The Church of Sancta Sophia, Constantinople: A Study of Byzantine Building

MARBLE MASONRY

§ 1. BUILDING PROCEDURE.

The method and sequence of the building operations as followed by the Byzantines seem to have been very much as follows. After the form of the building had been more or less decided, the first thing necessary was to collect marble monolithic shafts. At S. Sophia the eight verde-antique shafts match one another very closely; they are all of one length, and vary from 7½ to 8 diameters in proportion. The four pairs of porphyry shafts in the exedras differ much more; and, as we have remarked, those in the western exedras seem to be made up of separate drums. The proportions of these vary from less than 7 diameters on one side to 8½ on the other. The great monoliths are the largest known, and of nearly normal classic proportion, so we can readily see that it was necessary to have a certain knowledge where such marbles might be quarried or otherwise obtained, before even the foundations were prepared, for the columns decided the heights and points of support of the building. These once assured, the body of the structure was proceeded with as a brickwork shell without further dependence on the masons, who were only required to prepare bases and capitals, and then the cornices; everything else was completed as a brick “carcase.”

At S. Sophia the main square piers are in fact stone, but this was only for strength, not because they were to be seen finally, any more than the rough brick.

The building completed in this form we must remember was made up of vast masses of thin bricks, of which the mortar occupied probably a half of the aggregate; this had to thoroughly settle down and dry before the rest of the marble masonry was inserted, and the wall casings applied. The marble work, however, was all the while being prepared, and, the building once ready, the windows were inserted as screens in the openings previously left; marble jambs and lintels for the doors were placed in position also, with windows above them filling out to the brick arches. The walls were then sheeted with their marble covering, the vaults were overlaid with mosaic, and the pavement was laid down. In this way, as the bricklayers had not to wait for the masons, the carcase was completed in the shortest possible time; and by reserving the application of the marble until the structure was dry and solid, it was possible to bring together unyielding marble and brickwork that must have settled down very considerably.

§ 2. MARBLE QUARRIES.

Much confusion exists as to the marbles of which the ancient writers speak; this has been occasioned necessarily by wrong identifications when but few ancient quarries had been recovered, and most unnecessarily by a persistence in using antique names for modern varieties, long after the true provenance has been discovered, when the ancient marbles are not “in the market.” It is the Italian names that have been corrupted in this way, and it would be a great advantage if they were discarded in England, or better still, used only in conjunction with the geographical names. In this case as the Italian names are descriptive, and, as many varieties of marble are found in the same or neighbouring quarries, we should get a safe nomenclature. Synnadan would thus be qualified as Pavonazzetto or Fior de Persico, and the banded varieties from Carystian, Proconnesian, or modern quarries might without confusion be called cipollino.

In endeavouring to identify the marbles mentioned by the ancient writer on S. Sophia, we have made use of Salzenberg’s notes to the Poem of the Silentiary, and of the researches of Garofalo,[350] Corsi,[351] and C. O. Müller;[352] and we have also been helped by the practical knowledge of Mr. W. Brindley. The account of ancient marbles easily accessible in Professor Middleton’s Ancient Rome, 1892, is substantially an extract from Corsi.

Porphyry.—The “porphyry powdered with bright stars” of the poet is used for the columns of the exedras, and for some of the panels on the walls. The Anonymous author states that these columns came from a temple of the Sun, but the Silentiary says “they loaded the boats on the bosom of the Nile,” and there seems no reason to doubt that the columns came direct from the porphyry quarries at Mons Porphyrites in Egypt. This porphyry mountain is at Djebel Dochan, twenty-five miles north-east from Thebes. Lepsius[353] seems to prove that the quarries were worked as long as the Nile canal remained open; and ships still sailed on the canal till the appearance of Islam. Letronne[354] gives details of the method of transit. The porphyry was brought from the quarry to the Red Sea, and then by the Nile canal to the Lower Nile, and hence into the Mediterranean.

On this evidence we would say that the porphyry used at Constantinople in Justinian’s reign was quarried for the purpose, and not brought from Roman buildings.

