V. TRIUMPHAL ARCHES, TOMBS, AND OTHER BUILDINGS…

V. TRIUMPHAL ARCHES, TOMBS, AND OTHER BUILDINGS

CONTENTS.

Arches at Rome; in France—Arch at Trèves—Columns of Victory—Tombs— Minerva Medica—Provincial tombs—Eastern tombs—Domestic architecture— Spalato—Pompeii—Bridges—Aqueducts.

Triumphal Arches were among the most peculiar of the various forms of art which the Romans borrowed from those around them, and used with that strange mixture of splendour and bad taste which characterises all their works.

These were in the first instance no doubt borrowed from the Etruscans, as was also the ceremony of the triumph with which they were ultimately associated. At first they seem rather to have been used as festal entrances to the great public roads, the construction of which was considered one of the most important benefits a ruler could confer upon his country. There was one erected at Rimini in honour of an important restoration of the Flaminian way by Augustus; another at Susa in Piedmont, to commemorate a similar act of the same Emperor. Trajan built one on the pier at Ancona, when he restored that harbour, and another at Beneventum, when he repaired the Via Appia, represented in the preceding woodcut (No. 215). It is one of the best preserved as well as most graceful of its class in Italy. The Arch of the Sergii at Pola in Istria seems also to have been erected for a like purpose. That of Hadrian at Athens, and another built by him at Antinoë in Egypt, were monuments merely commemorative of the benefits which he had conferred on those cities by the architectural works he had erected within their walls. By far the most important application of these gateways, in Rome at least, was to commemorate a triumph which may have passed along the road over which the arch was erected, and perhaps in some instances they may have been erected beforehand, for the triumphal procession to pass through, and of which they would remain memorials.

The Arch of Titus at Rome is well known for the beauty of its detail, as well as from the extraordinary interest which it derives from having been erected to commemorate the conquest of Jerusalem, and consequently representing in its bassi-rilievi the spoils of the Temple. From the annexed elevation, drawn to the usual scale, it will be seen that the building is not large, and it is not so well proportioned as that at Beneventum, represented in the preceding woodcut, the attic being overpoweringly high. The absence of sculpture on each side of the arch is also a defect, for the real merit of these buildings is their being used as frameworks for the exhibition of sculptural representations of the deeds they were erected to commemorate.

In the later days of the Empire two side arches were added for foot-passengers, in addition to the carriage-way in the centre. This added much to the splendour of the edifice, and gave a greater opportunity for sculptural decoration than the single arch afforded. The Arch of Septimius Severus, represented to the same scale in Woodcut No. 217, is perhaps the best specimen of the class. That of Constantine is very similar and in most respects equal to this—a merit which it owes to most of its sculptures being borrowed from earlier monuments.

More splendid than either of these is the Arch at Orange. It is not known by whom it was erected, or even in what age: it is, however, certainly very late in the Roman period, and shows a strong tendency to treat the order as entirely subordinate, and to exalt the plain masses into that importance which characterises the late transitional period. Unfortunately its sculptures are so much destroyed by time and violence that it is not easy to speak with certainty as to their age; but more might be done than has hitherto been effected to illustrate this important monument.

At Rheims there is an arch which was probably much more magnificent than this. When in a perfect state it was 110 ft. in width, and had three openings, the central one 17 ft. wide by 40 ft. high, and those on each side 10 ft. in width, each separated by two Corinthian columns. From the style of the sculpture it certainly was of the last age of the Roman Empire, but having been built into the walls of the city, it has been so much injured that it is difficult to say what its original form may have been.

Besides these there is in France a very elegant single-arched gateway at St. Rémi, similar to and probably of the same age as that at Beneventum; another at Cavallon, and one at Carpentras, each with one arch. There is also one with two similar arches at Langres; and one, the Porta Nigra, at Besançon, which shows so complete a transition from the Roman style that it is difficult to believe that it does not belong to the Renaissance.

There still remains in France another class of arches, certainly not triumphal, but so similar to those just mentioned that it is difficult to separate the one from the other. The most important of these are two at Autun, called respectively the Porte Arroux and the Porte St. André, a view of which is given in Woodcut No. 218. Each of these has two central large archways for carriages, and one on each side for foot-passengers. Their most remarkable peculiarity is the light arcade or gallery that runs across the top of them, replacing the attic of the Roman arch, and giving a degree of lightness combined with height that those never possessed. These gates were certainly not meant for defence, and the apartment over them could scarcely be applied to utilitarian purposes; so that we may, I believe, consider it as a mere ornamental appendage, or as a balcony for display on festal occasions. It appears, however, to offer a better hint for modern arch-builders than any other example of its class.

Even more interesting than these gates at Autun is that called the Porta Nigra at Trèves; for though far ruder in style and coarser in detail, as might be expected from the remoteness of the province where it is found, it is far more complete. Indeed it is the only example of its class which we possess in anything like its original state. Its front consists of a double archway surmounted by an arcaded gallery, like the French examples. Within this is a rectangular court which seems never to have been roofed, and beyond this a second double archway similar to the first. At the ends of the court, projecting each way beyond the face of the gateway and the gallery surmounting it, are two wings four storeys in height, containing a series of apartments in the form of small basilicas, all similar to one another, and measuring about 55 ft. by 22. It is not easy to understand how these were approached, as there is no stair and no place for one. Of course there must have been some mode of access, and perhaps it may have been on the site of the apse, shown in the plan (Woodcut No. 219), which was added when the building was converted into a church in the Middle Ages. These apartments were probably originally used as courts or chambers of justice, thus realising, more nearly than any other European example I am acquainted with, the idea of a gate of justice.

Notwithstanding its defects of detail, there is a variety in the outline of this building and a boldness of profile that render it an extremely pleasing example of the style adopted; and though exhibiting many of the faults incidental to the design of the Colosseum, it possesses all that repetition of parts and Gothic feeling of design which give such value to its dimensions, though these are far from being contemptible, the building being 115 ft. wide by 95 in height to the top of the wings.

There probably were many similar gates of justice in the province, but all have perished, unless we except those at Autun just described. I am convinced that at that place there were originally such wings as these at Trèves, and that the small church, the apse of which is seen on the right hand (woodcut No. 220), stands upon the foundations of one of these. A slight excavation on the opposite side would settle this point at once. If it could be proved that these gateways at Autun had such lateral adjuncts, it would at once explain the use of the gallery over the arch, which otherwise looks so unmeaning, but would be intelligible as a passage connecting the two wings together.

Another form also is that of an arch at the entrance of a bridge, generally bearing an inscription commemorative of its building. Its purpose is thus closely connected with that of the arches before mentioned, which commemorate the execution of roads. Most of the great bridges of Italy and Spain were so adorned; but unfortunately they have either been used as fortifications in the Middle Ages, or removed in modern times to make room for the increased circulation of traffic. That built by Trajan on his noble bridge at Alcantara in Spain is well known; and there exists a double-arched bridge at Saintes, in the south of France. The most elegant and most perfect specimen, however, of this class is that of St. Chamas in Provence, represented in woodcut No. 221. It consists of two arches, one at each end of the bridge, of singular elegance of form and detail. Although it bears a still legible inscription, it is uncertain to what age it belongs, probably that of the Antonines: and I would account for the purity of its details by referring to the Greek element that pervades the south of France. Whether this is so or not, it is impossible not to admire not only the design of the whole bridge with its two arches, but the elegance with which the details have been executed.

Used in this mode as commencements of roads, or entrances to bridges, or as festal entrances to unfortified towns, there are perhaps no monuments of the second class more appropriate or more capable of architectural expression than these arches, though all of them have been more or less spoiled by an incongruous order being applied to them. Used, however, as they were in Rome, as monuments of victory, without offering even an excuse for a passage through them, the taste displayed in them is more than questionable: the manner, too, in which they were cut up by broken cornices and useless columns placed on tall pedestals, with other trivial details highly objectionable, deprive them of that largeness of design which is the only true merit and peculiar characteristic of Roman art, while that exquisite elegance with which the Greeks knew so well how to dignify even the most trivial objects was in them almost entirely lost.

COLUMNS OF VICTORY.

Columns of Victory are a class of monuments which seem to have been used in the East in very early times, though their history it must be confessed is somewhat fragmentary and uncertain, and they seem to have been adopted by the Romans in those provinces where they had been employed by the earlier inhabitants. Whatever the original may have been, the Romans were singularly unsuccessful in their application of the form. They never, in fact, rose above the idea of taking a column of construction, magnifying it, and placing it on a pedestal, without any attempt to modify its details or hide the original utilitarian purpose for which the column was designed. When they attempted more than this, they failed entirely in elaborating any new form at all worthy of admiration. The Columna Rostrata, or that erected to celebrate naval victories, was, so far as we can judge from representations (for no perfect specimen exists), one of the ugliest and clumsiest forms of column it is possible to conceive.

Of those of Victory, one of the most celebrated is that erected by Diocletian at Alexandria. A somewhat similar one exists at Arsinoë, erected by Alexander Severus; and a third at Mylassa in Caria. All these are mere Corinthian columns of the usual form, and with the details of those used to support entablatures in porticoes. However beautiful these may be in their proper place, they are singularly inappropriate and ungraceful when used as minarets or single columns.

There are two in Rome not quite so bad as these, both being of the Doric order. Had the square abacus in these been cut to a round form, and ornamented with an appropriate railing, we might almost have forgotten their original, and have fancied that they really were round towers with balconies at the top. The great object of their erection was to serve as vehicles for sculpture, though, as we now see them, or as they are caricatured at Paris and elsewhere, they are little more than instances of immense labour bestowed to very little purpose. As originally used, these columns were placed in small courts surrounded by open porticoes, whence the spectator could at two or perhaps at three different levels examine the sculpture at his leisure and at a convenient distance, while the absurdity of the column supporting nothing was not apparent, from its not being seen from the outside. This arrangement is explained in woodcut No. 200, which is a section through the basilica of Trajan, showing the position of his column, not only with reference to that building, but to the surrounding colonnade. The same was almost certainly the case with the column of Marcus Aurelius, which, with slight modifications, seems to have been copied from that of Trajan; but even in the most favourable situations no monuments can be less worthy of admiration or of being copied than these.

A far better specimen of this class is that at Cussi, near Beaune, in France. It probably belongs to the time of Aurelian, but it is not known either by whom it was erected or what victory it was designed to celebrate; still that it is a column of victory seems undoubted; and its resemblance to columns raised with the same object in India is quite striking.

The arrangement of the base serving as a pedestal for eight statues is not only elegant but appropriate. The ornament which covers the shaft takes off from the idea of its being a mere pillar, and at the same time is so subdued as not to break the outline or interfere with constructive propriety.

The capital, of the Corinthian order, is found in the neighbourhood used as the mouth of a well. In its original position it no doubt had a hole through it, which being enlarged suggested its application to its present ignoble purpose, the hole being no doubt intended either to receive or support the statue or emblem that originally crowned the monument, but of that no trace now remains.

There cannot be a more natural mode of monumental expression than that of a simple upright stone set up by the victors to commemorate their prowess and success. Accordingly steles or pillars erected for this purpose are found everywhere, and take shapes as various as the countries where they stand or the people who erected them. In Northern Europe they are known as Cath or battle-stones, and as rude unhewn monoliths are found everywhere. In India they are as elegant and as elaborately adorned as the Kutub Minar at Delhi, but nowhere was their true architectural expression so mistaken as in Rome. There, by perverting a feature designed for one purpose to a totally different use, an example of bad taste was given till then unknown, though in our days it has become not uncommon.

TOMBS.

In that strange collection of the styles of all nations which mingled together makes up the sum of Roman art, nothing strikes the architectural student with more astonishment than the number and importance of their tombs. If the Romans are of Aryan origin, as is generally assumed, they are the only people of that race among whom tomb-building was not utterly neglected. The importance of the tombs among the Roman remains proves one of two things. Either a considerable proportion of Etruscan blood was mixed up with that of the dominant race in Rome, or that the fierce and inartistic Romans, having no art of their own, were led blindly to copy that of the people among whom they were located.

Of the tombs of Consular Rome nothing remains except perhaps the sarcophagus of Scipio; and it is only on the eve of the Empire that we meet with the well-known one of Cæcilia Metella, the wife of Crassus, which is not only the best specimen of a Roman tomb now remaining to us, but the oldest architectural building of the imperial city of which we have an authentic date. It consists of a bold square basement about 100 ft. square, which was originally ornamented in some manner not now intelligible. From this rose a circular tower about 94 ft. in diameter, of very bold masonry, surmounted by a frieze of ox-skulls with wreaths joining them, and a well-profiled cornice: two or three courses of masonry above this seem to have belonged to the original work; and above this, almost certainly, in the original design rose a conical roof, which has perished. The tower having been used as a fortress in the Middle Ages, battlements have been added to supply the place of the roof, and it has been otherwise disfigured, so as to detract much from its beauty as now seen. Still we have no tomb of the same importance so perfect, nor one which enables us to connect the Roman tombs so nearly with the Etruscan. The only addition in this instance is that of the square basement or podium, though even this was not unknown at a much earlier period, as for instance in the tomb of Aruns (Woodcut No. 176). The exaggerated height of the circular base is also remarkable. Here it rises to be a tower instead of a mere circular base of stones for the earthen cone of the original sepulchre. The stone roof which probably surmounted the tower was a mere reproduction of the original earthen cone.

Next in age and importance was the tomb of Augustus in the Campus Martius. It is now so completely ruined that it is extremely difficult to make out its plan, and those who drew and restored it in former days were so careless in their measurements that even its dimensions cannot be ascertained; it appears, however, to have consisted of a circular basement about 300 ft. in diameter and about 60 ft. in height, adorned with 12 large niches. Above this rose a cone of earth as in the Etruscan tombs, not smooth like those, but divided into terraces, which were planted with trees. We also learn from Suetonius that Augustus laid out the grounds around his tomb and planted them with gardens for public use during his lifetime. More like the practice of a true Mogul in the East than the ruler of an Indo-Germanic people in Europe.

This tomb, however, was far surpassed, not only in solidity but in splendour, by that which Hadrian erected for himself on the banks of the Tiber, now known as the Mole of Hadrian, or more frequently the Castle of St. Angelo. The basement of this great tomb was a square, about 340 ft. each way and about 75 ft. high. Above this rose a circular tower 235 ft. in diameter and 140 in height. The whole was crowned either by a dome or by a conical roof in steps, which, with its central ornament, must have risen to a height of not less than 300 ft. The circular or tower-like part of this splendid building was ornamented with columns, but in what manner restorers have not been quite able to agree; some making two storeys, both with pillars, some, one of pillars and the upper one of pilasters. It would require more correct measurements than we have to enable us to settle this point, but it seems probable that there was only one range of columns on a circular basement of some height surmounted by an attic of at least equal dimensions. The order might have been 70 ft., the base and attic 35 ft. each.

Internally the mass was nearly solid, there being only one sepulchral apartment, as nearly as may be in the centre of the mass, approached by an inclined plane, winding round the whole building, from the entrance in the centre of the river face.

Besides these there was another class of tombs in Rome, called columbaria, generally oblong or square rooms below the level of the ground, the walls of which were pierced with a great number of little pigeon-holes or cells just of sufficient size to receive an urn containing the ashes of the body, which had been burnt according to the usual Roman mode of disposing of the dead. Externally of course they had no architecture, though some of the more important family sepulchres of this class were adorned internally with pilasters and painted ornaments of considerable beauty.

In the earlier ages of the Roman Empire these two forms of tombs characterised with sufficient clearness the two races, each with their distinctive customs, which made up the population of Rome. Long before its expiration the two were fused together so thoroughly that we lose all trace of the distinction, and a new form of tomb arose compounded of the two older, which became the typical form with the early Christians, and from them passed to the Saracens and other Eastern nations.

The new form of tomb retained externally the circular form of the Pelasgic sepulchre, though constructive necessities afterwards caused it to become polygonal. Instead however of being solid, or nearly so, the walls were only so thick as was necessary to support the dome, which became the universal form of roof of these buildings.

The sepulchres of Rome have as yet been far too carelessly examined to enable us to trace all the steps by which the transformation took place, but as a general rule it may be stated that the gradual enlargement of the central circular apartment is almost a certain test of the age of a tomb; till at last, before the age of Constantine, they became in fact representations of the Pantheon on a small scale, almost always with a crypt or circular vault below the principal apartment.

One of the most curious transitional specimens is that found near San Vito, represented in Woodcut No. 226. Here, as in all the earlier specimens, the principal apartment is the lower, in the square basement. The upper, which has lost its decoration, has the appearance of having been hollowed out of the frustum of a gigantic Doric column, or rather out of a solid tower like the central one of the Tomb of Aruns (Woodcut No. 176). Shortly after the age of this sepulchre the lower apartment became a mere crypt, and in such examples as those of the sepulchres of the Cornelia and Tossia families we have merely miniature Pantheons somewhat taller in proportion, and with a crypt. This is still more remarkable in a building called the Torre dei Schiavi, which has had a portico attached to one side, and in other respects looks very like a direct imitation of that celebrated temple. It seems certainly, however, to have been built for a tomb.

Another tomb, very similar to that of the Tossia family, is called that of Sta. Helena, the mother of Constantine. If it is not hers, it belongs at any rate to the last days of the Empire, and may be taken as a fair specimen of the tombs of that age and class. It is a vast transition from the tomb of Cæcilia Metella, though, like all the changes introduced by the Romans, it shows the never-failing tendency to transfer all architectural embellishments from the exterior to the interior of every style of building.

It consists of a basement about 100 ft. square, containing the crypt. On this stands a circular tower in two storeys. In the lower storey is a circular apartment about 66 ft. in diameter, surrounded by eight niches; in the upper the niches are external, and each is pierced with a window. The dimensions of the tomb are nearly the same as those of Cæcilia Metella, and it thus affords an excellent opportunity of comparing the two extremes of the series, and of contrasting the early Roman with the early Christian tomb.

The typical example of a sepulchre of this age is the tomb or baptistery of Sta. Costanza, the daughter of Constantine (Woodcut No. 423.) In this building the pillars that adorned the exterior of such a mausoleum, for instance, as that of Hadrian, are introduced internally. Externally the building never can have had much ornament. But the breaks between the lower aisle and the central compartment, pierced with the clerestory, must have had a very pleasing effect. In this example there is still shown a certain degree of timidity, which does not afterwards reappear. The columns are coupled and are far more numerous than they need have been, and are united by a fragment of an entablature, as if the architect had been afraid to place his vault directly on the capitals. Notwithstanding these defects, it is a pleasing and singularly instructive example of a completed transformation, and is just what we miss in those secular buildings for which the Christians had no use.

Another building, which is now known as the Lateran Baptistery (Woodcut No. 422), was also undoubtedly a place of sepulture. Its erection is generally ascribed to Constantine, and it is said was intended by him to be the place of his own sepulture. Whether this is correct or not, it certainly belongs to his age, and exhibits all the characteristics of the architecture of his time. Here the central apartment, never having been designed to support a dome, is of a far lighter construction, an upper order of pillars being placed on the lower, with merely a slight architrave and frieze running between the two orders, the external walls being slight in construction and octagonal in plan.[188] We must not in this place pursue any further the subject of the transition of style, as we have already trespassed within the pale of Christian architecture and passed beyond the limits of Heathen art. So gradual, however, was the change, and so long in preparation, that it is impossible to draw the line exactly where the separation actually took place between the two.

TEMPLE OF MINERVA MEDICA.

One important building remains to be mentioned before leaving this part of the subject. It commonly goes by the name of the Temple of Minerva Medica, though this is certainly a misnomer.[189] Recently it has become the fashion to assume that it was the hall of some bath; no building of that class, however, was known to exist in the neighbourhood, and it is extremely improbable that any should be found outside the Servian walls in this direction; moreover, it is wanting in all the necessary accompaniments of such an establishment.

It is here placed with the tombs, because its site is one that would justify its being so classed, and its form being just such as would be applicable to that purpose and to no other. It is not by any means certain, however, that it is a tomb, though there does not seem to be any more probable supposition. It certainly belongs to the last days of the Roman Empire, if indeed it be not a Christian building, which I am very much inclined to believe it is, for, on comparing it with the Baptistery of Constantine and the tomb of Sta. Costanza, it shows a considerable advance in construction on both these buildings, and a greater similarity to San Vitale at Ravenna, and other buildings of Justinian’s time, than to anything else now found in Rome.

As will be seen from the plan and section (Woodcuts Nos. 228 and 229), it has a dome, 80 ft. in diameter, resting on a decagon of singularly light and elegant construction. Nine of the compartments contain niches which give great room on the floor, as well as great variety and lightness to the general design. Above this is a clerestory of ten well-proportioned windows, which give light to the building, perhaps not in so effective a manner as the one eye of the Pantheon, though by a far more convenient arrangement, to protect from the elements a people who did not possess glass. So far as I know, all the domed buildings erected by the Romans up to the time of Constantine, and indeed long afterwards, were circular in the interior, though, like the temple built by Diocletian at Spalato, they were sometimes octagonal externally. This, however, is a Polygon both internally and on the outside, and the mode in which the dome is placed on the polygon shows the first rudiments of the pendentive system, which was afterwards carried to such perfection by the Byzantine architects, but is nowhere else to be found in Rome. It probably was for the purpose of somewhat diminishing the difficulties of this construction that the architect adopted a figure with ten instead of eight sides.

This, too, is, I believe, the first building in which buttresses are applied so as to give strength to the walls exactly at the point where it is most wanted. By this arrangement the architect was enabled to dispense with nearly one-half the quantity of material that was thought necessary when the dome of the Pantheon was constructed, and which he must have employed had he copied that building. Besides this, the dome was ribbed with tiles, as shown in Woodcut No. 230, and the space between the ribs filled in with inferior, perhaps lighter masonry, bonded together at certain heights by horizontal courses of tiles where necessary.

