which occupies the centre; in other examples the decoration is engraved in the paste. In either case, whether moulded or engraved, the glaze accumulating in the hollows helps to accentuate the pattern. The paste as seen through the glaze where the latter is thin appears white, but where the glaze is absent, as on the foot, or where it is exposed by bubbles or other irregularities, the ground is seen to be of a peculiar reddish tint. By this test the Chinese claim to distinguish the older celadon, the true _martabani_, from the later imitations made at King-te-chen. The paste of these later copies is often artificially coloured on the exposed surface so that they may resemble the old ware (Hirth, _Ancient Porcelain_, pp. 21 _seq._). As for the KO YAO, the old ware of Sung times is said to have been first made in the twelfth century. The Chinese character with which ‘Ko’ is written means ‘elder brother.’ According to the books there were at this time at Lung-chuan two brother potters named Chang. The elder brother leaving the younger Chang to continue in the old ways, started to make a new ware distinguished by the crackle of its glaze. This was originally a thick, heavy ware, with the iron-red foot and white paste already noticed, but, as we have said, the name is now used for a large class of crackle ware with a glaze of celadon, of greyish white and especially of a yellowish stone colour. This porcelain with grey and yellowish crackle does not seem to have been so largely exported as the uncrackled celadon; bowls and jars of a similar ware have, however, been found in Borneo and in the adjacent islands.