Marmor Molossium.—“The marble that the land of Atrax yields,” is called elsewhere in the poem “Thessalian,” and, from the province in Thessaly where it was found, “Molossian.” Corsi and Garofalo both wrongly describe Molossian as Fior di Persico. The marble really is the brecciated serpentine and limestone, now called Verde Antico, the Lapis Atracius of the ancients, of which the eight great columns in the nave and many others are formed. Here again it has been said that these eight large columns were taken from a building at Ephesus, but the Silentiary says, “Never were such columns hewn from sea-washed Molossis,” and we can hardly doubt that they were quarried especially for S. Sophia, together with the rest of the enormous quantity used in the church. The quarries were near Atrax in Thessaly, and the marble is best named as by French writers, Thessalian green.

Lapis Lacedaemonius.—“The fresh green, like emerald, from Sparta,” was probably the porphyry quarried in Mount Taygetus in Laconia. This green porphyry, called by Corsi serpentino, is used in the opus sectile of S. Sophia. As a green porphyry is obtainable in Egypt, the former should be distinguished as Spartan.

Proconnesium.—“The hills of Proconnesus,” according to Paulus, “strewed the floor.” The same marble was also used for the columns in the upper aisles, for the eight square columns below, and for the capitals, door frames, window lattices and other structural parts; also for the plating of the lower arcade and other parts of the wall-surfaces, and as frames to the coloured marbles. It is a soft white, or white with gray-banded streaks. The quarries of Marmora are still worked. This marble was greatly prized in Classic times, and Pliny mentions that it was used at the palace of Mausolus, where, it is said, the method of plating brick walls with marble was first applied. It closely resembles gray Carystian but they should not be confounded.

“The Bosporus stone with white streaks on black,” used for the floor, was probably the ordinary limestone—black with white veins—used at Constantinople.

Marmor Carystium.—“The fresh green from Carystus,” is the marble now known as cipollino; it was quarried at Carystus, at the foot of Mount Ocha, in the island of Euboea. Its beautiful greenish white surface, marked with broad wavy lines of green or purplish gray, was often praised by the later classical writers. Its resemblance to the markings of a sliced onion is the origin of its name. Modern cipollino need not be confused with true Carystian marble, which the ancient material should always be named.

Marmor Phrygium.—“The marble hewn from the Phrygian land towards the Mygdonian heights,” spoken of as “many-coloured,” has been identified as the marble which came from Dokimion near Synnada in Phrygia. The descriptions by Statius and Claudianus of the deep red-veined marble of Synnada agree closely with the Phrygian and Mygdonian stone as described by Paulus. It is a brecciated marble of a rosy colour, slabs of which alternate with verde antique in the panelling of the side aisles of S. Sophia.

The quarries at Dokimion were visited by Leake and Texier, and a recent examination of them by M. Leonti[355] disclosed all shades of “violet and white, yellow, and the more familiar brecciated white and rose-red.” This beautiful material is best called Synnadan, as the modern Italian name Pavonazzetto is also used for the streaked marble quarried at Carrara.

Marmor Hierapolitanum.—“The stone from the sacred city Hierapolis.” This marble has been identified by Professor Ramsay.[356] It was found at Thiounta about ten miles N.W. of Hierapolis in Asia Minor. It is variegated like Synnadan, and was much used for sarcophagi; indeed Professor Ramsay says, “On every occasion when its use is mentioned, it was employed to make sarcophagi.” It was called by the name of the great city which is not far distant, “and to which doubtless orders from the outer world were sent. Similarly the marble found at Dokimion was always called Synnadic marble from the time of Strabo, yet Dokimion was thirty-two miles from Synnada.”

