454 RETURN TO THE MIRAN SITE [Chap. XII
Ilek, and that a route which required guarding must have passed along it. The importance of this fact must be emphasized in view of the much-discussed theory, first set forth by Dr. Heclin, which assumes that the whole of the Tarim drainage passed at that period through the Kuruk-darya beds into an earlier Lop-nor lying due south of the Lou-lan Site, and to which we have repeatedly had to refer above.' It would be impossible to reconcile with this theory the detailed account which Li Tao-ytian's commentary gives us of the Tarim's junction with the Charchan River and its subsequent course eastwards into the lake north of Yii-ni,8 and the definite chronological evidence afforded by the ruined fort of Merdek-tim saves us from making the attempt.
It proves that while the Lou-lan Site was still occupied and the Kuruk-darya still ' in being', a branch of the Tarim ran southwards and met the Charchan River, as it does at the present time, before passing into that portion of the ancient Lop lake-bed which now holds the Kara-koshun marshes north of the 1\liran tract. Where the point of junction then lay, what relative volumes of water this southern branch of the Tarim and the Kuruk-darya then carried, and similar questions, we cannot hope ever to be able to answer, considering the scantiness of our records and archaeological data as well as the physical conditions of a flat, deltaic region where water, wind, and sand are ever at work effacing surface evidence. One inference, however, seems to me necessary : if at the period preceding the abandonment of the Lou-lan Site, to which Li Tao-ytian's information undoubtedly goes back, the drainage of the Tarim Basin could fill both the Kuruk-darya and a southern river branch, whereas at present it certainly would not suffice for both, we can scarcely go wrong in assuming that the intervening period of roughly fifteen centuries must have seen a diminution of the total water-supply, i.e. desiccation, whatever its cause and its progress.
On January 7 I left Merdek-tim in order to move south again by the Tarim towards Charkhlik and Miran. Anxious as I was to start excavations at the latter place in full earnest, I could not forgo the chance of paying a visit era route to the ruined site which Captain Roborovsky's survey, as embodied in the Asiatic Trans-frontier Map of the Russian 'Topographical Department,' marked under the name of ' Ketek-sher ', i.e. Kötek-shahri, near the terminal course of the Charchandarya.10 The journey of six days which brought me to it has been described in my personal narrative." There is the less need here for details regarding it because its first portion lay along a route which keeps close to the main course of the Tarim, and is well known as it connects Charkhlik with Tikkenlik, Korla, and the other oases northward. The luxuriant growth of wild poplars, forming natural avenues, which we passed by the main bed of the river as far as Shirghechapkan, seemed like a vision of that distant age when the lines of fallen dead Toghraks we had so often crossed still flourished by the branches of the Kuruk-darya.
Doubly striking in contrast with it was the bareness of the steppe, covered with scanty reeds and tamarisks, over which the road led to the desolate station of Lop. Near it, in a network of lagoons and marshes, the present terminal course of the Charchan River loses itself, to effect its junction with the waters of the Tarim. Thus surrounded, the fishermen's hamlet known as Loft seemed fit to represent all the typical features of this dreary Lop region. As far as I could gather, the restricted use of the ancient local name Lop for this particular place is not of recent date. Perhaps it may in some way be connected with the fact that this otherwise insignificant locality is
7 See above, pp. 359, 361 sq.
S Cf. above, pp. 326 sq.
9 See Sheet xxi, scale 4o versts to I inch.
70 The term Kdlek-shaliri, literally meaning ' the town of dead trees ', is widely applied in the Tarim Basin to sites in and around the great drift-sand desert where remains, real or
imaginary, of ancient settlements exist. In the form [Shahri-]Kaiak it is found already in the Tarikh-i-Rashi&; see above, p. 32o, note 12.
" See Deserl Calha r, i. pp. 4 2 7 sqq. ; also Map No. 57 for the route followed.