Hindu Architecture in Central Asia and Khotan

HINDU ARCHITECTURE IN SERINDIA OR CENTRAL ASIA—IN ADDITION TO numerous literary references and travellers' accounts, archaelogical evidences, especially architectural and sculptural remains, have shown beyond doubt that Indian civilization marched further into Central Asia. 6 The area extends from the Hindukush valleys and the uppermost Oxus right across the whole length of the Tarim Basin to the province of Kan-su on the western marches of China proper." This region covers practically the whole of that vast drainageless belt between the Pamirs in the west and the Pacific watershed in the east, which, for close on a thousand years, formed the special meeting-ground of Chinese civilization, introduced by trade and political penetration, and of Indian culture propagated by Buddhism.' This desert region contains 'relics of that ancient civilization which the joint influences of Buddhist India, China, and Hellenized Near East had fostered in the scattered oases of those remote Central Asian passage lands.' KHOTAN The Chinese records make it beyond doubt that Buddhism established in Eastern (Chinese) Turkestan, Khotan, Kasghar, oasis of Yarkand, and Karghalik, had been imported, directly or indirectly, from India. The migration of Indian culture in Central Asia was further corroborated by various archaeological evidences, including the discovery of the famous birch-bark codex by Colonel Bower from Kucha, the find of birch-bark leaves (containing a Buddhist text in early Prakrit language and in Kharoshthi script) by M. Dutreuil de Rhins at Khotan, and the Sanskrit texts (in Brahmi script) discovered in the ruined shrines and monastic dwellings of Dandan-Uiliq and Endere by Sir M. Aurel Stein, pointing to the study of Sanskrit Buddhist literature in Khotan down to the end of the eighth century A.D. The Central Asian explorations of Aurel Stein have supplied ample materials to identify in this region the Indian architectural and 1 This region has been briefly designated by the French scholars as Serindia', depending on 'Learned popular ctymology.' Sir Aurel Stein, Introduction, Serindia, Vol. I. 294

sculptural objects, including numerous articles of furniture. These are predominantly of Buddhistic style, but the purely Brahmanical names of deities and temples are not altogether missing. The village schemes and town-planning also appear to have been of Hindu origin in their general plan and other essential features. The differences may be easily accounted for by the climate, sandy desert, soil condition, and other local causes. That the territory of Khotan had been under Chinese supremacy for a long time during the Han and the Tang dynasties, and that the influence of Chinese civilization strongly asserted itself, are proved beyond doubt by the finds of Chinese records in wood and paper, Chinese graffiti, coins, articles of industrial art, etc. The finds of Tibetan MSS. and graffiti in the ruined fort and temple of Endere would also show the Tibetan influences. Despite these varied foreign influences at work, Eastern Turkestan, including Khotan, 'possessed during the pre-Muhammadan epoch a well-defined civilization of its own.' The religious predeliction and possibly the superiority of the Indian style may, however, have induced the migration of Hindu architecture in Central Indian colonies as in Indo-China and in Indonesian islands. The remains consist of numerous mounds, stupas, viharas, sanctuaries, shrines, temples, palaces, pavilions, forts, and unrecognizable dwellings. 6 The top of the mound (Kurghan-Tim) stands 70 feet above the ground level of the neighbouring fields, but it rises in reality to fully 85 feet above the lowest course of masonry at present traceable.' The whole mound consisted originally of sundried bricks of large size, laid in regular courses with thin layers of mud plaster to act as binding material.' 'The far advanced ruin of the whole structure makes it impossible to ascertain its original dimensions and constructive features. There can be no doubt as to the remains being those of a stupa, built with a remarkably large dome, possibly of hemispherical shape. The base was square and arranged in three storeys . . its greatest length measuring about 160 feet from cast to west, compared with about 130 feet from north to south."¹ • The Mauri-Tim Stupa bears similar features.2 There are some common architectural similarities between these and the ruined stupas 1 Ancient Khotan, by Sir M. Aurel Stein, Vol. I, p. 74, photo no. 13, Plate (Vol. II) xxii. 2 Ibid, Ct. Mauri-Tim Stupa (pp. 81, 82, photo I, Vol. II). 295