Besides the lightness and variety which the base of this building derives from the niches, it is 10 ft. higher than its diameter, which gives to it that proportion of height to width, the want of which is the principal defect of the Pantheon. It is not known what the side erections are which are usually shown in the ground-plans, nor even whether they are coeval with the main central edifice. I suspect they have never been very correctly laid down.

Taking it altogether, the building is certainly, both as concerns construction and proportion, by far the most scientific of all those in ancient Rome, and in these respects as far superior to the Pantheon as it is inferior to that temple in size. Indeed there are few inventions of the Middle Ages that are not attempted here or in the Temple of Peace—but more in this than in the latter; so much so, indeed, that I cannot help believing that it is much more modern than is generally supposed.

As might be expected from our knowledge of the race that inhabited the European provinces of the Roman Empire, there are very few specimens of tombs of any importance to be found in them. One very beautiful example exists at St. Rémi, represented in the annexed woodcut (No. 231). It can hardly, however, be correctly called a tomb, but is rather a cenotaph or a monument, erected as the inscription on it tells us, by Sextus and Marcus, of the family of the Julii, to their parents, whose statues appear under the dome of the upper storey. There is nothing funereal either in the inscription or the form, nor anything to lead us to suppose that the bodies of the parents repose beneath its foundation.

The lower portion of this monument is the square basement which the Romans always added to the Etruscan form of tomb. Upon this stands a storey pierced with an archway in each face, with a three-quarter pillar of the Corinthian order at every angle. The highest part is a circular colonnade, a miniature copy of that which we know to have once encircled Hadrian’s Mole.

The open arrangement of the arches and colonnade, while it takes off considerably from the tomb-like simplicity appropriate to such buildings, adds very much to the lightness and elegance of the whole. Altogether the building has much more of the aspiring character of Christian art than of the more solid and horizontal forms which were characteristic of the style then dying out.

Another monument of very singular and exceptional form is found at Igel, near Trèves, in Germany. It is so unlike anything found in Italy, or indeed anything of the Roman age, that were its date not perfectly known from the inscription upon it, one might rather be inclined to ascribe it to the age of Francis I. than to the latter days of the Roman Empire.

The form is graceful, though the pilasters and architectural ornaments seem somewhat misplaced. It is covered with sculptures from top to bottom. These, however, as is generally the case with Roman funereal monuments, have no reference to death, nor to the life or actions of the person to whom the monument is sacred, but are more like the scenes painted on a wall or ornamental stele anywhere. The principal object on the face represented in the woodcut is the sun, but the subjects are varied on each face, and, though much time-worn, they still give a very perfect idea of the rich ornamentation of the monuments of the last age of the Empire.

The Tour Magne at Nîmes is too important a monument to be passed over, though in its present ruined state it is almost more difficult to explain than any other Roman remains that have reached our times. It consists of an octagonal tower 50 ft. in diameter, and now about 120 ft. high. The basement is extended beyond this tower on every side by a series of arches supporting a terrace to which access was obtained by an external flight of steps, or rather an inclined plane. From the marks in the walls it seems evident that this terrace originally supported a peristyle, or, possibly, a range of chambers. Within the basement is a great chamber covered by a dome of rubble masonry, to which no access could be obtained from without, but the interior may have been reached through the eye of the dome. From the terrace an important flight of steps led upwards to—what? It is almost impossible to refrain from answering, to a cella, like those which crowned the tomb temples of Assyria. That the main object of the building was sepulchral seems hardly doubtful, but we have no other instance in Europe of a tomb with such a staircase leading to a chamber above it.

That Marseilles was a Phœnician and then a Phocian colony long before Roman times seems generally to be admitted, and that in the Temple of Diana (Woodcuts Nos. 188 and 189) and in this building there is an Etruscan or Eastern element which can hardly be mistaken, and may lead to very important ethnographical indications when more fully investigated and better understood.

EASTERN TOMBS.

This scarcity of tombs in the western part of the Roman Empire is to a great extent made up for in the East; but the history of those erected under the Roman rule in that part of the world is as yet so little known that it is not easy either to classify or to describe them; and as nearly all those which have been preserved are cut in the rock, it is sometimes difficult—as with other rock-cut objects all over the world—to understand the form of building from which they were copied.

The three principal groups of tombs of the Roman epoch are those of Petra, Cyrene, and Jerusalem. Though many other important tombs exist in those countries, they are so little known that they must be passed over for the present.

From the time when Abraham was laid in the cave of Machpelah until after the Christian era, we know that burying in the rock was not the exception but the general practice among the nations of this part of the East. So far as can be known, the example was set by Egypt, which was the parent of much of their civilisation. In Egypt the façades of their rock-cut tombs were—with the solitary exception of those of Beni Hasan[190]—ornamented so simply and unobtrusively as rather to belie than to announce their internal magnificence. All the oldest Asiatic tombs seem to have been mere holes in the rock, wholly without architectural decorations.

We have seen, however, how the Persian kings copied their palace façades to adorn their last resting-places, and how about the same time in Lycia the tomb-builders copied, first their own wooden structures, and afterwards the architectural façades which they had learned from the Greeks how to construct. But it was not till the Roman period that this species of magnificence extended to the places enumerated above; when to such an extent did it prevail at Petra as to give to that now deserted valley the appearance of a petrified city of the dead.

The typical and most beautiful tomb of this place is that called the Khasné or Treasury of Pharaoh—represented in elevation and section in the annexed woodcuts, Nos. 233 and 234. As will be seen, it consists of a square basement, adorned with a portico of four very beautiful Corinthian pillars, surmounted by a pediment of low Grecian pitch. Above this are three very singular turrets, the use and application of which it is extremely difficult to understand. The central one is circular, and is of a well-understood sepulchral form, the use of which, had it been more important, or had it stood alone, would have been intelligible enough; but what are the side turrets? If one might hazard so bold a conjecture, I would suggest that the original from which this is derived was a five-turreted tomb, like that of Aruns (Woodcut No. 176), or that of Alyattes at Sardis, which in course of time became translated into so foreign a shape as this; but where are the intermediate forms? and by whom and when was this change effected? Before forming any theories on this subject, it will be well to consider whether all these buildings really are tombs. Most of them undoubtedly are so; but may not the name el Deir, or the Convent, applied by the Arabs to one of the principal rock-cut monuments of Petra, be after all the true designation? Are none of them, in short, cells for priests, like the viharas found in India? All who have hitherto visited these spots have assumed at once that everything cut in the rock must be a tomb, but I am much mistaken if this is really the case with all.

To return, however, to the Khasné. Though all the forms of the architecture are Roman, the details are so elegant and generally so well designed as almost to lead to the suspicion that there must have been some Grecian influence brought to bear upon the work. The masses of rock left above the wings show how early a specimen of its class it is, and how little practice its designers could have had in copying in the rock the forms of their regular buildings.

A little further within the city is found another very similar in design to this, but far inferior to it in detail and execution, and showing at least a century of degradation, though at the same time presenting an adaptation to rock-cut forms not found in the earlier examples.

A third is that above alluded to, called el Deir. This is the same in general outline as the two former—of an order neither Greek nor Roman, but with something like a Doric frieze over a very plain Corinthian capital. In other respects it presents no new feature except the apparent absence of a door, and on the whole it seems, if finished, to deserve its name less than either of the other two.

Perhaps the most singular object among these tombs, if tombs they are, is the flat façade with three storeys of pillars one over the other— slightly indicated on the left of the Corinthian tomb in Woodcut No. 235. It is like the proscenium of some of the more recent Greek theatres. If it was really the frontispiece to a tomb, it was totally unsuitable to the purpose, and is certainly one of the most complete misapplications of Greek architecture ever made.

Generally speaking, the interiors of these buildings are so plain that travellers have not cared either to draw or measure them; one, however, represented in the annexed woodcut (No. 236), is richly ornamented, and, as far as can be judged from what is published, is as unlike a tomb as it is like a vihara. But, as before remarked, they all require re-examination before the purpose for which they were cut can be pronounced upon with any certainty.

The next group of tombs is that at Jerusalem. These are undoubtedly all sepulchres. By far the greater number of them are wholly devoid of architectural ornament. To the north of the city is a group known as the Tombs of the Kings, with a façade of a corrupt Doric order, similar to some of the latest Etruscan tombs.[191] These are now very much ruined, but still retain sufficient traces of the original design to fix their date within or subsequently to the Herodian period without much possibility of doubt. A somewhat similar façade, but of a form more like the Greek Doric, found in the Valley of Jehoshaphat, bears the name of the Sepulchre of St. James.

Close to this is a square tomb, known as that of Zechariah, cut in the rock, but standing free. Each face is adorned with Ionic pillars and square piers at the angles, the whole being crowned with a pyramidal roof. Perhaps this building should properly be called a cenotaph, as it is perfectly solid, and no cave or sepulchral vault has been found beneath it, though judging from analogies one might yet be found if properly looked for. A tomb with an architectural façade, similar to that of the so-called Tomb of the Judges, does exist behind it cut in rock, and is consequently of more modern construction. It may be to mark this that the architectural monolith was left.

Close to this is another identical with it in as far as the basement is concerned, and which is now popularly known as the Tomb of Absalom; but in this instance the pyramid has been replaced with a structural spire, and it is probable when this was done that the chamber which now exists in its interior was excavated.

One of the remarkable points in these tombs is the curious jumble of the Roman orders which they present. The pillars and pilasters are Ionic, the architraves and frieze Doric, and the cornice Egyptian. The capitals and frieze are so distinctly late Roman, that we can feel no hesitation as to their date being either of the age of Herod or subsequent to that time. In an architectural point of view the cornice is too plain to be pleasing if not painted; it probably therefore was so treated.

Another class of these tombs is represented by the so-called Tomb of the Judges (Woodcut No. 241). These are ornamented by a tympanum of a Greek or Roman temple filled with a scroll-work of rich but debased pattern, and is evidently derived from something similar, though Grecian in design. In age it is certainly more recent than the so-called Tomb of Zechariah, as one of precisely similar design is found cut into the face of the rock out of which that monument was excavated.

The third group is that of Cyrene, on the African coast. Notwithstanding the researches of Admiral Beechey and of M. Pacho,[192] and the still more recent explorations of Messrs. Smith and Porcher, above referred to (p. 285), they are still much less perfectly known to us than they should be. Their number is immense, and they almost all have architectural façades, generally consisting of two or more columns between pilasters, like the grottoes of Beni-Hasan, or the Tomb of St. James at Jerusalem. Many of them show powerful evidence of Greek taste, while some may be as old as the Grecian era, though the greater part are undoubtedly of Roman date, and the paintings with which many of them are still adorned are certainly Roman in design. Two of them are illustrated by Woodcuts Nos. 165 and 166: one as showing more distinct evidence of Greek taste and colour than is to be found elsewhere, though it is doubtful if it belongs to the Grecian period any more than the so-called Tomb of St. James at Jerusalem; the other, though of equally uncertain date, is interesting as being a circular monument built over a cave like that at Amrith (Woodcut No. 122), and is the only other example now known. None of them have such splendid architectural façades as the Khasné at Petra; but the number of tombs which are adorned with architectural features is greater than in that city, and, grouped as they are together in terraces on the hill-side, they constitute a necropolis which is among the most striking of the ancient world. Altogether this group, though somewhat resembling that at Castel d’Asso, is more extensive and far richer in external architecture.[193]

Time has not left us any perfect structural tombs in all these places, though there can be little doubt but they were once numerous. Almost the only tomb of this class constructed in masonry known to exist, and which in many respects is perhaps the most interesting of all, is found in Asia Minor, at Mylassa in Caria. In form it is something like the free-standing rock-cut examples at Jerusalem. As shown in the woodcut (No. 242), it consists of a square base, which supports twelve columns, of which the eight inner ones support a dome, the outer four merely completing the square. The dome itself is constructed in the same manner as all the Jaina domes are in India (as will be explained hereafter when describing that style), and, though ornamented with Roman details, is so unlike anything else ever built by that people, and is so completely and perfectly what we find reappearing ten centuries afterwards in the far East, that we are forced to conclude that it belongs to a style once prevalent and long fixed in these lands, though this one now stands as the sole remaining representative of its class.

Another example, somewhat similar in style, though remotely distant in locality, is found at Dugga, near Tunis, in Africa. This, too, consists of a square base, taller than in the last example, surmounted by twelve Ionic columns, which are here merely used as ornaments. There were probably square pilasters at the angles, like that at Jerusalem (Woodcuts Nos. 238, 239), while the Egyptian form of the cornice is similar to that found in these examples, though with the omission of the Doric frieze.

It apparently originally terminated in a pyramid of steps like the Mausoleum at Halicarnassus, and a large number of structural tombs which copied that celebrated model. Nothing of this now remains but the four corner-stones, which were architecturally most essential to accentuate the weak lines of a sloping pyramid in such a situation. Taken altogether, perhaps no more graceful monument of its class has come down to our days than this must have been when complete.

Besides these there are in Algeria two tombs of very great interest, both from their size and the peculiarity of their forms. The best known is that on the coast a short distance from Algiers to the westward. It is generally known as the Kubr Roumeïa, or Tomb of the Christian Virgin— a name it acquired from its having four false doors, each of a single stone divided into four panels, and the stile between them forming a cross, which has consequently been assumed to be the Christian symbol. The building itself, which is circular, and as nearly as may be 200 ft. in diameter, stands on a square platform measuring 210 ft. The perpendicular part is ornamented by 60 engaged columns of the Ionic order, and by the four false doors just mentioned; above this rose a cone—apparently in 40 steps—making the total height about 130 ft. It is, however, so ruined that it is very difficult to feel sure about its exact dimensions or form.

From objects and scribblings of various kinds found in the interior, it appears to have remained open till nearly the time of the Moslem conquest, but shortly afterwards to have been closed, and to have defied all the ingenuity of explorers till a passage was forced in 1866 by Messrs. MacCarthy and Berbrugger, acting under the orders and at the expense of the late Emperor Napoleon III.[194] The entrance was found passing under the sill of the false door on the east from a detached building standing outside the platform, and which seems to have been originally constructed to cover and protect the entrance. From this a winding passage, 560 ft. in length, led to the central chamber where it is assumed the royal bodies were once deposited, but when opened no trace of them remained, nor anything to indicate who they were, nor in what manner they were buried.

The other tomb, the Madracen, is very similar to this one, but smaller. Its peristyle is of a sort of Doric order, without bases, and surmounted by a quasi-Egyptian cornice, not unlike that on the Tomb of Absalom at Jerusalem (Woodcut No. 240), or that at Dugga (Woodcut No. 243). Altogether its details are more elegant, and from their general character there seems no reason for doubting that this tomb is older than the Kubr Roumeïa, though they are so similar to each other that their dates cannot be far distant.[195]

There seems almost no reason for doubting that the Kubr Roumeïa was the “Monumentum commune Regiæ gentis” mentioned by Pomponius Mela,[196] about the middle of the first century of our era, and if so, this could only apply to the dynasty that expired with Juba II., A.D. 23, and in that case the older monument most probably belonged to the previous dynasty, which ceased to reign with Bocchus III., 33 years before the birth of Christ.

One of the most interesting points connected with these Mauritanian tombs is their curious similarity to that of Hadrian at Rome. The square base, the circular colonnade, the conical roof, are all the same. At Rome they are very much drawn out, of course, but that arose from the “Mole” being situated among tall objects in a town, and more than even that, perhaps, from the tendency towards height which manifested itself so strongly in the architecture of that age.

The greatest similarity, however, exists in the interior. The long winding corridor terminating in an oblong apartment in the centre is an identical feature in both, but has not yet been traced elsewhere, though it can be hardly doubted that it must have existed in many other examples.

If we add to these the cenotaph at St. Rémi (Woodcut No. 231), we have a series of monuments of the same type extending over 400 years; and, though many more are wanted before we can fill up the gaps and complete the series, there can be little doubt that the missing links once existed which connected them together. Beyond this we may go still further back to the Etruscan tumuli and the simple mounds of earth on the Tartar steppes. At the other end of the series we are evidently approaching the verge of the towers and steeples of Christian art; and, though it may seem the wildest of hypotheses to assert that the design of the spire of Strasbourg grew out of the mound of Alyattes, it is nevertheless true, and it is only non-apparent because so many of the steps in the progress from the one to the other have disappeared in the convulsions of the interval.

DOMESTIC ARCHITECTURE.

We know, not only from the descriptions and incidental notices that have come down to us, but also from the remains found at Pompeii and elsewhere, that the private dwellings of the Romans were characterised by that magnificence and splendour which we find in all their works, accompanied, probably, with more than the usual amount of bad taste.

In Rome itself no ancient house—indeed no trace of a domestic edifice— exists except the palaces of the Cæsars on the Palatine Mount, and the house of the Vestal Virgins[197] at its foot; and these even are now a congeries of shapeless ruins, so completely destroyed as to make it difficult even for the most imaginative of restorers to make much of them. The extent of these ruins, however, coupled with the descriptions that have been preserved, suffice to convince us that, of all the palaces ever built, either in the East or the West, these were probably the most magnificent and the most gorgeously adorned. Never in the world’s history does it appear that so much wealth and power were at the command of one man as was the case with the Cæsars; and never could the world’s wealth have fallen into the hands of men more inclined to lavish it for their own personal gratification than these emperors were. They could, moreover, ransack the whole world for plunder to adorn their buildings, and could command the best artists of Greece, and of all the subject kingdoms, to assist in rendering their golden palaces the most gorgeous that the world had then seen, or is likely soon to see again. The whole area of the palace may roughly be described as a square platform measuring 1500 ft. east and west, with a mean breadth of 1300 ft. in the opposite direction. Owing, however, to its deeply indented and irregular outline, it hardly covers more ground than the Baths of Caracalla.

Recent excavations have laid bare nearly the whole of the western portion of this area, and have disclosed the plan of the building, but all has been so completely destroyed that it requires considerable skill and imagination to reinstate it in its previous form. The one part that remains tolerably perfect is the so-called house of Livia the wife of Augustus, who is said to have lived in it after the death of her husband. In dimensions and arrangement it is not unlike the best class of Pompeian houses, but its paintings and decorations are very superior to anything found in that city. They are, in fact, as might be expected from their age and position, the finest mural decorations that have come down to us, and as they are still wonderfully perfect, they give a very high idea of the perfection of art attained in the Augustan age, to which they certainly belong.

That part of the palace on the Palatine which most impresses the visitor is the eastern half, which looks on one hand to the Amphitheatre, on the other to the Baths of Caracalla, and overhangs the Circus Maximius. Though all their marble or painted decorations are gone, the enormous masses of masonry which here exist convey that impression of grandeur which is generally found in Roman works. It is not of Æsthetic beauty arising from ornamental or ornamented construction, but the Technic expression of power and greatness arising from mass and stability. It is the same feeling with which we contemplate the aqueducts and engineering works of this great people; and, though not of the highest class, few scenes of architectural grandeur are more impressive than the now ruined Palace of the Cæsars.

Notwithstanding all this splendour, this palace was probably as an architectural object inferior to the Thermæ. The thousand and one exigencies of private life render it impossible to impart to a residence—even to that of the world’s master—the same character of grandeur as may be given to a building wholly devoted to show and public purposes. In its glory the Palace of the Cæsars must have been the world’s wonder; but as a ruin deprived of its furniture and ephemeral splendour, it loses much that would tend to make it either pleasing or instructive. We must not look for either beauty of proportion or perfection of construction, or even for appropriateness of material, in the hastily constructed halls of men whose unbounded power was only equalled by the coarse vulgarity of their characters.

SPALATO.

The only palace of the Roman world of which sufficient remains are still left to enable us to judge either of its extent or arrangements is that which Diocletian built for himself at Spalato, in Dalmatia, and in which he spent the remaining years of his life, after shaking off the cares of Empire. It certainly gives us a most exalted idea of what the splendour of the imperial palace at Rome must have been when we find one emperor— certainly neither the richest nor the most powerful—building, for his retirement, a villa in the country of almost exactly the same dimensions as the Escurial in Spain, and consequently surpassing in size, as it did in magnificence, most of the modern palaces of Europe.

It is uncertain how far it resembles or was copied from that in Rome, more especially as it must be regarded as a fortified palace, which there is no reason to believe that at Rome was, while its model would seem to have been the prætorian camp rather than any habitation built within the protection of the city walls. In consequence of this its exterior is plain and solid, except on the side next the sea, where it was least liable to attack. The other three sides are only broken by the towers that flank them, and by those that defend the great gates which open in the centre of each face.

The building is nearly a regular parallelogram, though not quite so. The south side is that facing the sea, and is 592 ft. from angle to angle; the one opposite being only 570 in length;[198] while the east and west sides measure each 698 ft., the whole building thus covering about 9½ English acres.

The principal entrance to the palace is on the north, and is called the Golden Gate, and, as represented in the annexed woodcut (No. 247), shows all the peculiarities of Roman architecture in its last stage. The horizontal architrave still remains over the doorway, a useless ornament, under a bold discharging arch, which usurps its place and does its duty. Above this, a row of Corinthian columns, standing on brackets, once supported the archivolts of a range of niches—a piece of pleasing decoration, it must be confessed, but one in which the original purpose of the column has been entirely overlooked or forgotten.

Entering this portal, we pass along a street ornamented with arcades on either side, till exactly in the centre of the building this is crossed at right angles by another similar street, proceeding from the so-called Iron and Brazen Gates, which are similar to the Golden Gate in design, but are far less richly ornamented.

These streets divided the building into four portions: those to the north are so much ruined that it is not now easy to trace their plan, or to say to what purpose they were dedicated; but probably the one might have been the lodgings of the guests, the other the residence of the principal officers of the household.

