162 HISTORICAL NOTICES OF KHOTAN [Chap. VII
north-east where it leaves the present limits of the oasis, and this trend must have been even more pronounced at some earlier period, if the broad ` Sai ', which is passed in the desert area between the sites of Ak-sipil and Rawak and which I traced northward as far as Jumbe-Kum, marks indeed, as I believe it does, an ancient bed of the river 19. The extensive ` Tati ' around the ruin of Arka-kuduk-Tim lies just north of Hanguya, and must have formed an important site of this tract in pre-Muhammadan times 20. A look at my map shows that, assuming the Yurungkâsh to have followed during this period the more easterly course just referred to, the description of Hanguya as being south of the Khotan river would not have been altogether inappropriate.
The territorial division which the legend indicates between the settlements of Kustana's Chinese followers and those of the Indian exiles seems rather fantastic at first glance. Yet a réference to the real topography of Khotan helps to clear up the meaning. Though unable myself to consult the original text, I may hazard the conjecture that by ` the lower side of the U-then river ' and ` the upper bank of the river ' (shel-tchu gong-ma) which the legend distinguishes, the two branches of the Khotan Darya, viz. the Yurung-kash and the Kara-kash, are really intended. Whatever the equivalent Khotanese terms may have been which the Tibetan text is evidently endeavouring to reproduce with the literalness usual among Tibetan translators, such an interpretation is clearly suggested by the subsequent mention of the space ` between the two', scil. rivers (shel-tchu dâus), in which the followers of Kustana and of Yaéas are said to have settled indiscriminately. As the former were supposed to have come from China, i. e. the east, it was natural that tradition should have assigned to them seats on the eastern branch of the Khotan river, i. e. the Yurung-kash, while for the Indians, whose approach to Khotan could only be assumed to have lain from the west (Karakorum route), the land west of the Kara-lash would appear equally appropriate. In short, I am inclined to believe that the story about the threefold division of the territory is directly based on the geographical fact of the Khotan oasis being separated by the two rivers which traverse it into three main portions, viz. the one east of the Yurung-kash, the one west of the Kara-kash, and the interfluvial tract 21.
On comparing the Tibetan legend with the story as heard and recorded by Hsüan-tsang, we cannot fail to be struck by the close agreement they show in those points which for us are of essential interest. Both versions ascribe the origin of the Khotan kingdom to a joint settlement of colonies drawn from China on the one side and India on the other ; both make the chief of the eastern colony assume the royal power over the amalgamated population ; and in both we find the events leading to the foundation of the kingdom connected with the time of Moka. The only important difference between the two versions concerns just that part of the tradition in which we have already recognized an element of pure folklore, viz., the eponymous legend of Kustana. Whether the form which the Tibetan version gives to this part of the legend was known in the time of Hsüan-tsang or not, must remain uncertain. But on internal grounds it can scarcely be doubted that it represents a later development ; for it is evident that the name Kustana must have been invented in the first place as a ` popular
19 This ` Sai', with its streak of scrub, is shown on the map and was followed by me southwards to near the village of Suya. It is quite possible that the old river-bed which the jade pits of Kalta-Kumat, near Tam-öghil village, indicate (see Hedin, Reisen in Z.A., p. 28), has its continuation further north in this ` Sai'.
70 Regarding this large ruined area north of Hanguya, see below, chap. xiv.
2' It deserves to be noted that the GoSrnga-vyâkarana, which gives the story of the first settlement of Khotan in a briefer form, refers to the Indian immigrants as being divided from Kustana's Chinese subjects by a stream ; see Rockhill, Life of the Buddha, p. 234 note. For extracts of this text see Mr. Thomas' App. E. The same should be consulted also for a different version of the main passage which distinguishes a lower' and an upper' river.]