of Takhtaband in Buner and of Bahar near ancient Taxila,1 in the Manikyala Stupa, and in the numerous stupas of the Kabul valley.² It is clear that there exists a close agreement in regard to the general architectural arrangement between all Turkestan stupas and the corresponding structures in the Kabul valley and on the Indian north-west frontier. The development of stupa architecture in India, as elucidated by M. Foucher,³ will further show that the Turkestan stupas conform closely to an Indian type. 'The stupas of Mauri-Tim, Topa-Tim, Niya, Endere, and Rawak all show the dome, which is the essential and unvarying feature of every stupa, raised on a square base, and this is again arranged in three storeys. The relative proportion between these storeys varies considerably in the several structures, and so also do the shape of the dome and the relative height of the cylindrical portion or drum which intervenes between the top storey of the base and the copula proper. But the square shape of the base and its three-fold arrangement are constant features, and the former in any case is characteristic also of the great majority of stupas in the border lands of India and Afghanistan. On the other hand the round base, which belongs to an earlier stage of stupa construction, is represented in those territories only by a few examples, and seems completely absent in Eastern Turkestan.' 6 'The king's new monastery,' comprising the splendid stupa and temple of which Fa-hien has left a detailed description, were situated a little to the west of the Khotan capital. Seven or eight lea to the west of the city there is what is called the king's new monastery, the building of which took eighty years, and extended over three reigns. It may be 250 cubits in height, rich in elegant carving and inlaid work, covered above with gold and silver, and finished throughout with a combination of all the precious substances. Behind the tope there has been built a Hall of Buddha, of the utmost magnificence and beauty, the beams, pillars, venetianed doors and windows being all overlaid with leaf-gold. Besides these, the apartments for the 1 Stein, Report on Archaeological Tour with Buner Field Force, p. 40, Plate viii, Indian Antiquary, 1900, p. 145. 2 Cunningham, Archaeological Survey Report, Vol. V, Plate xxii. 3 Foucher, L'Art de Gandhara, I, pp. 62-98. Stein, Ancient Khotan, p. 83, note 12. 296 5 Ibid, p. 83.

monks are imposingly and elegantly decorated, beyond the power of words to express. "1 From the great height mentioned,' concludes Sir Aurel Stein, it is clear that the structure was of wood, like the famous Vaisravana temple in the capital itself."1 Stein identifics this great shrine with the So-mo-jc convent of Hiuen-Tsang and says that it must be located at the present village of Somiya, close to the west of Yot Kan." 12 ' Hicun-Tsang has left unusually full descriptions of various places of Buddhist worship outside the capital of Khotan. The nearest among these sanctuaries was the convent of So-mo-jc, with a stupa a hundred feet high in its centre a little over a mile to the west of the royal city. Hieun-Tsang ascribes its origin to a legend which really means that some architects were imported from India for its construction. The legend relates that: 'one time an Arhat coming from a distant foreign land had taken up his abode there in the middle of a wood. The miraculous light spread around by his spiritual power was noticed by the king as he stopped at night in a double-storeyed pavilion of his palace. Having been informed of its cause the king proceeded to the holy man and respectfully invited him into the palace. On the Sramana refusing to leave the wood, the king full of reverence built a convent for him and a stupa. When afterwards the king had procured a quantity of sacred relics and regretted not to have been able to insert them under the stupa, the Arhat directed him to have the precious objects enclosed successively in receptacles of gold, silver, copper and stone. When this had been done and the relics had been transported by the king and his chief officers on an ornamented car to the convent, the Arhat raised the stupa on the palm of his hand and held it while the king's workmen dug a place for the sacred deposit. Then, on the work being accomplished, the Arhat once more lowered the stupa to its original position without any damage.³ 1 Fa-hien's Travels, Translation, Legge, pp. 19 sq.; Stein, Ancient Khotan, I pp. 194. 2 Stein refers to Memoirs, II, pp. 235 sqq., Si-yu-ki, II, pp. 316 sq. : Remusat, Ville de Khotan, pp. 50 sqq., Watters, II, p. 297. 3 This legend, and a somewhat similar one told by Hiuen-Tsang of a stupa near Kapisa, (Memoirs, I, pp. 45 sq., Si-yu-ki, I, p. 60) have been discussed by M. Foucher, L'Art de Gandhara, I, p. 52, with reference to the light they throw on the purpose of stupa construction. Stein, ibid, p. 223. 297

