CHAPTER V

THE FAÇADE

The breaking up of the wall-face into horizontal zones was a device familiar to the ancient East. In the main gateway of Sargon’s palace at Khorsâbâd the wall is divided into a high orthostatic podium, decorated with reliefs, and a brick superstructure diversified by vertical flutes and rectangular recesses.[271] In the interior of the palace, the court of the ḥaram shows a similar disposition, except that the podium is of enamelled brick, not of stone.[272] The upper part of the walls is in no case preserved. On Assyrian reliefs it is not uncommon to find a horizontal band along the top of the walls below the crenellations;[273] but the nature of the upper zone or zones in decorated façades such as those of Khorsâbâd is a matter of conjecture. Concerning Chaldaean wall decoration we have little evidence. The building on the Wuswas mound at Warka, of which Loftus published a sketch,[274] has recently been re-examined by Dr. Jordan, who believes it to be post-Babylonian.[275] The walls of the temple of Bel at Niffer were decorated with shallow buttresses, while the gates bore resemblance, both in plan and decoration, to the gates of Khorsâbâd.[276] The gateway of Gudea at Tellôh has the same doubly recessed rectangular niches that have been noted at Khorsâbâd, but they do not seem to have been grouped in panels, and the plinth is reduced to insignificant proportions.[277] It is significant that in the post-Babylonian construction at Tellôh both the rectangular niche and the flute are present, and it may be surmised that the walls of Wuswas, with their recessed and fluted panels placed one above the other, represent an ancient scheme. It is a scheme which may be compared with that of the façade of Ctesiphon (see below, p. 134). At intervals groups of recessed niches are carried up continuously to the height of two registers of panels, just as in the two lower zones at Ctesiphon the engaged columns embrace two registers of arched niches. But at Ctesiphon we have architectural forms borrowed from Hellenism instead of the surface decoration (recess and flute) of Chaldaea and Assyria.

The orthostatic construction was used in Hittite architecture at Zindjirli, Boghâz Keui, and Sakcheh Geuzu. Mr. Hogarth has found it at Carchemish and Baron Oppenheim at Râs al-’Ain.[278] But in all these buildings, Babylonian, Hittite, and Assyrian, there was no attempt to ornament the façade with the similitude of plastic architectural forms. The elements of such ornament were not indeed lacking, but they appear in isolated examples and were not applied to the wall-face in a continuous decorative system. Side by side with stelae and altars adorned with fluted motives akin to those of the façades[279] there are instances of mock architecture in relief. An Assyrian stela upon a slab found at Quyundjik and now in the British Museum will serve as an illustration (Fig. 11). Two pilasters carry an architrave consisting of a double fillet and a band of crenellations; between and behind the pilasters an arched niche, placed in counterfeited perspective, frames a hunting scene. It is an early example of the application of the third dimension to architectural ornament, and it conveys the impression of plastic architecture in two planes. As Professor Delbrück observes, by the addition of free-standing columns placed before the pilasters, we should have here a motive familiar to Graeco-Roman façades.[280] The archivolt, of which the enrichment is expressed at Quyundjik in the terms of a shallow fillet, appears at Khorsâbâd, with enamelled brick enrichment, over a doorway,[281] and also upon reliefs.[282] All the methods of decorating the face of the arch which were known to antiquity are found on the Assyrian monuments. The podium façade is oriental, for it was used in Assyria and in Persia. Pre-Greek is the employment of blind openings; in the Persepolitan palaces a blind niche is placed in every intercolumniation, and in plastic architecture an open gallery or loggia was common to Egypt and to Assyria.[283] In pre-Hellenic Egypt and western Asia there is, however, no example of a continuous series of arches in relief, though the continuous treatment of decoration on the wall-face is typical of Babylonian architecture from the earliest time, and it remained only to apply it to true architectural motives instead of to the purely decorative motives of Chaldaea and Assyria. That these last were mainly based upon the outward aspect of primitive wooden structures, I do not doubt, but at the remote date at which we first know them they had already lost all structural significance. The step from pattern to imitative architecture must have been taken at an early stage in the Hellenistic East. Seleucid buildings which have vanished are reflected in the stupas of Hellenistic India, where the surfaces are adorned with blind openings between engaged piers, and in the rock-cut temples, where the decorative scheme of the façade is a podium carrying a colonnade in relief.[284] In Egypt rows of niches are present in the interior of tombs,[285] and an early example of the same motive can be seen in the gateway at Perge, a city which lay under the direct influence of Antioch.[286] The lightening of the massive wall by means of niches and blind openings can be traced through pre-Greek architecture in Mesopotamia (Assyrian palaces and temples) and in Egypt (from the Eighteenth Dynasty and even earlier) down to the Achaemenid period. The systematic application of this principle to the wall-face, and its union with imitative architecture in relief as a decorative scheme took place, as far as can be determined at present, in the Hellenistic age.

In the third and in the second century B.C. the division of the wall into two zones by means of a moulding appears at Delos, Priene, Magnesia, and other parts of western Asia,[287] and a little later it is found in what is known as the incrusted style at Oscan Pompeii. The lower zone consists of unpainted stucco decoration representing a stone wall, composed of one or of two rows of orthostatae, and above them several courses of dressed stones. The upper zone, which was at first undecorated (it represented space, the upper air), takes on later the semblance of a colonnaded gallery[288] in imitation of the open galleries characteristic of Eastern Hellenistic architecture.[289] The podium façade carrying an open arcade is, as Professor Delbrück is careful to point out, in origin different from the galleried wall, but in façade schemes the two run together so as to be almost indistinguishable. The theme is represented in relief upon the façade of the Bouleuterion at Miletus[290] and frequently in Pompeii, where, however, the engaged columns do not stand upon a podium.[291] Behind the columns, both at Delos and in the Pompeiian examples, the wall is still divided into two zones by a moulding. In all cases it is a theme which stands as a representation in relief of plastic architecture, of deep colonnades such as those which were to be seen on the Mausoleum at Halicarnassus.[292] The blind order of the Ephebeum at Priene may be cited as another striking example of imitative architecture.[293] Similarly the superimposition of one blind order upon another, a decorative motive so familiar in Roman theatres and amphitheatres, finds its prototype in the colonnades of Hellenistic stoae, such as those erected by Attalus in Athens and in Pergamon.[294]