The whole of the southern half of the building was devoted to the palace properly so called. It contained two temples, as they are now designated. That on the right is said to have been dedicated to Jupiter, though, judging from its form, it would appear to have been designed rather as the mausoleum of the founder than as a temple of that god. On the assumption that it was a temple it has been illustrated at a previous page.[199] Opposite to it is another small temple, dedicated, it is said, to Æsculapius.

Between these two is the arcade represented in Woodcut No. 185, at the upper end of which is the vestibule—circular, as all buildings dedicated to Vesta, or taking their name from that goddess, should be. This opened directly on to a magnificent suite of nine apartments, occupying the principal part of the south front of the palace. Beyond these, on the right hand, were the private apartments of the emperor, and behind them his baths. The opposite side is restored as if it exactly corresponded, but this is more than doubtful; and, indeed, there is scarcely sufficient authority for many of the details shown in the plan, though they are, probably, on the whole, sufficiently exact to convey a general idea of the arrangements of a Roman imperial palace.

Perhaps, however, the most splendid feature in this palace was the great southern gallery, 515 ft. in length by 24 in width, extending along the whole seaward face of the building. Besides its own intrinsic beauty as an architectural feature, it evinces an appreciation of the beauties of nature which one would hardly expect in a Roman. This great arcade is the principal feature in the whole design, and commands a view well worthy the erection of such a gallery for its complete enjoyment.

POMPEII AND HERCULANEUM.

Failing to discover any example of domestic architecture in Rome, we turn to Pompeii and Herculaneum, where we find numerous and most interesting examples of houses of all classes, except, perhaps, the best; for there is nothing there to compare with the Laurentian villa of Pliny, or with some others of which descriptions have come down to us. Pompeii, moreover, was far more a Grecian than a Roman city, and its buildings ought to be considered rather as illustrative of those of Greece, or at least of Magna Græcia, than of anything found to the northward. Still these cities belonged to the Roman age, and, except in taste and in minor arrangements, we have no reason to doubt that the buildings did resemble those of Rome, at least to a sufficient extent for illustration.

With scarcely an exception, all the houses of Pompeii were of one storey only in height. It is true that in some we find staircases leading to the roof, and traces of an upper storey, but where this latter is the case the apartments would appear to have been places for washing and drying clothes, or for some such domestic purpose rather than for living or even sleeping rooms. All the principal apartments were certainly on the ground floor, and as an almost inevitable corollary from this, they all faced inwards, and were lighted from courtyards or atria, and not from the outside; for, with a people who had not glass with which to glaze their windows, it was impossible to enjoy privacy or security without at the same time excluding both light and air, otherwise than by lighting their rooms from the interior. Hence it arose that in most instances the outside of the better class of houses was given up to shops and smaller dwellings, which opened on to the street, while the residence, with the exception of the principal entrance, and sometimes one or two private doors that opened outwards, was wholly hidden from view by their entourage.

Even in the smallest class of tradesmen’s houses which opened on the street, one apartment seems always to have been left unroofed to light at least two rooms on each side of it, used as bedrooms; but as the roofs of all are now gone, it is not always easy to determine which were so treated.

It is certain that, in the smallest houses which can have belonged to persons at all above the class of shopkeepers, there was always a central apartment, unroofed in the centre, into which the others opened. Sometimes this was covered by two beams placed in one direction, and two crossing them at right angles, framing the roof into nine compartments, generally of unequal dimensions, the central one being open, and with a corresponding sinking in the floor to receive the rain and drainage which inevitably came through it. When this court was of any extent, four pillars were required at the intersection of the beams, or angles of the opening, to support the roof. In larger courts eight, twelve, sixteen, or more columns were so employed, often apparently more as decorative objects than as required by the constructive necessities of the case, and very frequently the numbers of these on either side of the apartment did not correspond. Frequently the angles were not right angles, and the pillars were spaced unequally with a careless disregard of symmetry that strikes us as strange, though in such cases this may have been preferable to cold and formal regularity, and even more productive of grace and beauty. Besides these courts, there generally existed in the rear of the house another bounded by a dead wall at the further extremity, and which in the smaller houses was painted, to resemble the garden which the larger mansions possessed in this direction. The apartments looking on this court were of course perfectly private, which cannot be said of any of those looking inwards on the atrium.

The house called that of Pansa at Pompeii is a good illustration of these peculiarities, and, as one of the most regular, has been frequently chosen for the purpose of illustration.

In the annexed plan (Woodcut No. 248) all the parts that do not belong to the principal mansion are shaded darker except the doubtful part marked A, which may either have been a separate house, or the women’s apartments belonging to the principal one, or, what is even more probable, it may have been designed so as to be used for either purpose. B is certainly a separate house, and the whole of the remainder of this side, of the front, and of the third side, till we come opposite to A, was let off as shops. At C we have the kitchen and servants’ apartments, with a private entrance to the street, and an opening also to the principal peristyle of the house.

Returning to the principal entrance or front door D, you enter through a short passage into the outer court E, on each side of which are several small apartments, used either by the inferior members of the household or by guests. A wider passage than the entrance leads from this to the peristyle, or principal apartment of the house. On the left hand are several small rooms, used no doubt as sleeping apartments, which were probably closed by half-doors open above and below, so as to admit air and light, while preserving sufficient privacy, for Roman tastes at least. In front and on the right hand are two larger rooms, either of which may have been the triclinium or dining-room, the other being what we should call the drawing-room of the house. A passage between the kitchen and the central room leads to a verandah which crosses the whole length of the house, and is open to the garden beyond.

As will be observed, architectural effect has been carefully studied in this design, a vista nearly 300 ft. in length being obtained from the outer door to the garden wall, varied by a pleasing play of light and shade, and displaying a gradually increasing degree of spaciousness and architectural richness as we advance. All these points must have been productive of the most pleasing effect when complete, and of more beauty than has been attained in almost any modern dwelling of like dimensions.

Generally speaking the architectural details of the Pompeian houses are carelessly and ungracefully moulded, though it cannot be denied that sometimes a certain elegance of feeling runs through them that pleases in spite of our better judgment. It was not, however, on form that they depended for their effect; and consequently it is not by that that they must be judged. The whole architecture of the house was coloured, but even this was not considered so important as the paintings which covered the flat surfaces of the walls. Comparing the Pompeian decoration with that of the baths of Titus, and those of the House of Livia, the only specimens of the same age and class found in Rome, it must be admitted that the Pompeian examples show an equally correct taste, not only in the choice but in the application of the ornaments used, though in the execution there is generally that difference that might be expected between paintings executed for a private individual and those for the Emperor of the Roman world. Notwithstanding this, these paintings, so wonderfully preserved in this small provincial town, are even now among the best specimens we possess of mural decoration. They excel the ornamentation of the Alhambra, as being more varied and more intellectual. For the same reason they are superior to the works of the same class executed by the Moslems in Egypt and Persia, and they are far superior to the rude attempts of the Gothic architects in the Middle Ages; still they are probably as inferior to what the Greeks did in their best days as the pillars of the Pompeian peristyles are to the porticoes of the Parthenon. But though doubtless far inferior to their originals, those at Pompeii are direct imitations of true Greek decorative forms; and it is through them alone that we can form even the most remote idea of the exquisite beauty to which polychromatic architecture once attained, but which we can scarcely venture to hope it will ever reach again.

One curious point which has hitherto been too much overlooked is, that in Pompeii there are two perfectly distinct styles of decoration. One of these is purely Etruscan, both in form and colour, and such as is only found in the tombs or on the authentic works of the Etruscans. The other is no less essentially Greek, both in design and colour: it is far more common than the Etruscan form, and is always easily to be distinguished from it. The last-mentioned or Greek style of decoration may be again divided into two varieties; one, the most common, consisting of ornaments directly copied from Greek models; the other with a considerable infusion of Roman forms. This Romanised variety of Greek decoration represents an attenuated and lean style of architecture, which could only have come into fashion from the continued use of iron or bronze, or other metallic substances, for pillars and other architectural members. Vitruvius reprobates it; and in a later age Cassiodorus speaks of it in a manner which shows that it was practised in his time. The general adoption of this class of ornament, both at Pompeii and in the baths of Titus, proves it to have been a very favourite style at that time. This being the case, it must have either been a representation of metallic pillars and other architectural objects then in use, or it must have been copied from painted decorations. This is a new subject, and cannot be made clear, except at considerable length and with the assistance of many drawings. It seems, however, an almost undoubted fact that the Romans did use metal as a constructive material. Were it only that columns of extreme tenuity are represented in these paintings, we might be inclined to ascribe it to mere incorrect drawing; but the whole style of ornament here shown is such as is never found in stone or brick pillars, and which is only susceptible of execution in metal. Besides this, the pillars in question are always shown in the decorations as though simply gilt or bronzed, while the representations of stone pillars are coloured. All this evidence goes to prove that a style of art once existed in which metal was generally employed in all the principal features, all material traces of which are now lost. The disappearance of all remains of such a style is easily accounted for by the perishable nature of iron from rust, and the value and consequent peculation induced by bronze and similar metals. We are, moreover, aware that much bronze has been stolen, even in recent days, from the Pantheon and other buildings which are known to have been adorned with it.

Another thing which we learn from these paintings is, that though the necessities of street architecture compelled these city mansions to take a rectilinear outline, whenever the Roman architects built in the country they indulged in a picturesque variety of outline and of form, which they carried perhaps as far as even the Gothic architects of the Middle Ages. This indeed we might have expected, from their carelessness in respect to regularity in their town-houses; but these were interiors, and were it not for the painted representations of houses, we should have no means of judging how the same architects would treat an exterior in the country. From this source, however, we learn that in the exterior arrangements, in situations where they were not cramped by confined space, their plans were totally free from all stiffness and formality. In this respect Roman taste coincided with that of all true architecture in all parts of the world.

Each part of the design was left to tell its own tale and to express the use to which each apartment was applied, though the whole were probably grouped together with some reference to symmetry. There is certainly nothing in these ancient examples to justify the precise regularity which the architects of the Renaissance introduced into their classical designs, in which they sought to obliterate all distinction between the component parts in a vain attempt to make one great whole out of a great number of small discordant fragments.

BRIDGES AND AQUEDUCTS.

Perhaps the most satisfactory works of the Romans are those which we consider as belonging to civil engineering rather than to architecture. The distinction, however, was not known in those earlier days. The Romans set about works of this class with a purpose-like earnestness that always ensures success, and executed them on a scale which leaves nothing to be desired; while at the same time they entirely avoided that vulgarity which their want of refinement allowed almost inevitably to appear in more delicate or more ornate buildings. Their engineering works also were free from that degree of incompleteness which is inseparable from the state of transition in which their architecture was during the whole period of the Empire. It is owing to these causes that the substructions of the Appian way strike every beholder with admiration and astonishment; and nothing impresses the traveller more, on visiting the once imperial city, than the long lines of aqueducts that are seen everywhere stretching across the now deserted plain of the Campagna. It is true they are mere lines of brick arches, devoid of ornament and of every attempt at architecture properly so called; but they are so well adapted to the purpose for which they were designed, so grand in conception, and so perfect in execution, that, in spite of their want of architectural character, they are among the most beautiful of the remains of Roman buildings.

The aqueducts were not, however, all so devoid of architectural design as those of the Campagna. That, for instance, known as the Pont du Gard, built to convey water to the town of Nîmes in France, is one of the most striking works of antiquity. Its height above the stream is about 180 ft., divided into two tiers of larger arches surmounted by a range of smaller ones, giving the structure the same finish and effect that an entablature and cornice gives to a long range of columns. Without the introduction of one single ornament, or of any member that was not absolutely wanted, this arrangement converts what is a mere utilitarian work into an architectural screen of a beauty hitherto unrivalled in its class.

The aqueducts of Segovia and Tarragona in Spain, though not perhaps so grand, are quite as elegant and appropriate as this; and if they stood across a line of well wooded and watered valleys, might form as beautiful objects. Unfortunately the effect is much marred by the houses and other objects that crowd their bases. Both these rise to about 100 ft. above the level of their foundation in the centre. That of Segovia is raised on light piers, the effect of which is perhaps somewhat spoiled by numerous offsets, and the upper tier is if anything too light for the lower. These defects are avoided at Tarragona, the central arches of which are shown in Woodcut No. 251. In this example the proportion of the upper to the lower arcade is more perfect, and the whole bears a character of lightness combined with constructive solidity and elegance unrivalled, so far as I know, in any other work of its class. It wants, however, the grandeur of the Pont du Gard; for though its length is about the same, exceeding 800 ft., it has neither its height nor the impression of power given by the great arches of that building, especially when contrasted with those that are smaller.

The Roman bridges were designed on the same grand scale as their aqueducts, though from their nature they of course could not possess the same grace and lightness. This was, however, more than compensated by their inherent solidity and by the manifestation of strength imparted by the Romans to all these structures. They seem to have been designed to last for ever; and but for the violence of man, it would be hardly possible to set limits to their durability. Many still remain in almost every corner of the Roman Empire; and wherever found are easily recognised by the unmistakable impress of Roman grandeur which is stamped upon them.

One of the most remarkable of these is that which Trajan erected at Alcantara, in Spain, represented in the annexed woodcut. The roadway is perfectly level, as is generally the case in Roman bridges, though the mode by which this is obtained, of springing the arches from different levels, is perhaps not the most pleasing. To us at least it is unfamiliar, and has never, I think, been adopted in modern times. In such a case we should either have made the arches all equal—a mistake, considering their different heights—or have built solidly over the smaller arches to bring up the level, which would have been a far greater error in construction than the other is in taste. The bridge consists of six arches, the whole length of the roadway being 650 ft.; the two central arches are about 100 ft. span; the roadway is 140 ft. above the level of the stream which it crosses. The piers are well proportioned and graceful; and altogether the work is as fine and as tasteful an example of bridge-building as can be found anywhere, even in these days of engineering activity.

The bridge which the same Emperor erected over the Danube was a far more difficult work in an engineering point of view; but the superstructure being of wood, resting only on stone piers, it would necessarily have possessed much less architectural beauty than this, or indeed than many others.

These examples of this class of Roman works must suffice; they are so typical of the style that it was impossible to omit them altogether, though the subject scarcely belongs in strictness to the objects of this work. The bridges and aqueducts of the Romans richly deserve the attention of the architect, not only because they are in fact the only works which the Romans, either from taste or from social position, were enabled to carry out without affectation, and with all their originality and power, but also because it was in building these works that the Romans acquired that constructive skill and largeness of proportion which enabled them to design and carry out works of such vast dimensions, to vault such spaces, and to give to their buildings generally that size and impress of power which form their chief and frequently their only merit. It was this too that enabled them to originate that new style of vaulted buildings which at one period of the Middle Ages promised to reach a degree of perfection to which no architecture of the world had ever attained. The Gothic style, it is true, perished at a time when it was very far from completed; but it is a point of no small interest to know where and under what circumstances it was invented. We shall subsequently have to trace how far it advanced towards that perfection at which it aimed, but to which it never reached. Strangely enough, it failed solely because of the revival and the pernicious influence of that very parent style to which it owed its birth, and the growth and maturity of which we have just been describing. It was the grandeur of the edifices reared at Rome in the first centuries of the Empire which so impressed the architects of the fifteenth and sixteenth centuries, that they abandoned their own beautiful style to imitate that of the Romans, but with an incongruity which seems inevitably to result from all imitations, as contrasted with true creations, in architectural art.

VI. PARTHIAN AND SASSANIAN ARCHITECTURE

CONTENTS.

Historical notice—Palaces of Al Hadhr and Diarbekr—Domes—Palaces of Serbistan—Firouzabad—Tâk Kesra—Mashita—Rabbath Ammon.

CHRONOLOGY.

Parthians subject to Persia B.C. 554 Seleucus Nicator 301 Arsaces 250 Mithridates 163-140 Mithridates II 124-89 Palace of Al Hadhr built (about) A.D. 200 End of Parthian Empire 227 ---------- Ardeshir, or Artaxerxes, establishes Sassanian dynasty 226 Tiridates 286-342 Serbistan (about) 350 Bahram Gaur begins to reign 420 Firouzabad (about) 450 Khosru Nushirvan begins to reign 531 Khosru Nushirvan builds palace at Ctesiphon (about) 550 Khosru Purviz Chosroes 591 Palace at Mashita 614-627 Battle of Cadesia 636

There still remains one other style to be described before leaving the domain of Heathendom to venture into the wide realms of Christian and Saracenic art with which the remainder of these two volumes is mainly occupied. Unfortunately it is not one that was of great importance while it existed, and it is one of which we know very little at present. This arises partly from the fact that all the principal buildings of the Sassanian kings were situated on or near the alluvial plains of Mesopotamia and were therefore built either of sun-burnt or imperfectly baked bricks, which consequently crumbled to dust, or, where erected with more durable materials, these have been quarried by the succeeding inhabitants of these fertile regions. Partly also it arises from the Sassanians not being essentially a building race. Their religion required no temples and their customs repudiated the splendour of the sepulchre, so that their buildings were mainly palaces. One of these, that at Dustagird, is described by all contemporary historians[200] as one of the most gorgeous palaces of the East, but its glories were ephemeral: gold and silver and precious hangings rich in colour and embroidery made up a splendour in which the more stable arts of architecture had but little part, and all perished in an hour when invaded by the victorious soldiers of Heraclius, or the more destructive hosts of Arabian invaders a few years afterwards. Whatever the cause however, never was destruction more complete. Two or three ruined palaces still exist in Persia and Mesopotamia. A fragment known as the Tâk Kesra still remains to indicate the spot where Ctesiphon once stood, but the site of Dustagird is still a matter of dispute. So little in fact remains that we should hardly be able to form an idea of what the style really was, but for the fortunate discovery of a palace at Mashita in Moab, which seems undoubtedly to have been erected by the last great king of this dynasty, and which is yet unsurpassed for beauty of detail and richness of ornament by any building of its class and age.

As nearly as may be, one thousand years had elapsed since the completion of the palaces at Persepolis and Susa and the commencement of this building, and for the great part of that period the history of Persian or Central Asian architecture is a blank. The Seleucidæ built nothing that has come down to our times. The Parthians, too, have left us little, so that it is practically only after a hiatus of nearly six centuries, that we again begin to feel that the art had not entirely perished in the populous countries of Central Asia; but even then our history recommences so timidly and with buildings of such uncertain dates as to be very far from satisfactory.

One of the oldest buildings known as belonging to the new school is the palace of Al Hadhr, situated in the plain, about thirty miles from the Tigris, nearly west from the ruins of Kaleh Shergat.

The city itself is circular in plan, nearly an English mile in diameter, and surrounded by a stone wall with towers at intervals, in the centre of which stands a walled enclosure, nearly square in plan, about 700 ft. by 800. This is again subdivided into an outer and inner court by a wall across its centre. The outer court is unencumbered by buildings, the inner nearly filled with them.[201] The principal of these is that represented in plan on Woodcut No. 253. It consists of three large and four smaller halls placed side by side, with various smaller apartments in the rear. All these halls are roofed by semicircular tunnel-vaults, without ribs or other ornament, and they are all entirely open in front, all the light and air being admitted from the one end.

There can be little doubt that these halls are copies, or intended to be so, of the halls of the old Assyrian palaces; but the customs and requirements of the period have led the architect on to a new class of arrangements which renders the resemblance by no means apparent at first sight.

The old halls had almost invariably their entrances on the longer side, which with a vault required very thick external walls as abutments. This was obviated in Al Hadhr by using the halls as abutments the one to the other like the arches of a bridge; so that, if the two external arches were firm, all the rest were safe. This was provided for by making the outer halls smaller, as shown in the elevation (Woodcut No. 254), or by strengthening the outer wall. But even then the architect seems to have shrunk from weakening the intermediate walls by making too many openings in them. Those which do exist are small and infrequent; so that there is generally only one entrance to each apartment, and that so narrow as to seem incongruous with the size of the room to which it leads.

The square apartment at the back would seem to have been a temple, as the lintel over the entrance doorway (which faces the east) is carved with the sun, the moon, and other religious emblems; and the double wall round may have contained a stair or inclined plane leading to an upper storey, or to rooms which certainly existed over the smaller halls at least.

All the details of the building are copied from the Roman—the archivolts and pilasters almost literally so, but still so rudely executed as to prove that it was not done under the direct superintendence of a Roman artist. This is even more evident with regard to the griffins and scroll-work, and the acanthus-leaves which ornament the capitals and friezes. The most peculiar ornament, however, is the range of masks carried round all the archivolts of the smaller arches. Of the nineteen voussoirs of the larger arches, seven of them, according to Ross and Ainsworth, had figures carved on them in relief of angels, or females, apparently in the air, and with feet crossed and robes flying loose, possibly emblematic of the seven planets. Even tradition is silent regarding the date of these remarkable ruins; the town was besieged unsuccessfully by Trajan in 116 A.D., and it is recorded to have been a walled town containing a temple of the sun noted for its rich offerings. This is probably the square building at the back of the great hall on the left of the palace, and the existence of the carved religious emblems on the lintel suggest that the palace was erected in front at a later period. Professor Rawlinson, in his notes on the great monarchies,[202] suggests about 200 A.D. as the probable date, and ascribes its erection to the monarchs of the Parthian dynasty. There is no doubt that the execution of the masonry with its fine joints is of a totally different character from that which is found in Sassanian buildings, which comes more under the head of rubble masonry, and was entirely hidden, in the interior at least, by stucco. The ornament also is of a rich character, Roman in its design, but debased Greek in its execution. Mr. Loftus, during his researches in Chaldea, discovered at Wurka (the ancient Erech in Mesopotamia), a large number of ornamental details, in stone and in plaster, of precisely the same character as those found at Al Hadhr. Among these remains he found a griffin resembling those carved on the lintel of the square temple before referred to, and quantities of Parthian coins, so that it is fair to assume that Al Hadhr belongs to that dynasty.