The Tibetan form of the legend is slightly different. It is stated in the Annals of Li-yul that king Vijayavirya, the eighth successor of Vijaya-sambhava, under whom Buddhism was believed to have been first introduced, 'one day while looking out of srog-mukhar (life-fort) he perceived a light brilliant as gold and silver at the spot where now stands the Hgum-stir Chaitya. When the king learned that the Buddha had foretold that at that spot a vihara would be built, he called to his presence the Buddhist Buddhaduta, and having made him his spiritual adviser, ordered him to direct the building of the Hgum-stir Vihara.' Stein emphatically asserts that it refers to the same shrine and the Buddha-duta corresponds to Hiuen-Tsang's Arhata.1 The buildings of the ancient Khotan capital are referred to in all Chinese records, and comprise magnificent temples, palaces and pavilions of which, however, the architectural details are not available. In a passage of the Yini, composed by Hui-lin in the eighth or ninth century,2 reference is made to an important building. It shows that the famous temple of Vaisravana, which the legends related by Hiuen-Tsang and the Annals of Li-yul represent as the oldest shrine of the kingdom, stood in the city of Yu-tien, and that it was a tower built in wood seven storeys high. The god was supposed to reside on the summit. Hiuen-Tsang particularly extols the richness of the temple as it existed in his time.'³ C This building is apparently of Indian origin as it bears the name of an Indian deity. 6 'As regards the royal palace,' says Sir Aurel Stein,³ we learn from the account of the Chinese mission sent in 938 that it comprised a number of buildings all facing to the cast, and among them a pavilion called that of the seven phoenixes. An carlier notice, found in the Liang Annals, specially mentions the frescoes adorning the king's palace. 5 That the city was enclosed by walls we know from Hicun-Tsang. He adds that they were of no great height.. The ancient fortifications still existing at the sites of Endere, 1 Stein, Ancient Khotan, I, p. 223. 2 M. S. Levi, Notes Chinoises L'Inde, p. 39. 3 Stein, Ancient Khotan, I, p. 202. 4 Remusat, Ville de Khotan, p. 80, where the name of the palace is given as Chinise-tien. 5 Remusat, loc. cit., p. 16. 298

aradong, Ak-sipil show that we may safely assume those walls to have consisted mainly of ramparts of stamped loess. At the gate of the city there used to take place the solemn reception by the king and the ladies of his court of the procession of great image cars of which Fa-hien has left us so vivid a description."1 The Indian equivalent of the pavilion of seven phoenixes (Naga and Nagini?) and of the Chinese name, Chin-isc-tien, of the palace are not quite clear. But the castern orientation (facade), the fortified plan of the walled city with rampart around, and the procession of the great image cars are familiar objects of Hindu architecture described in great detail in the Manasara and other Silpa-sastras. ' In a convent known as the Ti-chia-po-fu-na Sangharama (Dirghabhavana) and situated a little over 10 lea to the south-west of the capital, the pilgrim (Hiuen-Tsang) was shown the statue of a standing Buddha which was supposed to have miraculously come to this spot from Kucha. A Khotan minister exiled to the latter state had secured his repatriation by assiduous worship of this image. The minister then built for it this convent.' Stein identifies it with the Ziarat of Bowa-kambar,' which he found to consist of a large square cemetery enclosing the high mud-built tomb of a saint. The level of the cemetery near its centre lics fully 12 feet below the surrounding fields. . . . A grove of fine old trees faces the eastern C entrance. 12 By far the most imposing structure among the extant ruins of the Khotan region is the Rawak Stupa and Vihara. The greater part of the stupa has gone, only its high base and enclosing quadrangle remaining. The vihara court formed a great rectangle, measuring 163 feet inside on its south-western and north-eastern faces, and 141 feet on its shorter sides towards the north-west and south-east. It was enclosed by a wall about 3 feet 6 inches thick, solidly built of sundried bricks. At the south corner . . this wall rose to a height of 11 feet, but was probably once still higher.³. The gate in the south-east face, 8 feet to 10 feet wide, was the only entrance to the court.' 1 Legge, Travels of Fa-hien, p. 19. • • 2 Stein, Ancient Khotan, 1, pp. 225-226. 3 The bricks used in it, as well as in the stupa, measured 20 by 14 inches with a thickness of 3 to 4 inches, practically identical with the dimensions of the Ak-sipil bricks. 299

