1254 RU INED SITES EAST AND NORTH OF KHOTAN [Chap. XXXI
Museum.11 They proved to contain a considerable portion of the Central-Asian version of the Saddharmapundarika, and have since been published by Professor de la Vallée Poussin.12 He recognizes a close resemblance in palaeographic features between this manuscript and the leaves of the Vajracchedikâ text which was recovered from a monastic dwelling of Dandân-oilik and which Dr. I-Ioernle considers as of the seventh or eighth century.13
This dating, conjectural as it must be in view of the many uncertainties still besetting the history of Brâhmi writing in Central Asia, would have an important bearing on the chronology of the site—if we were justified in a priori assuming that the shrine F. xi' dated from the same period as the other remains of Farhâd-Beg-yailaki. But grave doubts against this assumption are raised by the position of the shrine, which, in spite of its close vicinity, essentially differs from that of the other ruins. While these are found invariably on wind-eroded terraces or else on flat ground which has the same level as such terraces situated close by, the shrine F. xii is built on a tamarisk-cone the top of which at the time of construction must have risen at least 20 feet, if not more, above the general ground-level of the site. About the peculiar nature of the position thus chosen there can be no possible doubt,14 and it alone suffices to set this ruin F. xi' quite apart from the rest. I know no parallel for it except that of modern Mazars with adjoining Muhammadan burial-places in-the Khotan region, which are frequently found placed on high tamarisk-cones or other conspicuous eminences on the edge of the desert.
It cannot be mere chance that quite a number of these places of modern local worship, often marked only by a collection of high poles bedecked with votive rags, Yak tails, etc., are to be found close to old sites. Thus, in the very vicinity of the site with which we are concerned here, we have the Ulûgh-mazâr and the Mazâr of Lachin-atâ connected with the remains of ancient Pi-mo." In the midst of the ` Tatis ' of Farhacl-Beg-yailaki itself there rises such a modest Mazar (see Plate 56 between iv and v) said to have been visited from ` Old Domoko ', while further away it will suffice to mention the much-frequented Mazars of Imam Ja`far Sadiq and Imam `Asim in positions exactly corresponding with reference to the Niya Site and the Ak-sipil remains respectively.10 The analogy thus presented and what we know in the Khotan region about all the essential features of local worship reaching back far into Buddhist times are bound to raise the surmise that we may have in F. x11, not a ruin contemporary with the other remains of Farhad-Beg-yailaki, but a shrine raised in this ` Mazâr' fashion near the old site after it had been abandoned to the desert, and thus, perhaps, centuries later.
In view of this possibility it will be safest, when examining the question as to the approximate date of the site, to begin by considering only those chronological indications which the finds made at the site in general and apart from F. xii furnish. A brief review will show, I believe, that these indications are in close accord among themselves, and that they all point to a date of abandonment distinctly earlier than the period of Tang dominion in the Tarim Basin. In the first place importance may be claimed for the evidence of the coins. Apart from the Wu-chu coin unearthed on the floor of the temple cella F. ni, eight other Chinese coins were found at the site, all on the surface of wind-eroded ground, four of them near the ruined house I, three near the monastic dwelling II, and
11 For specimens, see PI. CXLIV, F. xii. 7.
12 Cf. L. de la Vallée Poussin, J.R.A.S., 1911, pp. 1067 sqq.; see also Hoernle, Appendix F.
12 Cf. Ancient Kholan, i, p. 295, Dr. Hoernle's note vi, on Nos. '4, 15, 17, 18. Prof. de la Vallée Poussin, J.R.A.S., 191 I, pp. îo68, nevertheless calls attention to certain characters which seem more archaic.
1' The unusual thickness of the walls in F. xii, wholly out
of proportion to the size of the structure, is directly due to this position, the loose sand (reek loess dust) accumulated round the tamarisk growth in such cones affording no secure foundation.
15 Cf. Ancien/Khotan, i. pp. 461 sqq.; Maps Nos. 27. n. 4; 31. A. 4.
16 For Imam `Asim Mazâr, see Map No. 27. A. 4 ; for Imam Ja`far Sadiq, cf. Ancient Kholan, i. p. 313.