Another building which merits more attention than has hitherto been bestowed upon it, is now used as the great mosque at Diarbekr. The ancient portions consist of the façades only of two palaces, the north and the south, which face one another at a distance of some 400 feet, and form the boundaries of the great court (Woodcut No. 255). They are apparently erected with materials taken from some more ancient building, and whilst the capitals and friezes are of debased Roman character, the carved shafts of the north palace (Woodcut No. 257) resemble in the plaster design ornaments found at Wurka.

As will be seen in Woodcut No. 256, which represents the façade of the South Palace, the openings of the ground storey are spanned by arches of two different forms; and those of the upper storey by lintels carried on corbels with relieving arch over; the latter a Byzantine treatment; the former of a very much later date, and probably Saracenic: above the openings and under the frieze are Cufic inscriptions. On the whole there seems little doubt that the building we now see was erected, as it now stands, at the age of the Cufic inscriptions,[203] whatever they may be, but that the remains of some more ancient edifice was most skilfully worked up in the new. Till, however, the building is carefully examined by some thoroughly competent person, this must remain doubtful. The building is rich, and so interesting that it is to be hoped that its history and peculiarities will before long be investigated.

With the accession of the Sassanians, A.D. 223, Persia regained much of that power and stability to which she had been so long a stranger. The capture of the Roman Emperor Valerian by the 2nd king of the race, A.D. 260, the Conquest of Armenia and victories over Galerius by the 7th (A.D. 296), and the exploits of the 14th King, Bahram Gaur, his visit to India and his alliance with its kings, all point to extended power abroad; while the improvement in the fine arts at home indicates returning prosperity and a degree of security unknown since the fall of the Achæmenidæ.

These kings seem to have been of native race, and claimed descent from the older dynasties: at all events they restored the ancient religion and many of the habits and customs with which we are familiar as existing before the time of Alexander the Great.

As before remarked, fire-worship does not admit of temples, and we consequently miss that class of buildings which in all ages best illustrates the beauties of architecture; and it is only in a few scattered remains of palaces that we are able to trace the progress of the style. Such as they are, they indicate considerable originality and power, but at the same time point to a state of society when attention to security hardly allowed the architect the free exercise of the more delicate ornaments of his art.

The Sassanians took up the style where it was left by the builders of Al Hadhr; but we only find it after a long interval of time, during which changes had taken place which altered it to a considerable extent, and made it in fact into a new and complete style.

They retained the great tunnel-like halls of Al Hadhr, but only as entrances. They cut bold arches through the dividing walls, so as to form them into lateral suites. But, above all, they learnt to place domes on the intersections of their halls, not resting on drums, but on pendentives,[204] and did not even attempt to bring down simulated lines of support to the ground. Besides all these constructive peculiarities, they lost all trace of Roman detail, and adopted a system of long reed-like pilasters, extending from the ground to the cornice, below which they were joined by small semicircular arches. They in short adopted all the peculiarities which are found in the Byzantine style as carried out at a later age in Armenia and the East. We must know more of this style, and be able to ascribe authentic dates to such examples as we are acquainted with, before we can decide whether the Sassanians borrowed the style from the Eastern Romans, or whether they themselves were in fact the inventors from whom the architects of the more western nations took the hints which they afterwards so much improved upon.

The various steps by which the Romans advanced from the construction of buildings like the Pantheon to that of the church of Sta. Sophia at Constantinople are so consecutive and so easily traced as to be intelligible in themselves without the necessity of seeking for any foreign element which may have affected them. If it really was so, and the architecture of Constantinople was not influenced from the East, we must admit that the Sassanian was an independent and simultaneous invention, possessing characteristics well worthy of study. It is quite certain too that this style had a direct influence on the Christian and Moslem styles of Asia, which exhibit many features not derivable from any of the more Western styles.

A few examples will render this clearer than it can be made in words. The plan and section (Woodcuts Nos. 258 and 259) of a small but interesting palace at Serbistan will explain most of the peculiarities of the style. The entrances, it will be observed, are deep tunnel-like arches, but the centre is covered by a dome resting on pendentives. In the palace of Firouzabad these are constructed by throwing a series of arches across the angles, one recessed behind the other, the lower ones serving as centres for those above, until a circular base for the dome has been obtained; but here in Serbistan they do not seem to have known this expedient: the lower courses run through to the angle, and the upper ones are brought forward in so irregular and unscientific a way as to suggest that for their support they placed their reliance almost entirely on the tenacious qualities of the mortar. That which, however, would have formed the outer arch of the pendentive is wrought on the stone down almost to the springing, as if the builder of Serbistan had seen regular arched pendentives of some kind, but did not know how to build them. This is the more remarkable because, as we shall see later on, they knew how to construct semi-domes over their recesses or square niches, and in regular coursed masonry; if they had applied these to the angles, they would have invented the squinch, a kind of pendentive employed in Romanesque work in the south of France. The dome is elliptical, as are also the barrel vaults over the entrances, the recesses in the central hall, and the vaults over the lateral halls. In these lateral halls piers are built within the walls, forming a series of recesses; these either have transverse arches thrown across them where the lofty doorways come, or are covered with semidomes in regular coursed masonry, the angles being filled in below them with small arches. The lower portions of the piers consist of circular columns about six feet high, behind which a passage is formed. The builders thus obtained the means of counteracting the thrust of the vault, without breaking the external outline by buttresses and without occupying much room on the floor, while at the same time these projections added considerably to the architectural effect of the interior. The date of the building is not correctly known, but it most probably belongs to the age of Shapour, in the middle of the fourth century.

The palace at Firouzabad is probably a century more modern, and is erected on a far more magnificent scale, being in fact the typical building of the style, so far at least as we at present know.

As will be seen in the plan, the great central entrance opens laterally into two side chambers, and the inner of these into a suite of three splendid domed apartments, occupying the whole width of the building. Beyond this is an inner court, surrounded by apartments all opening upon it.

As will be perceived from Woodcut No. 261, representing one of the doorways in the domed halls, the details have nothing Roman about them, but are borrowed directly from Persepolis, with so little change that the style, so far as we can now judge, is almost an exact reproduction, except that the work is only surface ornament in plaster, and is an irregular and a degraded copy of the original stone features at Persepolis. The opening also is spanned by a circular arch under the lintel of the Persian example, the former being the real constructive feature, the latter a decorative imitation. The portion of the exterior represented in Woodcut No. 262 tells the same tale, though for its prototype we must go back still further to the ruins at Wurka—the building called Wuswus at that place (see p. 165) being a palace arranged very similarly to these, and adorned externally by panellings and reeded pilasters, differing from these buildings only in detail and arrangement, but in all essentials so like them as to prove that the Sassanians borrowed most of their peculiarities from earlier native examples.

The building itself is a perfectly regular parallelogram, 332 ft. by 180, without a single break, or even an opening of any sort, except the one great arch of the entrance; and externally it has no ornament but the repetition of the tall pilasters and narrow arches represented in Woodcut No. 262. Its aspect is thus simple and severe, but more like a gigantic Bastile than the palace of a gay, pavilion-loving people, like the Persians.

Internally the arrangement of the halls is simple and appropriate, and, though somewhat too formal, is dignified and capable of considerable architectural display. On the whole, however, its formality is perhaps less pleasing than the more picturesque arrangements of the palace at Serbistan last described.

Another century probably elapsed before Khosru (Nushirvan) commenced the most daring, though certainly not the most beautiful ever attempted by any of his race; for to him we must ascribe the well-known Tâk Kesra (Woodcuts Nos. 263, 264), the only important ruin that now marks the site of the Ctesiphon of the Greeks—the great Modain of the Arabian conquerors.

As it is, it is only a fragment of a palace, a façade similar in arrangement to that at Firouzabad, but on a much larger scale, its width being 312 ft., its height 105 to 110, and the depth of the remaining block 170 ft. In the centre is a magnificent portal, the Aiwan, or Throne room of the palace, vaulted over with an elliptical barrel vault and similar to the smaller vestibules of Serbistan and Firouzabad; the lower portion of the arch, the springing of which is about 40 ft. from the ground, is built in horizontal courses up to 63 ft. above the ground, above which comes the portion arched with regular voussoirs; by this method not only was an enormous centering saved, but the thrust of that portion built with voussoirs was brought well within the thickness of the side walls. It is probable that the front portion of the arch, about 20 ft. in depth, was built on walls erected temporarily for that purpose; the remainder of the vault, however, was possibly erected without centres, the bricks being placed flatwise and the rings being inclined at an angle of about 10° towards the back of the front arch. The tenacious quality of the mortar was probably sufficient to hold the bricks in their places till the arch ring was complete, so that the centering was virtually a template only, giving the correct form of the ellipse, and constructed with small timbers so as to save expense. A similar method of construction was found by Sir Henry Layard in the drain vaults at Nimroud, and it exists in the granaries built by Rameses II. in the rear of the Rameseum at Thebes. The lower or inner portion of the great arch is built in four rings of bricks or tiles laid flatwise, two of which are carried down to the springing of the whole arch: above these in the upper portion of the arch comes a ring 3 feet in height, regularly built in voussoir-shaped bricks breaking joint, on the surface of which are cut a series of seventeen foils, the whole being crowned by a slightly projecting moulding. These have nothing to do with the construction, and are simply a novel method of decoration carved after the arch was built.

The wall flanking the great arch on either side is decorated with buttress shafts and blind arches, which are partially constructive, and intended to support and strengthen those portions of the wall which were simply screens, or to resist the thrust of the walls of the vaulted chambers behind, consisting of one storey only. Decoratively they divide up the front and were apparently introduced in imitation of the great Roman amphitheatres. The position occupied by these semi-detached shafts on the first storey (resting on the ledge left by the greater thickness of wall of the lower storey), which are not in the axes of those below, proves that the Sassanian architect thought more of their constructive value as buttresses, than of their architectural value as superimposed features.

Though it may not perhaps be beautiful, there is certainly something grand in a great vaulted entrance, 72 ft. wide by 85 ft. in height and 115 in depth, though it makes the doorway at the inner end and all the adjoining parts look extremely small. It would have required the rest of the palace to be carried out on an unheard-of scale to compensate for this defect. The Saracenic architects got over the difficulty by making the great portal a semidome, and by cutting it up with ornaments and details, so that the doorway looked as large as was required for the space left for it. Here, in the parent form, all is perfectly plain in the interior, and painting alone could have been employed to relieve its nakedness, which, however, it never would have done effectually.[205]

The ornaments in these and in all the other buildings of the Sassanians having been executed in plaster, we should hardly be able to form an idea of the richness of detail they once possessed but for the fortunate discovery of a palace erected in Moab by Khosru Purviz, the last great monarch of this line.[206]

As will be seen from the woodcut (No. 265), the whole building is a square, measuring above 500 ft. each way, but only the inner portion of it, about 170 ft. square, marked E E, has been ever finished or inhabited. It was apparently originally erected as a hunting-box on the edge of the desert for the use of the Persian king, and preserves all the features we are familiar with in Sassanian palaces. It is wholly in brick, and contains in the centre a triapsal hall, once surmounted by a dome on pendentives like those at Serbistan or Firouzabad. On either side were eight vaulted halls with intermediate courts almost identical with those found at Eski Bagdad[207] or at Firouzabad. So far there is nothing either remarkable or interesting, except the peculiarity of finding a Persian building in such a situation, and in the fact that the capitals of the pillars are of that full-curved shape which are first found in the works of Justinian, which so far helps to fix the date of the building.

It seems, however, that at a time when Chosroes possessed all Asia and part of Africa, from the Indus to the Nile, and maintained a camp for ten years on the shores of the Bosphorus, in sight of Constantinople, that this modest abode no longer sufficed for the greatest monarch of the day. He consequently determined to add to it the enclosure above described, and to ornament it with a portal which should exceed in richness anything of the sort to be found in Syria. Unfortunately for the history of art, this design was never carried out. When the walls were raised to the height of about twenty feet, the workmen were called off, most probably in consequence of the result of the battle of Nineveh in 627; and the stones remain half hewn, the ornament unfinished, and the whole exactly as if left in a panic, never to be resumed.

The length of the façade—marked A A in plan, Woodcut No. 265—between the plain towers, which are the same all round, is about 170 ft.,[208] the centre of which was occupied by a square-headed portal flanked by two octagonal towers. Each face of these towers was ornamented by an equilateral triangular pediment, filled with the richest sculpture. In that shown in Woodcut No. 267, two large animals are represented facing one another on the opposite sides of a vase, on which are two doves, and out of which springs a vine which spreads over the whole surface of the triangle, interspersed with birds and bunches of grapes. In another panel one of the lions is represented with wings, evidently the last lineal descendant of those found at Nineveh and Persepolis, and in all are curious hexagonal rosettes, carved with a richness far exceeding anything found in Gothic architecture, but which are found repeated with very little variation in the Jaina temples of western India.

The wing walls of the façade are almost more beautiful than the central part itself. As on the towers, the ornamentation consists of a series of triangles filled with incised decorations and with rosettes in their centres; while, as will be observed in Woodcut No. 265, the decoration in each panel is varied, and all are unfinished. The cornice only exists at one angle, and the mortice stones never were inserted that were meant to keep it in its place. Enough however remains to enable us to see that, as a surface decoration, it is nearly unrivalled in beauty and appropriateness. As an external form I know nothing like it. It is only matched by that between the arches of the interior of Sta. Sophia at Constantinople, which is so near it in age that they may be considered as belonging to the same school of art.

Notwithstanding the incomplete state in which this façade was left, there does not seem much difficulty in restoring it within very narrow limits of certainty. The elevation cannot have differed greatly from that shown in Woodcut No. 269, on the preceding page. In the first place there must have been a great arch over the entrance doorway—this is de rigueur in Sassanian art, and this must have been stilted or horse-shoed, as without that it could not be made to fit on to the cornice in the towers, and all the arches in the interior take, as I am informed, that shape. Besides this there is at Takt-i-Gero[209] a Sassanian arch of nearly the same age and equally classical in design, which is, like this one, horse-shoed to the extent of one-tenth of its diameter; and at Urgub, in Asia Minor, all the rock-cut excavations which are of this or an earlier age have this peculiarity in a marked degree.[210]

Above this, the third storey, is a repetition of the lowest, on half its scale—as in the Tâk Kesra,—but with this difference, that here the angular form admits of its being carried constructively over the great arch, so that it becomes a facsimile of an apse at Murano near Venice,[211] which is adorned with the spoils of some desecrated building of the same age, probably of Antioch or some city of Syria destroyed by the Saracens. Above this the elevation is more open to conjecture, but it is evident that the whole façade could not have been less than 90 ft. in height, from the fact that the mouldings at the base (Woodcut No. 265) are the mouldings of a Corinthian column of that height, and no architect with a knowledge of the style would have used such mouldings four and a half feet in height, unless he intended his building to be of a height equal at least to that proportion. The domes are those of Serbistan or of Amrith (Woodcut No. 122); but such domes are frequent in Syria before this age, and became more so afterwards.

The great defect of the palace at Mashita as an illustration of Sassanian art arises from the fact that, as a matter of course, Chosroes did not bring with him architects or sculptors to erect this building. He employed the artists of Antioch or Damascus, or those of Syria, as he found them. He traced the form and design of what he wanted, and left them to execute it, and they introduced the vine—which had been the principal “motif” in such designs from the time of Herod till the Moslem invasion—and other details of the Byzantine art with which Justinian had made them familiar from his buildings at Jerusalem, Antioch, and elsewhere. Exactly the same thing happened in India six centuries later. When the Moslems conquered that country in the beginning of the thirteenth century they built mosques at Delhi and Ajmere which are still among the most beautiful to be found anywhere. The design and outline are purely Saracenic, but every detail is Hindu, but, just as in this case, more exquisite than anything the Moslems ever did afterwards in that country.

Though it thus stands almost alone, the discovery of this palace fills a gap in our history such as no other building occupies up to the present time. And when more, and more correct, details have been procured, it will be well worthy of a monograph, which can hardly be attempted now from the scanty materials available. Its greatest interest, however, lies in the fact that all the Persian and Indian mosques were derived from buildings of this class. The African mosques were enlargements of the atria of Christian basilicas, and this form is never found there, but it is the key to all that was afterwards erected to the eastward.

The palace of Rabbath Ammon (Woodcuts Nos. 270, 271), also in Moab, consists of a central court open to the sky, and four recesses or transepts, one on each face; two of these are covered with elliptical barrel vaults, and two with semidomes carried on pendentives. The decoration of this palace is similar to that found at Mashita, but not so rich in design or so good in its execution.

The remains of two other palaces have been found in Persia, one at Imumzade, which consists of a dome on pendentives, and a second, called the Tag Eiran, made known to us by M. Dieulafoy, and published in his work on the ancient art of Persia.[212] The latter is probably a late example, for it shows a considerable advance in construction, and is lighted by clerestory windows between the brick transverse arches which span the hall. The plan consisted of a central hall, covered over by a dome carried on pendentives, and two wings; of the original building, only one of these wings remains, and two sides of the central hall, in both cases up to the springing of the real arch, the lower courses being horizontal as in the arch at Ctesiphon.

In the dearth of Sassanian buildings there is one other monument that it is worth while quoting before closing this chapter. It is an archway or grotto, which the same Chosroes cut in the rock at Takt-i-Bostan, near Kermanshah (Woodcut No. 272). Though so far removed from Byzantine influence it is nearly as classical as the palace at Mashita. The flying figures over the arch are evident copies of those adorning the triumphal arches of the Romans, the mouldings are equally classical, and though the costumes of the principal personages, and of those engaged in the hunting scenes on either hand, partake more of Assyria than of Rome, the whole betrays the influence of his early education and the diffusion of Western arts at that time more than any other monument we know of. The statue of Chosroes on his favourite black steed “Shubz diz,” is original and interesting, and, with many of the details of this monument, it has been introduced into the restoration of Mashita.

This, it must be confessed, is but a meagre account of the architecture of a great people. Perhaps it may be that the materials do not exist for making it more complete; but what is more likely is that they have not yet been looked for, but will be found when attention is fairly directed to the subject. In the meanwhile what has been said regarding it will be much clearer and better understood when we come to speak of the Byzantine style, which overlapped the Sassanian, and was to some extent contemporary with it.

II. CHRISTIAN ARCHITECTURE

INTRODUCTORY.

If a line were drawn north and south from Memel on the shores of the Baltic to Spalato on the Adriatic, it would divide Europe into nearly equal halves. All that part lying to the west of the line would be found to be inhabited by nations of Celtic or Teutonic races, and all those to the eastward of it by nations of Sclavonic origin, if—as we must do—we exclude from present consideration those fragments of the effete Turanian races which still linger to the westward, as well as the intrusive hordes of the same family which temporarily occupy some fair portions to the eastward of the line so drawn.

This line is not of course quite straight, for it follows the boundary between Germany on the one hand, and Russia and Poland on the other as far as Cracow, while it crosses Hungary by the line of the Raab and separates Dalmatia from Turkey. Though Sclavonic influences may be detected to the westward of the boundary, they are faint and underlie the Teutonic element; but to the eastward, the little province of Siebenburgen, in the north-east corner of Hungary, forms the only little oasis of Gothic art in the desert of Panslavic indifference to architectural expression. Originally it was a Roman, afterwards a German, colony, and maintained its Gothic style throughout the Middle Ages.[213]

From Spalato the line crosses the Adriatic to Fermo, and then following very closely the 43rd parallel of latitude, divides Italy into two nearly equal halves. Barbarian tribes settled to a certain extent to the northward of this boundary and influenced the style of architecture in some degree; while to the southward of it, their presence can with difficulty be detected, except in a few exceptional cases, and for a very limited time.

Architecturally all the styles of art practised during the Middle Ages to the westward and northward of this boundary may be correctly and graphically described as the Gothic style, using this term in a broad sense. All those to the eastward may with equal propriety be designated as the Byzantine style of art.

Anterior, however, to the former there existed a transitional style known as Romanesque, but which was virtually at first nothing more than debased Roman. It was, in fact, a modification of the classical Roman form which was introduced between the reigns of Constantine and Justinian, and was avowedly an attempt to adapt classical forms to Christian purposes. At first the materials of ancient buildings sufficed for its wants, and if after the 4th century the style did not lapse into absolute barbarism it was due to the influence which the Proto-Byzantine style began to exert and to the magnificent works erected by Greek artists at Parenzo and Grado in Dalmatia, at Ravenna, Milan, and even in Rome herself. To the eastward of the line of demarcation the transition was perfected under the reign of Justinian (A.D. 527 to 564), when it became properly entitled to the name of Byzantine. To the westward, in Italy and the south of France, this first phase of the Romanesque continued to be practised till the 6th or 7th centuries; but about that time occurs an hiatus in the architectural history of Western Europe, owing to the troubles which arose on the dissolution of the Roman Empire and the irruption of the Barbarian hordes. When the art again reappeared, it was strongly tinctured by Barbarian influences, and might with propriety be designated the Gothic style, the essential characteristic being that it is the architecture of a people differing from the Romans or Italians in blood, and, it need hardly be added, differing from them in a like ratio in their architectural conceptions.

