Sec. iv] THE TOGA-DONG MOUNDS AND THE KERIYA—KHOTAN ROUTE 467
the Bazars by treasure-seekers'. But only three small objects of uncertain origin had been secured (see below Ker. 001—o03). Among them was the small intaglio, Ker. 003 (Plate X LI X), showing the figures of a man and a running deer, and the bronze finial, Ker. ooi (see Plate LI), which in its decorative motif so closely resembles the styles used in Indian sürmaclâns of the present day as to suggest doubts as to its antiquity. The scarcity of these acquisitions seemed to me an indication that the Tatis, which I had heard reported at Kara-khan, to the northwest of Keriya town, and at Tokhtal, near the western edge of the oasis, and to the south of the Khotan road, could not be of great extent or importance. For visits to them I was unable to spare time.
Keriya town itself is not believed by the people to be a place of great age, and has probably acquired importance only since it was made the head quarters of the separate district of Keriya, newly created after Yaqûb Beg's rebellion, and now officially styled Yü-t`ien. The oasis itself is, no doubt, of ancient date ; but there is nothing to indicate that it was in old times larger or more important than the string of oases which still extend in an almost unbroken line from Chira to Kara-kir, and are linked with Keriya by ground that, as the subsequent observations will- show, is for a large part capable of cultivation. I am unable to trace any distinct reference to Keriya in the early Chinese texts accessible to me in translations.
The territory of Yü-mi , which is mentioned by the Chien Han shu among the small
states to the east of Khotan, and subsequently referred to under the slightly different and probably more correct form of Han-mi if •' in the Wei lio and Tang Annals, has indeed been identified by a modern Chinese writer with Keriya'. But the topographical indications furnished by the historical texts make it appear far more probable that this territory comprised the whole
t of the oases between Chira and Keriya, thus corresponding roughly to Marco Polo's ` Province
of Pein ', and the tract of which Hsüan-tsang's P`i-mo was the chief place. In the Han Annals'
i notice Yii-mi, then independent of Khotan, is described as being 390 li to the east of the
latter 2. This, assuming that as elsewhere the measurement is taken from capital to capital, points to a location about Gulakhma-Domoko rather than near the town of Keriya. The "1'`ang
i Annals, which, as we have seen, enumerate Han-mi among the small territories annexed to
Yü-t`ien 3, place it to the east of the river of Chien-11-114. The latter itself, being placed 300 li east of Yti-tien, can only be the river of Chira, the distance of which from Yatkan, about 6o miles, exactly corresponds to the above measurement. We are not told by the notice of the Tang Annals where in particular was the position of the town of Ta-11--li, its chief place, also known as Chü-mi, or earlier as Ning mi. But seeing what we have ascertained above as to the long-continued importance of the town which Sung Yün called Han-mo and Hsüan-tsang 71-mo, the location of Ta-té-li at the same site seems distinctly more probable than at the modern Keriya.
On April 2 I started back to Khotan by forced marches. The first, which brought me to Kara-kir, gave me ample opportunity to observe the abundance of spring water which feeds the numerous marshes skirted by the road between the western edge of the Keriya oasis and the Kara-kir stream, and which further on runs to waste in the Shivul Darya and its terminal swamps.
0 My companions from Keriya, among them the local official whom the attentive Amban had
deputed to escort me, declared that the springs were perennial, and believed that with the