or lavender turquoise under slightly varying conditions of firing, and according to the presence or absence of an infinitesimal proportion of iron or copper oxide in the body or glaze. Even with their empirical methods the old Chinese potters must have soon discovered the conditions which favoured certain results, but in the meantime quite a number of apparently different wares would have emerged from the same factory, and yet, in spite of local peculiarities, a general relationship might be observed in productions of different districts. So that when one Chinese writer compares the Ju ware to the Ch´ai, another the Kuan to the Ju, another the Ko to the Kuan, and another the Lung-ch´üan to the Ko, it is not necessary to assume that these porcelains were all grass-green celadons because we happen to know that that colour was the prevailing tint of the Lung-ch´üan ware. The Ch´ai and the Lung-ch´üan may have been as far apart as lavender and celadon green, and the chain of relationship linked up by the Chinese writers still hold firm. [Illustration: Fig. 1 Fig. 2 Fig. 3 Plate 15.--Sung Wares. Fig. 1.--Peach-shaped Water Vessel, dark-coloured biscuit, smooth greenish grey glaze. (?) Ju or Kuan ware. Length 5¼ inches. _Eumorfopoulos Collection._ Figs. 2 and 3.--Shallow Cup with flanged handle, and covered box, opalescent grey glaze. Kuan or Chün wares. Length of cup 7½ inches. Diameter of box 6 1/16 inches. _Rothenstein Collection._] No one but an experienced potter can speak with confidence of the methods by which the varying colour effects in the Sung glazes were obtained, but it is quite certain that the Sung potters were not ignorant of the value of such colouring agents as the oxides of iron, copper, cobalt, and perhaps even of antimony. Green, blue, yellow, and brown glazes, which owed their tint to these minerals, had appeared some centuries before on the T´ang wares. But to what extent the men of Sung made deliberate use of these oxides is another question. It is certain, for instance, that the green celadon owed its colour to the presence of iron oxide, but whether that was a natural element in the clay of certain districts, or whether it was introduced in the glaze by the admixture of ferruginous clay, is not always clear. Again, those bursts of contrasting colour, usually red, which enrich the opalescent grey and lavender glazes, are most readily explained by the local presence of copper or iron oxide in an appreciable quantity. No doubt these effects were at first accidental, but it is certain that observation and experiment eventually taught the potters to produce them systematically. Otherwise, how explain the appearance of these colours in symmetrical splashes? The _flambé_ glazes of the eighteenth century are known to have been produced by means of copper oxide, and it is not unreasonable to infer its presence in similar effects at an early date. But it is equally certain that many of the changing