CENTRAL ASIA • • HINDU ARCHITECTURE IN 'The centre of the quadrangle is occupied by the imposing stupa base which rises in three storeys to a height of 22 feet above the floors of the court. The lowest storey, 78 feet square and 7 feet high, rests on a plinth of four steps showing aggregate clevation of 3 feet. The second storey is 45 feet square, with a height of 9 feet. It is surmounted by a circular drum, 3 feet high and receding on the top, which serves as a plinth for another circular drum forming part of the stupa dome." As the masonry of the dome is intact to a height of about 8 feet, the exact elevation of the dome and the shape of the cupola cannot be determined. But the other indications will show that, like other monuments. it is also of the Indian style. This expectation is fulfilled by the other members of the stupa. * Considerable variation and originality was introduced into the ground-plan by a series of bold projections on each face of the base supporting well-proportioned flights of steps. Through these projections the ground-plan of the base has assumed the shape of a symmetrically developed cross, each of the four arms of which extends about 52 feet on the lowest level, as measured from the centre line of the stupa. The broad flights of steps which occupied the centre of each of the four faces of the base, and, carried by the projecting portions, led up without a break from the level of the court to the very foot of the dome must have been an imposing architectural feature."1 This obviously points to the Svastika plan of the Manas ara architecture. The extraordinary architectural features and imposing dimensions of Rewak monuments are quite in keeping with the rich series of relief sculptures decorating the walls of the enclosing vihara. The main adornment of the walls, both towards the court and outside, consisted throughout of rows of relief statues in stucco over life-size. All the large reliefs represented Buddhas or Bodhisattvas ; but from the varying poses, accessories, etc., still recognizable, a number of groups could be distinguished, arranged apparently with some regard to symmetrical disposition. Between the colossal statues at the frequent intervals were smaller representations of attendant divinities or saints. In numerous instances the walls were further decorated with elaborate plaques in stucco, forming halocs 1 Stein, Ancient Khotan, pp. 483, 484, 485. 300

above the heads of figures, or, in some cases, where sufficient space had been allowed, even with complete aureoles in relief around them. . . . The whole of the relief work had been originally coloured, but the layers of paint had in most cases peeled off except where well-protected in drapery folds, etc. Thus the greatest portions of the stucco images presented themselves in the red ground colour of the clay in which they have been modelled.'¹ " "The total number of individual reliefs, most of them over lifesize, amounted to ninety-one.' In the poses of these reliefs Sir Aurel Stein has identified the Abhaya-pani-mudra like in the Dandan-Uiliq and Kighillik reliefs, and the Vara-mudra and Bhumisparsa-mudra, differing only by the pose of the hand with the palm turned towards or inwards. There are images of colossal, and small Buddhas and Bodhisattvas, both standing and seated with wave-lined draperies and characteristic nimbus, vajras, and other features. There are dvarapalas and female busts and heads of Yakshas and Yakshinis. So far as the date of the Rawak remains is concerned, Sir Aurel Stein cautiously says that the fact of the Rawak sculptures approaching their Gandhara models much nearer than those of Dandan-Uiliq or Endere, and the total absence among them of any of those images with multiple limbs, etc., which characterize the later pantheon of Northern Mahayana Buddhism, may at once be accepted as a proof that they are older. But beyond this it would scarcely be safe to draw any further chronological conclusions from the evidence of the artistic remains themselves, seeing how scanty our data are for the chronology of the Graeco-Buddhist sculpture in Gandhara itself, and how little we know as yet of the historical development of its offshoot transplanted to Eastern Turkestan."³ The ruins of Endere comprise the remains of a fort together with the remains of a stupa and a temple. The fort is circular in shape, a variety described in the Manas ara Silpa-sastra. It is surrounded by a 30 feet wide rampart varying in height from 5 to 25 feet, with masonry parapet of 5 feet 6 inches height and 3 feet thickness at the entrance to the south. Behind the parapet ran a platform. The gate, 18 feet wide, was flanked on either side by a small square bastion or tower, projecting about 20 feet beyond the outer foot of the 1 Stein, Ancient Khotan, p. 486. 2 See photos nos. 61, 62, 63, 64, 65, 66, 67, 68; ibid, I, pp. 490-99. 3 Stein, ibid., I, p. 500. 301

