_Dome_, &c.--Agathias tells us that when Justinian rebuilt the dome it was made higher, and that large alterations were made to the sustaining arches on the north and south sides. Salzenberg cites Theophanes and Zonaras who give the increase of height as twenty and twenty-five feet respectively. If we examine the longitudinal section we shall see that the great semidomes of the hemicycles and the apsoid of the bema show much less of their curvature outside than the present central dome. The windows in these do not stand above a cornice, but are pierced through the vaults at middle height; the domical surface being unbroken by any cornice from springing to crown. The cupola of the baptistery is also continuous with the pendentives. A dome of this kind, however, continuing the pendentives, would seem to be impossibly flat, and would be some thirty feet less than the present height--see A in Fig. 4, the existing dome rising to B. If a curve between these two be obtained by lowering the crown of the dome about fifteen feet to C, it may be noticed that a straight line tangential to the curve of the eastern apsoid, and also to the great semidome would form similar contact with the dome. Salzenberg, understanding an account of Cedrenus as to a strengthening of the abutments of the dome to refer to the great buttress masses which rise above the gynaeceum roof, considers that the external parts of these masses were additions made at the time of Justinian’s restoration. These great vertical piles are so essential to the structure, to the logical beauty of the design, and to the staircase service of the building; moreover the preparation for them beneath is so adequate, that we cannot accept this suggestion, and therefore follow Choisy in considering them original. Now Choisy, examining the external base of the dome where it forms a square, found that the four angles had been increased, and that it did not originally form a square, but rose above the piers and the lateral arches as shown in Fig. 34, and in Fig. 37, where the first base is shown by hatching and the additions by dotted lines, A A. “This alteration,” he writes, “is not hypothetical. I verified the entire absence of bond between the first base of the dome and the added work” (p. 138). These additions were built on the lateral arches, and on the top of the piers, altering the form shown in our Fig. 35 to the present form given by Salzenberg. That Choisy is right, is borne out by seeing the resemblance of treatment that there would have been between the growth of the dome on the north and south and the semidome on the west (see Fig. 34). [Illustration: FIG. 34.--View of Vaulted System of S. Sophia, adapted from Choisy.]