at Civray, and Carcassonne (St- Nazaire). An improvement on this design, in view of the illumination of the nave, consistsin giving to the vaulting of the aisles the form of a “rampant” arch,as at Silvacanne, and from this it was but a step to the arrangement by whichthe section took the form of a simple quadrant as at Parthenay-le-Vieux,Preuilly, and Fontfroide. This method of quadrant vaulting, as Viollet-le-Ducand others have observed, provides a kind of continuous internal “flyingbuttress”, though it is by no means certain that the idea of the flyingbuttress in the Gothic architecture of Northern France was actually suggested bythese Southern buildings. In point of stability.
the hall-churches of theeleventh century leave nothing to be desired. Their great defect is want oflight. And this defect almost equally affects a class of buildings which may bedescribed as two-storied hall-churches, and which are found principally, if notexclusively, in Auvergne and its neighbourhood. These are furnished, like a fewof the Roman basilicas and certain Byzantine churches, with a gallery, which isnot a mere triforium contrived in the thickness of the walls, but a chamber ofequal dimension with the aisle. This arrangement not only affords additionalspaces but also, by reason of the greater height of the edifice, might seem tofacilitate the provision of a more liberal supply of light, unimpeded byneighbouring buildings. This last mentioned advantage is, however, almostentirely negatived by the circumstance that, in this class of buildings, eachbay of the gallery is subdivided by means of coupled or grouped arches, so thatthe additional obstruction offered to the passage of the light almost entirelycounterbalance the possible gain through additional fenestration.
We say”the possible gain” because, in fact, the galleries of these churchesare but sparingly provided with windows. In these churches (which to the Englishreader should be of special interest by reason of their affinity in point ofconstruction to the Westminster cathedral) the aisle is usually cross-vaulted,while the gallery has a quadrant vault abutting in the wall of the nave justbelow the springing of the transverse arches. The most noteworthy examples arefound at Clermont-Ferrand (Notre Dame du Port), Issoire (St-Paul), and Conques. To the same family belongs moreover, the great church of St-Sernin at Toulousealready mentioned, which is distinguished from those previously named by havinga double aisle. At Nevers the church of St-Etienne resembles those at Clermont,Issoire, and Conques, except that it is provided with a range of upper windowswhich break through the barrel-vaulting, somewhat after the fashion whichafterwards became so common in Italy in churches of the Renaissance period. Theinherent shortcomings of the barrel-vault, especially when used as a roof forthe nave of an aisled church, have been sufficiently illustrated.
Thesedisadvantages, so far as structural stability and fenestration are concerned,might indeed be overcome by adopting the system of a succession of transversebarrel-vaults, such as are seen in the unique instance of the church of St-Philibertat Tournus. Such a construction is, however, “ponderous and inelegant, andnever came into general use” (Moore, Gothic Architecture, 42). The systemof cross-vaulting, which has now to be considered, may be regarded as acombination of longitudinal with transverse barrel-vaulting, inasmuch as it maybe described as consisting of a central barrel which is penetrated orintersected by a series of transverse vaults, corresponding of course to thesuccessive bays or compartments of the nave. The advantages of cross-vaultingare threefold. In the first place the total amount of the outward lateral thrustis very greatly diminished, since one half of it is now replaced by longitudinalthrusts, which, being opposed in pairs, neutralize one another. Secondly, allthat is left of the lateral thrust, as well as the longitudinal thrusts, and thewhole of the vertical pressure instead of being distributed throughout the wholelength of the building, is now collected and delivered at definite points,namely the summits of the columns or pillars.