Professor Delbrück is of opinion that the impulse towards decorating the wall-face with the similitude of plastic architecture was quickened by Greek painting, which, from the fourth century B.C. onward, gained an increasing mastery in the representation of spatial dimensions. Plastic examples of the phase of development represented by the Boscoreale frescoes might be expected in the second century B.C., and in fact there were at that period mock colonnades in relief, such as the Ephebeum at Priene. The cutting away of the wall-face by means of niches was foreshadowed in Hellenistic art; the lightening of the wall-mass by niches has been noticed in the gate at Perge and the tombs of Alexandria, while windows in the intercolumniations were of frequent occurrence.[295] It is possible, as Professor Delbrück suggests, that in Hellenistic Mesopotamia decoration by means of blind openings, whether doors, windows, or niches, won a great popularity because it was based on pre-Hellenic tradition, and it is interesting to observe that the only early examples of the arched niche, which is the leading motive at Ukhaiḍir, are to be found in western Asia.[296] But the systematic application of these principles to the façade was accomplished only in the latest phases of Hellenistic art, and we may perhaps owe it to Roman builders. In the intercolumniations of the decorated zone niches, arcades and windows take the place of the traditional moulding,[297] and the upper wall is broken by a row of arches or of windows.[298] On inner walls a double row of niches is sometimes accompanied by stucco incrustation,[299] while the podium is decorated with engaged columns.[300]

It remained for the Imperial period to complete the development. Orders of columns were placed in zones one above the other; niches of richer type occupied the surface of the wall, and not infrequently they were placed one within the other; rounded and rectangular niches followed one another in a rhythmic sequence; columns and piers stood out in higher relief and the podium and architrave were broken above and below them. Gradually the orders and niches lost their original significance; they were looked upon merely as decorative motives, and as such followed a development of their own. They lent to the wall-surface an ever-increasing movement and rhythm as their forms grew richer and freer. This evolution can be seen upon the walls of Roman buildings which are yet standing; if in the cities of the eastern Mediterranean most of the monuments have fallen, the elements of their composition have been found and put together, as in the Nymphaeum at Miletus,[301] or the theatre at Ephesus,[302] and similar decoration can still be studied upon the walls of Ba’albek.[303] But in western Asia, and notably in Syria, the old classical love of unbroken wall-surfaces died hard--perhaps it may be said to have survived long into the Middle Ages in the smooth faces of dressed stone which give so much dignity to the Mohammadan buildings of Damascus and Aleppo. Older and simpler decorative forms continued to rule when in Rome the evolution had gone on to other stages. The façade of the Nabataean temple at Sî’, for example, echoes in free-standing architecture the features of the relief decoration of the Ephebeum at Priene.[304] In the temenos of the basilica at Apamea (second century A.D.) the solid outer wall has disappeared, and its place is taken by a series of piers with rectangular openings between, but in the basilica itself the treatment of the wall is still of an extremely simple character.[305] The temenos wall of the temple at Palmyra is treated with the old formal severity. At Bâqirḥâ and at Isriyyeh the walls are unbroken save by shallow pilasters,[306] a simplicity which rivals that of the pre-Roman tomb of Ḥamrath at Swaidâ.[307] At Mushennef and at Qanawât pilasters are set at the angles, and the rest of the wall is undecorated.[308] In the pre-Roman temple at Swaidâ, niches, in imitation of small doors, are placed on either side of the single entrance;[309] at ‘Atîl a double order of niches, the lower rectangular, the upper rounded and arched, occupy the same position, but the walls of the cella are without even the customary pilasters;[310] in the Qaisariyyeh at Shaqqâ a genuine opening flanks the doorway on either side, but the façade is otherwise unadorned.[311] In the Philippeion at Shahbâ the side niches are omitted and there are no pilasters except at the angles; rounded and rectangular niches are employed on the interior walls of the palace, and on either side of the interior doorways of the bath, but in all other respects the latter building is noticeable for the entire absence of decoration upon its walls;[312] and as late as the sixth century angle pilasters set upon a podium were considered a sufficient decoration for the walls of the exquisite tomb at the southern Dânâ,[313] while the porticoes of house and stoa are models of severity.[314]