Marmor Iassense.—The “Iassian, with slanting veins of blood-red on livid white,” was used for the phiale. Corsi identifies this with Porta Santa, but Porta Santa, Garofalo says, came from Chios, and this conclusion we believe is now accepted. Garofalo thought Iassian to be the same as the Carian marble mentioned by Porphyrogenitus in his Life of Basil the Macedonian, and says it was quarried on the island quite close to the coast of Caria. A “stone mingled with streaks of red” is also mentioned by Paulus as brought from “the Lydian Creek.” Possibly the port of Iassus is again intended. The ordinary Lapis Lydius was a black touchstone. The “rosy cipollino,” in which wide bands of deep red alternate with white, used in the panelling of the aisles does not seem to be mentioned specifically by Paulus; unless this is the Iassian marble to which his words would very well apply. A variety of rosy cipollino, the splendidly figured red and white marble, is obtained in Laconia.

Marmor Numidicum.—“The stone, nurtured in the hills of the Moors, crocus colour glittering like gold,” is the beautiful warm yellow African marble from Semittu Colonia, about fifty miles from Tunis, so highly prized by the Romans, and now called giallo antico. It is used in S. Sophia in the sectile work.

Marmor Celticum.—“The product of the Celtic crags, like milk poured on a flesh of glittering black,” has been identified as the Bianco e Nero Antico, quarried in the Pyrenees.[357] The black marble with white streaks, which occurs in some of the panels in the nave, is probably the one to which the poet refers.

Onychites.—“The precious onyx” mentioned by the poet is the alabastrites or onychites of the ancients. It is the oriental alabaster (aragonite) used in the horizontal bands of the nave, and some of the panels. It is a translucent, fibrous stalagmite formation, generally of a clear honey-colour. Some of the varieties are strongly veined with white, and others are much darker. Large ancient quarries of this Egyptian alabaster have been discovered on the east bank of the Nile.

Paulus appears to make no mention of the dusky black with dull golden veins used in the bema apse, which closely resembles the “Porto Venere” quarried at Spezzia.

The marble blocks were roughly hewn into shape with picks while still attached to the rock, and were then separated by the aid of metal wedges. Many objects discovered show that they were sometimes completed at the quarry, at other times the blocks were roughly brought to the sizes and forms required. The quarries appear to have been officially inspected. Texier found many architectural fragments and blocks at Dokimion bearing the signs of the inspectors of the block. Professor Ramsay writes: “The route from Dokimion to the coast is commercially almost the most important in Asia Minor. The road along which the enormous monolithic columns were transported passed through Synnada, where the central office for managing the quarries was situated.”

Ephesus and Alexandria were most important centres for the working and export of marble, of which such an enormous quantity was required by the Byzantine builders. The method of slicing up the blocks into veneer is described by an Eastern pilgrim, Nasiri Khusrau, in 1047. He says: “In the city of Ramlah there is marble in plenty ... they cut the marble here with a toothless saw which is worked with Mekkah sand.” This sand he tells us came from Haifa near Acre (Pal. Pilgrims’ Text Soc. Compare Pliny, Hist. Nat. xxxvi.)

§ 3. APPLICATION OF MARBLE.

At S. Sophia the application of the thin sheathing and incrustations (the “crustae” of Pliny) of the “delectable variety” of marbles is made in many ways. First there are the large sheets of the grayish Proconnesian, opened out side by side “so that the veining of one follows from the next.” Then the richer varieties are set in bands and panels with narrow notched fillets between them, and still more precious slabs are framed round with carved margins of white. Over the doors entering the aisles at the west there are panels with especially wide and rich borders of meanders growing from chalices. The large panels are very often of two pieces with matched veining. Fig. 46 shows one of a row of strongly veined panels from the narthex with the frieze above. All the wall plating is arranged with delightful variety as to size, and in the alternate placing of light against dark, so that there is no rigidity or over-accurate “setting out.”

