About the later Ptolemaic time and through the Roman age the main work in glass is that of minute mosaics (fig. 122). They were built up with glass rods, heated until they half fused together, and then drawn out so as to produce a great length of much reduced section. Thus patterns of extreme delicacy were produced; and one single piece of construction could be cut across into a hundred slices, each repeating the whole design. The patterns are sometimes purely Egyptian, as _ankh_ and _uas_ alternately, but more usually Roman, such as heads and flower patterns. Such mosaics were mounted in jewellery, or, on a coarser scale, set in large designs for caskets and temple furniture. The characteristic of Roman times is the use of blown glass. The cups, bottles, and vases were nearly all blown, often with threads woven around, dabs attached and impressed, or patterns stamped while soft. The feet of cups were modelled into form while pasty, the tool marks showing plainly upon them. Ornamental stamps were pressed on soft lumps put on the sides of vases. Such stamps became used for official marks, and in early Arab times they registered the substance for which the glass measure was intended, also the amount of the capacity, and the maker’s name in many cases. Another main development of Byzantine and Arab glass was for weights, usually to test gold and silver coins, but also for larger amounts up to a pound. These weights bear the stamps of the Byzantine epochs in a few cases, but are found by the hundred of the VIIIth to Xth centuries, and by the thousand of the Xth to XIth centuries, dying out at the early crusading age. We now turn to the purely technical side, to describe the process of manufacture in the time of Amenhotep IV, about 1370 B.C., when it is best known to us, from the remains of the factory at Tell-el-Amarna. A clear glass could be produced, which was usually not quite colourless, but sufficiently so to take up various colours. It was free of lead and borates, and consisted of pure silica from crushed quartz pebbles, and alkali doubtless from wood ashes. It was fused in pans of earthenware. This glass was coloured by dissolving the blue or green frit in it, or mixing other opaque colours. Samples were taken out by pincers to test the colour at different stages. The whole mass was fairly fused, and then left to get cold in the earthen pan, which was about four or five inches across, and held half an inch to an inch deep of the glass. When cold the pan was chipped away, the frothy top of the glass was chipped off, and lumps of pure glass were obtained free from sediment and scum. A lump of glass thus purified was heated to a pasty state, and patted into a cylindrical form, then rolled under a bar of metal, which was run diagonally across it, until it was reduced to a rod about the size of a lead pencil, or rather less. Such a rod was then heated, and drawn out into “cane” about ⅛ inch thick. Every vase was built up from such cane.