The fantastic variety which characterized the late Hellenistic and the Roman Imperial age must be sought for in south-west Asia in another group of monuments. The influence of Alexandria dominates over the tomb façades of Petra, and was felt even in the earlier tombs at Madâin Ṣâliḥ.[315] With the latter I am not immediately concerned, except in so far as they help to determine the date of the Petra tombs. It is enough to notice that the local oriental forms, the pylon tombs with a band or bands of crenellated ornament, or with a staircase motive at the angles, dropped out of fashion during the first half of the first century after our era, and that in the first century A.D. Hellenistic forms had invaded the Ḥedjr tombs.[316] The gable tomb and the columned façade, which Domaszewski has christened the Roman temple tomb, do not indeed appear at Madâin Ṣâliḥ, but the fully developed aedicula, with quarter-columns in the antae, is found there as early as the year A.D. 31 in the tabernacle which frames the doorway,[317] and the tabernacle, both with a gable and with an archivolt, was employed in Arabia at an early date for votive niches.[318] It is therefore unnecessary, as Puchstein has pointed out, to assign such gable tombs at Petra as date from a period before the Roman occupation (i.e. before A.D. 106) to some fortuitous Greek influence,[319] since the type was familiar to the stone-cutters of an earlier period. Not later than the middle of the first century A.D. a second order of dwarf columns was placed in the attic (the earliest dated example is tomb F 4 at Madâin Ṣâliḥ, A.D. 63-64), but it is instructive to note that the appearance of a new form does not imply the elimination of older types. At Madâin Ṣâliḥ all the different variations continue to exist side by side, and there is an example of the primitive pylon tomb with a single band of crenellations, the unmitigated copy of an Arabian house for the living turned into a house for the dead, which is dated as late as the year A.D. 74,[320] just as the Egyptian gorge is found side by side with, and indeed upon the same tombs as, a fully developed Ionic entablature. The Roman temple tomb of Petra is predicted in the dwarf piers of the attic (which are of frequent occurrence at Madâin Ṣâliḥ) inasmuch as they imply a corresponding series of engaged piers in the wall below. A single example of this so-called temple tomb exists at Madâin Ṣâliḥ, but without the piers in the attic; it is probably to be dated in the middle of the first century A.D.[321] The engaged column, in contradistinction to the engaged pier, is employed at Madâin Ṣâliḥ only in the antae of the tabernacles; at Petra it takes its place among the main supports of the façade. At Petra, too, the plastic freedom of late Hellenistic architectural forms makes itself felt. Broken podiums are found upon wall paintings of the second style at Boscoreale, though their architectural counterpart cannot be pointed out at so early a date; broken entablatures are present in late Hellenistic work at Alexandria, but not elsewhere in the Greek cultural sphere at the same period.[322] Both these features, together with the preference for engaged columns instead of piers, are common at Petra, and they are like sign-posts pointing to the source whence the stone-cutters of Petra drew their inspiration. There are, it is true, early examples of the broken architrave in Italy in the triumphal arches of Rimini (27 B.C.) and Aosta (25 B.C.), but the systematic use of broken podium and entablature is one of the distinctive features of the later Imperial period. In the Lion Tomb at Petra, which recalls the tabernacle of the tomb F 4 at Madâin Ṣâliḥ, architrave, frieze, and cornice are broken over the angle columns and piers. In the tombs of the second century the principle is carried further; architrave, frieze, and cornice are all broken, and the system is extended to the plinth-like member which is interposed between the entablature and the dwarf order of the attic, and, when the façade reaches a second story, to the upper entablature also.[323] In the Corinthian tomb, the Dair, and the Khazneh a second order is superimposed upon the first. In each case a tholos occupies the centre of the upper story and the pairs of flanking columns are crowned by a broken pediment. In the Dair an engaged pier and quarter-column fill out the façade on either side (Plate 82, Fig. 2). In the Corinthian tomb the lower zone is complete in itself (Plate 82, Fig. 1). The engaged columns stand upon a high plinth and carry a broken architrave composed of frieze and cornice only; the dwarf piers are placed upon a broken plinth with a moulded cornice, which is interrupted above the central door by a moulded archivolt. The dwarf columns carry a complete entablature, architrave, frieze, and cornice, and a low broken pediment occupies the centre of the façade. Above this structure the second order, with its tholos, stands upon a moulded plinth. In the Storied tomb the lower order carries a complete entablature and a broken attic which contains the gables and archivolts of the doors; upon a plinth with a moulded cornice rises a second order bearing an entablature; a second plinth, itself divided by a horizontal moulding, carries a dwarf order which is crowned by a third entablature (Fig. 30). Yet another order crowned the tomb, but it was built, not rock-cut, and little of it remains. The tholos in these façades is a Hellenistic motive, though it is known to us at an early period only from wall paintings and from literary sources.[324] To the multiplication of horizontal decorations earlier Nabataean tombs had shown a strong inclination. The double band of crenellations in the pylon tombs of Madâin Ṣâliḥ and of Petra, the double attic of the so-called Ḥedjr tombs in both places, point the way to such compositions as the Storied tomb. Everywhere a strong centralization rules the scheme of the façade. It is rare to find more than one door; where doors are placed in the flanking intercolumniations they are insignificant in size, as in the Corinthian tomb. In the Dair (Plate 82, Fig. 2), mock windows occupy the outer intercolumniations. In the Storied tomb, where there are four doors, the two central entrances are higher than the others, and, in the upper story, the central intercolumniation is wider than those on either side. But the long unbroken lines of the horizontal mouldings give an exceptional monotony to this façade. Usually a gable or archivolt, breaking into the attic, emphasizes the centre of the façade and is re-echoed in the pediment, with its central acroterion which crowns the whole, while in the tholos tombs the centralization is even more strongly underlined. The angles are commonly in antis, with a quarter-column set against the corner pier. The archivolt is conspicuous by its absence. It is never used except in exchange for the pediment over aediculae, and, exceptionally, over mock windows, as for example in the lower zone of the Dair.

The same insistence upon horizontality is to be observed in the façades of Ctesiphon and Ukhaiḍir; but the effect is produced in a different manner. No doubt it is difficult to do justice to the horizontal members in these buildings, owing to the fact that, from the perishable nature of the material, they have suffered complete destruction, but it can safely be conjectured that they were never of much importance to the general effect. The space left between the decorated zones is too small to admit of the full entablature, attic, and podium which separate the lower order from the first upper order in the Storied tomb, or even of the entablature and podium which are interposed between the upper order and the order of dwarf columns. The multiplication and the breaking of horizontal members in Western Hellenistic monuments are discarded in Mesopotamia, and with them vanishes much of the significance of the façade. The zone decoration becomes a pattern composed of innumerable groups of architraved and arched divisions, set one within the other, so as to cover the whole surface of the wall. Where exigency demands, real doors and windows may be placed in the niches; the zones may correspond to a certain extent with the structural division of the building into stories; but the main intention of the architect is to cover his wall with continuous motives which are not dependent upon the structure and must fit into it as best they can. It is the traditional surface decoration of the ancient East, disguised in the new dress which it had borrowed from Hellenism.

No better example of the oriental practice can be found than in the façade of Ctesiphon. The north wing and the face of the great central arch have fallen, but they are preserved in M. Dieulafoy’s photograph[325] (Plate 83). The façade is divided into three zones, but organic connexion between them is lacking. Each zone, in either wing, is subdivided into two horizontal registers. The lower register of the lowest zone consists of wide arches separated by pairs of engaged columns which are carried up to the top of the zone. The width of the intercolumniations bears no relation to the width of the wing; a space remains over at the outer end which is awkwardly filled by two small blind arched niches, placed one above the other. The upper register is occupied by groups of three niches; in each group the central niche is wider than the other two, and each niche is flanked by engaged colonnettes. At the outer end there is no room to complete the pattern, and the outer flanking niche is omitted. The lower zone breaks off abruptly here against a plain pylon-like wall, and at the inner end it is not organically connected with the great archway which forms the centre of the façade. Single engaged columns divide the middle zone into five compartments. They are not placed above the pairs of engaged columns of the lower zone, nor yet in the centre of the lower intercolumniations, but purely in accordance with the dictates of the pattern which covered the middle zone. It, too, is subdivided into two horizontal registers. In the lower register there are five pairs of niches, with three engaged colonnettes between. At the inner end the pair must have been incomplete owing to lack of space; at the outer end the engaged column is omitted for the same reason. In what relation the triple colonnettes stood to the niche arches is not clear. They were not regarded as necessary to the arch, for on the outer side of each pair they are absent, and the same applies to the colonnettes and arches in the upper register of this zone. These groups consist of three niches of equal size, with a pair of colonnettes between the central and the flanking niches. In the third zone the upper of the two registers has almost entirely disappeared; it is obvious, however, that the two registers were not welded together by engaged columns. In the lower register the arched niches, separated by engaged colonnettes, are conceived without any thought of the division of the wall below them, and, from the fragment of the upper register which remains, it would seem that the niches which adorned it were equally independent of the niches of the lower register. Into this confusion breaks the huge central arch, cutting short the pattern at the inner end of the wings just as the pylon wall cuts it short at the outer end. Yet the gigantic size of the façade and the even repetition of the arches in each register gives to the eye a sense of orderly grouping, and draws the whole into an apparent symmetry which an analysis of the details proves to be lacking in reality.