Besides this constant change of size, colour, and arrangement, there is a great variety in the surface treatment. We have the shallow channelling into continuous mouldings of the skirtings, some portion of which has a stiff fret sunk in the surface in addition. Then there are panels on either side of the great door, and on the faces of the projections from the great piers in the aisles, coming just above the eye, (Fig. 48) of plain russet-red or brown which bear severe abstract patterns, made out by slight sinking into the surface. The centre in some cases is overlaid with an oval or square of another precious material such as red or green porphyry or the “onyx”; the whole of the sunk portions may have been filled by inlays, or in some the sinking alone may have formed the design. The upper part of the bema is incrusted with slabs patterned in this way, and here the sunk portions are entirely inlaid; several parts of this are represented by Salzenberg. In this work “casements” are sunk into the rosso or other deep coloured field, and green porphyry and other materials, set off by yellowish-white lines and spaces are inlaid in geometrical panels, or friezes of stiff foliage.

Our Fig. 47 shows the arrangement of the marble plating on the great piers towards the middle compartment of the aisles; in this we have shown one of the enriched panels now only sunk, as inlaid. Fig. 48 gives outlines of others of these panels. The marble used in the aisles is as follows. First comes the moulded skirting of white Proconnesian, then a 3′·3″ band of the streaked variety of the same marble. A band of verde antique 2′·0″ wide follows, above which is a row of slabs alternately verde antique and Synnadan. A second similar row of slabs comes above a band of rosy cipollino. The frieze below the cornice is of marble sectile work. The passages through the piers are lined with slabs of streaked Proconnesian marble, nearly fourteen feet high.

The gynaeceum has two bands at the bottom and an upper band of rosy cipollino; the wall space between is covered with a row of vertical slabs of streaked Proconnesian, except the central space on north side where the slabs are of rosy cipollino. In the spandrils of gynaeceum arcade at the west are roundels of oriental alabaster.

Directly over the Royal Door is a very beautiful arrangement of decorated slabs. First there is an immense upright piece of verde antique in the middle, ten or twelve feet high, with two lateral horizontal pieces making a great cross, in the quarters of which are panels with sunk and inlaid designs. At the head of the cross is a fifth panel which displays a still richer form of decoration. It represents a vaulted recess or ciborium between the columns of which hang curtains, looped back, and displaying a dark field. Here is the matrix of a cross which was probably of silver; right and left of the cross are other matrices, in which were set crowns or other objects, not to be determined from below. The two upper lateral panels have sunk geometrical designs. The lower pair are inlaid; their centres are charged with circles, above and below which are pairs of dolphins. These inlaid designs are made out in porphyry and green, which are separated by white lines and spaces which shine out bright, and are probably of mother of pearl like similar inlaid panels of this date around the apse at Parenzo. These panels at Parenzo are so much like those of S. Sophia that we do not doubt they were sent from Constantinople. There are very similar panels in the baptistery at Ravenna.

Finally we have the enriched surfaces of the two ranges of arcade spandrils. The upper row being sectile work of coloured morsels put together to form a pattern of scrolls and foliage, and the lower series having the surface entirely sculptured with the exception of discs of precious substance which are set in them.

This uttermost splendour is quiet and soft in its result. The surface of course has not that mechanically even, repellently smooth, painfully fitted appearance of modern work. The planes are waved under the hand sawing, and the face is smooth but hardly polished. The colour in consequence, gray and russet rising to full yellow, green and reds, veined, waved, and flowered in all manner of gradations and lovely combinations, vibrates with a wonderful “bloom” which doubtless owes much to age; but it is very probable that the marble was polished with wax encaustic which was so generally used for finishing surfaces by ancient workers. The wax deepens and mellows the colour and leaves a dull pleasant polish. We suppose the method followed was that recommended by Vitruvius for the encaustic polishing of coloured stucco walls. “Lay on with a brush a coat of melted Punic wax tempered with oil; then with a brazier of hot charcoal heat all the waxed surface, forcing the wax to melt in an even way over the whole surface; finally rub the wall with a wax candle and then polish it with a clean linen cloth just in the way the nude marble statues are treated. This practice is called γάνωσις by the Greeks.” Felix Fabri, who travelled in Palestine at the end of the fifteenth century, describes the rows of costly columns at Bethlehem, “and they are polished with oil so that a man can see his face in them as in a mirror.”