The term “Gothic,” however, is so generally adopted throughout Europe to designate the style in which the intersecting vault with pointed arches is the main characteristic, that to depart from it, even when subdivided into round arched and pointed arched Gothic, would only lead to confusion. It would therefore seem better to retain the nomenclature usually employed in modern architectural works, and to class all the phases of the transitional style between the Roman and the Gothic periods under the broad title of Romanesque. This would include what we have termed Early Christian——Lombardi——Rhenish——those phases of the style which in Italy and France are influenced by Byzantine detail——the pure Romanesque or Romance of the south of France——the Norman style in Italy, Sicily, and the North of France, and——Saxon and Norman in our own country. The attempt to restrict the term Romanesque within the confines of the 6th and 7th centuries, which was formerly attempted, has proved to be illusory, as it has never been recognised by any student of architecture. At the same time it is not necessary to insist on the term when describing its various phases, and when they are better known under other terms. It is, however, of importance, when writing a general history of all styles, to keep strictly to some definite system, and not to adopt the nomenclature which has in some cases been given by persons writing monographs of the style of their own particular country. The Germans, for instance, are inclined to call the architecture of such cathedrals as Spires, Worms, etc., by the absurd name of Byzantine, though no features in them have ever been borrowed from the Eastern capital, nor do they resemble the buildings of that part of Europe.

The title Gothic, which was originally invented as a term of reproach, and which was applied to the imaginary work of the western Barbarians who at one time overthrew the western Empire and settled within its limits, has no architectural or ethnological value, it being impossible to point out any features, much less buildings, which the Goths introduced, and which are not to be more correctly attributed to Roman or Byzantine artists. If we except the tomb of Theodoric, all the works in Ravenna are scarcely to be distinguished from the basilicas of the Eastern Empire, and only embody such modifications as the material of the country and a certain influence of debased Roman architecture in Italy would naturally exert. The churches and thermæ which Theodoric is said to have restored in Rome have no characteristics which are not found in other buildings of the same class before his reign, and even in Spain and the south of France, which was occupied more or less continuously by the Visigoths for more than two centuries, there are no features which they could claim to have invented.

The term Gothic, therefore, is misplaced, but inasmuch as the Goths never invented any style, there is not likely, if this fact is recognised, to be any confusion in its adoption.

The chief difficulty which presents itself in any attempt to classify the work of the Romanesque and the Gothic styles is that of drawing a line of demarcation between the two. It is not sufficient to take the pointed arch, for in France a pointed arched barrel vault preceded the round arched vault; and in the East, as we know, the pointed arch made its appearance at a much earlier period: that characteristic, therefore, must not be too rigidly insisted upon.

Beyond this general classification, the use of local names, when available, will always be found most convenient. First, the country, or architectural province, in which an example is found should be ascertained, so that its locality may be marked, and if possible with the addition of a dynastic or regal name to point out its epoch. When the outline is sufficiently marked, it may be convenient, as the French do, to speak of the style of the 13th century[214] as applied to their own country. The terms they use always seem to be better than 1st, or 2nd, Middle Pointed, or even “Geometric,” “Decorated,” or “Perpendicular,” or such general names as neither tell the country nor the age, nor even accurately describe the style, though when they have become general it may seem pedantic to refuse to use them. The system of using local, combined, and dynastic names has been followed in describing all the styles hitherto enumerated in this volume, and will be followed in speaking of those which remain to be described; and as it is generally found to be so convenient, whenever it is possible it will be adhered to.

In order to carry out these principles, the division proposed for this part of the subject is—

1st. To begin the history of Christian Art by tracing up the successive developments of the earliest perfected style, the Byzantine, in the countries lying to the eastward of the boundary line already defined. Owing to the greater uniformity of race, the thread of the narrative is far more easily followed to the eastward than we shall find to the westward of the line. The Byzantine empire remained one and undivided during the Middle Ages; and from that we pass by an easy gradation to Russia, where the style continued to be practised till Peter the Great superseded it by introducing the styles of Western Europe.

2nd. To treat of the early Christian style as it prevailed in Italy, down to the age of Charlemagne, so long, in fact, as it remained a debased Roman style influenced only by its connection with the Eastern Empire. Continuing our description of the various phases of the style as practised in Italy and in Istria and Dalmatia (the two countries with which she was so intimately connected) down to the revival of classic architecture: subdividing it into those sections which are suggested by the predominant influence of Lombardic, Byzantine, or Gothic art, and keeping as far as possible to a chronological sequence.

3rd. To take up the Romanesque style in France, and to follow it through its various phases whilst it was being gradually absorbed in the predominant impetus given to its successor, the Gothic style, by the adoption of the pointed arch in intersecting vaulting during the 12th century, and then its subsequent development in succeeding centuries, till it perished under Francis I.

If this arrangement is not quite logical, it is certainly convenient, as it enables us to grasp the complete history of the style in the country where most of the more important features were invented and perfected. Having once mastered the history of Gothic art in the country of its birth, the sequence in which the other branches of the style are followed become comparatively unimportant. The difficulty of arranging them does not lie so much in the sequence as in the determination of what divisions shall be considered as separate architectural provinces. In a handbook, subdivision could hardly be carried too far; in a history, a wider view ought to be taken. On the whole, perhaps, the following will best meet the true exigencies of the case:—

4th. Belgium and Holland should be taken up after France as a separate province during the Middle Ages, while at the same time forming an intermediate link between that country and Germany.

5th. Though not without important ethnographical distinctions, it will be convenient to treat all the German-speaking countries from the Alps to the Baltic as one province. If Germany were taken up before France, such a mode of treatment would be inadmissible; but following the history of the art in that country, it may be done without either confusion or needless repetition.

6th. Scandinavia follows naturally as a subordinate, and, unfortunately, not very important, architectural subdivision.

7th. From this we pass by an easy gradation to the British Islands, which in themselves contain three tolerably well-defined varieties of style, popularly known as the Saxon, the Norman, or round-arched, and the Gothic, or pointed-arched style of Architecture.

8th. Spain might have been made to follow France, as most of its architectural peculiarities were borrowed from that country; but some too own a German origin, while on the whole the new lessons to be learned from a study of her art are so few, that it is comparatively unimportant in what sequence the country is taken, and therefore it has been found more convenient to place her last.

I. INTRODUCTORY

CHRONOLOGY.

DATES. Constantine founds Constantinople A.D. 324 First Council of Nice 325 Julian the Apostate 361 Theodosius the Great 379 Theodosius II. 408 Marcian 450 Fall of Western Empire 476 Justinian I. 527 Justin II. 565 Heraclius 610 The Hejira 622

The term Byzantine has of late years been so loosely and incorrectly used—especially by French writers on architecture—that it is now extremely difficult to restrict it to the only style to which it really belongs. Wherever a certain amount of coloured decoration is employed, or a peculiar form of carving found, the name Byzantine is applied to churches on the Rhine or in France; although no similar ornaments are found in the Eastern Empire, and though no connection can be traced between the builders of the Western churches and the architects of Byzantium, or the countries subject to her sway.

Strictly speaking, the term ought only to be applied to the style of architecture which arose in Byzantium and the East after Constantine transferred the government of the Roman Empire to that city. It is especially the style of the Greek Church as contradistinguished from that of the Roman Church, and ought never to be employed for anything beyond its limits. The only obstacle to confining it to this definition occurs between the ages of Constantine and Justinian. Up to the reign of the last-named monarch the separation between the two churches was not complete or clearly defined, and the architecture was of course likewise in a state of transition, sometimes inclining to one style, sometimes to the other. After Justinian’s time, the line may be clearly and sharply drawn, and it would therefore be extremely convenient if the term “Greek architecture” could be used for the style of the Greek Church from that time to the present day.

If that term be inadmissible, the term “Sclavonic” might be applied, though only in the sense in which the Gothic style could be designated as Teutonic. Both, however, imply ethnographic distinctions which it would not be easy to sustain. The term “Gothic” happily avoids these, and so would “Greek,” but for the danger of its being confounded with “Grecian,” which is the proper name for the classical style of the ancient Greeks. If the employment of either of these terms is deemed inadvisable, it will be necessary to divide the style into Old and New Byzantine—the first comprehending the three centuries of transition that elapsed from Constantine to the Persian war of Heraclius and the rise of the Mahomedan power, which entirely changed the face of the Eastern Empire,—the second, or Neo-Byzantine, including all those forms which were practised in the East from the reappearance of the style, in or after the 8th century, till it was superseded by the Renaissance.

Thus divided, the true or old Byzantine style might be regarded as the counterpart of the early Romanesque or debased Roman style, except that, owing to the rapid development in the East, the former culminated in the erection of Sta. Sophia (A.D. 532-558); the Eastern Empire thus forming a style of its own of singular beauty and perfection, which it left to its Sclavonic successors to use or abuse as their means or tastes dictated. The Western Empire, on the contrary, was in a state of decay ending in a débâcle, from which the Romanesque style only partially emerged during the reign of Charlemagne and his successors with a new revival in the 11th century.

Though the styles of the East and the West became afterwards so distinctly separate, we must not lose sight of the fact, that during the age of transition (324-622) no clear line of demarcation can be traced. Constantinople, Rome, and Ravenna were only principal cities of one empire, throughout the whole of which the people were striving simultaneously to convert a Pagan into a Christian style, and working from the same basis with the same materials.[215] Prior to the age of Constantine one style pervaded the whole empire. The buildings at Palmyra, Jerash, or Baalbec, are barely distinguishable from those of the capital, and the problem of how the Pagan style could be best converted to Christian uses was the same for all. The consequence is, that if we were at present writing a history which stopped with the beginning of the 7th century, the only philosophical mode of treating the question would be to consider the style as one and indivisible for that period; but as the separation was throughout steadily, though almost imperceptibly, making its way, and gradually became fixed and permanent, it will be found more convenient to assume the separation from the beginning. This method will no doubt lead to some repetition, but that is a small inconvenience compared with the amount of clearness obtained. At the same time, if any one were writing a history of Byzantine architecture only, it would be necessary to include Ravenna, and probably Venice and some other towns in Italy and Sicily, in the Eastern division. On the other hand, in a history devoted exclusively to the Romanesque styles, it would be impossible to omit the churches at Jerusalem, Bethlehem, or Thessalonica, and elsewhere in the East. Under these circumstances, it is necessary to draw an arbitrary line somewhere; and for this purpose the western limits of the Turkish Empire and of Russia will answer every practical purpose. Eastward of this line every country in which the Christian religion at any time prevailed may be considered as belonging to the Byzantine province.

During the first three centuries of the style (324-622) it will be convenient to consider the whole Christian East as one architectural province. When our knowledge is more complete, it may be possible to separate it into several, but at present we are only beginning to see the steps by which the style grew up, and are still very far from the knowledge requisite for such limitations, even if it should hereafter be discovered that a sufficient number exist. All the great churches with which Constantine and his immediate successors adorned their new capital have perished. Like the churches at Jerusalem and Bethlehem, they were probably constructed with wooden roofs and even wooden architraves, and thus soon became a prey to the flames in that most combustible of capitals. Christian architecture has been entirely swept off the face of the earth at Antioch, and very few and imperfect vestiges are found of the seven churches of Asia Minor. Still, the recent researches of De Vogüé in Northern Syria,[216] and of Texier in Thessalonica[217] show how much unexpected wealth still remains to be explored, and in a few years more this chapter of our history may assume a shape as much more complete than what is now written, as it excels what we were compelled to be content with when the Handbook was published, 1855.

Since therefore, under present circumstances, no ethnographic treatment of the subject seems feasible, the clearest mode of presenting it will probably be to adopt one purely technical.

For this purpose it will be found convenient, first, to separate the Neo-Byzantine style from the older division, which, in order not to multiply terms, may be styled the Byzantine par excellence; the first chapter extending from Constantine, 324, to the Hejira, 622; and the second from that time to the end of the Middle Ages.

In reference to the ecclesiastical architecture of the first division, it is proposed to treat—

First, of churches of the basilican or rectangular forms, subdividing them into those having wooden, and those having stone roofs.

Secondly, to describe circular churches in the same manner, subdividing them similarly into those with wooden roofs, and those with stone roofs or true domes.

This subdivision will not be necessary in speaking of the Neo-Byzantine churches, since they all have stone roofs and true domes.

With regard to civil or domestic architecture very little can at present be said, as so little is known regarding it, but we may hope that, a few years hence, materials will exist for an interesting chapter on even this branch of the subject.

II. BASILICAS

CONTENTS.

Churches at Bethlehem, Jerusalem, and Thessalonica—Rectangular Churches in Syria and Asia Minor, with wooden roofs and stone vaults.

Basilicas may be subdivided into two classes—that in which the nave is divided from the side-aisles by pillars, carrying either entablatures or arches, as the most purely Romanesque—and that which has piers supporting arches only, and is transitional between the first style and the more original forms which were elaborated out of it.

Of the former class one of the most authentic and perfect is that erected at Bethlehem by Helena, the mother of Constantine, in front of the cave of the Nativity. The nave seems to be a nearly unaltered example of this age, with the advantage over the contemporary churches at Rome, that all its pillars and their capitals were made for the places they occupy, whereby the whole possesses a completeness and justness of proportion not found in the metropolis. Its dimensions, though sufficient for effect, are not large, being internally 103 ft. across, by 215 ft. east and west. The choir with its three apses does not seem to be part of the original arrangement, but to have been added by Justinian when he renovated—Eutychius says rebuilt—the church. My impression is that a detached circular building, external to the basilica, originally contained the entrance to the cave. The frescoes were added apparently in the 11th or 12th century.[218]

One of the principal points of interest connected with this church is, that it enables us to realise the description Eusebius gives us of the basilica which Constantine erected at Jerusalem in honour of the Resurrection. Like this church it was five-aisled, but had galleries; the apse also was on a larger scale than could well have been possible in the Bethlehem church, and adorned with twelve pillars, symbolical of the Apostles.

Of this building nothing now remains, and the only portion which could be claimed as part of Constantine’s work is the western wall of the Rotunda, which to a height of 15 to 20 ft. was cut out of the solid rock in order to isolate the Holy Sepulchre in the centre. The so-called tombs of Absalom and Zachariah in the valley of Jehoshaphat were detached in a similar way from the rock behind them.[219]

THESSALONICA.

As before mentioned, it is to Constantinople, or Alexandria, or Antioch, that we should naturally look to supply us with examples of the style of the early transition, but as these fail, it is to Thessalonica alone—in so far as we now know—that we can turn. In that city there are two ancient examples. One, now known as the Eski Djuma or old mosque (Woodcut No. 274), may belong to the 5th century, though there are no very exact data by which to fix its age. It consists of a nave, measuring, exclusive of narthex and bema, 93 ft. across by 120 ft.—very much the proportion of the Bethlehem church, but having only three aisles, the centre one 48 ft. in width. The other church, that of St. Demetrius, is larger, but less simple. It is five-aisled, has two internal transepts, and various adjuncts. Altogether it seems a considerable advance towards the more complicated form of a Christian church. Both these churches have capacious galleries, running above the side aisles, and probably devoted to the accommodation of the women. The date of St. Demetrius is most probably among the first years of the sixth century.[220] The general ordinance of the columns will be understood from the woodcut (No. 276). Generally they are placed on elevated square or octagonal bases, or pedestals, as in the tepidaria of the Thermæ in Rome, and all have a block (known as the dosseret), placed above the capital, which is supposed to represent the entablature of the Roman example, but is probably an original feature inserted over the capital to support the springing of the arch. In this form it is found very generally in the 5th and 6th centuries, after which it fell into disuse, an increased depth being given to the abacus of the capital to take its place.

So far as we now know, there is only one church of this class at Constantinople—that known as St. John Studius,—a three-aisled basilica, 125 ft. long by 85 in width externally. Its date appears to be tolerably well ascertained as A.D. 463, and from this circumstance, as well as its being in the metropolis, it shows less deviation from the classical type than the provincial examples just quoted. The lower range of columns supporting the gallery still retain the classical outline and support a horizontal entablature (Woodcut No. 277); the upper supporting arches have very little resemblance to the classical type, and are wanting in the architrave block or dosseret, which in fact never seems to have been admired in the capital.

SYRIA AND ASIA MINOR.

The country where—so far at least as we at present know—the Byzantine Basilica was principally developed was Northern Syria. Already in De Vogüé’s work on Central Syria some dozen churches are indicated having the aisles divided from the naves by pillars supporting arches. One of these only—that at Soueideh—has five aisles, all the rest three. Almost all have plain semicircular apses, sometimes only seen internally, like those mentioned further on (page 510), but sometimes also projecting, as was afterwards universally the fashion. Two at least have square terminations (Kefr Kileh and Behioh), but this seems exceptional. Most of them are almost the size of our ordinary parish churches—100 ft. by 60 or thereabouts—and all belong to the three centuries—the 4th, 5th, and 6th—of which this chapter especially treats.

The church at Baquoza may serve as a type of the class both in plan and section (Woodcuts Nos. 278, 279). Its dimensions externally are 60 ft. by 105; and besides the narthex—not shown in the section—it has four lateral porches. It has also two square chapels or vestries at the end of the aisles—an arrangement almost universal in these churches.

The most remarkable of the group, however, is that of St. Simeon Stylites, at Kalat Sema’n, about 20 miles east of Antioch. Its dimensions are very considerable, being 330 ft. long, north and south, and as nearly as may be, 300 ft. east and west, across what may be called the transepts. The centre is occupied by a great octagon, 93 ft. across, on a rock in the centre of which the pillar of that eccentric saint originally stood. This apparently was never roofed over, but stood always exposed to the air of heaven.[221]

The greater part of the conventual buildings belonging to this church still remain in a state of completeness,—a fact which will be startling to those who are not aware how many of the great religious establishments of Syria still stand entire, wanting only the roofs, which were apparently the only parts constructed of wood.

The whole of the buildings at Kalat Sema’n seem to have been completed within the limits of the 5th century, and not to have been touched or altered since they were deserted, apparently in consequence of the Mahomedan irruption in the 7th century. The most curious point is that such a building should have remained so long in such a situation, unknown to the Western world; for the notices hitherto published have been meagre and unsatisfactory in the extreme, and De Vogüé is only able to state that it was visited and described by the historian Evagrius in the year 560 A.D.

In the same province we find also the earliest examples of the use of pier arches in a church to separate the nave from the aisles. These seem to have been currently used in Northern Syria in the 6th century, though not found in the West—at least not used in the same manner—for several centuries later. Generally three such arches only were employed in the length of the nave, and they consequently left the floor so open and free, that it is very questionable if in churches of limited dimensions the introduction of a much larger number by the Gothic architects was an improvement. Taking it altogether, it is probable that such a church as that at Roueiha (Woodcut No. 282) would, if literally reproduced, make a better and cheaper church for an English parish than the Mediæval models we are so fond of copying. A considerable amount of perspective effect is obtained by throwing two transverse arches across the nave, dividing it into three compartments, each including four windows in the clerestory; and the whole design is simple and solid in a degree seldom surpassed in buildings of its class. Its dimensions are 63 ft. by 150 over all externally.

In many of these churches the transverse arches of the nave are omitted; and when, as at Qalb Louzeh (Woodcut No. 284), the clerestory is accentuated by roofing shafts, the same effect of perspective is obtained by other means, and perhaps as successfully. It is very interesting, however, to find that as early as the 6th century the architects were thoughtfully feeling their way towards those very principles of design which many centuries afterwards enabled the Gothic architects to produce their most successful effects. The introduction of four windows over each great arch, and of a rooting-shaft between each to support the beams of the roof, was a happy thought, and it is wonderful it was so completely lost sight of afterwards.

It is probable that the apse (Woodcut No. 284) was originally adorned with paintings or mosaics, or at least that it was intended it should be so ornamented; but even as it is, it is so well proportioned to the size of the church, and to its position, and so appropriately ornamented, that it is better than most of those found in Roman basilicas; and, for a small church, is a more dignified receptacle for the altar than either the French chevet or the English chancel.

Did our limits admit of it, it would be not only pleasant but instructive to dwell longer on this subject; for few parts of our inquiry can be more interesting than to find that, as early as the 6th century, the Roman basilica had been converted into a Christian church, complete in all its details, and—internally at least—in a style of architecture as consistent and almost as far removed from its classical prototype as the Mediæval Gothic itself.

Externally, too, the style was becoming independent of classical models, though hardly in the same degree. The porches of the churches were generally formed in two storeys, the lower having a large central arch of admission, the upper consisting of a colonnade which partially hid, while it supported, an open screen of windows that admitted a flood of light into the nave just in the position where it was most effective. Without glass or mullions such a range of windows must have appeared weak, and would have admitted rain; but when sheltered by a screen of pillars, it was both convenient and artistic.

This mode of lighting is better illustrated at Babouda, where it is employed in its simplest form. No light is admitted to the chapel except through one great semicircular window over the entrance, and this is protected externally by a screen of columns. This mode of introducing light, as we shall afterwards see, was common in India at this age, and earlier, all the Chaitya caves being lighted in the same manner; and for artistic effect it is equal, if not superior, to any other which has yet been invented. The light is high, and behind the worshipper, and thrown direct on the altar, or principal part of the church. In very large buildings it could hardly be applied, but for smaller ones it is singularly effective.

The external effect of these buildings though not so original as the interior, is still very far removed from the classical type, and presents a variety of outline and detail very different from the simplicity of a Pagan temple. One of the most complete is that at Tourmanin (Woodcut No. 287), though that at Qalb Louzeh is nearly as perfect, but simpler in detail. For a church of the 6th century it is wonderful how many elements of later buildings it suggests; even the western towers seem to be indicated, and, except the four columns of the gallery, there is very little to recall the style out of which it arose.

There are considerable remains of a wooden-roofed basilica at Pergamus, which may be even older than those just described; but having been built in brick, and only faced with stone—the whole of which is gone—it is difficult to feel sure of the character of its details and mouldings. It had galleries on either side of the nave, but how these were supported or framed is not clear. It may have been by wooden posts or marble pillars, and these would have either decayed or been removed. The two square calcidica or vestries, which in the Syrian churches terminate the side-aisles, are here placed externally like transepts, and beyond them are two circular buildings with domical roofs and square apses. What their use was is, however, doubtful. In fact, we know so little of the architecture of that age in Asia Minor that this building stands quite exceptionally; and very little use can be made of it, either as throwing light on other buildings, or as receiving illustration from their peculiarities. But seeing how much has been effected in this direction of late, we may fully hope that this state of isolation will not long remain.