Ukhaiḍir, separated from Ctesiphon by an interval of some 500 years, shows a sensible advance. The north façade of the court is not indeed centralized, nor is it symmetrically placed in the wall of the three-storied block, but the two lower zones are organically connected with one another. The seven blind niches of the lower order correspond with those of the second order. In the second order the breaking up of the zone into registers is still adhered to, but since an archivolt has taken the place of the architrave of Ctesiphon, the principle is not so strongly marked. It works only within the arched niches. That it is substantially the same is, however, apparent from the fact that at Ukhaiḍir, as at Ctesiphon, the lower register consists of groups of two small niches, the upper register of groups of three, the central niche being the largest. The seven large niches of the second order are separated by a cluster of four columns; in the spandrels of the arches there are niches containing windows. The pylon-like wall of Ctesiphon is represented by a battered wall at Ukhaiḍir, but instead of sloping back and forming horizontal ledges, its perpendicular face seems to have been divided at intervals by horizontal bars of masonry. There is no space between the zones for important horizontal mouldings. Dr. Reuther in his reconstruction (_Ocheïdir_, Plate 25) places a plain masonry balcony along the narrow platform formed by the summit of the second zone. It is, however, conjectural, and in my opinion it lays a stress upon the horizontal divisions between the zones which is contrary to the spirit of the decorative scheme. In the upper zone the plain wall is in far better accord with the classical treatment of wall-surfaces than are the restless nichings of Ctesiphon, and it enhances the value of the rich orders below it. But it is not regarded, like the plain wall of early Hellenistic decoration, as representing space, the upper air;[326] it is rather the gallery wall of ancient Assyrian and early Hellenistic architecture. It is confined by an upper row of arched niches, each one, so far as can be determined in their ruined condition, placed within a rectangular frame of engaged columns and architrave, like the niches upon the outer fortification wall of the palace. And here we have the system that dominates Ctesiphon, the column and architrave framing arched niches. In the upper zone of the Ukhaiḍir façade symmetry has vanished. The long crowning row of niches calls attention to the fact that the decorated lower zones of the façade do not stand in the centre of the wall, and the doorways of the third zone bear no more relation to the arches below them than the perpendicular divisions of the Ctesiphon wall bear relation to one another. Another similarity exists between the two buildings. The arches of the second zone at Ukhaiḍir are decorated not with the mouldings of the classical archivolt, but with the cusp of the great arch at Ctesiphon. So far as I am aware the earliest example of this cuspidated ornament in monumental architecture is at Ctesiphon. It appears in northern Syria in the fifth century A.D., when it can be seen both with the cusps pointing inward[327] and with the cusps pointing outward.[328] In the latter form it bears a close resemblance to the broken palmette of late Graeco-Roman ornament,[329] and its origin is probably to be sought in oriental Hellenism, but whether it was developed in the Syrian or in the Mesopotamian regions I cannot determine. It became a common motive in Syrian architecture during the sixth century,[330] where it is used in both forms, but in the Mesopotamian sphere it is almost always inverted, as at Ctesiphon. We have it at Ukhaiḍir, not only in the façade but also on the arches of the mosque doorways and possibly in the lîwân arches in the courts.[331] In exactly the same form it is employed in the early Abbâsid buildings of Sâmarrâ,[332] and there is another notable example of its use over the doorway of the mosque at Ḥarrân, where an outward-pointing cusp is used (Plate 84, Fig. 2). In the mosque at Mayâfârqîn it is found inverted on the elaborate arches which cover the miḥrâb niches, on the relieving arches over the doors of the outer north wall (Plate 84, Fig. 3), and on the blind niches above. This part of the wall belongs to the earlier portion of the building, which is ascribed, in an inscription round the dome, to the Ortokid Alpi (A.D. 1152-1176). It is a common feature of Ortokid decoration at Diyârbekr,[333] and in the first half of the thirteenth century it is seldom absent from the lintels of Christian churches and Mohammadan mosques in Môṣul and the surrounding districts,[334] nor yet, in the late fourteenth and early fifteenth centuries, from the lace-like decoration of the arches in the mosques at Ḥasan Kaif[335] (Plate 84, Fig. 1). Other examples in late Mohammadan architecture are too numerous to be mentioned. I select the few which I have quoted because they are little known.

In attempting a reconstruction of the Ukhaiḍir façade (Plate 85) I have sought some guidance from the representation of a Sasanian fortress which is to be seen upon a silver dish, now in the possession of the Kais. Archäol. Kommission of St. Petersburg[336] (Plate 86, Fig. 2). It has been assigned to the beginning of the Sasanian period. The façade depicted bears some interesting analogies to that of Ukhaiḍir. It is divided into two stories. In the lower story the lower zone consists of eight arched niches, the arches borne on tall engaged columns without capitals. The archivolts are decorated with three fillets and a small oval motive is placed in the spandrels. Above the arches there is a cornice composed of two simple horizontal mouldings with a band of spirals between them. I surmise that these spirals, which seem to be singularly out of place in a monumental façade, were put in to fill up the space and have no warrant in any actual building. The gateway occupies the centre of this zone. A wide archway, set in a rectangular frame, covers two narrow arched doors. Within the semicircle of the embracing arch there is a shallow calotte decorated with broken concentric rings. The archivolt is outlined by a moulding which is carried up continuously round the rectangular frame. Within this frame a horizontal moulding is laid above the arch. This scheme of archivolt and rectangular frame with a continuous moulding is common in Syria and Mesopotamia.[337] The crowning member of the portal breaks the line of the cornice. It consists of a frieze carved in relief with a human (or divine?) head and bust, and a cornice bearing a row of cusps. The upper zone of the lower story is less easy to describe in terms of architecture. There is a frieze (or dwarf order?) decorated with four groups of six flutings or engaged colonnettes and five groups of four circles, each circle containing a quatrefoil. The cornice is composed of two bands, the first decorated with alternate circles and rhomboids, the second with diagonal brickwork. A projecting hourd is placed at either end of the building, and between the hourds the top of the wall is battlemented. These crenellations form a parapet to the gangway which runs along the base of the second tower-like story. Upon the gangway stand eight figures, seven of whom are blowing trumpets. Behind them the wall is plain, but the upper part is decorated first with a band of half-florettes, then with a row of arched niches, each niche being set within a rectangular frame, and finally with a band of diagonal brickwork. The summit of the wall is battlemented and a wooden hourd projects from either side. The lower zone of the lower story corresponds very fairly with the lowest zone at Ukhaiḍir. The schematized horizontal bands of the second zone bear little or no relation to real architecture, but the upper story is set back, as at Ukhaiḍir, and the battlemented parapet of the gangway is a very probable solution for the parapet of the Ukhaiḍir gangway. The upper story, with its plain wall and its row of niches is the same in both façades, and the upper battlements may safely be restored at Ukhaiḍir.