In regard to the wall plating we wish especially to point out the extremely easy way in which it is applied, without thought of disguise. The slabs of great size are placed vertically, entirely the reverse of solid construction; moreover the slabs of the finer panels are opened out side by side so that the veinings appear in symmetrical patterns. At the angles the lap shows in the most open way; while it is mitred where restored. The best account of the actual methods of fixing the marble slabs to walls by metal clamps which notch into the edges of the sheet before the adjoining one is fixed, is given by Professor Middleton, who figures an example of the second century from Rome which might belong to S. Sophia.

§ 4. MARBLE MASONRY.

After more than a thousand years of working marble through one complete development, Greek builders, by considering afresh the prime necessities of material, and a rational system of craftsmanship, opened the great quarry of ideas in constructive art which is exhaustless. In a hundred years architecture became truly organic, features that had become mere “vestiges” dropped away, and a new style was complete; one, not perhaps so completely winning as some forms of Gothic, but the supremely logical building art that has been.

If anywhere this vitalising had not been completed, it would have been in the more decorative forms; but here we find no mere exercise in applying architectural orders, everything is as real and fresh as in the structure. Having the Corinthian and Ionic capitals before their eyes and without forgetting or rejecting them, the Byzantine builders invented and developed an entirely fresh group of capitals fitted in the most perfect way for arched brick construction. As Mr. Freeman has said (Hist. Essays, iii. p. 61) of the new architecture: “The problem was to bring the arch and column into union—in other words to teach the column to support the arch.” This was done by shaping the block of marble which formed the capital so that a simple transition from the square block to the circle of the column was formed. When they were sculptured, and most of them are most elaborately sculptured, the general form is not altered but the carving enriches the surface only. The new “Impost capital” is found throughout the great cistern generally known as that of Philoxenus which is usually referred to the time of Constantine. In their study of the vaulted cisterns of Constantinople Forchheimer and Strzygowski have contributed much that is new to our knowledge of the architecture of the city and show that the evidence is entirely against this theory, which was propounded by Gyllius, whom more recent writers have been content to copy. This cistern, known to the Turks as Bin Bir direk (thousand and one columns), they identify with a great cistern which the Paschal Chronicle says was built by Justinian in 528. We believe with them that the architecture of the cistern agrees entirely with what we might expect as an outcome of the special circumstances in the time of the great building era. “Bin Bir direk exhibits the highest development of the art of cistern building, and it thus in its particular sphere resembles S. Sophia; like it the boldness of its construction was never again equalled by the Byzantines. It would be an explanation of the bold achievement if it might be assumed that Anthemius proved his capability in this subterranean work before he made his supreme effort in S. Sophia. Technical features, however, make it seem probable that the builder was an Alexandrine.”

“It is of the widest significance for the history of Byzantine art that here throughout the new ‘impost capital’ is employed in its plainest constructive form. It seems not improbable that the daring builder of the cistern was the first to make use of this form of capital which completely broke with classical tradition and is in such perfect accord with the exigencies of arch-architecture.” This is to go too far; for if the cistern is rightly referred to 528 it is probable, as we shall show, that the impost capital had at that time been for many years in use.

At S. Sophia the four main varieties of the new capital are all found. In the cistern the change of form is made by rounding away the angles at the bottom without reference apparently to any geometrical idea; but in other capitals which belong essentially to this type the method seems to have been that explained in Fig. 53 which represents the form of the caps of the lamp pillars on the front of the western gynaeceum. They are most delicately carved with a network of ornament, but the general form is undisturbed as we have explained. The plain capitals of the west window and the isolated sculptured capital Salzenberg found in the north aisle are also of this form, which we shall call the Impost Capital type I. The profile can be made convex or inflected, we are only speaking of the simplest method of changing the form from a circle to a square.