One other church of the 4th century is known to exist—at Nisibin. It is a triple church, the central compartment being the tomb of the founder, the first Armenian bishop of the place. Though much ruined, it still retains the mouldings of its doorways and windows as perfect as when erected, the whole being of fine hard stone. These are identical in style with the buildings of Diocletian at Spalato; and as their date is well known, they will, when published, form a valuable contribution to the information we now possess regarding the architecture of this period.

CHURCHES WITH STONE ROOFS.

All the buildings above described—with the exception of the chapel at Babouda—have wooden roofs, as was the case generally with the basilicas and the temples of the classical age. The Romans, however, had built temples with aisles and vaulted them as early as the age of Augustus, as at Nîmes, for instance (Woodcut No. 189), and they had roofed their largest basilicas and baths with intersecting vaults. We should not therefore feel surprised if the Christians sometimes attempted the same thing in their rectangular churches, more especially as the dome was always a favourite mode of roofing circular buildings; and the problem which the Byzantine architects of the day set themselves to solve was—as we shall presently see—how to fit a circular dome of masonry to a rectangular building.

One of the earliest examples of a stone-roofed church is that at Tafkha in the Hauran. It is probably of the age of Constantine, though as likely to be before his time as after it. Its date, however, is not of very great importance, as its existence does not prove that the form was adopted from choice by the Christians: the truth being that, in the country where it is found, wood was never used as a building material. All the buildings, both domestic and public, are composed wholly of stone—the only available material for the purpose which the country afforded. In consequence of this, when that tide of commercial prosperity which rose under the Roman rule flowed across the country from the Euphrates valley to the Mediterranean, the inhabitants had recourse to a new mode of construction, which was practically a new style of architecture. This consisted in the employment of arches instead of beams. These were placed so near one another that flat stones could be laid side by side from arch to arch. Over these a layer of concrete was spread, and a roof was thus formed so indestructible that whole towns remain perfect to the present day, as originally constructed in the first centuries of the Christian era.[222]

One example must suffice to explain this curious mode of construction. The church at Tafkha is 50 ft. square, exclusive of the apse. It is spanned by four arches, 7 ft. 6 in. apart. On each side are galleries of flat slabs resting on brackets, as shown in Woodcuts Nos. 289, 291, which again are supported by smaller transverse arches. At one side is a tower, but this is roofed wholly by bracketing, as if the architect feared the thrust of the arch even at that height.

The defect of this arrangement as an architectural expedient is the extreme frequency of the piers, 8 or 10 ft. being the greatest distance practicable; but as a mechanical expedient it is singularly ingenious. More internal space is obtained with a less expenditure of material and danger from thrust than from any mode of construction—wholly of stone— that we are acquainted with; and with a little practice it might no doubt be much improved upon. The Indian architects, as we shall presently see, attempted the same thing, but set about it in a diametrically opposite way. They absolutely refused to employ the arch under any circumstances, but bracketed forward till the space to be covered was so limited that a single stone would reach across. By this means they were enabled to roof spaces 20 or 25 ft. span without arches, which is about the interval covered with their aid at Tafkha.[223]

Another circumstance which renders these Hauran examples interesting to the architectural student is that they contain no trace or reminiscence of wooden construction or adornment, so apparent in almost every other style. In Lycia it is absurdly so. In Egypt, in Greece, in India, in Persia—everywhere, in fact—we can trace back the principal form of decoration to a wooden original; here alone all is lithic, and it is probably the only example of the sort that the whole history of architecture affords.

If there are any churches in the Byzantine province of the age of which we are treating, whose naves are roofed by intersecting vaults, they have not yet been described in any accessible work; but great tunnel-vaults have been introduced into several with effect. One such is found at Hierapolis, on the borders of Phrygia (Woodcut No. 293). It is divided by a bold range of piers into three aisles, the centre one having a clear width of 45 ft. 6 in. The internal dimensions of the church are 177 ft. by 115. There are three great piers in the length, which carry bold transverse ribs so as to break the monotony of the vault, and have between them secondary arches, to carry the galleries.

There is another church at the same place, the roof of which is of a somewhat more complicated form. The internal length, 140 ft., is divided into three by transverse arches; but its great peculiarity is that the vault is cut into by semi-circular lunettes above the screen side-walls, and through these the light is introduced. This arrangement will be understood from the section (Woodcut No. 295). Taken altogether, there is probably no other church of its age and class in which the vault is so pleasingly and artistically arranged, and in which the mode of introducing the light is so judicious and effective.

The age of these two last churches is not very well ascertained. They probably belong to the 5th, and are certainly not later than the 6th, century; but, before we can speak with certainty on the subject, more examples must be brought to light and examined. From our present knowledge it can hardly be doubted that a sufficient number do exist to complete the chapter; and it is to be hoped they will be published, since a history of vaults in the East, independent of domes, is still a desideratum.

III. CIRCULAR OR DOMICAL BUILDINGS

CONTENTS.

Circular Churches with wooden roofs and with true domes in Syria and Thessalonica—Churches of St. Sergius and Bacchus and Sta. Sophia, Constantinople—Domestic Architecture—Tombs.

At the time of the erection of the churches described in the last chapter, a circular domical style was being simultaneously elaborated in the East, which not only gave a different character to the whole style, but eventually entirely superseded the western basilican form, and became an original and truly Byzantine art.

Constantine is said to have erected a church at Antioch which, from the description given by Eusebius, was octagonal in plan.

On Mount Gerizim, on or near the site of the Samaritan temple, Justinian built an octagonal church showing in its multifold chapels a considerable advance towards Christian arrangements; it has, however been so completely destroyed that only its foundation can now be traced, from which the plan (Woodcut No. 296) was measured and worked out by Sir Charles Wilson.

At Bosra in the Hauran there is a church of perfectly well-ascertained date—A.D. 512—which, when more completely illustrated, will throw considerable light on the steps by which a Pagan temple was transformed into a Christian church. It is a building externally square, but internally circular (Woodcut No. 297). The central space is 91 ft. in diameter, and was evidently covered with a wooden roof, according to M. de Vogüé, supported on eight piers. The interest of the plan consists in its showing the progress made in adapting this form to Christian purposes, and it is to be hoped that further investigation may enable us to supply all the steps by which the transformation took place. De Vogüé is of opinion that there was a central dome carried on piers and columns similar to the church of SS. Sergius and Bacchus in Constantinople, with aisles round and gallery over them, the latter covered with a timber roof, the holes in which the rafters were fixed being still visible. Owing to want of lateral support the dome fell down, and at a later period a small basilica church was erected within the enclosure in front of the apse; the proximity of the piers of this church suggests that it was covered with stone slabs according to the custom of the country. The inscription over the principal entrance door states that the church was dedicated to SS. Sergius and Bacchus, and was completed in the 400th year of Bosra (511-512 A.D.). Another example exists at Kalat Sema’n, in Northern Syria, and presents a combination of an octagonal with a rectangular church very common in Armenia and Georgia. As is generally the case there, they are very small in dimensions, the whole group only measuring 120 ft. by 73. Their actual destination is not known, but M. de Vogüé suggests that the triapsal arrangement in the octagonal building points to its having been erected as a baptistery. This group is situated about 200 yards from the main buildings illustrated in Woodcut (No. 280).

CHURCHES WITH DOMES.

Whether the dome of the Pantheon at Rome (p. 320) was erected in the time of the Antonines, or before the time of Augustus, as was formerly supposed, it is evident that the Romans had conquered the difficulties of domic construction long before the transference of the seat of power to Byzantium; the Pantheon being, up to this hour, the largest (single) dome ever constructed by the hand of man. Simple and grand as it undoubtedly is, it had several glaring defects in its design which the Byzantines set themselves to remedy. The first was that twice the necessary amount of materials was consumed in its construction. The second, that the mode of lighting by a hole in the roof, which also admitted the rain and the snow, was most objectionable before the invention of glass. The third, that a simply circular plan is always unmeaning and inconvenient. A fourth, that a circular building can hardly, by any contrivance, be made to fit on to any other buildings or apartments.

In the Minerva Medica (Woodcut No. 229) great efforts were made, but not quite successfully, to remedy these defects. The building would not fit on to any others, and, though an improvement on the design of the Pantheon, was still far from perfect.

The first step the Byzantines made was to carry the dome on arches resting on eight piers enclosing an octagon A (Woodcut No. 300); this enabled them to obtain increased space, to provide nave, choir, and transepts, and by throwing out niches on the diagonal lines, virtually to obtain a square hall in the centre. The difference between the octagon and circle is so slight, that by corbelling out above the extrados of the arches, a circular base for the dome was easily obtained B. The next step was to carry the dome on arches resting on four piers, and their triumph was complete when by the introduction of pendentives— represented by the shaded parts at D (Woodcut No. 301), they were enabled to place the circular dome on a square compartment. The pendentives and dome thus projected formed part of a sphere, the radius of which was the half-diagonal of the square compartment. Constructively it would probably have been easier to roof the space by an intersecting vault; and even if of 100 or 150 ft. span it would without difficulty have been effected. The difference between the intersecting vault and the dome (as shown in Woodcuts 302 and 303; the former the tomb of Galla Placidia, built 450 A.D., the latter the chapel of St. Peter Crysologus attached to the archiepiscopal palace of about the same date, and both in Ravenna) is perhaps the most striking contrast the history of architecture affords between mechanical and ornamental construction. Both are capable of being ornamented to the same extent and in the same manner; but the difference of form rendered the dome a beautiful object in itself wholly irrespective of ornament, whereas the same cannot always be said of the intersecting barrel vault. Altogether, the effect would have been architecturally so infinitely inferior, that we cannot but feel grateful to the Byzantines that they persevered, in spite of all mechanical temptations, till they reached the wonderful perfection of the dome of Sta. Sophia.

Among the earliest domical churches found in the East is that of St. George at Thessalonica. It is also, perhaps, the finest example of its class belonging strictly to that group which has been designated above as the Eastern Romanesque.

As will be seen from the plan it is a circular apartment, 79 ft. in diameter, surrounded by walls 20 ft. in thickness, into which are cut seven great niches; two apparently serving as entrances, opposite one of which is a bema or presbytery of considerable importance and purely Christian form. The dome is hemispherical, pierced at its base by eight semi-circular lunettes, and externally covered and concealed by a wooden roof. This form of roof is first found in the West at Nocera dei Pagani (p. 547), but the dome there is only half the diameter of this one, and of a very different form and construction. The dome of St. George’s retains its internal decorations, which are among the earliest as well as the most interesting Christian mosaics in existence.[224] The architecture presented in them bears about the same relation to that in the Pompeiian frescoes which the Jacobæan does to classical architecture, and, mixed with Christian symbols and representations of Christian saints, makes up a most interesting example of early Christian decoration.

No inscriptions or historical indications exist from which the date of the church can be fixed. We are safe, however, in asserting that it was erected by Christians, for Christian purposes, subsequently to the age of Constantine. If we assume the year 400 as an approximate date we shall probably not err to any great extent, though the real date may be somewhat later.

How early a true Byzantine form of arrangement may have been introduced we have no means of knowing; but as early as the year 285—according to De Vogüé—we have a Kalybe[225] at Omm-es-Zeitoun, which contains all the elements of the new style. It is square in plan, with a circular dome in its centre for a roof. The wing walls which extend the façade are curious, but not singular. One other example, at least, is found in the Hauran, at Chaqqa, and there may be many more.

Still, in the Hauran they never seem quite to have fallen into the true Byzantine system of construction, but preferred one less mechanically difficult, even at the expense of crowding the floor with piers. In the church at Ezra, for instance, the internal octagon is reduced to a figure of sixteen sides before it is attempted to put a dome upon it, and all thought of beauty of form, either internally or externally, is abandoned in order to obtain mechanical stability—although the dome is only 30 ft. in diameter.

As the date of this church is perfectly ascertained (510) it forms a curious landmark in the style just anterior to the great efforts Justinian was about to make, and which forced it so suddenly into its greatest, though a short-lived, degree of perfection.

CONSTANTINOPLE.

As before mentioned, all the churches of the capital which were erected before the age of Justinian, have perished, with the one exception of that of St. John Studius mentioned above (page 421). This may in part be owing to the hurried manner in which they were constructed, and the great quantity of wood consequently employed, which might have risked their destruction anywhere. It is, however, a curious, but architecturally an important, fact that Byzantium possessed every conceivable title to be chosen as the capital of the Empire, except the possession of a good building-stone, or even apparently any suitable material for making good bricks. Wood seems in all times to have been the material most readily obtained and most extensively used for building purposes, and hence the continual recurrence of fires, from before the time of Justinian down to the present day. That monarch was the first who fairly met the difficulty; the two churches erected during his reign, which now exist, are constructed wholly without wood or combustible materials of any sort—and hence their preservation.

The earliest of these two, popularly known as the “Kutchuk Agia Sophia,” or lesser Sta. Sophia, was originally a double church, or more properly speaking two churches placed side by side, precisely in the same manner as the two at Kalat Sema’n (Woodcut No. 298). The basilica was dedicated to the Apostles Peter and Paul; the domical church, appropriately, to the Martyrs Sergius and Bacchus. The former has entirely disappeared, from which I would infer that it was constructed with pillars and a wooden roof.[226] The latter remains very nearly intact. The frescoes and mosaics have, indeed, disappeared from the body of the church, hidden, it is to be hoped, under the mass of whitewash which covers its walls—in the narthex they can still be distinguished.

The existing church is nearly square in plan, being 109 ft. by 92 over all, exclusive of the apse, and covering only about 10,000 sq. ft. It has consequently no pretensions to magnificence on the score of dimensions, but is singularly elegant in design and proportion. Internally, the arrangement of the piers of the dome, of the galleries, and of the pillars which support them, are almost identical with those of St. Vitale at Ravenna, but the proportions of the Eastern example are better, being 66 ft. in height by 52 in diameter, while the other, with the same diameter, is nearly 20 ft. higher, and consequently too tall to be pleasing.

The details of this church are generally well designed for the purposes to which they are applied. There is a certain reminiscence of classical feeling in the mouldings and foliage—in the latter, however, very faint. The architrave block (No. 313) here seems almost to have superseded the capital, and what was once a classical entablature has retained very little of its pristine form (No. 314), and indeed was used constructively only, for the support of a gallery, or some such mechanical requirement. The arch had entirely superseded it as an ornamental feature long before the age of Justinian.

STA. SOPHIA.

Although the building just described, and others that might be quoted, probably contain the germs of all that is found in Sta. Sophia, they are on so small a scale that it is startling to find Justinian attempting an edifice so grand, and so daring in construction, without more experience than he appears to have obtained. Indeed so exceptional does this great structure appear, with our present knowledge, that we might almost feel inclined at first sight to look upon it as the immediate creation of the individual genius of its architect, Anthemius of Thralles; but there can be little doubt that if a greater number of contemporary examples existed we should be able to trace back every feature of the design to its origin. The scale, however, on which it was carried out was certainly original, and required great boldness on the part of the architect to venture upon such a piece of magnificence. At all events, the celebrated boast of its founder on contemplating his finished work was more than justified. When Justinian exclaimed, “I have surpassed thee, O Solomon,” he took an exaggerated view of the work of his predecessor, and did not realize the extent to which his building excelled the Jewish temple. The latter was only equal to a small church with a wooden roof supported by wooden posts, and covering some 7200 sq. ft. Sta. Sophia covers ten times that area, is built of durable materials throughout, and far more artistically ornamented than the temple of the Jews ever could have been. But Justinian did more than accomplish this easy victory. Neither the Pantheon nor any of the vaulted halls at Rome equal the nave of Sta. Sophia in extent, or in cleverness of construction, or in beauty of design. Nor was there anything erected during the ten centuries which elapsed from the transference of the capital to Byzantium till the building of the great mediæval cathedrals which can be compared with it. Indeed it remains even now an open question whether a Christian church exists anywhere, of any age, whose interior is so beautiful as that of this marvellous creation of old Byzantine art.

The original church of Sta. Sophia which had been erected by Constantine was, it seems, burnt to the ground in the fifth year of Justinian, A.D. 532, when he determined to re-erect it on the same spot with more magnificence and with less combustible materials. So rapidly were the works pushed forward, that in six years it was ready for dedication, A.D. 537. Twenty years afterwards a portion of the dome fell down in consequence of an earthquake; but this damage was repaired, and the church re-dedicated, A.D. 563, in the form, probably very nearly, in which we now find it.

In plan it closely approaches an exact square, being 235 ft. north and south by 250 east and west, exclusive of the narthex and apse. The narthex itself is a splendid hall, 205 ft. in length internally, by 26 ft. wide, and two storeys in height. Beyond this there is an exo-narthex which runs round the whole of the outer court, but this hardly seems to be part of the original design. Altogether, the building, without this or any adjuncts which may be after-thoughts, covers about 70,000 sq. ft., or nearly the average area of a mediæval cathedral of the first class.

Externally the building (Woodcut No. 316) possesses little architectural beauty beyond what is due to its mass and the varied outline arising from the mechanical contrivances necessary to resist the thrust of its internal construction. It may be that, like the early Christian basilicas at Rome, it was purposely left plain to distinguish it from the external adornment of Heathen temples, or it may have been intended to revêt it with marble, and add the external ornament afterwards. Before we became acquainted with the ornamental exteriors of Syrian churches, the former theory would seem the more plausible, though it can hardly now be sustained; and when we consider that the second dedication only took place the year before Justinian’s death, and how soon troublous times followed, we may fairly assume that what we now see is only an incomplete design. Whatever may be the case with the exterior, all the internal arrangements are complete, and perfect both from a mechanical and an artistic point of view. In such a design as this, the first requirement was to obtain four perfectly stable arches on which the dome might rest. The great difficulty was with the two arches running transversely north and south. These are as nearly as may be 100 ft. span and 120 high to the crown, and 10 ft. on the face. Each of them has a mass of masonry behind it for an abutment, 75 ft. long by 25 ft. wide, only partially pierced by arches on the ground and gallery floor; and as the mass might have been carried to any height, it ought, if properly constructed, to have sufficed for an arch very much wider and more heavily weighted than that which it supports. Yet the southern wall is considerably bulged, and the whole of that side thrown out of the perpendicular. This probably was the effect of the earthquake which caused the fall of the dome in 559, since no further settlement seems to have taken place. The longitudinal arches presented no difficulty. The distance between the solid parts of the piers was 75 ft., and this was filled up with a screen wall supporting the inner side of the arch; so, unless that was crushed, the whole was perfectly stable. Pendentives between these four arches ought not to have presented any difficulties. It would, however, have been better, from an architectural point of view, if they had been carried further up and forward, so as to hang a weight inside the dome to counteract the outward thrust, as was afterwards so successfully practised at Beejapore.[227] As it is, the dome rests rather on the outer edge of the system, without sufficient space for abutment. In itself the dome is very little lower than a hemisphere, being 107 ft. across by 46 ft. in height. Externally, it would have been better if higher; for internal effect this is sufficient. Its base is pierced by forty small windows, so small and so low as not to interfere in any way with the apparent construction, but affording an ample supply of light—in that climate at least—to render every part of the dome bright and cheerful.

Beyond the great dome, east and west, are two semi-domes of a diameter equal to that of the great dome, and these are again cut into by two smaller domes, so that the building, instead of being a Greek cross, as usually asserted, is only 100 ft. across in the centre and 125 wide beyond the central space each way. There is a little awkwardness in the way in which the smaller semi-domes cut into the larger, and the three windows of the latter are unconnected with any other part of the design, which is unpleasing, but might easily be remedied in a second attempt. These very irregularities, however, give a variety and appropriateness to the design which has probably never been surpassed. A single dome of the area of the central and two semi-domes would not have appeared nearly so large, and would have overpowered everything else in the building. As it is, the eye wanders upwards from the large arcades of the ground floor to the smaller arches of the galleries, and thence to the smaller semi-domes. These lead the eye on to the larger, and the whole culminates in the great central roof. Nothing, probably, so artistic has been done on the same scale before or since. In these arrangements Sta. Sophia seems to stand alone.

If, however, the proportions of this church are admirable, the details are equally so. All the pillars are of porphyry, verd antique, or marbles of the most precious kinds. The capitals are among the most admirable specimens of the style. It will be remembered that the governing line of a classical Corinthian capital is a hollow curve, to which acanthus-leaves or other projecting ornaments were applied. When the columns were close together, and had only a beam to support, this form of capital was sufficient; but when employed to carry the constructive arches of the fabric its weakness became instantly apparent. Long before Justinian’s time, the tendency became apparent to reverse the curve and to incise the ornament. In Sta. Sophia the transition is complete; the capitals are as full as elegance would allow, and all the surfaces are flat, with ornaments relieved by incision. In the lower tier of arches (Woodcut No. 318) this is boldly and beautifully done, the marble being left to tell its own story. In the upper tier, further removed from the eye, the interstices are filled in with black marble so as to ensure the desired effect.

All the flat surfaces are covered with a mosaic of marble slabs of the most varied patterns and beautiful colours; the domes, roofs, and curved surfaces, with a gold-grounded mosaic relieved by figures or architectural devices. Though much of the mosaic is now concealed, enough is left to enable the effect of the whole to be judged of, and it certainly is wonderfully grand and pleasing. The one thing wanting is painted glass, like that which adorns the Dome of the Rock at Jerusalem, to render this building as solemnly impressive as it is overpoweringly beautiful.

