The Development of Buddhist Art in South India

By Devaprasad Ghosh


The Indian Historical Quarterly


Vol 4:4, December, 1928, p 724-740



p. 724 Amaravati Stupa Amaravati is picturesquely situated on the south bank of the Krsna River close by the modern town of Dharanikota, ancient Dhanyakataka, the capital of Maha-Andhra, about eighteen miles west of Bezwada. The earliest stupa was raised under the patronage of the Andhras about 200 B.C., of which a few archaic sculptures have survived, but most of the exquisite marbles which survive to-day belong to a subsequent restoration about four centuries later. The great Buddhist stupa of Amaravati which was once unrivalled by any other Indian structure of its class in form, dimension and decorative grandeur p. 725 shared no better fate than the rest of the ancient monuments. "When Huen-tsang visited the place in the year 639 A.D. it had already been deserted for a century, but he speaks of its magnificence and the beauty of its site in more glowing terms than he applies to almost any other monument in India.''(1) From this time onward the monument gradually began to decay and fall into ruins. Towards the end of the eighteenth century, the famous mound, the upper part of which rose in a turreted shape encased with bricks to the height of 20ft. with a diameter of about 90 ft. at the top, was locally known as Dipaldinne or "Hill of Lights". Colonel Mackenzie who went to the site in 1797 found to his great chagrin that just a year before, the local Raja Venkatadri Naidu had discovered and disemboweled the mound in a fruitless search after hidden treasures; he afterwards caused a reservoir to be dug in the centre and used the priceless marble slabs in building the new temple of Amaresvara and the flight of steps to the adjacent tank of Sivaganga. Some of the slabs were utilised by the Mussalmans in their mosques, after 'carefully divesting of every carving by rubbing them on harder stones, to prevent, as it is said, any pollution arising to Muhammadan faith from idolatrous substances'.(2) Mackenzie revisited it in 1816, when as a result of excavation he recovered some 130 slabs, made drawings of them and prepared a ground-plan of the stupa. The place was next visited by Sir Walter Elliot in 1845; but in the meantime 70 pieces of sculptures left behind in the open had been carried away by the enterprising villagers and burnt into lime! (3) It is deplorable that even the Government Public Works Engineers were equally guilty of such acts of vandalism.(4) The slabs excavated by Sir Walter were transhipped to England and now adorn the grand stair-case of the British Museum. The next excavation was undertaken by Mr. Sewell, but it was reserved for Dr. Burgess to make a shifting and scientific examination of the spot in 1882-83 and incorporate his findings in a voluminous report, In the first decade of the 20th century, the work was continued, with ______________________ 1. Fergusson, History of Indian and Eastern Architecture, 2nd ed., London, 1910, vol.I, p.123 2. Burgess, The Buddhist Stupas of Amaravati and Jaggayyapeta, London, 1887, p. 15. 3. Sewell, Report on the Amaravati Tope, London, 1880, p. 67. 4. Madras Govt.,Orders No. 467, 30 April, 1888, p. 15. p. 726 valuable results, by Mr. Rea of the Arclaeological Survey. The sculptures which are now in India after surviving the ruthless vandalism through the ages are shared by the Museums of Madras and Calcutta. An inscription of the reign of Palumavi Vasisthiputra tells us that the Amaravati stupa was known as the Mahacaitya or 'Great Caitya' of the Holy One belonging to the Caitika School. A stupa or Caitya has its origin in the primitive burial mound of both the Arya and the Asura.(1) In the vicinity of Amaravati itself, there are numerous funeral tumulii, surrounded by rude stone circles, of remote antiquity, which served as the prototypes of the later stately structures in stone or brick. The stupa at Amaravati was not a commemorative monument like the ones at Sarnath or Nagarahara, neither was it a hollow Caitya containing some relic, as the earlier stupas at Sanchi, Sonari and Manikyalado. It was a solid structure and rested within a square stone casket, on the top of the dome, in conformity with the convention of the day. The circular base of the stupa was 162 ft. in diameter, perhaps only 6 ft. high, supporting a frieze and cornice, and was faced with marble slabs possessing the richest carvings and characterised by the most delicate treatments, depicting miniature representations of the stupa itself and interposed by panels elaborately carved with scenes from the life of Buddha and the Jatakas. It is very difficult to ascertain whether the dome rose directly from the drum or rested upon several receding