Thirdly and lastly, a perfectlydeveloped system of cross-vaulting makes it possible so to heighten theclerestory windows that their archivolts shall reach the utmost interior heightof the building, and so to broaden them that their width between reveals mayapproximate very closely to the interval between column and column below. Bythese improvements (as ultimately realized in the perfected Gothic of thethirteenth century) the somewhat rudimentary design of the ancient Romanbasilica may be said to have reached the highest development of which it iscapable. The gradual development of cross-vaulting it is to be observed, did nottake place in those districts of Southern and Central France which had alreadybecome the home of the barrel-vault and to a less degree of the cupola, butfirst in Lombardy then in Germany, and finally in Northern France and inEngland. In these countries the evolution of the Romanesque timber-roofedbasilican church had — with local variations of course — reached a far moreadvanced stage than was ever attained in these regions in which the adoption ofbarrel-vaulting at a relatively early date had in a manner put a check onarchitectural progress.
And it is noteworthy that in Lombardy and Germany, whencross-vaulting was first adopted, its development was far less complete than inNorthern France, and that in like manner the advance towards perfection was bothless rapid and less complete in Normandy than in Picardy and the Ile-de-France. These two districts were the last to adopt the system, but it was here that itwas within the brief space of less than fifty years (1170-1220), brought to itsfinal perfection. The reason may probably have been, as Dehio and von Bezoldsuggest, that the architects of the Ile- de-France, in the days of PhilipAugustus and St. Louis, were less trammelled than those of Normandy by thetraditions of a school. The comparative lack of important architecturalmonuments of an earlier date left them, say these writers, a more open field fortheir inventive enterprise (op. cit.
I, 418). The simplest form ofcross-vaulting is of course that which is formed by the intersection of twocylindrical barrel-vaults of equal span. And this, without the use of ribbedgroining, was the method mostly adopted by the Roman builders in their civicedifices. In the case of a pillared or columned church, however, this method hadits disadvantages.
In particular, having regard to the dimensions of the aisleand its vaulting, the builders of Northern Europe had all but universallyadopted the plan of so spacing the columns and pillars which flank the nave thatthe intervals between them should be one-half the width of the church. Now theonly means by which an equal height could be given to vaults of unequal span wasthe use of the pointed arch; and so it came about that the pointed arch wasadopted, not primarily for aesthetic reasons, but rather for constructivepurposes. And the same is to be said of the use of ribbed groining. The medievalbuilders, who, as has been said above, possessed neither a tenacious mortar northe command of an abundant supply of rough labour, and who therefore could not– even had they wished it — have adopted the massive concrete masonry of theRomans, were driven by the very necessities of the case to aim at the same timeto depend for stability not on the cohesion of the materials, but on thereduction of thrusts to a minimum, and on their skilful transmission to pointswhere they could be effectively resisted. It was, then, plainly desirable tosubstitute for a vaulting of uniform thickness a framework of ribs on which acomparatively thin layer of stones (cut to the requisite curvature) could belaid, and as far as possible to lighten the whole construction by moulding theribs and likewise the columns which supported the vaulting.
The same principleof aiming at lightness of construction led to the elimination, as far aspossible, of arches of the nave. This was done by the enlargement of the windowsand the development of the triforium, till the entire building, with theexception of the buttresses, and of the spandrels below the triforium, became agraceful framework of grouped shafts and interlacing ribs (Moore, op. cit. , 17). The final stage in the evolution of architecture of the pointed arch was not,however, reached, until, for the solid Romanesque buttresses, which rested onthe vaulting of the aisles, and which were not only clumsy but often provedinadequate for their purpose, the genius of the Gothic builders hit upon theepoch-making device of the flying buttress. By means of this device the thrustof the main vaulting was not, indeed, as has been too often said, “met by acounter-thrust”, but was transmitted to the solid buttresses, mostlyweighted with pinnacles, which were now built outwards to a great distance fromthe aisles, and the spaces between which were sometimes utilized, and might withadvantage have been more often utilized, for a range of lateral chapels.