At Ctesiphon the capitals and bases (if bases there were) of the columns and colonnettes were moulded in stucco and have disappeared. Bases seem to have been absent from the slender engaged columns on the outer walls of Firûzâbâd and Sarvistân, but at both places the state of the ruins renders an exact determination of such details difficult. The engaged columns seem to rest upon a low plinth. The decoration in those palaces is, however, far more nearly connected with oriental than with occidental tradition. We have not much information as to Sasanian capitals. The columns and double columns of the inner rooms at Sarvistân are covered by rectangular imposts,[338] and de Morgan gives a drawing of a stucco capital from Shirwân.[339] It is scarcely necessary to allude to the famous impost-capitals of Bîsutûn and Iṣfahân, which belong, in all probability, to the end of the sixth or the beginning of the seventh century. They show far greater skill in the handling of the rectangular impost than the capitals at Sarvistân, but whether they are a natural development out of the latter, or borrowed directly from Byzantine art, existing material does not enable us to decide.[340] The latter theory seems to be the more probable, and it is supported by the fact that the evolution of the Mesopotamian capital did not proceed upon the Bîsutûn-Iṣfahân lines. At Ukhaiḍir there is a reversion to the simple impost of Sarvistân, nor did the development there go beyond the elementary impost-capital of rooms 30 and 40. The capitals of the swelling columns on the north façade of the central court may have been more like those of Bîsutûn and Iṣfahân, but unfortunately they are completely ruined. At a later date, in the church of Mâr Ṭahmâsgerd at Kerkûk (eighth or ninth century), the scheme of the Sarvistân halls is repeated, but the pairs of columns are without capitals or bases, and the colonnettes of the niches in the spandrels are similarly treated (Plate 75, Fig. 1). I should be inclined to reconstruct all the columns and engaged columns at Ukhaiḍir and Sarvistân, and possibly at Ctesiphon also, without bases.

On the western side of the Syrian desert the evolution of the capital is different. The engaged capitals at Madâin Ṣâliḥ and Petra show a marked tendency towards the Corinthian. Like the capitals of the Kôm al-Shukâfa oasis[341] and capitals on Pompeiian frescoes of the second style, they have the Corinthian form and the Corinthian rosette upon the abacus, not indeed worked out into a true rosette, but left in the shape of a simple boss. In the second-century façades at Petra, such as the Corinthian tomb and the Khazneh, this tendency reaches full expression. The replacing of the architrave by the archivolt created a structural need which was satisfied by the introduction of the impost-capital, and we find the latter both at Mshattâ[342] and at Muwaqqar,[343] the capitals at Muwaqqar being closely related to the Bîsutûn-Iṣfahân type. With these stone-carved capitals, the brick and plaster capitals of Ukhaiḍir, so far as they are preserved, are little concerned. The further history of the Muwaqqar capitals must be sought, in the realm of Mohammadan art, at Sâmarrâ and in the mosque of Ibn Ṭulûn.[344]

New to Mesopotamian architecture are the clustered columns in the middle zone of the Ukhaiḍir façade. No doubt they are not essentially different from the triple supports between the arches of the second zone at Ctesiphon; but at Ukhaiḍir they are given a true architectural meaning, the central pair carries the wall, the flanking columns carry the cusped arches; moreover they are set in different planes, the central pair standing in front of the flanking columns. The effect produced is almost Gothic, a foreshadowing of the clustered piers of Armenian churches.[345] It was a scheme which was not to remain sterile in early Mohammadan art. Clustered piers carried the roof of the great mosque at Sâmarrâ[346] and the arcades of the mosque at Ibn Ṭulûn.

The first great distinction, then, between the second-century façades of Petra and the third-century façade of Ctesiphon is that the mock architecture at Petra is organically coherent, whereas at Ctesiphon it is incoherent, i.e. it is a pattern covering the wall-face rather than a simulation of plastic construction. The second great distinction is the systematic use of the archivolt at Ctesiphon for all the secondary intercolumniations in the wings. It is perhaps not without importance to observe that the same change from architrave to archivolt took place, though at a rather later date, in the stone-building regions of western Asia. In Syria, for example, the arched window almost entirely replaced the rectangular window in the course of the fifth century.[347] In the lower and central zones of Ctesiphon the arches are framed by groups in a rectangle composed of engaged piers and architraves; in the upper zone this system is abandoned. The principle of the arched niche within a rectangular frame appears, as has been seen, as early as Assyrian stelae, but for the use of the motive in a continuous series upon the façade there is, so far as I am aware, no example earlier than the Tabularium.[348] In the Augustan age it is found upon the Porta Praetoria at Aosta,[349] and thenceforward it governs the decorative scheme of Roman city gateways. Whether it was derived from Hellenistic Alexandria, together with the whole city gateway type, as Schultze surmises;[350] or whether it was evolved out of such wooden superstructures as gave birth to the decoration upon the Etruscan gates at Perugia;[351] or whether it was a specifically Roman (Stadtrömisch) conception, it is impossible to say. Nor does it signify. We know it as Roman, not only in the gateways, but also in the theatres and amphitheatres of the Roman empire, and I cannot doubt that the perfected Roman scheme is at least as directly responsible for Mesopotamian wall-surface decoration as is the western Asiatic development of Hellenistic façades. The gateway at Aosta, the Storied tomb at Petra, may well be taken as representing the immediate progenitors of Ctesiphon.