terraces like the Gandhara, Further Indian or Indonesian specimens. But there was no balustrade to encircle the procession path at the base of the drum as on the great stupa at Sanchi. The great marble dome of Amaravati, unlike the short and stunted dome of Sanchi, rose to a considerable height of 90 ft. (twice that of Sanchi ) and was more or less bulging in form. In this respect it presented a contrast to the stilted hemispheres of the earlier northern examples and was more akin to the soaring forms of the Ceylonese dagobas, 'The domical part was covered with stucco, and with wreaths and medallions either executed in relief or painted'." The marble panels were also 'covered' originally with thin plaster,coloured and gilt. _______________________ 1. A very illuminating article on the 'Stupas or Caityas' has been recently contributed by Mr. R.D. Banerjee (vide Modern Review, Calcutta, Feb. 1928). 2. Fergusson, op. cit., p.80. p. 727 Thus the conception of the whole thing, profound and majestic, was matched by an exterior at once brilliant and dazzling. As all traces of the great stupa have been wiped away from the site, we cannot help looking at one of the numerous panels representing the miniature stupa in order to gain an idea of the original one (see plate). The very first thing that strikes us, and which is visible nowhere in northern India, is the five tall stelae 'above the front slab, which slightly projects from the base of the dagoba--the bases are square and sometimes ornamented with carvings of Cakra, Bodhi Tree and Dagoba; the shafts are octagonal, and they have square carved capitals'.(1) The existence of these novel features on the great stupa is attested by the discovery by Dr. Burgess of a number of these pillars at the Jaggayyapeta stupa 30 miles north-west of Amaravati, of which we have already spoken.(2) In an inscription they are called 'Aryaka Khambhe'. That this was a common feature of the Kalinga Stupas is proved by the recurrence of this element also in the stupas at Bhattiprolu and Ghantasala. These projecting pedestals with the enigmatical columns, on the four cardinal points of each stupa, may correspond to the four shrines in the stupas at Sanchi and Bharhut, and the niches for the Dhyani Buddhas in the dagobas of Ceylon and the Caityas of Nepal. With the march of time the number of these chapels went on increasing; at Sarnath they are doubled while Borobudur simply bristles with them. Other slabs invariably present us with another peculiar feature, viz., a dwarf figure standing on each side of the gate, holding a tray on his head.(3) Their constant occurrences lead us to believe that in the original structure they represented statues in the round, bearing trays to receive the offerings of the visitors. Dr. Burgess opines, 'No example of them has been found and the only analogue I know of, is a similar small figure bearing a basin by the doorjamb of the cave at Lonad of the Thana district near Kalyan."(4) But we think a closer examination of the extant monuments may yet reveal such figures and in fact there are such at Karli and in Orissa. A pair of vases with flowers ___________________ 1. Burgess, op. cit,, P. 72. 2. Ghosh, Development of Buddhist Art in South India, Indian Historical Quarterly, Sept, 1927, p. 502 3. Burgess, op. cit,, Plate XXXI, Figs. 6 and 7. 4. Ibid., p. 72. p. 728 (mangalakalasa?) prominently placed at the entrance is another regular feature of the sculptured slabs. The appearance of two slender pillars or free-standing lats with small Caitya capitals, crowned sometimes with plenty of Chatas, one on each side of the entrance within the enclosure, is also remarkable The paucity of such examples in the northern stupas is striking; and if they occur at all (as for example at Sarnath and Sanchi) they are situated outside and not inside the rail. The actual presence of these columns in the great stupa, is supported by the excavations at Jaggayyapeta and Bhattiprolu. They have also a close affinity with innumerable concentric lats, still standing round the Thuparama and Lankarama dagobas in Ceylon--a perpetual enigma to the generations of archaeologists. The Rail.--The most singular feature of the early Buddhist and Jaina stupas is the rail, upon which the artist devoted his most scrupulous attention and lavished all the splendour he could conceive. We are aware of the extant rails at Bharhut, Bodh Gaya, Sanchi and Mathura and we know too their wealth of decoration, but the remarkable rail at Amaravati has far surpassed them all in the magnificence of elegant carvings and the marvellous display of intrinsic merit. The ornamental detail is simply staggering in its profusion and afford a striking contrast to the plain and simple rail of the great stupa at Sanchi. The great rail at Amaravati was about 600 ft. in circumference and 14 ft. in height with a procession path 13 ft. broad, intervening between it and the base. It was more than twice the dimension of the rail at Bharhut. The Tibetan historian Taranatha records that the great Buddhist Acarya Nagarjuna, the founder of the Madhyamika School 'surrounded the great shrine of Dhanyakataka with a railing.