• • • circular rampart. To the west of the gate there is a guard-room.1 Following the rules of the Silpa-sastra, the temple is situated just in the centre, which is technically known as the Brahmapitha. 'The cella of the temple formed a square of 18 feet 4 inches inside, enclosed by walls of timber and plaster having a thickness of 10 inches. The entrance lay to the cast. The enclosing passage, the walls of which were of similar construction was 5 feet wide, its plastered floor lay 3 feet higher than that of the cella.' 'It appears probable that, for purposes of lighting, the cella was provided with a raised roof in the centre." This would look like some form of Hindu sikhara. The four corners of the cella were occupied by plaster images, almost wholly detached and standing each on a basement to represent an open lotus with the petals pointing downwards. By the side of the statues in the northeast and south-east corners there were probably figures of lokapalas or guardians of regions.' 'The centre of the cella was occupied by a massive octagonal base or platform, 9 feet long and 7) feet broad. The base proper rose to a height of 2 feet 8 inches above the floor. On it four life-size relief statues in stucco must have stood, the lotus-shaped pedestals of three being still more or less intact.' * On the north-east facet the fresco remains showed two rows of seated male figures, seven in each, apparently Buddhas or Bodhisattvas." 6 The temple is surrounded by groups of small buildings to the north, east, south-west and north-west at a distance of some 50 feet. The purpose of all these buildings is not definitely established. Sir Aurel Stein surmizes one of these in the north to be a grain store. In the eastern group there is a half-open hall and a number of other rooms of various sizes and purposes. This is stated to be undoubtedly the main structure within the area enclosed by the ramparts, and judging from the size of its halls and rooms, and the general arrangement of its plan,' Sir Aurel Stein concludes that it had served as the residence of the officer and establishment, for the protection of whom the fort must be supposed to have been primarily intended."³ There appears to have been some dwelling-houses to the south-east also, which had distinctly an upper storey. 6 Sir Aurel Stein is doubtful if the whole structure could have been suggested by a Chinese Ya-men, because the disposition of the rooms, 1 See Plate xxxvi, Ancient Khotan, by Stein, Vol. II. 2 Ibid., I, p. 424. 3 Ibid., 1, p. 432. 302

etc., is not compatible with Chinese architectural convention.1 It would look more like an Indian fort with a public temple in the centre and with other features noted above. The Endere stupa, situated outside the fort towards the northwest, ' consisted of a square base, approximately oriented with its corners, and of a cylindrical dome rising above it. The base rose in three storeys according to the canonical arrangement previously explained, the lowest storey measuring 27 feet, with a height of 1 1/2 feet. The next storey, receding 2 feet from the first and 6 feet high, formed the main portion of the base; above, the third, only 1 feet high, also receded by 2 feet. The dome had a diameter of 16 feet, rising with its broken top to a height of 14 feet." The date or rather the period of the antiquities in Eastern Turkestan has been fixed by Sir Aurel Stein with some definiteness. For this purpose he has taken into consideration not only the remains of ancient town-planning, buildings, sculptures, industrial arts, but also the languages and scripts (Kharoshthi, Indian Brahmi, Central Asian Brahmi, Tibetan, Chinese, and Hebrew) of the numerous manuscripts (and inscriptions) he has explored. 6 • • • The remains of the Rawak Vihara enable us better than any other ruins . . . to realize what the plan and decorative aspect must have been of large Buddhist shrines in ancient Khotan. The affinity which the Rawak reliefs show in style and most details of execution with the Gracco-Buddhist sculptures of Gandhara is far closer than that we have had occasion to observe in the plastic remains brought to light elsewhere. These considerations make it all the more important to determine the period, if only within approximate limits, to which the ruined vihara belongs. The fact of the Rawak sculptures approaching their Gandhara models much nearer than those of Dandan-Uiliq or Endere, and the total absence among them of any of those images with multiple limbs, etc., which characterize the later pantheon of Northern Mahayana Buddhism, may at once be accepted as a proof that they are older. But beyond this it would scarcely be safe to draw any further chronological conclusion from the evidence of the artistic remains themselves, seeing how scanty our data are for the chronology of Graeco-Buddhist sculpture in Gandhara itself, and how little we know as yet of the historical development of its offshoot transplanted to Eastern Turkestan."3 1 Ancient Khotan, I, p. 432. 2 Ibid., I, p. 437. ³ Ibid., I, p. 500. 303