Five hundred years later, in round figures, comes Ukhaiḍir--five hundred years of architectural growth and of fairly continuous intercourse with the West. The architrave has vanished from the principal orders; it is retained only to form the old rectangular framework for the small niches at the top of the wall. Symmetry and organic cohesion rule over the two lower zones. But in the details of its composition there is nothing at Ukhaiḍir which might not have been foretold from the façade of Ctesiphon.

The lower zone of the north façade forms part of the decorative scheme of the central court as a whole. The central court resembles, as has been observed by Dr. Reuther, a Greek peristyle with engaged columns in place of free standing columns; the southern side is, however, treated as a separate façade, the façade of the lîwân. The principal feature was necessarily the wide arched opening of the lîwân itself. There is nothing new here; we have it at Ctesiphon, combined with Hellenistic wings; we have it at Firûzâbâd, without side doors, and at Sarvistân and at Hatra.

Hatra, though in plan it is no less purely oriental than Ctesiphon, shows direct Western influence far more strongly than the southern Mesopotamian or the Persian palaces. Dr. Herzfeld has compared its triple-arched façade, wherein the central arch surpasses the flanking arches in height and width, with that of the triumphal arch,[352] and the comparison is apt. So far as my knowledge goes, the triple-arched scheme appears for the first time in the Assyro-Persian cultural sphere at Hatra, and it is accompanied there by strongly Hellenized details of decoration, which distinguish it from the older oriental palaces to which it is related in plan. This Hellenized decoration is present in all other Parthian ruins, and it is not surprising that it should be so. The Parthians wrested their empire from a Greek dynasty. The Mesopotamia which they conquered was a part of Asiatic Greece; it was more closely linked to Greek culture than it had ever been linked before, or was ever to be linked again. The Hellenistic triple-arched scheme fitted the lîwân plan admirably, inasmuch as it provided the great opening which was essential to the lîwân hall. But it implied the placing of doors in the two flanking chambers, and this was done for the first time at Hatra. The side doors were an innovation which was not accepted without hesitation. It was not adopted in the façade of Firûzâbâd, where Hellenistic influence is almost entirely lacking. To a great extent the Sasanians stand for a reaction against Hellenism. A fresh wave of orientalism flows back into Mesopotamia with their conquest, and they went far to complete the severance with the West which the Parthians had begun when they overthrew the Seleucids. But the Greek domination, together with the fitful occupation of parts of northern Mesopotamia by Roman armies, left an indelible mark. Moreover, the Sasanian frontiers marched with those of Rome, and the interpenetration of the two civilizations was inevitable. It is felt in the façade of Ctesiphon. Though the triple-arched scheme is not present there, the provision of independent doors to the side chambers was a convenience; it was used at Firûzâbâd in the lîwân group at the back of the posterior façade; it was used at Ctesiphon, and thereafter it was not to disappear. With it the triple-arched façade came into favour. It formed part of the truly oriental façade of Sarvistân; no doubt it existed at Qaṣr-i-Shîrîn; it exists at Ukhaiḍir, but it is there completely re-orientalized. The ṭarmah-lîwâns bear a faint resemblance to the Hellenistic motive; in the lîwâns of courts C and G the likeness fades; in the south façade of the central court it is gone altogether and the side doors are no more necessary to the scheme than they were at Ctesiphon. In place of the triumphal arch façade we have the lîwân façade which dominates the architecture of Persia and of India. The central hall is raised above the flanking vaults and this raised vault implies a lifting of the central part of the façade. Dr. Reuther conjectures that a rectangular frame was given to the central arch, and since that is the stereotyped form of the lîwân façade of a later date, I have adopted his view. Moreover, some such device must have been used at Hatra. There, too, the vault of the lîwân rises above the flanking vaults, and Dr. Andrae, in his reconstruction of the façade, has given it a rectangular frame (Fig. 31). But at Hatra the arched opening of the lîwân was considerably lower than its vault and need not necessarily have broken the horizontal lines of the façade. It must, however, be borne in mind that something very like the later lîwân façade must have existed at Hatra, as it existed at Ukhaiḍir. Flandin and Coste, in their restorations of Sarvistân (_Voyage en Perse_, Plate 29), give a true lîwân façade to the principal entrance and to the side lîwân, and indeed their section indicates the vault of the side lîwân as springing so high that the façade must have been raised to correspond. The lîwân arch has been given in these restorations the same rectangular frame which has been conjectured to have existed at Hatra and at Ukhaiḍir. At Ukhaiḍir, as at Ctesiphon, the wings are decorated by blind arcades, two of which, for the sake of convenience, are broken by doors. The arcades are shallower than those which are carried round the other three sides of the court; the capitals of the columns, as Dr. Reuther has pointed out, must have been different from the other engaged capitals, since the shafts swell outwards towards the top;[353] and the calottes which cover the niches are adorned with Hazârbâf, the interwoven motive common in oriental woodwork.[354] The great arch of the lîwân is carried by pairs of engaged columns set in antis, and this is the arrangement which was usually adopted in the later lîwân façades. We have seen it in the tombs of Madâin Ṣâliḥ and of Petra. On either side there is a narrow arched niche which has the appearance of buttressing the central arch; beyond these follow three arched niches of wider span, the innermost on either side being slightly narrower than the others. The engaged column of the lîwân arch is joined to the quarter-column of the small flanking niche by a straight wall-face, on the same principle as that which is employed in the central supports of the ṭarmah-lîwâns of courts B and H. The result is in plan a double column, similar to the double columns which carry the arcades of every early Christian church in central Anatolia.[355] I saw one of these double columns in a graveyard at Raqqah, where it is used as a tomb-stone. They are foreshadowed in the Nabataean façade at Si’ in the Ḥaurân.[356]

The triple-arched façade must have been popular in the early Abbâsid period. It is found in the Bait al-Khalîfah at Sâmarrâ, where it is as pronounced as it was at Hatra. It was present in the two main façades of the audience chambers at Balkuwârâ.[357] But the single arched motive was to play an equally important part in Mohammadan architecture, a part of which an early (perhaps the earliest) indication is to be seen at Ukhaiḍir. On the north wall of the great hall the central feature is the great arch with its shallow calotte. Within this frame is set the smaller arched opening of the door. Here, as Fergusson has observed,[358] is the ‘perfectly satisfactory solution of a problem which has exercised the ingenuity of architects of all ages’. It has always been manifest ‘that to give a large building a door at all in proportion to its dimensions was, to say the least of it, very inconvenient. Men are only six feet high and they do not want portals through which elephants might march. It was left, however, for the Saracenic architects completely to get over the difficulty. They placed their portals--one or three, or five, of moderate dimensions--at the back of a semi-dome. This last feature thus became the porch or portico, and its dimensions became those of the portal, wholly irrespective of the size of the opening. No one, for instance, looking at this gateway (south gate of Akbar’s mosque at Fatehpur Sîkrî) can mistake that it is a doorway, and that only, and no one thinks of the size of the openings that are provided at its base. The semi-dome is the modulus of the design, and its scale that by which the imagination measures its magnificence’. The same principle rules over two of the smaller doorways of Ukhaiḍir, the doors at the outer ends of the corridor 5-6.