(1) Colonel Mackenzie in 1797 Was responsible for starting the theory that the stupa was surrounded by two rails--one inner and another outer. The error persisted with veteran archaeologists like Fergusson and Burgess, not to speak of Elliot and Sewell. It was only about two decades ago that Burgess acknowledged and rectified the mistake. 'From some misunderstanding of the first accounts' he added, 'it was supposed that the Amaravati Stupa had an inner ___________________ 1. Schiefner's Taranatha's Geschichte des Buddhismus, p.72; JASB. vol.LI, pp. 119; Indian Antiquary, vol. XII, p. 88. p. 729 rail; this was a mistake; the inner circle of sculptures was the facing of the base of the stupa'.(1) The rail at Amaravati resembled its predecessors in the principal features; but the plinth was richly carved with a frieze of running boys and animals, grotesquely treated. The rectangular pillars were as usual edged off into shallow flutes. They were decorated with half lotus discs at the top and the bottom, and circular discs in the middle inserted with a full-grown lotus or a scene, in the usual manner. But the most typical characteristic about these pillars, is the complete absence of the large standing human representations, occupying the entire surface of the uprights, such as the graceful statues of Yaksas and Yaksinis of Bharhut, Bodh Gaya and the dancing girls of Mathura. They have entirely disappeared and their place is occupied by greatly magnified and richly carved lotus discs, curling leaves carefully corrugated, comical Ganas and an enormous variety of scenic sculptures. The preference for group composition, as opposed to single figures, is very obvious in the swarming of the space between the discs--which was generally left bare and unadorned in the earlier days by vivid and animated delineation of the Jatakas and other incidents. The three cross-bars were each embellished with a beautiful lotus disc with concentric bands of petals, the most elaborate of its kind ever made, and all different. On the massive coping, the meandering creeper of Bharhut was replaced by a long wavy roll, carried by moving human figures and dwarfs and interspersed with symbols in the loops. The marvellous change which has taken place in the sphere of ornamentation has already been noticed in the previous chapter on Ornamental Representation.(2) On the whole the inner side of the rail, covered with scenes full of life and movement, was decorated with greater beauty and elaboration than the exterior. The Amaravati rail has a close resemblance to the rail of Stupa no. 2 at Sanchi, in excellence of carving and richness of detail. The decorative tendency which was strongly evident at Sanchi became more pronounced at Amaravati. The lotus medallions grew larger in size and became more prominent (those at the top and bottom were often three-quarters and not half) till at last they reached their climax in the rail of the Gautamiputra cave, Nasik, where the pillars and cross-bars were adorned with full discs only. "The discs were _______________ 1. Fergusson, op. cit., revised by Burgess, pp. 119f. 2. IHQ., Sept, 1927, pp. 486-91. p. 730 multiplied till the pillars almost became evanescent quantities in the composition." In spite of all these arresting details, we are confronted with the rather astonishing fact that the four openings piercing the great rail at the cardinal points, were not adorned with the beautiful towering Toranas, such as we find at Bharhut and Sanchi. The sides of the entrance are shown instead as coming out in a 'rude sort of perspective and terminating in neat pillars with bases and capitals, crowned by figures of lions; at the angles too, above the roll, on each side is a lion.'(1) One such lion lying prostrate near the west gate yielded to the spade of Sir Walter Elliot.(2) While the reliefs abound with representations of such structures over the city and palace gates, the conspicuous absence of the characteristic Toranas from the great rail is mysterious indeed. Architectural Representations Now that we have a picture of the stupa in the height of its glory; let us proceed to discuss the various forms of architectural representations from the extant remains in relief as well as in the round. Dwellings and Palaces.