Although the Indian relics in Central Asia consist mostly of Buddhistic shrines, scanty remains of dwellings and forts are not altogether missing. In fact, in certain sites the existing dwellinghouses built according to old traditions are the only evidence of the past. "Outside it (old Kafir house at Dawawish) looked at first sight like a large heap of stones. But closer inspection showed walls far more solid than usual in these parts, built of uncut but well-set slabs of stone. The most striking feature inside was a large central room or hall, showing elaborate carving on its massive pillars and along one wall decorative panelling in deodar. . . . . . The ornamentation chiefly consisted of a diaper of four-petalled flowers, closely resembling in style those . . . from the ancient wood-carvings of the Niya site and from Gandhara relievos. . . . The square pillars supported large corbels ending in elaborate volutes. The opening in the centre of the roof (called aiwan or kumal in Khowar), which alone admits light and air, showed the typical construction. It consists of successive overlapping courses of massive beams gradually reducing the square opening.'1 • • • The remains of another dwelling-house have been found within cleven yards of a shrine. 'It consisted of a room built with walls of sundried bricks, and a small room and verandah adjoining it on the north. Both were constructed with timbers and plaster walls. The room, about 17 by 12 feet, still retained its substantially built clay fireplace, and in the corner beside it a plastered sitting platform. The roof was probably made of rafters and recd bundles." At Niya site the remains of several other dwelling-houses have been discovered. In one of these 'the main room showed on three sides a sitting platform, 4 feet broad and 15 inches high, faced with sundried bricks 17 by 12 by 3 inches. There was found here a roughly carved round central pillar and a plain double-bracket, 8 feet long and 7 inches wide, with ends curving upwards. In another room there was a wooden pillar, 5 feet high and oval in section, with a longer axis of 9 inches and the sides showing sixteen facets. The top, both in front and on the back, was decorated in rough bold chip-carving with a design showing a vase from which issued two long curving stems ending in broad leaves and fruit pendants'³. ¹ Serindia, I, pp. 35-36, Fig. 15, 16. • 2 Ibid., I, p. 199. 3 Ibid., p. 239. 304