The arched niche, either blind or pierced with doors or windows, is used at Ukhaiḍir to complete the decoration of the north wall of the great hall. Blind niches with a rectangular frame stand on either side of the central calotte, while above it the three niches are pierced by windows. Here and in all other examples at Ukhaiḍir, the opening, simulated or real, is covered by a shallow calotte. In the central court the single niche at the south-east corner is potentially a doorway; it is covered by a fluted semi-dome (compare the doubtful example at Mshattâ, above, p. 118). In the same manner the niches on the two side walls of room 32 are potentially windows; at Karkh, where they are similarly placed, but in outer walls, they are actually pierced by window openings. The single niche motive is found in room 140, where, however, the niche is unusually shallow. That the form of such niches as those of the great hall and of rooms 31 and 32 is Hellenistic is not open to a moment’s doubt. Out of the countless classical parallels I may cite the aedicula upon the east façade of the basilica at Shaqqah.[359] The archivolt at Shaqqah is carried on colonnettes, the semi-dome is fluted, and the addition of a pediment, in the true Graeco-Roman style of Syria, involves the doubling of the colonnettes. The purely decorative character of the aedicula may well be compared with that of the niches on either side of the central calotte in the great hall. Dr. Reuther draws an apt parallel between the placing of the niches in the great hall and the placing of the niches in the building on the citadel at ‘Ammân,[360] and he calls attention to the fact that at ‘Ammân the colonnettes have neither capital nor bases and that the archivolts of one of the pairs of niches in room 32 are decorated with a zigzag ornament analogous to that of ‘Ammân. All these points help to prove the Mohammadan origin of the building on the citadel. It is not, however, strictly correct to describe the colonnettes either at ‘Ammân or at Ukhaiḍir as being without capitals. They are all provided with a small impost block. In room 32 a strikingly oriental motive is introduced into the niches on the side walls. The spear-shaped ornament in the centre of each niche was familiar to Assyrian decoration. Whether it had, or had not, its origin in the spear-shaped loopholes of fortified walls,[361] it is used for purely ornamental purposes in Assyrian decorative crenellations at Assur and in Parthian crenellations at Warka.[362] It was common in a similar position during the Achaemenid period,[363] and was carried on into later Mohammadan work, with the difference that the whole niche was given a spear-shaped or trifoliate heading[364] (Plate 75, Fig. 1). Nor are the recessed rosettes of the stucco decoration at Ukhaiḍir connected with Hellenistic types; they have affinities with the rosette motives of Assyrian fresco and enamelled brick,[365] but the floret shape of the Assyrian rosette disappears with the perspective treatment. In a cruder form the rosette of Ukhaiḍir is used at Mâr Ṭahmâsgerd. Here it is not recessed but cut deeply into the wall, and its effect is produced solely by the resultant shadow. The crenellated motive of the stucco work in the mosque has its counterpart in the ornamental crenellations of Assyria and Persia, but it is used at Ukhaiḍir with singular freedom. The crenellations are combined so as to form recessed rhomboids; they are even applied to the archivolt in the two doorways of corridors 5 and 6.[366] Save for the rosettes, all the stucco decoration at Ukhaiḍir is of an architectural character--that is to say that it imitates plastic construction such as crenellations, arched and columned openings; or else it is an elaboration of structural details, such as the squinch or the transverse arch. Sometimes it is actually called into being by structural processes, as in the horizontal ridges of the vaults in the mosque and room 31. The motives placed on the summits of the vaults in rooms 31 and 32 are reminiscent of coffering, and I have little doubt that their origin is to be sought in the Hellenistic scheme of ceiling decoration. It is, however, interesting to note that Western forms are more obscured at Ukhaiḍir than in buildings of a later Mohammadan period. The stucco coffers of the vaults at Sâmarrâ stand very close to classical types,[367] whereas the coffers at Ukhaiḍir are employed in a manner foreign to classical conceptions. This must be largely due to the fact that in the great palaces at Sâmarrâ Western artificers were at work, while in the comparatively unimportant desert retreat oriental workmen and oriental ideas had the upper hand, yet I would suggest that the differences between Ukhaiḍir and Sâmarrâ indicate a considerable difference in date. In the ninth century Western influence was stronger in Mesopotamia than it was in the preceding age, when the arts were still held closely in the thrall of Sasanian tradition. Consequently we find at Sâmarrâ capitals inspired by the Corinthian acanthus capital, and among the wall decorations the Hellenistic vine motive plays a conspicuous part.[368] We have yet to learn that the flowing vine, so essential to Coptic decoration and to that of the Hellenistic coast-lands, was a feature of Sasanian architectural ornament. It occurs in monuments of the Umayyad period which were directly under the sway of Hellenistic Syria, such as Mshattâ and the miḥrâb of the Khâṣakî Djâmi’,[369] but except for sporadic examples in Parthian architecture, where the Hellenizing tendencies of the decorations are indisputable,[370] its systematic use on Babylonian soil begins (so far as the evidence goes) at Sâmarrâ in the middle of the ninth century, and there it was the artificers, not the work of their hands, which were imported. I do not deny that in comparison with the Sâmarrâ palaces Ukhaiḍir is a crude product of local workmanship, wherein it is natural to expect a closer adherence to local tradition; but it is important to point out how close that adherence is, and how well it corresponds with recorded examples of Mesopotamian and Persian decoration earlier than the Umayyads, whereas the decoration in the same regions, but at a later period, diverges widely from the older schemes. The divergence is due, in my estimation, to the diffusion of Western influence when the western and the eastern provinces of the khalifate were drawn together under the Abbâsids and all quarters of their empire contributed to their constructions. In the ninth century we find Mesopotamian architecture in Cairo and Coptic decoration in Sâmarrâ. I regard the oriental character of Ukhaiḍir as indicative not only of its isolated position, beyond the direct course of international civilization and arts, but also as typical of the primitive age during which it arose.