--From the sculptured slabs we can find that the ordinary dwelling places were really oblong shaped huts with barrel-vaulted roofs which unlike the curvilinear forms of Bharhut, Sanchi and Bodh Gaya are more or less semi-circular in shape. This may be a peculiar South Indian feature and differs strikingly from the square-thatched houses of Bengal, Behar and Orissa and other early sculptures of the North. Other small detached huts show that they were crowned with circular domed roofs. These instances may lead us to infer that the South had dispensed with all angularity in construction of the roofs of the poor and the common. The few instances of single and double storied palaces, buildings and shrines, carved here as well as on the Jaggayyapeta slabs, with their railed verandahs, caitya windows and arched roofs with finials --which were continued till the time of the Mahaballipur Rathas-- reveal no dissimilarity between them and their northern prototypes. To make the scenes inside visible they are shown in a sort of conventional perspective. Most of the buildings represented are distinctly ___________________ 1. Burgess, op. cit., p, 70. 2. Ibid., Plate XLV, Fig. 7. 3. Ibid., XXVII, Fig. I. p. 731 modern in character as Fergusson conjectures; and the practice of setting up wooden architecture was prevalent in South India till comparatively recent times. As at Sanchi, the difference in material of domestic architecture from that of civic and military architecture is distinctly shown in the brick construction of the latter. The palace buildings are usually surrounded by high walls on all the four sides with two or more entrance ways. Over these gateways, there are high spires or flag-towers, where sentinels were stationed and where also play bands or Mangalavadya, pipe and music, both in the mornings and the evenings. Such places are now found in all Muhammadan palaces or Nowbatkhanas. The construction may be laid out square or circular in accordance with the taste of the kings or owners of the grounds, or it may be even laid out in the form of a semi-circle as in the Karmuka form of town-plan.'(1) Most of the above features were recorded by the artists in the panels. Fortifications.-There is also complete agreement between the southern and northern examples of fortifications. A comparison of the reliefs of Amaravati with the architraves of Sanchi gateways, will make this apparent in the identical forms of high and broad brickwalls, massive palisades, strong gateways, lofty towers bristling with turrets and pinnacles set with the usual Buddhist Caitya-window facades, strongly built watch-towers, tiers stories each supreimposed on the other, adorned with hanging balconies and numerous strategic windows facilitating the discharge of arrows from safe from safe quarters, and other apartments invariably fringed with the rail pattern and crowned with gable-shaped roofs. it must be admitted, however, that it is very difficult to distinguish between a fortress and a palace proper, as in those days every royal abode was a military stronghold and vice- versa.(2) Temples.--The method of building temples and shrines does not seem to have made much progress since the days of Bharhut. The object of adoration was usually placed and worshipped in a courtyard generally flanked on three sides only by buildings (vide Asoka's temple at Bodh Gaya, carved on a Bharhut pillar), or within separate structures either oblong or square, but generally open and _________________ 1. Rajagrhalaksmanam, Manasara, ch. xi. Translated by lyer in Indian Architecture, Madras, 1921, vol. iii, Bk. I, ch. XI. 2. Burgess, op. cit., Plate xxv, Fig. 2, and xxvii, Fig. 2. p. 732 surrounded by pillars. Indeed one may be easily led to ascribe the shrines represented on some of the earliest slabs of Amaravati to the Bharhut railings. Gateways.--Although we have no evidence as to actual gateways guarding the entrances of the stupa itself, the reliefs afford us with copious examples. Two different kinds of Toranas can be noticed. One type represents two square and carved pillars surmounted by cushion capitals and crowned with crouching animal figures like those at Bharhut, which in their turn support a superstructure of a very broad, solid semi-circular architrave without any volute ends.(1) The second type, occurring more frequently, has exceedingly slender and often plain, square shafts, rising from pot-bases and crowned with or without cushion capitals, There are the usual two or three architraves with volute ends but entirely bare, each ranged above the other, the gaps being linked by vertical posts. A few of them are carved with geometrical patterns.(2) The difference in appearance of the Amaravat Toranas from those of Bharhut, Sanchi and Mathura, lies in the architraves of the former being more curved and the volute ends correspondingly curled up to a greater degree. It can also be noted that perhaps the gateways were not so lavishly enriched with marvellous bas-reliefs as those of Sanchi. Pillars and Pilasters.