As already mentioned, many fragments of pottery were included in the important finds made by Sir Aurel Stein in his excavations in Turfan. Unfortunately, many of the sites have little evidential value, because they have clearly been revisited at comparatively late periods; but there are a few localities which ceased to be inhabited as early as the Sung dynasty, and which furnished fragments of glazed pottery and porcellanous wares. I only mention those sites which, as far as these finds are concerned, were not vitiated by the occurrence of obviously recent wares. On one site named Ushaktal, supposed to have been abandoned in the Sung dynasty, if not before, were fragments of greenish brown celadon with combed ornament on the body, such as was certainly made in Corea and probably in China as well. The same site produced opalescent glazes of the Chün and Yüan type. The site of Vash-shahri, which was "occupied probably down to the eleventh or twelfth century," produced a number of interesting fragments (1) with buff and grey stoneware bodies and glazes of the opalescent Chün and Yüan kinds, (2) the same body with emerald green crackled glaze, (3) celadon glazes over carved ornament, (4) speckled olive brown glaze resembling the later "tea dust," (5) opaque dark brown glaze, (6) speckled dark purplish brown glazes, (7) thick greenish glaze evenly dappled with pale bluish grey spots. Early wares found on the mixed sites, such as Kan Chou and Hsi Yung ch´êng, which were occupied down to Sung times but evidently visited later, include carved white porcelain and creamy white ware of the _t´u Ting_ class, and several kinds of _Tz´ŭ Chou_ wares, the _graffiato_, as well as the black painted. But evidence from excavations of this kind is always open to the objection that the ruins may have been visited later, and the broken pottery dropped by subsequent explorers. This objection, however, cannot reasonably be offered to more than a small proportion of the objects found, and these finds, though not in themselves conclusive, may be regarded, at any rate, as valuable corroboration of existing theories. FOOTNOTES: [Footnote 251: [Chinese] is an alternative form of [Chinese].]