Materials for the study of early Mohammadan decoration are still so scanty that the difficulty of assigning exact dates to such as we possess is great. It is enhanced by the fact that the workmen of the first khalifs must have been of non-Arab extraction. The Arab invaders, pouring in out of deserts which were innocent of monumental constructions, had nothing to contribute to architecture or to the arts. So far as we know them in the pre-Mohammadan period they had not created an art of their own. Along the trade-routes, the rock-cut tombs of Madâin Ṣâliḥ and of Petra exhibit, without salient divergence, the artistic principles of Hellenized Egypt and Hellenized Syria, while concerning the older Arab civilizations in the southern parts of the peninsula we have as yet no evidence save that of inscriptions. The Mohammadan conquerors employed the workmen of their predecessors, and according to the nature of their own traditions, these workmen might raise a palace with a basilical hall, like Mshattâ, or a palace entirely composed of lîwân groups like Ukhaiḍir; they might cover their walls with Hellenistic fresco, as at Qṣair ‘Amrah, or with ornament derived mainly from the ancient East, as again at Ukhaiḍir. The variations of this period were due to individual idiosyncrasy, or rather to individual training; there is no reason why they should be taken to denote a chronological distinction. A hundred and fifty years later this heterogeneous material had been welded together and the Islâmic _Weltkunst_ was beginning to take shape. Sâmarrâ, in the eastern part of the Abbâsid dominions, the mosque of Ibn Ṭulûn in the western part, re-echo one another; artistic conceptions are not only interchangeable, they are the same; and though, all through the history of the arts of Islâm, local peculiarities, based on local conditions and traditions, continue to differentiate one region from another, it is not the differences but the similarities which are the most striking. They go hand in hand with the singular solidarity of Islâm, with the uninterrupted intercourse between remote parts of the Mohammadan world, with the ceaseless passage of travellers and scholars from the western limits of Europe on the one hand to the eastern limits of Asia on the other. This intercourse was quickened, as the Prophet had intended that it should be, by the institution of the annual pilgrimage. The mosque of Ibn Ṭulûn is not an isolated example of a direct borrowing by one region from another. The gates of al-Mehdiyyeh in Tunis were copied from the gates of Raqqah.[371] It is impossible to explain the curious niching of the walls of the eleventh-century palace of the Menâr, to take another Tunisian example, except by a comparison with the wall-surface decoration of Babylonia and Assyria.[372] I am fully aware that a long period of time had elapsed between the fall of the Mesopotamian empires and the erection of the Menâr, and that it would be vain to attempt to establish a continuous sequence of buildings between them, but I would point out that the Parthians, when they reconstructed the Babylonian palace at Tellôh, reproduced the Babylonian wall decoration so closely that de Sarzec was persuaded that the ruins of their palace belonged to the Chaldaean age.[373]

Turn again to the fortress of the Bani Hammâd and you will find the cusp motive of Syria and Mesopotamia repeated on its arches;[374] and at the palace of Medînat al-Zahrâ in Spain (end of the tenth century) we have the plaster decorations of the walls of Sâmarrâ carried out in a style which betrays their Coptic and classical parentage,[375] though they are not devoid of characteristic motives, such as the palmette tree and the continuous pattern, which are rooted in oriental tradition.[376] In the same ruins the workers in stone have borrowed alike from Byzantium and from Mesopotamia; some of the continuous geometrical patterns are closely allied to those of Sâmarrâ,[377] while the free use of the crenellated motive may be compared with its use at Ukhaiḍir (Plate 87). The earliest Mesopotamian examples of such patterns as these are Parthian (Plate 86, Fig. 1).

One of the structural features of Ukhaiḍir has a value which is not only structural but also decorative. I allude to the use of masonry tubes between parallel barrel vaults. Obviously it is a scheme which was born of the systematic use of the vault. It is to be found at Hatra, where it appears in some of the tombs.[378] The same system is present at Firûzâbâd, where there was a masonry tube between the barrel vaults of the side chambers of the entrance lîwân and the domed chamber.[379] In later Mohammadan architecture I have found masonry tubes at Khân al-Khernîna above Tekrît.[380] A second device for the lightening of the wall mass between parallel barrel vaults is employed at Ukhaiḍir in the east annex and in the buildings to the north of the palace. It takes the form of a number of narrow tubes. I saw it also in a fourteenth-century khân at the foot of the Djebel Sindjâr (Plate 88, Fig. 1), a khân which is famous for the dragon reliefs on its doorway,[381] and in a mosque of the early fifteenth century at Ḥasan Kaif (Plate 88, Fig. 2). The decorative importance of the first scheme, the large single tube, lies in the effect which its opening produces on the façade. This can be observed in the courts on the ground floor at Ukhaiḍir, as well as in the court on the upper story of the gate-house. The arched openings of the tubes between the arched doors of the lîwân and its side chambers form an essential part of the façade, and they are retained when vault and tube are alike absent. The existence of tube openings in the façades round the central court, the ṣaḥn, of the mosque of Ibn Ṭulûn is sufficient to show that the Egyptian mosque was copied from a vaulted prototype (Plate 89, Fig. 1). I do not doubt that it was modelled on the vaulted buildings of Mesopotamia, though vault and tube are absent from its structure. The great mosque at Sâmarrâ was not vaulted; unfortunately the data are insufficient to determine the scheme of the façades of its ṣaḥn. Nor was the mosque of Abû Dulaf vaulted; it had a flat roof carried on arches, like Ibn Ṭulûn; but the tube openings appear in the form of niches on the façades of the ṣaḥn (Plate 89, Fig. 2). As at Ibn Ṭulûn, they have become purely decorative. I do not know whether there are tubes between the vaults of the Bait al-Khalîfah at Sâmarrâ, but the openings are simulated upon the façade by shallow blind niches. The same system holds good in the ṣaḥn façades of the Azhar at Cairo, a building which has no other connexion with Mesopotamian architecture than this traditional use of a decorative motive, the true significance of which had long been forgotten.