--Apart from the pillars which serve architechtonic purpose, freestanding sculptured lats can be observed on many of the slabs. There are some with cushion capitals and inverted steps bearing Cakras and other Buddhist symbols(3)The slender columns within the enclosure, which 'at once remind us of the Asoka lat in Northern India and Iron pillar at Delhi' and specially the rows of pillars round the Anuradhapur stupas in Ceylon supporting the same cushion capital and inverted slabs, have miniature dagobas always placed on top of them. This is perhaps the first instance where a Caitya constitutes the crowning emblem instead of the usual animal or other familiar northern conventions. The Ceylonese capitals of the particular type are either topped with a knob or with a flat surface. None of the pillars which must have stood at the gates, remains in its entirety--only fragments have been found. The earlier types were plain and carved with rail pattern and other Buddhist symbols ------------------------ 1. Burgess, op. cit., Plate v, Fig. 2. 2. Ibid., XVIII, Fig. 2. 3. Ibid., V, Fig. 2. p. 733 while the later examples were adorned with the figures of the Buddha and other sculptures. The shafts were square, octagonal or Some of the broken fragments of pilasters betray crude craftsmanship and antique characteristics noted below. Dr. Burgess judiciously observes, "These slabs so closely resemble those round the Jaggayyapeta stupa that we cannot mistake in ascribing them to the same age.(1) They must have belonged to the early stupa. Like the Nasik and Junar pillars, the base consists of three thin slabs supporting a vase, carved with leaf and bead pattern. This clearly indicates that in ancient times the original wooden shafts were inserted into metal pots to preserve them from decay and injury. This theory has been strengthened by the recent discovery of the bronze shoe of a column at Balawat in Assyria, which points to the frequent use of this particular method in Assyria, Mesopotamia and Persia. Remarkable plastic examples of this type are found at Bodh Gaya, Khandagiri (Ananta Gumpha) and Gautamiputra cave at Nasik. But unlike the western prototypes, a projecting member, carved with dwarfs or hybrid creatures and ornamented with the old battlements and other motifs, stands over the neck of the body. The shafts, the edges of which are slantingly cut off like those of the Bharhut pillars, are adorned with half lotus discs at each and a full one in the middle; and closely resemble the pilasters in the Pithalkora Vihara. But no pillar with cushion capital, first encountered at Kanheri and so often sculptured in the reliefs, has been discovered. Generally the double carve of the bell-shaped or lotus capital, is very slight and do not possess the graceful sweep of the Asokan capitals;neither it is boldly modelled with soft drooping flutes nor facaded like the Karli examples. The flutes, on the contrary, are extremely crude and shallow, like the Bhaja specimens and their significance is further reduced by the intersection of bands of lotus leaves and beads. Absolutely smooth capitals, parallel with those found in Nasik caves are not, however, rare. The necking consists of the bead and reel pattern which supports the terraced superstructure, surmounted by a pair of winged animals seated in juxtaposition, similar to those prevalent in other regions.(3) So we may conclude that the typical characteristics of pillars and _________________________ 1. Burgess, op. cit., p. 94. 2. Ibid., Plate XLIV, nos. 5, 6; LIV, no. 2. p. 734 pilasters during the 2nd and 1st centuries before the Christian era, are almost identical throughout India, whether in the north (Bharhut, Sanchi, Mathura), in the west (Bhaja, Pithalkora, Junar, Karli, Nasik, etc.), in the east (Bodh Gaya, Udayagiri, Khandagiri, etc.) or in the south (Amaravati, Jaggayyapeta etc.). Mr. Havell remarks "The lotus and vase pillar, besides being one of the most ancient Indian architectural orders, is also the most frequently used. It is found at all periods."(1) The Ruins of Sankaram and Ramatirtham Let us now take leave of the Krsna district and proceed a little higher up the Kalinga country. There are two isolated hills covered all over with monolithic and structural Buddhist remains, very close to the village of Sankaram, in the Vizagapatam district of the Madras Presidency. "The monuments," says Mr. Alexander Rea, "are the earliest of their class in the South of India and constitute one of the most remarkable groups of Buddhist remains in the Presidency. Indeed the only other known site in the South, where monolithic remains exist in any considerable number, is that of the Seven Pagodas, and though the Sankaram site is not to be compared with it in point of extent, it takes precedence as regards the age of the monuments."