Sec. iii] KERIYA, NIYA, AND IMAM JA`FAR SÀDIQ 311
the mighty snow-covered wall of the outer Kun-lun range to the south and the high dunes of the desert northward. The barren uniformity of stony ` Sai ' or low sand-dunes over which the route leads is broken only by the small oasis of Ui-toghrak, east of Keriya, and the tiny halfway settlement of Yes-yulghun s. At Niya I was glad to find myself once more on Hsüan-tsang's track, for there can be no doubt that the present small oasis corresponds to the pilgrim's town of Ni jang. He reached it after entering the desert east of 131-mo, which, as already noted, must be located near Uzun-Tati, north of the present oases of Chira and Gulakhma. The distance of 200 li between the two places is certainly under-estimated, even allowing for the reduced length of the more direct route north of the present one, which Hsüan-tsang is
likely to have followed. But apart from this measurement, his account of Ni fang, 0,
points quite clearly to the present Niya 4. ` This town is 3 or 4 li in circumference ; it is situated in a great marsh ; the soil of the marsh is warm and soft, so that it is difficult to march on it. It is covered with reeds and tangled herbage, and one sees neither roads nor tracks. There is only the route leading to the town which is practicable to some extent. On this account those coming and going must pass by this town. The king of Ch`ü-sa-ta-na [Khotan] has placed there the guard of his eastern frontier.'
The extensive marshes, which, as the map shows, are found on both sides of the Niya river just north of the present cultivated area and the main village of Niya, correspond exactly to the pilgrim's description. The route from the east winds through them for a distance of close on four miles, and westwards also the traveller would have to pass through them if he followed, as Hsüan-tsang in all probability did, a track a little more northerly than the present Niya—Keriya route. Swamps of less width, but equally thickly covered with reeds and tangled herbage ', accompany the river-course for about twelve miles further north, and between them numerous small lakes are passed. Neither the stony ` Sai ' commencing south of Niya nor the forest belt further north offers similar features, and it hence appears to me safe to conclude that Hsüan-tsang's town of Ni jang must have lain within the zone just indicated, though its location may not have been exactly the same as that of the present Niya villages. The latter, counting altogether about Soo houses 6, formed the easternmost of the smaller oases which were included in the Khotan District until the recent constitution of Keriya as a separate administrative unit extending to Charchan. Thus Niya may well be said to have retained to our own days the character of a frontier station, as noted by Hsüan-tsang. Ni jang, by which the Hsi yü-chi and the ` Life ' render the old name of Niya, has been supposed by Julien to reproduce a form approximately sounding *Ninya, because the character 0 in other transcriptions is used to represent the Indian nya, na, jna °. The supposition that the second syllable of the old name began with a nasal sound would be greatly strengthened if my suspicion should prove right that by Nina, a locality repeatedly mentioned in one of the Kharosthi documents from the Niya Site which Professor Rapson has published, our Niya is meant 7. The phonetic transition from a form *Nina into Niya could be accounted for by a parallel phonetic process in the case of Prâkrit languages 8.
' For a description see Ruins of Kholan, p. 340.
4 See Mémoires, ii. p. 246; Si yu-k:; transi. Beal, ii. p.324. a For a brief description comp. Hedin, Reisen in Z.-A., p. 215 ; also Ruins of Kholan, pp. 34 r sq.
e See Mémoires, ii. p. 517 ; Méthode pour transcrire, p. 115.
7 See Kharoshl tablet N. i. i 6 in Prof. Rapson's Specimens of Kharosthi Inscriptions, &c., p. i 4. This tablet, forming
together with N. i. 104 a double ` wedge ' document, refers to the dispatch of a certain Sameka as a messenger from Calmadana via Saca and Nina to Khotana ; comp. below, p. 326. I surmise that by Calmadana is meant the same locality which Hsüan-tsang calls Chih-mo-fo-na, and which, being placed ten marches to the east of Niya, manifestly corresponds to the present Charchan.
8 If Hsüan-tsang's So-mo-j1, discussed above, pp. 223 sqq.,