Sta. Sophia is so essentially different from the greater number of churches, that it is extremely difficult to institute a comparison between them. With regard to external effect, Gothic cathedrals generally excel it; but whether by accident or by the inherent necessity of the style is by no means so clear. In so far as the interior is concerned, no Gothic architect ever rose to the conception of a hall 100 ft. wide, 250 ft. in length, and 180 ft. high, and none ever disposed each part more artistically to obtain the effect he desired to produce. Where the Byzantine style might profit from the experience subsequently gained by Gothic architects is in the use of mouldings. The one defect in the decoration of Sta. Sophia is that it depends too much on colour. It would have been better if the pier-arches, the window-frames, and the string-courses generally had been more strongly accentuated by moulding and panellings, but this is a slight defect among so many beauties.

A comparison with the great Renaissance cathedrals is more easy, but results even more favourably to the Byzantine example. Two of these have domes which are considerably larger—St. Peter’s at Rome and Sta. Maria at Florence being each 126 ft.; St. Paul’s, London (108), is within a foot of the same diameter, all the rest are smaller.[228] This, however, is of less consequence than the fact that they are all adjuncts to the design of the church. None of them are integral or supported by the rest of the design, and all tend to dwarf the buildings they are attached to rather than to heighten the general effect. With scarcely an exception also all the Renaissance cathedrals employ internally great sprawling pillars and pilasters, designed for external use by the Romans, which not only diminish the apparent size of the building but produce an effect of unreality and sham utterly fatal to true art.

In fact, turn it as we will, and compare it as we may with any other buildings of its class, the verdict seems inevitable that Sta. Sophia— internally at least, for we may omit the consideration of the exterior, as unfinished—is the most perfect and most beautiful church which has yet been erected by any Christian people. When its furniture was complete the verdict would probably have been still more strongly in its favour; but so few of the buildings described in these pages retain these adjuncts in anything like completeness that they must be withdrawn from both sides and our remarks be confined to the architecture, and that only.

The church of Sta. Sophia at Thessalonica, according to Greek tradition, was built by Justinian in the latter part of his reign.[229] It is a church of considerable dimensions, measuring 140 ft. east and west by 118 ft. in width, with a dome 33 ft. in diameter. It possesses also an upper gallery, and its arrangements generally are well considered and artistic. There does not seem to be any documentary evidence of its age, but judging from the details published in Texier, the date ascribed to it seems probable. This has been further established lately from an inscription found in the apse, which as well as the dome still retain their ancient mosaics; the inscription is incomplete, but Messrs. Duchesne and Bayet, in an appendix to their work on Mount Athos, ascribe it to the second half of the 6th century. The church possesses one special characteristic: above the pendentives is a low drum, circular internally,[230] in which windows are pierced, but which, externally, is carried up square: by this means the angle piers are well weighted and are thus enabled to resist more effectually the thrust of the arches carrying the pendentives. The two side walls also, which in Sta. Sophia at Constantinople were built almost flush with the inner arch, leaving outside a widely-projecting arch thrown across between the buttresses to carry the buttresses of the dome, are here placed flush with the outside of the arch, thus giving increased space to the interior.

DOMESTIC ARCHITECTURE.

The publication of the Count De Vogüé’s book has enabled us to realise the civil and domestic architecture of Syria in the 5th and 6th centuries with a completeness that, a very short time ago, would have been thought impossible. Owing to the fact that every part of the buildings in the Hauran was in stone, and that they were suddenly deserted on the Mahomedan conquest, never, apparently, to be re-occupied, many of the houses remain perfectly entire to the present day, and in Northern Syria only the roofs are gone.

Generally they seem to have been two storeys in height, adorned with verandahs supported by stone columns, the upper having a solid screen-fence of stone about 3 ft. 6 in. high, intended apparently as much to secure privacy to the sleeping apartments of the house as protection against falling out. In some instances the lower storey is twice the height of the upper, and contained the state apartments of the house. In others, as in that at Refadi (Woodcut No. 320), it seems to have been intended for the offices. In the plan of a house at Moudjeleia (Woodcut No. 321) the principal block of the house is in two storeys, with portico on ground floor and verandah over. The buildings at the back with their courtyard were probably offices, and those in front by the side of the main entrance warehouses or stores.

In some instances one is startled to find details which we are accustomed to associate with much more modern dates; as, for instance, this window (Woodcut No. 322) from the palace at Chaqqa, which there seems no reason whatever for doubting belongs to the 3rd century— anterior to the time of Constantine! It looks more like the vagary of a French architect of the age of Francis I.

The building known as the Golden Gateway at Jerusalem and attributed to Justinian, bears in its details many striking resemblances to those of the 5th and 6th centuries in Central Syria, illustrated in De Vogüé’s book. It is situated on the east side of the Haram enclosure, and consists of a vestibule divided by columns into two aisles of three bays each vaulted with a cupola[231] carried on arches, between which and the capitals of the columns is found the Byzantine dosseret already referred to. Within the eastern doorways (said to have been blocked up by Omar) are two huge monoliths 14 ft. 6 in. and 11 ft. respectively, the doorposts of an earlier gateway. Externally, on the entrance fronts (east and west), the entablature of the pilasters is carried round the circular-headed doorways which they flank; the earliest instance of this development is found in the Palace of Diocletian at Spalato, and there is a second example in the Roman gateway to the Mosque of Damascus, which probably suggested the idea to the Byzantine builders; the sharp stiff foliage of Greek type with which the ornament is carved on the Golden Gate agrees in style and character with that in the church of St. Demetrius at Thessalonica dating from the commencement of the 6th century.

Of similar style and character are the arch-moulds of the double gate on the south wall of the Haram, and the cupolas of the interior vestibule, the columns carrying them however being probably of earlier date and possibly part of the substructure of Herod’s temple. The surface decoration of these cupolas is similar to that found in Central Syria.

The sepulchral remains of Syria, both structural and rock-cut, seem nearly as numerous as the dwellings of the living, and are full of interest, not only from their frequently bearing dates, but from their presenting new types of tombs, or old types in such new forms as scarcely to be recognizable.

The oldest example, that of Hamrath in Souideh, dates from the 1st century B.C., and consists of a tomb 28 ft. square decorated with semi-detached Doric columns; the roof is gone, but it was probably covered with one of pyramidal form like the tomb of Zechariah (Woodcut No. 238).

The tomb of Diogenes at Hass (Woodcut No. 326), also square, consisted of two storeys, with a portico on the ground storey on one side, and a peristyle on all four sides of the upper storey, above which rose the central walls carrying a pyramidal roof, not stepped, as in the Mausoleum at Halicarnassus, but with projecting bosses on each stone. The same class of roof is found on other tombs, being adopted probably as the simplest method of covering over the tomb; these tombs date from the 4th and 5th centuries, and in all cases the sepulchral chambers within them are vaulted with large slabs of stone carried on stone ribs.

Besides these, there is another class of tomb apparently very numerous, in which the sepulchral chamber is below the ground, with vaulted entrance rising to form a podium on which columns either two or four in number are erected;[232] in the latter case the columns bearing an entablature with small pyramidal roof; in the former a fragment of architrave only, the two columns being sometimes tied together one-third of the way down by a stone band with dentils carved on it: these tombs are, many of them, dated, and belong to the 2nd and 3rd centuries.

With our present limits it is only possible to characterize generally the main features of the Byzantine style, and to indicate the sources from which further information may be obtained. In the present instance it is satisfactory to find that ample materials now exist for filling up a framework which a few years ago was almost entirely a blank. Any one who will master the works of De Vogüé, or Texier, or Salzenberg, and other minor publications, may easily acquire a fair knowledge of the older Byzantine style of architecture. Once it is grasped it will probably be acknowledged that there are few more interesting chapters than that which explains how a perfect Christian Church like that of Sta. Sophia was elaborated out of the classical edifices of ancient Rome. It will also probably be found that there are few more instructive lessons to be learnt from the study of architectural history than the tracing of the various contrivances which were so earnestly employed, during the first two centuries of Christian supremacy, in attaining this result.

IV. NEO-BYZANTINE STYLE

CONTENTS.

Sta. Irene, Constantinople—Churches at Ancyra, Trabala, and Constantinople—Churches at Thessalonica and in Greece—Domestic Architecture.

Santa Sophia at Constantinople was not only the grandest and most perfect creation of the old school of Byzantine art, but it was also the last. It seems as if the creative power of the Empire had exhausted itself in that great effort, and for long after it the history is a blank. We always knew that the two centuries which elapsed between the ages of Constantine and Justinian were ages of great architectural activity. We knew that hundreds, it may be thousands, of churches were erected during that period. With the two subsequent centuries, however, the case seems widely different. Shortly after Justinian’s death, the troubles of the Empire, the Persian wars of Heraclius, and, more than either, the rise of the Mahomedan power in the East, and of the Roman pontificate under Gregory the Great in the West—all tended so to disturb and depress the Byzantine kingdom as to leave little leisure and less means for the exercise of architectural magnificence. It is therefore hardly probable that we shall ever be in a position to illustrate the 7th and 8th centuries as we now know we can the 5th and 6th. Still, building must have gone on, because when we again meet the style, it is changed. One of the very earliest churches of the new school is that of Sta. Irene at Constantinople, rebuilt as we now find it by Leo the Isaurian (A.D. 718-740). It differs in several essential particulars from the old style, and contains the germ of much that we find frequently repeated. The change is not so great as might have taken place in two centuries of building activity, but it is considerable. In this church we find, apparently for the first time in a complete form, the new mode of introducing the light to the dome through a perpendicular drum, which afterwards became so universal that it serves to fix the age of a building in the East with almost as much certainty as the presence of a pointed arch does that of a building in the West. As this invention is so important, it may be well to recapitulate the steps by which it was arrived at.

The oldest mode of lighting a dome is practised in the Pantheon (Woodcut No. 191), by simply leaving out the central portion. Artistically and mechanically nothing could be better, but before the invention of glass it was intolerably inconvenient whenever much rain or snow fell. A change therefore was necessary, and it is found in the tomb or temple of Marcellus, built during the reign of Constantine on the Via Prenestina at Rome. It consists simply of boring four circular holes through the dome a little above its springing. The next step is seen at Thessalonica in the church of St. George (Woodcut No. 305). There eight semi-circular lunettes are pierced in the dome, at its springing, and answer the purpose very perfectly. The system culminated in Sta. Sophia, where forty windows introduce a flood of light without its ever falling on the eyes of the spectator. After this it seems to have been considered desirable not to break the hemisphere of the dome, but to place the windows in a perpendicular circular rim of masonry—called the drum—and to introduce the light always through that. Externally there can be no doubt but that this was an improvement; it gave height and dignity to the dome in small churches, where, without this elevation, the feature would have been lost. Internally, however, the advantage is problematical: the separation of the dome from its pendentives destroyed the continuity of the roof, and introduced the stilted effect so objectionable in Renaissance domes. In the Neo-Byzantine churches the dome became practically a skylight on the roof, the drum increasing in height and the dome diminishing in dignity as the style progressed. As all the churches are small, the feature is unobjectionable; but in larger edifices it would have been found difficult to construct it, and the artistic result would hardly have been pleasing, even had this difficulty been got over. Be this as it may, its value as a chronometric landmark is undoubted.

As a rule it may generally be asserted that, in all Christian domes erected during the old Byzantine period, the light is introduced by openings in the dome itself.[233] After that time, the light is as generally admitted through windows in the drum, the dome itself being cut into only in the rarest possible instances.

If these views are correct, the church of St. Clement at Ancyra is a transitional specimen subsequent to Sta. Sophia, because the dome is raised timidly (Woodcut No. 328) on a low drum pierced with four small windows; but it is anterior to Sta. Irene, because the dome is still pierced with twelve larger windows, after the manner of Sta. Sophia and the older churches. All the details of its architecture, in so far as they can be made out, bear out this description. They are further removed from the classical type than the churches of Justinian, and the whole plan (Woodcut No. 329) is more that which the Greek church afterwards took than any of the early churches show. Its greatest defect—though the one most generally inherent in the style—is in its dimensions. It is only 64 ft. long, over all externally, by 58 ft. wide. Yet this is a fair average size of a Greek church of that age.

Another church, very similar, is found at Myra, dedicated to St. Nicholas. It exceeds that of St. Clement in size, and has a double narthex considerably larger in proportion, but so ruined that it is difficult to make out its plan, or to ascertain whether it is a part of the original structure, or a subsequent addition. The cupola is raised on a drum, and altogether the church has the appearance of being much more modern than that at Ancyra.

A third church of the same class, and better preserved, is found at Trabala in Lycia. It is of the same type as St. Clement, and similar in its arrangements to Sta. Sophia, except in the omission of the semi-domes, which seem never to have been adopted in the provinces,[234] and indeed may be said to be peculiar to the metropolitan church. Notwithstanding the beauty of that feature, it appears to have remained dormant till revived by the Turks in Constantinople, and there alone.

In this example there are two detached octagonal buildings, either tombs or sacristies; a form which, except in large detached buildings, does not seem to have been so common as the circular, till after the time of Justinian.

Returning to the capital, we find one other remarkable peculiarity of the Neo-Byzantine style in the attempt to allow the external surface of an ordinary tunnel-vault to retain its form without any ridge whatever. It can hardly be doubted that this is artistically a mistake. With domes it was early felt to be so, and consequently we always find a flower or pinnacle in iron, or some such ornament, marking the centre. In this the Saracenic architects were especially successful—all their domes possess a central ornament sufficient to relieve them, and generally of the most beautiful proportions. With the extrados of a circular vault, however, it is even worse than with a dome. A roof is felt to be a contrivance to keep off the rain. It may be more or less sloping, according to the materials of which it is constructed; but to make one part of each ridge sloping, and the central portion flat, is a discord that offends the eye, besides looking weak and unmeaning. A pointed arch would avoid the evil, but a reverse or ogee curve is perhaps the most pleasing. In the Neo-Byzantine age, however, between the 8th and the 12th centuries, the eye seems to have got accustomed to it. It is common in the East, especially at Constantinople and at Venice. In St. Mark’s and elsewhere it became so familiar a form that it was copied and continued by the Renaissance architects even to the end of the 16th century.

One of the best illustrations of these peculiarities is the church of Moné tés Choras at Constantinople, now converted into a mosque and called Kahriyeh Djamisi. The older part of it seems to belong to the 11th century, the side-aisles to the 12th, and though small, it illustrates the style perfectly. The porch consists of five arches covered with an intersecting vault, visible both externally and internally. The last two bays are covered with cupolas which still retain their mosaics internally, and those of singular beauty and brilliancy, though, owing to the constructive defects of the intermediate parts, the wet has leaked through, and the mosaics have mostly peeled off. Externally the front is ornamented with courses of stones alternating with two or three layers of tiles, and even in its ruined state is effective and picturesque. Its principal interest is that it shows what was the matrix[235] of the contemporary church of St. Mark at Venice. Subsequent additions have much modified the external appearance of St. Mark, but there can be very little doubt that originally it was intended to be very like the façade shown in Woodcut No. 331.

Not far from Moné tés Choras there are two other churches of the same class and of about the same age. One, the Pantokrator, has been added to at various times so as to cover a large space of ground, but it consists consequently of small and ill-assorted parts. It retains, however, a good deal of its marble pavements and other features of interest. The other, known as the Fethîyeh Djamisi, is smaller and more complete, and possesses some mosaics of considerable beauty.

The best example of its class, however, in Constantinople is that known as the Theotokos. Like those just mentioned it is very small, the church itself being only 37 ft. by 45, and, though its double narthex and lateral adjuncts add considerably to its dimensions, it is still only a very small church. Some parts of it are as old as the 9th or 10th century, but the façade represented in Woodcut No. 333 is certainly not older than the 12th century. Taking it altogether, it is perhaps the most complete and elegant church of its class now known to exist in or near the capital, and many of its details are of great beauty and perfection.

It seems scarcely possible to suppose that the meagre half-dozen of small churches just enumerated are all that were erected in the capital between the death of Justinian and the fall of the city. Yet there is no evidence that the Turks destroyed any. Why should they? They converted them into mosques, finding them especially convenient for that purpose, and they have maintained them with singularly little alteration to the present day.

THESSALONICA.

This deficiency of examples in the capital is to some extent supplied by those which are found existing at Thessalonica. Three churches belonging to this age are illustrated in Texier and Pullan’s work.

The first of these is the church of Kazandjita Djami, dedicated to the Mother of God, a small church measuring only 53 ft. by 37, exclusive of the apse. Its date is perfectly ascertained—viz., 1028.

Next to these comes the church of Elias, A.D. 1054, and very similar to it in style is that of the Apostles (Woodcut No. 334), which we may consequently date with safety in the 11th century, from this juxtaposition alone, though there are several other examples which enable us to treat it as a characteristic type of the age. It is a pleasing and picturesque specimen of Byzantine brickwork. Like all the churches of the time, it is small, 63 ft. by 59 externally. In plan it very much resembles the Theotokos at Constantinople, but in elevation is taller and thinner; though whether this arises from any local peculiarity, or from some difference of age, is not clear. I suspect the former. The earthquakes of the capital may have induced a less ambitious form, as far as height is concerned, than was adopted in the provinces.

GREECE.

There can be little doubt but that, if a systematic search were made among the churches of Greece, many would be brought to light which would be most useful in completing our knowledge of the Neo-Byzantine style.[236] At Mount Athos there exists from twenty to thirty monasteries, each with its Catholicon or principal church and other chapels. Many of these are of ancient date, ranging between the 10th and 16th centuries, and although some of them may have been restored, in some cases rebuilt in later times, they have not yet been examined or illustrated by any competent architect. Brockhaus in his work[237] gives the plan of three churches, one of which, the Catholicon (dated 1043) of the Dochiariu Monastery (Woodcut No. 335), is further illustrated by a bird’s-eye view taken from a photograph. The domes and drums over the narthex and two eastern chapels would seem to be later additions, made either in consequence of the proximity of the buildings of the monastery which obscured the light obtainable from windows, or to show better the wall frescoes, which in the case of the narthex, where no windows ever existed, must have been quite dark at first. The oldest church (963 A.D.) apparently is that of the Protaton at Caryas, which consists of a short nave, a transept, and a long choir, and is wanting in that one feature which is supposed to be characteristic of a Byzantine church, viz., a dome; the whole building is covered like a basilica with a flat wooden roof, beneath which are clerestory windows. Photogravures or woodcuts are given of the churches of Chilandari (1197 A.D.), Xeropotamu (1028-34 A.D.), the Laura (963 A.D., but rebuilt under Turkish rule), and woodcuts from photographs in an interesting description of the Monasteries by Mr. A. Riley,[238] give a good general idea of the work to be found in Athos, from which it would seem that the chief interest centres in the sumptuous carvings of the icon and stalls,[239] and in the frescoes with which most of the interiors of the churches are painted.

For Greece proper we are dependent almost wholly on Couchaud[240] and Blouet.[241] So far as the illustrations go they suggest that there are no churches of such dimensions as would ensure dignity, nor are any so beautiful in outline or detail as to make us regret much that we do not know more about them. Still they are sufficiently original to be worthy of study, and when properly known may help to join together some of the scattered links of the chain which once connected the architecture of the West and East, but which is at present so difficult to follow out.

In Athens there are several churches of considerable interest, and not without architectural pretension. They are all small, however. The largest is that known as Panagia Lycodemo, or the church of St. Nicodemus, and is only 62 ft. long by 45 ft. wide over all. It seems also to be the oldest, since its dome is partially pierced with windows inside, though outside there is a distinctly marked drum (Woodcut No. 337). Notwithstanding the smallness of its dimensions, considerable effect is obtained internally by the judicious arrangement of the parts and the harmony of proportion which reigns throughout. The exterior is also pleasing, though the loss of the cornice gives an unfinished look to the whole, and there is a want of sufficient connection between the dome and the walls of the building to make them part of one composition.

A more beautiful and more interesting example is the church known as the Catholicon or Cathedral at Athens (Woodcut No. 338). It is a cathedral, however, only in a Greek sense, certainly not as understood in the Latin Church, for its dimensions are only 40 ft. by 25 over all externally. It is almost impossible to judge of its age from its details, since they are partly borrowed from older classical buildings, or imitations of classical forms, so fashioned as to harmonize with parts which are old. But the tallness of its dome, the form of its windows, and the internal arrangements, all point to a very modern date for its erection—as probably the 13th century as the 11th or 12th.

The church of the Virgin at Mistra in the Peloponnesus was built in the 13th century on a hillside overlooking the plain of Sparta, and partly with materials taken from the remains of the ancient city; but though it belongs possibly to the same age as the Catholicon at Athens, it differs considerably from it in style, and bears much more resemblance to the churches of Apulia and Sicily than either of those described above.

Where arcades are used externally in these Greek churches, they are generally supported by pillars of somewhat classical look (often old classic columns and capitals were used up), crowned by capitals of the square foliaged form, employed to support arches in the early styles all over Europe; and the windows, when divided, take merely the form of diminutive arcades. The Byzantines never attained to tracery; all their early windows are single round-headed openings. These were afterwards grouped together in threes and fives; and, as in the Gothic style, when they could be put under one discharging arch, the piers were attenuated till they became almost mullions, but always supporting constructive arches, without any tendency to run into interlacing forms like the Gothic. The universal employment of mural painting in Byzantine churches, and the consequent exclusion of painted glass, rendered the use of the large windows which the Gothic architects employed quite inadmissible; and in such a climate very much smaller openings sufficed to admit all the light that was required. Tracery would thus, in fact, have been an absurdity, and the windows were often filled in with transparent marble slabs pierced with holes, which were either glazed or occasionally even left open. The Byzantine architects sought to ornament their windows externally by the employment of tiles or colours disposed in various patterns, and often produced a very pleasing effect, as may be seen from the woodcut (No. 337) illustrating the apse of the Panagia Lycodemo at Athens, in the Hebdomon Palace (Woodcut No. 342), and other specimens already quoted.