[ * 4 PLATE XLV yzy A BUDDHIST STUPA OF CHIAO-TZU Pag 304

PLATE XLVI A CLOSE-UP VIEW OF THE CAVES OF THE THOUSand BuddhAS Page 305

he idea of the central pillar (griha-stambha), around and by the standard of which the whole house used to be built, is a peculiarity of Hindu architecture, which has been specially referred to in the Manasara and other Silpa-sastras. The round (oval) pillar with sixteen facets is also a peculiarity of Hindu architecture. The wooden pillars of Tang Fort, Endere site, are further examples of Hindu architecture. There is another larger house built partly in timber and plaster and partly with plastered rush walls. 'Room, i, to the north-east, served as an office and yielded sixteen well-preserved Kharoshthi tablets, mostly wedge-shaped or rectangular. The adjoining looks like a fireplace. Outside the large room, v, yet within the fenced enclosure of a court or garden, lay the remains of long twisted vine branches. The cattle-shed is on the north-east.'³ uncut stones ' • • The remains of the old forts are scanty. No clear idea of the whole is available. They appeared to have been constructed of large and to have formed an oblong of over 40 yards in length and about 17 yards across. '4 At the Wakhan region there are similar remains of three Kafir forts, which the natives. believe to have been erected by the Guebers or fire-worshippers.'5 They are known as Zamr-i-Atish-parast, Zangibar, and Kala-i-Kaka, the latter on the right bank of the Oxus." ' • The stupas of Chitral region are like those of Khotan. 'There are the characteristic three bases successively receding above them the high cylindrical drum; next, a projecting cornice surmounted by the proper stupa dome, which is approximately hemispherical in shape. Above the dome, in conventional outlines and drawn with a rather primitive attempt at perspective, appear the orthodox succession of 'umbrellas,' apparently seven, the lowest resting on supports which seem to slant outwards from the top of the dome. Through the last three of the umbrellas is seen passing the central pillar which in reality carried the whole series of umbrellas, and which, according to an early tradition, symbolized the beggar's staff planted on the top of Buddha's own stupa model." 1 Serindia, Fig. 70, pp. 236-37. 3 For plan and disposition of rooms, see ibid., plate 16. 5 Wood, Source of the Oxus, p. 218. 7 Ibid, I, pp. 37, see Fig. 5. ² Ibid., pp. 237, 238. 4 Ibid., p. 36. 6 • Stein, Serindia, I. p. 66. 305 t

The shrines of Khadalik resemble the Hindu temples. One of these as reconstructed¹ would look like the Linga type of temples described in the Silpa-sastras. It 'formed a quadrangle of which the external measurements were 75 on the east and west, and 73 on the north and south. The middle of this quadrangle contained a cella, measuring on the inside a little over 28 feet. The centre of the cella was occupied by a badly decayed base or platform, 10 by 9 feet. It rose in its completely broken state to a height of about 2 feet above the plastered floor. The space left between the outer walls and the cella, 21 feet on the north and south, and 20 on the other sides, seemed far too wide for the enclosing passage. The remains revealed manifest remains of intermediate walls which seem to have divided this outer space into no less than three concentric passages. This shrine faces southwards. 6 • " 22 The next shrine, which is Buddhist, consisted of a rectangular cella, measuring 25 by 24 feet inside and enclosed by a passage 8 feet wide on the north and 5 elsewhere, with a large hall adjoining the south passage. The walls are of sundried bricks and were strengthened by wooden posts. 3 In the cella have been found some two dozens miniature stupa models in clay, none higher than two inches. 'The structure immediately adjoining this shrine contained a single hall measuring some 47 by 42 fect. Its southern portion was occupied by a plastered platform, 15 feet wide, rising 10 inches above. the floor. The foot of the platform showed a bold moulding 3 inches wide.^{3 3} A few more shrines of the same type existed in this region. But the scanty remains do not give a clear idea of the whole. Their Indian origin is, however, assured by many indications, including the use of Kharoshthi script and an early Prakrit language for administrative purposes.4 The Miran site discloses two kinds of ancient monuments. 'The ruined fort of Miran, imposing as was the appearance, rises on the pebble-covered plain which extends, with a total width of 3 miles, from the belt of the vegetation watered by the present course of the Miran stream eastwards to an ancient river-bed. It is of irregular shape and construction. Its walls still rise to a considerable height. 1 See Fig. 6, Serindia, I, pp. 4-5. ³ Ibid., p. 157. 2 Ibid., I, p. 156. 4 Ibid., p. 243. 306

