Ties.—As a general rule tie rods or beams were used, either of iron or wood. In the latter case they were painted with leaf or fret ornaments, and were evidently considered as natural features. But large vaults are often found without such ties as in the narthex of the Pantokrator. Many churches have ties to the dome-arches, and none to the main vault; but it is difficult to lay down a fixed rule. The enormous amount of mortar in the walls must have made them yield to a certain degree when newly built, and some of the larger vaults would have been the better for rods.

Abutments.—The system of abutments in the Byzantine churches of the great period has been carefully studied by M. Choisy. [33] In early examples the dome springs directly from the pendentives on the inside, but is thickened externally over the haunches, producing a double curve and an apparent drum. This is seen very clearly in SS. Sergius and Bacchus. In S. Sophia the numerous windows are cut through this drum, so that it resembles rather a series of small abutments. The object was to support the crown of the dome by adding weight over the haunches. In both these churches the thrust of the dome and its supporting arches is taken by the two-storied galleries, which form, in fact, flying buttresses within the buildings, and are adapted to their architectural requirements. The square plan and the enormous size of the dome in S. Sophia demanded the great buttresses on the sides; while in SS. Sergius and Bacchus the eight buttresses show only on the outside of the dome and are not carried over the aisles as they are in S. Sophia. Below the roof the arches and piers of the galleries and aisles are arranged so as to carry the thrust to the external walls, and following the tradition of Roman vaulting all buttressing is internal. In S. Irene, where the true drum dome first appears, the buttresses between the windows of the dome still remain, though much reduced in size. A dome raised on a drum can evidently no longer exercise a thrust against the dome-arches; its thrust must be taken by the drum, and only its weight can rest on the arches.

The weight of the drum and dome rests on the pendentives and dome-arches. Their thrust is neutralized by the use of ties and by the barrel vaults of the cross arms, and these in their turn depend on the thickness of the walls. The lower buildings attached to the church in the form of side-chapels and the narthex also helped to stiffen and buttress the cross walls. The system is by no means perfect in these late churches. It was apparently found impossible to construct drum domes of any size, except at the extreme risk of their falling in, and probably it is for this reason that many of the larger domes in late churches, like SS. Peter and Mark, S. Theodosia, the Chora, have fallen. No system of chainage appears to have been used for domes in Constantinople.

Flying buttresses probably of the ninth century are used at the west end of S. Sophia. The double-flying buttress to the apse of the Chora does not bond with the building and is certainly not original. It may be set down as part of the Byzantine restoration of the church in the fourteenth century. In any case, such external flying abutments are alien to the spirit of Byzantine architecture, and may be regarded as an importation from the West. Flying buttresses, it may here be noted, are not uncommon in the great mosques of the city. They are found in Sultan Bayazid, Rustem Pasha, Sultan Selim, the Suleimanieh, and the Shahzadé. But they are generally trifling in size, and are rather ornaments than serious attempts to buttress the dome.

Walls.—The walls of the earlier churches are built of large thin bricks laid with mortar joints at least as thick as the bricks, and often of greater thickness. Stone is used only in special cases, as in the main piers of S. Sophia, but monolithic marble columns are an important part of the structure. In the later churches stone is used in courses with the bricks to give a banded effect, and herring-bone, diamond, and radiating patterns are frequently introduced. The palace of the Porphyrogenitus, the parecclesion of the Pammakaristos, and Bogdan Serai, exhibit this style of work. As illustrations of the method adopted in the construction of walls the following measurements may be given, the sizes being in centimetres:

Brick.Joint.
Parecclesion of the Pammakaristos .08.04
4 courses brick, 5joints .46
S. John in Trullo .03.07 to .09
Refectory of the Monastery of Manuel .04.04 to .06
4 course stone, 3joints .78
4 courses brick, 5joints .30
.0375.052
Bogdan Serai.035.035
.04.04
4 courses stone, 8joints .55 to .60
4 courses brick, 5joints .43 to .47
Sanjakdar, brick .045

Building Procedure.—The first step in the erection of a building was to obtain the necessary marble columns with their capitals and bases. These seem to have been largely supplied ready made, and Constantinople was a great centre for the manufacture and export of stock architectural features. Then the main walls were built in brick, the columns were inserted as required, the vaults were thrown, and the whole building was left to settle down. Owing to the enormous amount of mortar used this settling must have been very considerable, and explains why hardly a plumb wall exists in Constantinople, and why so many vaults show a pronounced sinking in at the crown or have fallen in and have been rebuilt. After the walls had set the marble facings, mosaic, and colour were applied and could be easily adapted to the irregular lines of the walls.

Byzantine architecture made little use of mouldings. The great extension of flat and spacious decoration rendered unnecessary, or even objectionable, any strong line composition. External cornices are in coursed brick, the alternate courses being laid diagonally so as to form the characteristic dentil. The richest form is that found in the Pammakaristos, S. Theodosia, and S. Thekla, where the small dentil cornice is supported on long tapering corbels, a design suggested by military machicolations.

PLATE IV.