Two capitals now used as mounting blocks outside the east porch, which we illustrate (Fig. 51), furnish us with a sculptured example of a similar capital in two stages of development, one of them never having been completed. We give here an outline of the blocked out capital, in which the method of workmanship may be plainly seen. First, the block was cut away below convexly to meet the circular shaft. In this state it exactly resembles the capitals of the cistern. Secondly, on this was marked a border all round the top; also centre lines running down each of the faces, about the centre point of each of which a circle of about seven inches diameter was drawn; and at the bottom the width for the necking was marked off. Thirdly, the intermediate spaces were sunk about two inches; the hollow of the abacus was formed; the necking, and edge of the circular discs were rounded. This brings the capital to the stage shown in the diagram, the point to be observed being that the abacus, boss, and necking lie in one surface, first obtained, and the rest in another face, sunk some two inches below the former. It cannot be doubted that the style of these capitals is contemporary with the work at S. Sophia, and the finished one bears a monogram which appears to read ΘΕΟΔΩΡΟΥ; it is, however, almost identical with that of Theodora, which occurs on the capitals of the interior. MM. Curtis and Aristarches,[358] who have written on these monograms, think it belonged to a portico, restored in 409 by an eparch called Theodoros. Work of this style was not done at that time, and these capitals possibly belonged to some of the outer courts of the church mentioned by Procopius. They resemble the great capitals so closely that they might almost be preliminary studies. The strips which are left down two sides of the capitals were customary in the capitals of a Byzantine colonnade, especially where screens were inserted between.

The two capitals in the loggia by the baptistery furnish a well-defined variety of the impost capital. The square at the top is here wrought into curves recalling the antique abacus. These are gathered together into the circle of the necking in a beautiful convex form which may be called the Melon type II., see Fig. 54.

We give in Fig. 50 an outline of the whole column of the great order in the interior of the church, and in Fig. 56 a diagram of the blocking out of the capital. The columns here and throughout the great church being monoliths of fine material, the supporting area is very small compared to the area of the arch imposts, which are of brick sheeted with marble. It will be seen that the projection is just that required by the impost, which springs directly from the outside edge.

The great capitals of S. Sophia are remarkable examples of the evolution of beautiful forms on the mason’s banker; the workman finding form in the stone block by the application of practical methods. The lower half of the capital is circular like the shaft, rising in a slightly swelling curve of a bowl; the upper part is square like the impost. The basis of form is that of a bowl with a tile placed above it, and is thus that of the Greek Doric. This type III. in which the circle does not pass by transition into the square impost, but changes abruptly, we may call the bowl and tile capital.

At S. Sophia the surface of the form obtained as shown in the figure is wrought into crisp acanthus and palm foliage; and is in many places, especially at the tips of the leafage and behind the monograms, entirely undercut. The cutting being so sharp, and the shadows so deep, while at the same time the general form with its broad gradation of light and shade is so little modified by surface modelling, the effect is almost that of inlaying black on white. The capitals of the columns standing in the aisles, and those of the first floor ranged against the central area, are similar to the great order, but simplified and reduced.

The columns of the aisles on the first floor have block capitals, with small volutes below; Fig. 57 will make the elementary form clear. This type IV. is really a Byzantine Ionic. The dual columns of west gallery have a capital in common, which is a variation of these, and the capitals of atrium were also similar. One capital of the north gallery is entirely different from all the rest, the block, not being carved all over continuously, is broken up into several horizontal lines of ornament.

For the capitals of the square pillars of ground floor, and others to the windows, we must refer to Salzenberg; they are all of the simple block form delicately sculptured.

Salzenberg also figures two capitals, now on the porphyry columns at the east porch. These are comparatively small, and may possibly have belonged to some position in the interior of the church, such as Justinian’s first ambo. The form is that of a basket with four doves perched on the rim, and crosses between. Doves associated with crosses symbolized the Church. Now in St. Clemente at Rome there are two capitals of this kind which belonged to the ciborium, set up as the inscription shows while Hormisdas was pope (514-523), they are figured by Cattaneo, Fig. 7, who says they obviously were sculptured by Greek chisels. It is thus extremely possible that ours may have been late additions to the pre-Justinian church, where they also may have belonged to the ciborium. Rohault de Fleury believed that this form of capital was intended to represent an offerings basket.