he two first wall-faces measure about 240 feet each. The west and south faces are considerably shorter, being about 168 and 200 feet in length. Projecting oblong bastions of irregular shape and varying dimensions protected the corners. It is, however, of comparatively late origin and of Tibetan type as disclosed by the finds of abundant Tibetan records."¹ The other monuments of Miran are of Indian origin and comprise Buddhist shrines, including stupa of the same pattern as in other sites, temple, remains of Buddhist sculptures, mural painting of a Buddhist legend in the cella, and Kharoshthi inscriptions. The main structure 'presented itself as a solid mound built of sundried bricks, oblong in shape, but showing no readily recognizable surface features. Two storeys, however, could at once be distinguished, and of these the lower are measuring about 46 feet on its longer sides and a little over 36 feet on the shorter. Its height was about 9 feet above the original level of the ground. On the top of this solid platform or base there rose a second storey, also oblong in shape. Its ground-plan measured about 17 by 15 feet. The corners of the whole structure were roughly orientated towards the cardinal points. The niches had once contained stucco statues in relievo, probably a little under life-size." This clearly proves that the architectural design of this decoration, which places relievo images in niches divided by Indo-Persian pilasters, is one directly borrowed from Graeco-Buddhist art. '3 • C The colossal heads in soft stucco also showed the influence of Graeco-Buddhist sculpture in modelling and proportions.4 The outer wall is lined with a row of colossal stucco images of Buddhas seated with cross-folded legs and usual pose (meditation, dhyanamudra) of the hands and drapery. The statues measured 7 feet to 7 feet 3 inches across the knees, which reached to a height of a little over a foot. The stucco bases occupied by the torsoes were 7 feet 6 inches long and 2 feet 4 inches wide, and rose to a height of 1 foot and 4 inches. The spaces dividing them were only 6 to 7 inches across in front, but widened towards the wall behind."5 Further evidence of their Indian origin was supplied by the discovery of a palm-leaf pothi (book) written in Sanskrit language and Brahmi characters of an early Gupta type. 1 Stein, Serindia, p. 456. 2 Ibid., p. 485. ³ Ibid., p. 486. 5 Ibid., p. 488. 4 See Fig. 121, 122, Serindia, Vol. I, pp. 486-87. 307

'The shrine, square outside but circular within, had once been surmounted by a dome and enclosed a small stupa in its centre. The interior of the circular cella had been lit by windows passing through the centre of the side walls, which were approximately orientated to the south, east, and north. The width of the windows was 2 feet 3 inches, and their lowest portions reached down to a level of 2 feet 8 inches above the floor. The masonry enclosing the dome cella was 4 feet thick at the windows, increasing to fully 10 feet where it filled the square corners. It consisted of sundried bricks mixed with straw, measuring 16 by 10 inches, with a thickness of 5 inches. The dome must once have surmounted the square cella, which was enclosed by an outside passage." "The peculiarity of this stupa lies in the circular shape of the base instead of the usual square one. . . . It may be assumed that the choice of the round base, amply attested among the stupas of India and the north-west border, was due to considerations of the space available within the small circular cella." "On the lowest division of the base, 9 feet in diameter and 9 inches in height, there is placed a receding drum, I foot 6 inches high, terminated both below and above by a uniform series of mouldings. Then follows another plain division, 9 inches high and of the same diameter as the lowest, and above it a succession of small step-like mouldings, with a total height of less than 6 inches. From this rises the cylindrical member, about 2 feet 3 inches high which carries the dome, both being 6 feet 8 inches in diameter. The projecting frieze-like moulding, about 4 inches thick, which intervenes between drum and dome, is a feature seen with equal clearness in the stupa carving of Pakhtardini, the wooden stupa models of Niya and Lonlan, etc.'s Considering the influence of the pradakshina custom it seems highly probable that the composition of the wall-paintings as a whole had its starting point on the left of the entrance to the cella and thus in the south-eastern segment.' There are various kinds of figures, representation of riding prince, saddle and horse-gear of mount, princess driving quadriga, white elephant in sylvan scene, prince leading elephant, figures of armed men, animals, etc., 1 Stein, Serindia, p. 493, Fig. 123, 128. 3 2 Ibid., p. 494, Fig. 126, 129, 130, 131. Ibid., 494. (This seems to be the very last outpost of Buddhist Central Asian monuments.) 308

-HUANG REGION representations of the story of Vessantara-jataka, Prince Vessantara's exile, Indian ascetics with bushy hair and beards carrying the long staffs and small bowl for water (kamandalu), lastly, on a piece of frieze, is represented a low throne of the Indian gadi type, covered with drapery.1 There are also figures like a girl playing on a mandoline, girl carrying decanter, bearded male head, bust of Indian prince, etc.2