Thesubject of Gothic architecture in its details is, however, one that needsseparate treatment, and for present purposes this very inadequate indication ofsome of the general principles involved in its development must suffice. THECIRCULAR CHURCH AND ITS DERIVATIVES It was stated at the outset of the articlethat all ecclesiastical architecture may be said to have been devel- oped fromtwo primitive germs, the oblong and the circular chamber. Of those very numerouschurches, principally, but by no means exclusively, Eastern or Italian, whichmay be regarded as the products of the second line of development, we shallspeak very briefly. That a circular chamber without any kind of annex wasunsuitable for the ordinary purposes of public worship is plain enough. And themost obvious modification of this rudimentary form was to throw out a projectingsanctuary on one side of the building, as in St.
George’s, Thessalonica, or inthe little church of S. Tommaso in Limine, near Bergamo. It was hardly lessobviously convenient to build a projecting porch or narthex on the oppositeside, as in St. Elias’s, also at Thessalonica, and to complete the cross bymeans of lateral projection, as in the sepulchral chapel of Galla Placidia atRavenna.
Thus it was that churches having the form of a Greek cross, as well asother varieties of what German authors call the Centralbau, may be said to owetheir origin to a very simple process of evolution from the circular domedbuilding. Among the almost endless varieties on the main theme may be hereenumerated: buildings in which a circular, or polygonal, or quadrilateral aisle,whether in one or more stories, surrounds the central space, buildings in which,though the principal open space is cruciform, and the whole is dominated by acentral cupola, the ground- plan shows a rectangular outline, the cross being,as it were, boxed within a square; and buildings in which one of the arms of thecross is considerably elongated, as in the Duomo at Florence, St. Peter’s inRome, and St. Paul’s in London. The last-named modification, it is to beobserved, has the effect of assimilating the ground-plan of those greatchurches, and of many lesser examples of the same character, to that of theRomanesque and Gothic cruciform buildings whose genealogical descent from thecolumned rectangular basilica is contestable. Among ecclesiastical edifices ofhistorical importance or interest which are either circular or polygonal, or inwhich the circular or polygonal centre predominates over subsidiary parts of thestructure, may be mentioned the Pantheon in Rome, St.
Sergius at Constantinople,S. Vitale at Ravenna, S. Lorenzo at Milan, the great baptisteries of Florence,Siena, and Pisa, and the churches of the Knights Templars in various parts ofEurope. St.
Luke at Stiris in Phocis, besides being an excellent typicalinstance of true Byzantine architecture, affords a good example of the”boxing” of a cruciform building of the Greek type, by enclosingwithin the walls the square space between the adjacent limbs of the cross. Practically, however, the full development of cruciform from circular buildingsbecame possible only when the problem had been solved of roofing a squarechamber with a circular dome. This has in some cases been done by first reducingthe square to an octagon, by means of “squinches” or “trompettes”,and then raising the dome on the octagon, by filling in the obtuse angles of thefigure with rudimentary pendentives or faced corbelling. But already in thesixth century the architect and builder of Santa Sophia had showed for all timethat it was possible by means of “true” pendentives, to support adome, even of immense size, on four arches (with their piers) forming a square.
The use of pendentives being once understood, it became possible, not only tocombine the advantages of a great central dome with those of a cruciform church,but also to substitute domical for barrel- vaulting over the limbs of the cross,as at S. Marco, Venice, St-Front, Périgueux,and S. Antonio, Padua, or even to employ domical vaulting for a nave dividedinto square bays, as in the cathedral at Angouleme and other eleventh centurychurches in Perigord, in S. Salvatore at Venice, in the London Oratory, and(with the difference that saucer domes are here employed) in the WestminsterCathedral. Nor should it be forgotten that in the nave of St.
