formerly inhabited by those very Tu-huo-lo whose extensive territories on the Middle Oxus the
pilgrim had previously visited and described.
There is nothing in the words used by Hsüan-tsang to imply that the name Tu-huo-lo is
recorded by him as a quasi-antiquarian designation of the tract. On the contrary, the analogy of the two remaining territories mentioned by him on his progress eastwards implies that he is here giving a name he actually heard from his local guides and informants. After marching six hundred li eastwards, the Hsi-yu-chi tells us, ` he arrived at the old Chê^^-mo-(o-na country (kingdom), which is precisely the territory of Chii-mo. The city walls are very lofty, but there are no inhabitants. Thence he went north-east for about a thousand li and reached the old Na fic po country (kingdom), which is the same as the territory of Lou-lan.' 4 The two ` kingdoms ' mentioned here correspond beyond all doubt to the present tracts of Charchan and Lop (south of Lop-nor), as I shall have occasion to show further on.r In each case the expression ` the old Chê-mo-t'o-na (Na-fu-po) country ' is followed by the ancient designations of these territories, as recorded for a much earlier period in the Former Han Annals.
This analogy makes it clear that in speaking of ` the old Tu-huo-lo country ' in the case of the Endere tract Hsüan-tsang does not wish to convey any special antiquarian information, but only to reproduce the name by which he heard the deserted site designated at the time. The fact that he does not quote here an earlier historical name for the locality is significant. It proves that Hsüantsang could trace as little as we can any special mention of the Endere tract in the Chinese historical records of the periods preceding his own. It is very probable that by using the expression ` old country (kingdom) ' Hsüan-tsang merely wishes to indicate that none of these three territories between Khotan and Lop-nor had any longer a chief of its own. It is a question which I must leave to a Sinologist definitely to answer. But what appears to me certain is that, as in the case of the Chê-mo-t`o-na (Charchan) and Na-fu-po (Lop), the name Tu-huo-lo was one taken by Hsüantsang from current local use.
In the absence of other historical references to the Endere Site any opinion as to the origin and character of the designation heard by Hstian-tsang must remain purely conjectural. But with this necessary reservation I may give expression here to a surmise which repeated personal experience under conditions rather similar to those of Hsüan-tsang's own passage over this ground has suggested to me. Again and again throughout the Tarim Basin I have heard ruins of all sorts spoken of, not by particular local names which may or may not be attached to them, but by terms which vaguely associated them with former invaders or rulers of the country. Thus general designations, such as Kalmak oilar, ` the Kalmaks' houses ', Kôiie-Khitai shawl,' township of the Old Khitais (Chinese)', Kône-Khitai tam, `walls of the Old Chinese ', are as common for ancient remains of any period as the ` Kafir-kots' and ` Kafir-kilas ' on the Indian North-west Frontier. Popular historical tradition being everywhere in Central Asia very limited in its range, it is only natural that such designations should be borrowed from the latest races whose rule over the country is still remembered.6 Is it not possible that an answer exactly similar in character to the one which any modern traveller might receive to his question about a ruined site passed en route by the desert edge was the basis for the name Tu-huo-lo as recorded by Hsüan-tsang at Endere ?
We know from Chinese historical records and Sung Yün's itinerary that the dominion of the Yeh-ta or Hephthalites, which lasted from about the middle of the fifth century down to
4 See Julien, 1116moires, ii. p. 247 ; Gutters, Yuan known as Eleuths or Zungars, whose dominion preceded the
Chwang, ii. 304 ; also Ancient Kho.an, i. p. 435. last Chinese conquest in the middle of the eighteenth century ;
6 See below, pp. 295-6, 321-2. cf. Elias and Ross, Târikh-i-Rashidi, p. 97.
By the Kalmaks are meant the Oirats, subsequently