The marble is fixed to the wall with a dark brown resin. In the opus sectile, pieces of coloured marble about a quarter of an inch thick were cut to the forms of the design, and then laid with their polished faces downward at the bottom of a mould; on this was poured a three-quarter inch backing of resin mixed with bits of stone and brick. When set, the slabs so formed were attached to the wall with cement. The large marble slabs are one to two inches thick, and, besides the cement, are fastened to the walls by iron [? bronze] clamps. The pavements of ground floor and gynaeceum are of white marble with dark gray stripes. [Proconnesian.] In the south-east angle of the square area under the dome is a square of marble mosaic, of which details are given in Salzenberg’s Plate xxii., Figs. 9-15. It is formed of a circular centrepiece of a gray brown granite, ten feet two inches in diameter, round which are arranged coloured marble discs of various sizes, set in a mosaic of marbles, with a little glass mosaic in the angles. In the centre of the west end of the gynaeceum is a square [of about twenty-four feet] in the pavement laid with slabs of “gray cipollino” [Proconnesian], having a border of verde antique, with a patterned edging[282] of giallo and rosso on one side, and giallo and serpentine on the other. [Between this and the parapet is a circular slab of verde antique four feet seven inches in diameter.] CHAPTER IX THE ANCIENT PRECINCTS AND EXTERNAL PARTS OF THE CHURCH _Palace._--Before entering on particulars of the exterior of the church, it will be well to have a clearer view of the edifices in its immediate neighbourhood as they appeared in the time of Justinian. The group of buildings of which the Augusteum was the centre was profoundly modified by the fire of the Nika Sedition, and by the building energy of the emperor. The researches of Labarte and Paspates have been almost entirely confined to the elucidation of the palace as it existed in the tenth century. A restoration of the relative position of the several parts of the palace, unless by the discovery of remains positive evidence is obtained, is certainly impossible; the attempt of Labarte was worth making, but Paspates, in bringing forward another scheme, seems only to have succeeded in showing how conjectural the whole matter is, although he speaks of certain scraps of walls as belonging to this or that part of the palace with as much confidence as if he had found them labelled. His work carries internal evidence of the greatest inexactness and confusion, and has proved most misleading, although his citations are valuable.