Occasionally we find in these churches projecting porches or balconies, and machicolations, which give great relief to the general flatness of the walls. These features are all marked with that elegance peculiar to the East, and more especially to a people claiming descent from the ancient Greeks, and possibly having some of their blood in their veins. Sometimes, too, even a subordinate apse is supported on a bracket-like balcony, so as to form a very pleasing object, as in the accompanying specimen from Mistra.

On the whole the Neo-Byzantine style may be said to be characterised by considerable elegance, with occasional combinations of a superior order; but after the time of Justinian the country was too deficient in unity or science to attempt anything great or good, and too poor to aspire to grandeur, so that it has no claim to rank among the great styles of the earth.[242] The old Byzantine style was elevated to a first-class position through the buildings of Justinian; but from his time the history of the art is a history of decline, like that of the Eastern Empire itself and of Greece, down to the final extinction both of the Empire and the style, under the successive conquests by the Venetians and the Turks. The only special claim which the Neo-Byzantine style makes upon our sympathies or attention is that of being the direct descendant of Greek and Roman art. As such, it forms a connecting link between the past and present which must not be overlooked, while in itself it has sufficient merit to reward the student who shall apply himself to its elucidation.

DOMESTIC ARCHITECTURE.

It is more than probable that very considerable remains of the civil or domestic architecture of the Neo-Byzantine period may still be recovered. Most of their palaces or public buildings have continued to be occupied by their successors, but the habits of Turkish life are singularly opposed to the prying of the archæologist. Almost the only building which has been brought to light and illustrated is the palace of the Hebdomon at Blachernæ in Constantinople, built by Constantine Porphyrogenitus (913-949). All that remains of it, however, is a block of buildings 80 ft. by 40 in plan, forming one end of a courtyard; those at the other end, which were more extensive, being too much ruined to be restored. The parts that remain probably belong to the 9th century, and consist of two halls, one over the other, the lower supported by pillars carrying vaults, the upper free. The façade towards the court (Woodcut 342) is of considerable elegance, being adorned by a mosaic of bricks of various colours disposed in graceful patterns, and forming an architectural decoration which, if not of the highest class, is very appropriate for domestic architecture.

One great cause of the deficiency of examples may be the combustibility of the capital. They may have been destroyed in the various fires, and outside Constantinople the number of large cities and their wealth and importance was gradually decreasing till the capital itself sunk into the power of the Turks in the year 1453.

V. ARMENIA

CONTENTS.

Churches at Dighour, Usunlar, Pitzounda, Bedochwinta, Mokwi, Etchmiasdin, and Kouthais—Churches at Ani and Samthawis—Details.

CHRONOLOGY.

Dates. Tiridates converted to Christianity by Gregory II. A.D. 276 St. Gregory confirmed as Pontiff by Pope Sylvester 319 Christianity proscribed and persecuted by the Persians 428-632 Fall of Sassanide dynasty. 632 Establishment of Bagratide dynasty under Ashdod 859 Greatest prosperity under Apas 928 Ashdod III. 951 Sempad II. 977-989 Alp Arslan takes Ani 1064 Gajih, last of the dynasty, slain 1079 Gengis Khan 1222

The architectural province of Armenia forms an almost exact pendant to that of Greece in the history of Byzantine architecture. Both were early converted to Christianity, and Greece remained Christian without any interruption from that time to this. Yet all her earlier churches have perished, we hardly know why, and left us nothing but an essentially Mediæval style. Nearly the same thing happened in Armenia, but there the loss is only too easily accounted for. The Persian persecution in the 5th and 6th centuries must have been severe and lasting, and the great bouleversement of the Mahomedan irruption in the 7th century would easily account for the disappearance of all the earlier monuments. When, in more tranquil times—in the 8th and 9th centuries—the Christians were permitted to rebuild their churches, we find them all of the same small type as those of Greece, with tall domes, painted with frescoes internally, and depending for external effect far more on minute elaboration of details than on any grandeur of design or proportion.

Although the troubles and persecutions from the 5th to the 8th century may have caused the destruction of the greater part of the monuments, it by no means follows that all have perished. On the contrary, we know of the church above alluded to (p. 428) as still existing at Nisibin and belonging to the 4th century, and there can be little doubt that many others exist in various corners of the land; but they have hardly yet been looked for, at least not by anyone competent to discriminate between what was really old and what may have belonged to some subsequent rebuilding or repair.

Till this more careful examination of the province shall have been accomplished, our history of the style cannot be carried back beyond the Hejira. Even then very great difficulty exists in arranging the materials, and in assigning correct dates to the various examples. In the works of Texier,[243] Dubois,[244] Brosset,[245] and Grimm[246] some forty or fifty churches are described and figured in more or less detail, but in most cases the dates assigned to them are derived from written testimony only, the authors not having sufficient knowledge of the style to be able to check the very fallacious evidence of the litera scripta. In consequence of this, the dates usually given are those of the building of the first church on the spot, whereas, in a country so troubled by persecution as Armenia, the original church may have been rebuilt several times, and what we now see is often very modern indeed.

Among the churches now existing in Armenia, the oldest seems to be that in the village of Dighour near Ani. There are neither traditions nor inscriptions to assist in fixing its date; but, from the simplicity of its form and its quasi-classical details, it is evidently older than any other known examples, and with the aid of the information conveyed in De Vogüé’s recent publications we can have little hesitation in assigning it to the 7th century.[247] The church is not large, being only 95 ft. long by 82 wide over all. Internally its design is characterised by extreme solidity and simplicity, and all the details are singularly classical in outline. The dome is an ellipse, timidly constructed, with far more than the requisite amount of abutment. One of its most marked peculiarities is the existence of two external niches placed in projecting wings and which were no doubt intended to receive altars. Its flanks are ornamented by three-quarter columns of debased classical design. These support an architrave which is bent over the heads of the windows as in the churches of Northern Syria erected during the 6th century.

Its western and lateral doorways are ornamented by horse-shoe arches, which are worth remarking here, as it is a feature which the Saracenic architects used so currently and employed for almost every class of opening. The oldest example of this form known is in the doorway of the building called Takt-i-Gero on Mount Zagros.[248] In this little shrine, all the other details are so purely and essentially classic that the building must be dated before or about the time of Constantine. The horse-shoe arch again occurs in the church at Dana on the Euphrates in 540.[249] At Dighour we find it used, not in construction but as an ornamental feature. The stilting of the arch was evidently one of those experiments which the architects of that time were making in order to free themselves from the trammels of the Roman semi-circular arch. The Saracens carried it much further and used it with marked success, but this is probably the last occasion in which it was employed by a Christian architect as a decorative expedient.

The six buttresses, with their offsets, which adorn the façade, are another curious feature in the archæology of this church. If they are integral parts of the original design, which there seems no reason to doubt, they anticipate by several centuries the appearance of this form in Western Europe.

One of the oldest and least altered of the Armenian churches seems to be that of Usunlar, said to have been erected by the Catholicos Jean IV. between the years 718 and 726. In plan it looks like a peristylar temple, but the verandahs which surround it are only low arcades, and have very little affinity with classical forms. These are carried round the front, but there pierced only by the doorway. The elevation, as here exhibited, is simple, but sufficiently expresses the internal arrangements, and, with an octagonal dome, forms, when seen in perspective, a pleasing object from every point of view. Both plan and design are, however, exceptional in the province. A far more usual arrangement is that found at Pitzounda in Abkassia, which may be considered as the typical form of an Armenian church. It is said to have been erected by the Emperor Justinian, and there is nothing in the style or ornamentation of the lower part that seems to gainsay its being his. But the plan is so like many that belong to a much later age, that we must hesitate before we can feel sure that it has not been rebuilt at some more modern date. Its cupola certainly belongs to a period long after the erection of Sta. Irene at Constantinople (Woodcut No. 327), when the dome pierced with tall windows had become the fashionable form of dome in the Byzantine school. Its interior, also, is unusually tall, and the pointed arches under the dome look like integral parts of the design, and when so employed belong certainly to a much more modern date. On the whole, therefore, it seems that this church, as we now see it, may have been rebuilt in the 9th or 10th century.

Whatever its date, it is a pleasing example of the style. Externally it is devoid of ornament except what is obtained by the insertion of tiles between the courses of the stone, and a similar relief to the windows; but even this little introduction of colour gives it a gay and cheerful appearance, more than could easily be obtained by mouldings or carving in stone.

The upper galleries of the nave and the chapels of the choir are also well expressed in the external design, and altogether, for a small church—which it is (only 137 ft. by 75)—it is as pleasing a composition as could easily be found.

The idea that the date of this church is considerably more modern than Dubois and others are inclined to assign to it, is confirmed by a comparison of its plan with that at Bedochwinta, which Brosset determines from inscriptions to belong to the date 1556-1575; and the knowledge lately acquired tends strongly to the conviction that this plan of church belongs to a later period in the Middle Ages, though it is difficult to determine when it was introduced, and it may be only a continuation of a much earlier form.

One other church of this part of the world seems to claim especial mention, that of Mokwi, built in the 10th century, and painted as we learn from inscriptions, between 1080 and 1125. It is a large and handsome church, but its principal interest lies in the fact that in dimensions and arrangement it is almost identical with the contemporaneous church of Sta. Sophia at Novogorod, showing a connection between the two countries which will be more particularly pointed out hereafter. It is now very much ruined, and covered with a veil of creepers which prevents its outward form from being easily distinguished.

As will be perceived, its plan is only an extension of the two last mentioned, having five aisles instead of three; but it is smaller in scale and more timid in execution. The church which it most resembles is that at Trabala in Syria (Woodcut No. 330), which is certainly of an earlier date than any we are acquainted with further east. Practically the same plan occurs at Athens (Woodcut No. 338), and at Mistra (Woodcut No. 339), but these seem on a smaller scale than at Mokwi, so that it may be considered as the typical form of a Neo-Byzantine church for four or five centuries, and it would consequently be unsafe to attempt to fix a date from its peculiarities.

Interesting as these may be in an historical point of view, the most important ecclesiastical establishment in this part of the world is that of Etchmiasdin. Here are four churches built on the spots from which, according to tradition, rose the two arches or rainbows, crossing one another at right angles, on which our Saviour is said to have sat when he appeared to St. Gregory. They consequently ought to be at the four angles of a square, or rectangle of some sort, but this is far from being the case. The principal of these churches is that whose plan is represented in Woodcut No. 353. It stands in the centre of a large square, surrounded by ecclesiastical buildings, and is on the whole rather an imposing edifice. Its porch is modern; so also, comparatively speaking, is its dome; but the plan, if not the greater part of the substructure, is ancient, and exhibits the plainness and simplicity characteristic of its age. The other three churches lay claim to as remote a date of foundation as this, but all have been so altered in modern times that they have now no title to antiquity.

The idea that the churches at Pitzounda and Bedochwinta must be comparatively modern is confirmed by comparing their plan with that of Kouthais, a church which there seems no reasonable ground for doubting was founded in 1007, and erected, pretty much as we now find it, in the early part of the 11th century. It has neither coupled piers nor pointed arches, but is adorned externally with reed-like pilasters and elaborate frets, such as were certainly employed at Ani in the course of the 11th century. The annexed elevation (Woodcut No. 355) of one of its windows exhibits the Armenian style of decoration of this age, but is such as certainly was not employed before this time, though with various modifications it became typical of the style at its period of greatest development.

ANI.

Even Etchmiasdin, however, sinks into insignificance, in an architectural point of view, when compared with Ani, which was the capital of Armenia during its period of greatest unity and elevation, and was adorned by the Bagratide dynasty with a series of buildings which still strike the traveller with admiration, at least for the beauty of their details; for, like all churches in this part of the world, they are very small. If, however, the cathedral at Ani is interesting to the architect from its style, it is still more so to the archæologist from its date, since there seems no reason to doubt that it was built in the year 1010, as recorded in an inscription on its walls. This, perhaps, might be put on one side as a mistake, if it were not that there are two beautiful inscriptions on the façade, one of which is dated 1049, the other 1059. To this we must add our knowledge that the city was sacked by Alp Arslan in 1064, and that the dynasty which alone could erect such a monument was extinguished in 1080. With all this evidence, it is startling to find a church not only with pointed arches but with coupled piers and all the characteristics of a complete pointed-arch style, such as might be found in Italy or Sicily not earlier than the 13th century. This peculiarity is, however, confined to the constructive parts of the interior. The plan is that of Pitzounda or Bedochwinta, modified only by the superior constructive arrangement which the pointed arch enabled the architects to introduce; and externally the only pointed arch anywhere to be detected, is in the transept, where the arch of the vault is simulated to pass through to the exterior.

In the plan and elevation of the building will be observed a peculiarity which was afterwards almost universal in the style. It is the angular recess which marks the form of the apses outside without breaking the main lines of the building. In the lateral elevation of this cathedral (Woodcut No. 358) they are introduced on each side of the portal where the construction did not require them, in order to match those at the east end. But in the Cathedral at Samthawis (Woodcut No. 359) they are seen in their proper places on each side of the central apse. Though this church was erected between the years 1050-1079, we find these niches adorned with a foliation (Woodcut No. 360) very like what we are accustomed to consider the invention of the 14th century in Europe, though even more elegant than anything of its class used by the Gothic architects.

At Sandjerli, not far from Ani, is another church, which from inscriptions translated by M. Brosset, and from sections given by him, appears to belong to the same date (1033-1044), and to possess coupled columns and pointed arches like those of the cathedral of Ani, which indeed it resembles in many points, and which renders the date above given highly probable.

The plans above quoted may probably be taken as those most typical of the style, but in no part of the world are the arrangements of churches so various. All being small, there were no constructive difficulties to be encountered, and as no congregation was to be accommodated, the architects apparently considered themselves at liberty to follow their fancies in any manner that occurred to them. The consequence is that the plans of Armenian churches defy classification; some are square, or rectangles of every conceivable proportion of length to breadth, some octagons or hexagons, and some of the most indescribable irregularity. Frequently two, three, or four are grouped and joined together. In some instances the sacred number of seven are coupled together in one design, though more generally each little church is an independent erection; but they are all so small that their plans are of comparatively little importance. No grandeur of effect or poetry of perspective can be obtained without considerable dimensions, and these are not to be found in Armenia.

There are also some examples of circular churches, but these are far from being numerous. Generally speaking they are tombs, or connected with sepulchral rites, and are indeed mere amplifications of the usual tombs of the natives of the country, which are generally little models of the domes of Armenian churches placed on the ground, though perhaps it would be more correct to say that the domes were copied from the tombs than the reverse.

The most elegant of all those hitherto made known is one found at Ani, illustrated in Woodcuts Nos. 361, 362. Notwithstanding the smallness of its dimensions, it is one of the most elegant sepulchral chapels known.

Another on a larger scale (Woodcut No. 363) is borrowed from Mr. Layard’s book. This tomb shows all the peculiarities of the Armenian style of the 11th or 12th century. Though so much larger, it is by no means so beautiful as the last mentioned tomb at Ani. In its ornamentation a further refinement is introduced, inasmuch as the reed-like columns are tied together by true love-knots instead of capitals—a freak not uncommon either in Europe at the same age, or in the East at the present day, but by no means to be recommended as an architectural expedient.

With scarcely an exception, all the buildings in the Armenian provinces are so small that they would hardly deserve a place in a history of architecture were it not for the ingenuity of their plans and the elegance of their details. The beauty of the latter is so remarkable that, in order to convey a correct notion of the style, it would be necessary to illustrate them to an extent incompatible with the scope of this work. In them too will be found much that has hitherto been ascribed to other sources. The annexed capital (Woodcut No. 364), for instance, would generally be put down as Saracenic of the best age, but it belongs, with a great deal more quite as elegant, to one of the churches at Ani; and the capital from Gelathi (Woodcut No. 365) would not excite attention if found in Ireland. The interlacing scrolls which occupy its head are one of the most usual as well as one of the most elegant modes of decoration employed in the province, and are applied with a variety and complexity nowhere else found in stone, though they may be equalled in some works illustrated by the pen.

Besides, however, its beauty in an artistic point of view, this basket pattern, as it is sometimes called, is still more so as an Ethnographic indication which, when properly investigated, may lead to the most important conclusions. The three following woodcuts, Nos. 366, 367, and 368, taken from churches at a now deserted village called Ish Khan, will serve to explain its more usual forms; but it occurs almost everywhere in the Armenian architectural province, and with as infinite a variety of details as are to be found with its employment in Irish manuscripts.

Out of Armenia it occurs in the church at Kurtea el Argyisch in Wallachia (Woodcut No. 385), and is found in Hungary and Styria, and no antiquary will probably fail to recognise it as the most usual and beautiful pattern on Irish crosses and Scotch sculptured stones. On the other hand it occurs frequently in the monolithic deepdans or lamp-posts and in the temples on the Canarese or West Coast of India, and in all these instances with so little change of form that it is almost impossible that these examples should be independent inventions. Still the gaps in the sequence are so great that it is very difficult to see how they could emanate from one centre. Few, however, who know anything of the early architecture of Ireland can fancy that it did come from Rome across Great Britain, but that it must have had its origin further east, among some people using groups of churches and small cells, instead of congregational basilicas. So far, too, as we can yet see, it is to the East we must look for the original design of the mysterious round towers which form so characteristic a feature of Irish architecture, and were afterwards so conspicuous as minars in the East, and nowhere more so than in Armenia. Recent researches, too, are making it more and more clear that Nestorian churches did exist all down the West Coast of India from a very early period, so that it would not be impossible that from Persia and Armenia they introduced the favourite style of ornament.

All this may seem idle speculation, and it may turn out that the similarities are accidental, but at present it certainly does not look as if they were, and if they do emanate from a common centre, tracing them back to their original may lead to such curious ethnological and historical conclusions that it is at all events worth while pointing them out in order that others may pursue the investigation to its legitimate conclusion.

Taken altogether, Armenian architecture is far more remarkable for elegance than for grandeur, and possesses none of that greatness of conception or beauty of outline essential to an important architectural style. It is still worthy of more attention than it has hitherto received, even for its own sake. Its great title to interest will always be its ethnological value, being the direct descendant of the Sassanian style, and the immediate parent of that of Russia. At the same time, standing on the eastern confines of the Byzantine Empire, it received thence that impress of Christian art which distinguished it from the former, and which it transmitted to the latter. It thus forms one of those important links in the chain of architectural history which when lost render the study of the subject so dark and perplexed, but when appreciated add so immensely to its philosophical interest.

VI. ROCK-CUT CHURCHES

CONTENTS.

Churches at Tchekerman, Inkerman, and Sebastopol—Excavations at Kieghart and Vardzie.

Intermediate between the Armenian province which has just been described and the Russian, which comes next in the series, lies a territory of more than usual interest to the archæologist, though hardly demanding more than a passing notice in a work devoted to architecture. In the neighbourhood of Kertch, which was originally colonised by a people of Grecian or Pelasgic origin, are found numerous tumuli and sepulchres belonging generally to the best age of Greek art, but which, barring some slight local peculiarities, would hardly seem out of place in the cemeteries of Etruria or Crete.

At a later age it was from the shores of the Palus Mœotis and the Caucasus that tradition makes Woden migrate to Scandinavia, bearing with him that form of Buddhism[251] which down to the 11th century remained the religion of the North—while, as if to mark the presence of some strange people in the land, we find everywhere rock-cut excavations of a character, to say the least of it, very unusual in the West.

These have not yet been examined with the care necessary to enable us to speak very positively regarding them;[252] but, from what we do know, it seems that they were not in any instance tombs, like those in Italy and many of those in Africa or Syria. Nor can we positively assert that any of them were viharas or monasteries[253] like most of those in India. Generally they seem to have been ordinary dwellings, but in some instances appropriated by the Christians and formed into churches.

One, apparently, of the oldest is a rectangular excavation at Tchekerman in the Crimea. It is 37 ft. in length by 21 in width, with hardly any decoration on its walls, but having in the centre a choir with four pillars on each face, which there seems no doubt was originally devoted to Christian purposes. The cross on the low screen that separates it from the nave is too deeply cut and too evidently integral to have been added. But for this it would seem to have been intended for a Buddhist vihara.

Under the fortress at Inkerman—facing the position held by our army— there is an excavation undoubtedly of Christian origin. It is a small church with side-aisles, apse, and all the necessary accompaniments. Beyond this is a square excavation apparently intended as a refectory, and other apartments devoted to the use of a monastic establishment. These again are so like what we find among the Buddhist excavations in India as to be quite startling. The one point in which this church differs from a Buddhist chaitya is that the aisle does not run round behind the altar. This is universally the case in Buddhist, but only exceptionally so in Christian, churches.

Close to Sebastopol is another small church cave with its accompanying monastery. This one is said to be comparatively modern, and if its paintings are parts of the original design it may be so, but no certain data are given for fixing the age of the last two examples. That under the fortress (Woodcut No. 371) seems, however, to be of considerable antiquity.

There is one which in plan is very like those just described at Vardzie, said to belong to the 12th century, and another, almost absolutely identical with a Buddhist vihara, at Kieghart in Armenia, which has a date upon it, A.D. 1288.

On the banks of the Kour, however, at Ouplous-Tsikhe and Vardzie, are some excavations which are either temples or monasteries, and which range from the Christian era downwards. These are generally assumed to be residences—one is called the palace of Queen Thamar—and they were evidently intended for some stately purpose. Yet they were not temples in any sense in which that term would be employed by the Greek or Roman world. Whatever their destination, these rock-cut examples make, when taken altogether, as curious a group of monuments as are to be found in this corner of Asia, and which may lead afterwards to curious archæological inferences. At present we are hardly in a position to speculate on the subject, and merely point to it here as one well meriting further investigation.