We have examined a handful of gold tesserae from S. Sophia through the kindness of Mr. James Powell. The cubes average a quarter of an inch in size, the glass is yellowish, slightly amethyst or dark green. The surface layer equals stout paper in thickness. At the back of the tesserae a dusty red appears, which under a glass proves to be of powdered tile. This roughens and adheres to the surface of the glass, which was evidently sanded with the powder while in a molten state, and of course before it was broken into morsels. The first purpose of this without doubt was to increase the hold of the cubes to the cementing material, but the reddening--almost like a coat of vermilion paint--may probably have assisted the gold to show out better than if the tesserae had been fixed without it into the perfectly white stucco which forms the bed. The cementing material was an inch or more in thickness, formed of lime with broken reed for binding, and a considerable amount of crushed white marble, in the part next the mosaic at least. § 3. GLASS, PLASTER AND PAINTING. The Romans probably largely used coloured glass for windows. The lattices were sometimes bronze or thin slabs of marble pierced into a pattern.[409] Sidonius († 484), describing the basilica of Tours, clearly mentions the patterned windows of green and sapphire glass.[410] It has been suggested that some of the windows at S. Sophia were filled with glass of brilliant colour. Theophilus, in his preface to the section of his work dealing with coloured glass, says, “I have approached the atrium of Holy Sophia, and beheld the chancel filled with every variety of divers colours.” He proceeds to describe windows of painted glass in which the pieces are united by leads: but assuredly, if coloured windows did exist in the apse of S. Sophia, the glass was inserted in pierced marble, like the plaster lattices of the Orientals. Beautiful windows of brilliant-hued glass exist in the mosques and turbehs. The Arab lattices show us what beautiful mosaics of jewels may be formed in this way; the singular charm of them is the spreading and blending of the colours, by reflection from the sides of the thick dividing bars; lumps of crystal seem to have been used occasionally in place of glass. Most beautiful ‘braided’ Byzantine lattices of marble are to be found at S. Mark’s which would be well characterised as θύραι δεδικτυομέναι which according to Lenoir was the name of these windows. If coloured glass was used in S. Sophia, we think it can only have been in small windows of this kind in the apse and conchs. Labarte thought, from the descriptions of Procopius and Paulus, that the windows were of white glass which allowed the rays of the sun to shine through unaltered. It is hardly possible to conceive of the great windows being of anything else than white glass.