444 THE ETSIN-GOL DELTA AND THE RUINS OF KHARA-KHOTO [Chap. XIII
by a:cutting., still stood to a height of close on 20 feet, including its. podium. Here and at a group of six. small Stûpas . placed close together on one platform to the so.uth,. as roughly shown in the plan, Pl. 18, the finds were restricted to quantities of miniature votive 'Stûpas, such as have already been described, and of small clay tablets, showing in relief Buddha seated within a trefoil halo on a lotus seat and flanked by a Caitya representation on each side. Specimens of two slightly varying types are described in the List, under K.K. v. 031-4, 049-52, &c. (Pl. LIII). These tablets, too, had been made from similar moulds and deposited, together with the model Stilpas, in the cavity round the central shaft of the Stûpas. Similar small votive reliefs had been found by me in 1907 at Wan-fo-hsia and the Turfân site of Sassik-bulak.1
But a more interesting discovery was made on clearing away the debris at the foot. of the three small badly decayed Stûpas (marked K.K. v. b in the plan, Pl. 18) which form a separate little group to the south of the larger Stûpa. Here a careful search brought to light packet after packet of well-preserved leaves from different Hsi-hsia texts, mostly.. written but some also block-printed (see Pl. CXXXVI, CXXXVII), and from large Tibetan Pôthis (Pl. CXXXI—CXXXIII). They all lay embedded in sand mixed with debris of bricks from the fallen masonry of the Stûpas. I noted that several intact convolutes contained folded leaves from different Pôthis and books. This suggested that detâched leaves had been originally deposited . at the Stûpa bases as votive offerings, after the fashion I had often observed at sites. such as Dandàn-oilik, Khadalik, Endere,2 and subsequently heaped up by the wind in sheltered places where they had been preserved first by accumulating sand and then by debris. A rough calculation made at the time of packing showed that the total of complete leaves with Hsi-hsia text exceeded a hundred, that of Tibetan leaves being about half that number, with a great quantity of fragments. Mixed up with these manuscript and block-print remains were some drawings and diagrams, also a piece of blue silk painted with a lotus pattern, K.K. v. b. oi, probably the remains of a votive banner.
Finds of an exactly similar character were made at a mound that marked a completely collapsed Stûpa, K.K. III, situated less than a hundred yards from the north-eastern corner of the town walls. The mound rose only to about io.feet above the gravel Sai, but still retained remains of the central wooden shaft. The slopes were covered with small votive Stûpas in clay, and on clearing the north foot of the base, numerous leaves, written and block-printed, in Hsi-hsia and Chinese, were brought to light. Tibetan writing was here rare and was found almost exclusively on the reverse of Hsi-hsia and Chinese papers. The finds included a small Chinese printed book. Fragments of painted silk banners, K. K. III. oi, 02-5, well drawn but faded by exposure, were also recovered here, besides numerous pieces of silk fabrics which probably also formed part of votive banners. Among other objects may be mentioned the clay mould of a rectangular pläque, K.K. III. 013 (Pl. LV), showing a seated Buddha in good modelling.
A structure quite different in type from these Stûpas and of far greater interest was the ruin, K.K. 11, which was pointed out to me on my arrival at the site as the place where Colonel Kozlov in 1908 had secured his great haul of manuscripts, paintings and other antiques.3 It was situated
1 Cf. Serindia, iii. pp. 1112, 1170 ; iv. Pl. CXXXIX.
2 See Anc. Khota;:, i. pp. 265, 274,425 ; Ser., i. pp. 155 sqq•
3 See Geogr. Journal, Sept. 19ro, pp. 306 sq. The account of this ruin and its contents, as given in Colonel Kozlov's paper, runs thus :
` Meanwhile investigations were undertaken to discover and concentrate our energies on new excavations, with the result that we turned to the tomb situated outside the fortress,
about Soo yards from the western wall, and on the bank of the dry channel. This tomb was called by us " the Grand ", and henceforth occupied all our time and attention. It yielded a large collection of books, rolls, manuscripts, and a quantity of specimens of Buddhist paintings executed in colours on thick linen cloth, on thin silken material, and on paper. Among a quantity of books and paintings, piled up in the most complete confusion, very interesting statuettes