To these Bird and Basket capitals, type V., may be added varieties of the great class of derivatives from the Corinthian of which this is in fact one. These were in general use before the block type of capital was developed. We will here only mention two of these acanthus capitals. Those in which the leaves are set upright on the stem of the shaft we will call Byzantine Corinthian and type VI. Those in which the leaves turn over and bend round the capital we will, with Mr. Ruskin, call “Wind-blown acanthus,” and type VII.

Distribution and Dates of Capitals.—We have referred before to our belief that Constantinople was a marble working centre from which sculptured marbles were dispersed to all parts of the Roman world. Having the chief types of Byzantine capitals before us it will be convenient to consider this more fully. We suppose that as white marble had to be bought in any case, the custom grew up of obtaining the capitals fully wrought. Importation was, of course, a general antique practice in regard to figure sculpture, columns, and other objects of marble. Proconnesian marble seems to have been the common stone of Constantinople so that it is used for the columns and capitals of the cisterns. We believe that careful examination of the capitals at Ravenna, Parenzo, and other Byzantine centres will show that they are in the main of this material. As to design the capitals lying neglected about the city, together with those in situ in the churches and cisterns, furnish a perfect museum of the types with which others dispersed through the whole area of the empire agree in the minutest particulars of design and workmanship.

To take the types we have mentioned:

Impost Capital, I.—This capital is found with the surface richly sculptured at S. Sergius. Capitals identical in form and decoration with the isolated capital of S. Sophia (Salz. Pl. xx., fig. 8) are found at Parenzo and in Jerusalem. The splendid examples of this type at S. Vitale, Ravenna, are well known; here the fretwork of sculpture is almost entirely relieved from the ground. We found an example absolutely similar at Constantinople. Mr. Ruskin’s “Lily Capital” which belongs to this group is found at S. Mark’s, at S. Vitale, at Parenzo, and at Alexandria. Another variety is covered all over with horizontal bands of zigzag fillets; an example rests in the Tchenli-Kiosk Museum, others are found at Athens, at Mistra, and a third now at S. Mark’s is figured in the Stones of Venice. The capitals at S. Sophia, Salonica, figured by Texier are probably the earliest of type I. to which an approximate date can be given; it was certainly in general use at the end of the fifth century.

Melon Form, II.—These magnificent eight-lobed capitals form the great order at S. Sergius, and are found at the church usually called Agia Theotokos. Similar capitals belong to the upper order at S. Vitale, and others are found at S. Mark’s. Some of the nave columns of S. Demetrius at Salonica have fine capitals of this type which although evidently derived from the last probably also originated in the fifth century.

Bowl Type, III.—These, the great capitals of S. Sophia, seem to have been especially designed for the metropolitan church: the beautiful palm foliage, however, with which they are sculptured is found again at Parenzo and on a capital in the Ravenna museum said to have been brought from Pomposa. The church at Parenzo was begun in 535.

Byzantine Ionic, IV.—These occur in their perfected form of block capital fully sculptured in S. Sergius and at the palace of Hormisdas in Constantinople, also in the upper order at S. Sophia, Salonica. Examples are also found at Venice.

In their earlier form of transition from the “Ionic with a plain dosseret” an immense number are found in the subterranean structures of Constantinople. An example has been found in Chalcis.[359]

Bird and Basket, V.—S. Sophia furnishes two examples, but there is no proof that they originally belonged to the building. Another example is in Cairo. That at S. Clemente, Rome, is signed with the name of John Mercurius; Piranesi figures a capital of this kind from the Palazzo Mattei, bearing a monogram which is indecipherable in his plate. Period, end of fifth century and beginning of sixth.

Byzantine Corinthian Type, VI.—These are of great variety; we will only mention one. In the portico of John Studius the acanthus leaves are doubled, one leaf lying over and within another, so that a double row of serrations is shown around the margins (see figure in Salz.). Similar capitals are found in S. Demetrius, Salonica, and at S. Mark’s, Venice. This particular form is probably nearly concurrent with the last, possibly a little earlier.

