Gold was also used largely for gilding both metals and wood. The gold leaf was often about a 5000th of an inch thick, weighing one grain to the square inch. Thus a pound’s weight of gold would cover about six feet square; and the gilding of doors and of the caps of obelisks as described is not at all unlikely. Silver was known to the Egyptians later than gold, as it is called “white gold”; and it was scarcer than gold in the early ages. Of the prehistoric time there is a cap of a jar, and a small spoon with twisted handle. A few silver amulets are known in the XIIth dynasty. In the XVIIIth dynasty silver became commoner, as the source in northern Syria which supplied the Hittites became accessible. The silver dishes of this age are rather thick, and not finely beaten. One bowl, probably of Ramesside date from Bubastis, has the brim turned inward like a modern anti-splash basin (fig. 115). It seems to have been made by spinning the metal, as thin vessels are now wrought. [Illustration: JEWELLERY 108. Silver bowls 109. Roman gold chain] The most elaborate style of silver work is that of the bowls from Mendes (fig. 108). These are entirely made by hammer work, and no moulds or matrices were used for the forms. But the finish of the surfaces is so fine that no trace of hammer or polishing is left. The design is derived from the fluted vases and bowls of the XVIIIth dynasty; the fluting was made deeper and stronger, and it was suppressed below, as it interfered with the using of the bowl, while round the sides it remained as deep bosses. The detail was all put in by the graving tool, the sinking round the central rosette, the hollows in the petals, and the outlines of the petals. There is no sign of punch-work. The number of ribs is, curiously, indivisible, being 18, 26, 28, and 30; these show that it was not divided either by triangles, hexagons, or repeated halving. Probably a suitable size of rib was designed, and then repeated an even number of times; and the divisions not being truly radial, show that eye-design was followed rather than geometrical scaling. CHAPTER IX METAL WORK