Paul’s, London,the architect had shown that domical vaulting is possible even when the bays ofnave or aisles are not square, but pronouncedly oblong. Indeed, if account betaken of the manifold disadvantages of barrel-vaulting as a means of roofing thenave of a large church, it may safely be said that the employment of some formof the dome or cupola is as necessary to the logical and structural perfectionof the architecture of the round arch as ribbed groining and the use of flyingbuttresses are necessary to the logical and structural perfection of thearchitecture of the pointed arch. SYSTEMS AND STYLES OF ARCHITECTURE IN RELIGIONTO MODERN NEEDS A word must now be said, in conclusion, as to the merits of theseveral systems and styles of architecture, more especially in relation to theneeds of our own day. Of systems, indeed, there are in truth only three, thetrabeate or that of which the horizontal lintel may be regarded as thegenerating element, and which of necessity postulates a timber roof; that of theround arch, which by virtue of the law of economy postulates, as has been said,the use of domical rather than barrel-vaulting and that of the pointed arch,which, if carried to perfection postulates ribbed groining and the use of theflying buttress. The second system, however, admits of two methods of treatmentwhich are sufficiently distinctive to be classed as two “styles”, viz. the neoclassical, or Renaissance, and the Byzantine, and which shall beparticularized presently.
Now the trabeate system, or that of the timber roof,may be very briefly dismissed. In the great majority of cases we must, indeed,of necessity be content with such a covering, for our churches; but no one wouldchoose a wooden roof who could afford a vaulted building. Again, the varioustypes of Romanesque architecture, with their imperfect and tentative methods ofvaulting, though historically of great interest, should be regarded as finallyout of court. On the other hands of the Gothic architecture of the thirteenthcentury as exemplified in the great cathedrals of Northern France and ofCologne, it mas be quite fearlessly asserted: that every single principle ofconstruction employed therein was the outcome of centuries of practicalexperience, in the form of successive and progressive attempts to solve theproblems of church vaulting; that the great loftiness of these buildings was notprimarily due (as has been sometimes suggested) to any mere Emporstreben, or”upward-soaring” propensity, but was simply the aggregate result ofgiving to the windows of the aisles and of the clerestory a height in suitableproportion to their width, and to the triforium a height sufficient to allow ofthe abutment of the aisle roof; and that every subsequent attempt to modify inany substantial particular, this perfected Gothic style, was of its natureretrogressive and decadent, as may be illustrated from the English perpendicularand the Italian and Spanish varieties of Gothic architecture. Nevertheless itmust be admitted that thirteenth-century Gothic, though perfect of its kind, hasits limitations, the most serious of which — in relation to modern needs — isthe necessarily restricted width of the nave.
When the architect of the Milancathedral attempted to improve on his French predecessors by exceeding theirmaximum width of fifty feet, and to construct a Gothic building with a navemeasuring sixty feet across it was found impossible, as the building proceeded,to carry out the original design without incurring the almost certain risk of acollapse, and hence it was necessary to depress the clerestory to its presentstunted proportions. Now under modern conditions of life, especially in the caseof a cathedral of first-class importance, a nave of far greater width is by allmeans desirable; and in order to secure this greater width it is necessaryeither to fall back on the unsatisfactory compromise of Italian or SpanishGothic, as illustrated in the cathedrals of Milan, Florence, or Gerona, or elseto adopt the principle of the round arch, combined, by preference, with domicalvaulting. This, as everyone knows, is what Mr. Bentley has done, with altogetherconspicuous success, in the case of the Westminster Cathedral. Of the design ofthis noble edifice it is impossible to speak here. But it may be worth while toindicate one main reason for the choice of the Byzantine rather than theneoclassic or Renaissance treatment of the round-arch system.
The principaldifference between the two is this: that, whereas the neoclassical style, by itsuse of pilasters, treats every pier as though it were a cluster of huge,flat-faced columns; the Byzantine boldly distinguishes between piers andcolumns, and employs the latter exclusively for the purposes which monolithicshafts are suited to fulfil, for instance the support of a gallery while thepiers in a Byzantine building make no pretence of being other than what theyare, viz. , the main supports of the vaulting. The Byzantine method ofconstruction was employed at Westminster has the further advantage that itbrings within the building the whole of the spaces between the buttressesthereby at the same time increasing the interior dimensions and avoiding theawkward appearance of ponderous external supports. Nor is the Byzantine style ofarchitecture suitable for a great cathedral alone; and one may venture to hopethat the great experiment which has been tried at Westminster will be fruitfulof results in the future development of ecclesiastical architecture.