Wind-blown Acanthus, VII., is represented at Constantinople by two examples forming bases for the posts of a wooden porch to a house near Gûl Jami, and another is found in the cistern usually called after Arcadius or Pulcheria. Absolutely similar capitals are found in S. Sophia, Salonica (circa 490) and one occurs at S. Demetrius. At Ravenna fine examples are dated by bearing the monogram of Theodoric. Others at S. Apollinare in Classe resemble the last so closely that we doubt their having been made specially for the church built in 534-549. An example was found in Chalcis with the Ionic capital just referred to and De Vogüé figures one from Syria. Period, say 425 to 525.

The seven most typical Byzantine orders were thus being wrought concurrently at the end of the fifth century, and it seems that the three last did not long outlast this century. The others in their central types probably did not continue in use much beyond the sixth century. After this time somewhat coarse varieties of Byzantine Corinthian, or Type I., were mostly used.

The evidence of the original block in the fully sculptured finished work which we find in the most characteristic examples of the Byzantine capitals is of primary importance in all marble sculpture, and differentiates the work of the chisel from being a mere stone model of a clay model which is practically what most modern sculpture has become. In many of these capitals the vertical strip shown in Fig. 55 left in the finished work furnishes a further suggestion of the block from whence they were hewn.

Shafts and Bases.—The usual theory that the Byzantines wrought but few new marble shafts does not bear scrutiny. Byzantine shafts have neckings of very slight projection, thus obviating the waste of labour and material of Roman work.[360] The shafts of the baptistery loggia at S. Sophia, figured by Salzenberg, furnish good examples; sometimes the necking, as to the square marble pillars, is a simple broad fillet of about a quarter of an inch projection. The hundred round shafts of S. Sophia exhibit a remarkable and beautiful structural expedient by which the necking is entirely suppressed, and bronze annulets surround the shaft under the capital and above the base; which prevent the shafts from sliding or splitting, and retain the lead beds from being forced out by the weight (see Choisy, p. 15). Large monolithic shafts were the more apt to split, as they had to be set up contrary to the direction of the quarry strata.

Fig. 58 represents these bronze zones in association with the great capitals and bases. The pedestals of the exedra columns A a, next figure, are worked together with the bases in one stone. In these profiles we again see how little the mouldings disturb the original form.

Responds.—A very remarkable feature in the interior, is the way in which the colour of the marble columns of the arcade is reflected as it were on the responds, where the arches fall on the great square piers. A strip of porphyry or verde antique, the width and height of the free shafts, is inlaid into the marble casing of the piers absolutely flush, the edge being only defined by a line of the notched fillet. A flat sculptured slab at the top echoes the capital, and a base slab of mouldings worked in a vertical plane ranges with the bases of the columns. Salzenberg’s plate does not render this feature properly, the “capital” is flat and has straight sides and instead of the “base” he shows a portion of the wall skirting. Fig. 59 shows this base in elevation (B), and section (C), ranging with the pedestals of the exedra, Columns (A). The way in which the sculptured and inlaid spandrils of the arcades stop against the plain veneering of the great piers is also most noteworthy.

Cornices and Skirtings.—We give here (Fig. 59, D and E) two profiles of the skirtings where the principle of working out of thin veneering-slabs is applied to moulded work. The parapet slabs of first floor are worked in a very similar way; Salzenberg shows design of front, and they bear flat lozenges between two crosses at the back. See Fig. 61.

The cornices of the interior, which really formed walks for the lamplighters, are made up of no regular combination of curves; they project steeply forward, the general slanting plane being little disturbed (A, Fig. 60); they are decorated with rows of acanthus, the curved tips of which catch the light in bright points. The cornice of aisle is given at B. We also give a profile of the door-head, which shows how the mouldings conform to a plane of least labour (C). By the jambs and heads being mitred together, the difficulty of working stop ends was also obviated. The mouldings are not sharp and accurate, as is suggested by Salzenberg’s engravings.