(2) The Eastern Hill, which is the higher of the two, is literally strewn with rock-cut caves and dagobas, the monoliths set upon platforms and terraces, rising in tiers over each other culminate in the dominating structure of a great stupa on the summit. The grandeur of Borobudur flashes across the mind when we visualize the almost identical arrangement and the imposing profile of the whole mass in its original and pristine glory. The remains can be classified into three main heads, viz., (I) rock-cut caves, (2) monolithic dagobas and structural stupas and (3) structural buildings for residential purposes. Rock-cut Caves.--The surface of the rock at places is hollowed out into a deep recess in order to provide a vertical wall with a platform before it--just in the manner of the peculiar rock-dwellings in Asia Minor, called the Syppilus. On the Eastern Hill, in one such wall, ___________________ 1. Havell, Handbook of Indian Art, London, 1920, p. 44. 2. Rea, A Buddhist Monastery on the Sankaram Hills, Vizaga-patam Dt." Arch. Sur. Ann. Rep., 1907-08, p. 149. p. 735 double caves are cut, one standing over the other and each of them is entered only by a single rectangular doorway. The facade is of the Behar caves, the Western Caitya halls or the neighbouring caves in the Godavari district. We sorely miss the much familiar Caitya-window, the Caityas, the Buddhist rail and other ornamental devices which decorated the facades of almost all the Buddhist rockcut caves of the early period. In Cave I, "over the door which is guarded by figures of Dvarapalas" like the Nasik caves, "weather-worn traces of an architrave can be traced which include two semi-circular pediments with a cornice over it." The usual place of the Caitya-windows is usurped by a semi-circular recess occupied by a large-seated image of the Buddha. Figures of the Buddha, sometimes with attendants, are also carved in niches beside the facade. The interior of this cave also differs materially in plan and construction from the early Buddhist Caitya halls. The Chamber, instead of being oblong in shape with an apsidal end, is absolutely square in dimension. It is further characterised by the absence of the double row of columns dividing the interior into a central nave and the two side-aisles as in the Western Caitya caves, The hall is, on the contrary, demarcated into twenty compartments by four cross rows of sixteen pillars. The columns are massive in proportion and do not resemble in any way the early types of pillars with a pot base, lotus or bell capital and animal superstructure. They belong to a different class altogether--having a square base, short octagon in the centre changing into sixteen sides upwards, these several unskilfully moulded neckings followed by a thin and small torus, surmounted again with square block. "Two central piers of the central square have a standing image apparently a Cauri-bearer, cut in the front of the base." Stranger still, a Caitya or rock-cut dagoba with a plainly moulded base, a circular dome and the remains of a tree, stands on a square platform which fills up the space between the four central piers and is situated in the middle of the cave instead of rising precipitously from the floor at the apsidal end of the hall, according to convention. So here we are confronted with the unique spectacle of the combination: a Caitya hall and a Vihara combined into one. Cave II above it consists of two apartments--one rectangular vestibule, and a shrine which is also rectangular and without pillars. "The walls of the vestibule are also carved with the Buddha and attendant images and some representations of the dagoba with strikingly bul- p. 736 bous domes" like some at Amaravati. Instead of the Caitya, there is a seated image of the Buddha on a pedestal on the back wall of Cave III. In Cave V, the type of pillar is identical with that of No. I; only it is more slender and has a fluted(?) or moulded torus. "There is a lotus patera at the top of each square and pediment at the top of the octagon." The principal cave on the Western Hill contains another novel feature, viz., the Caitya is placed in a square cavity in the middle of the chamber below the ground level. The ceilings of these caves are plain and flat. The walls and images were originally coated with plaster. Dagobas.--Almost all the dagobas, strewn about the hill and converging upward, are rock-cut monoliths. They are very crudely worked out and their forms are characterised by the utmost simplicity. The hemispherical ''auda" which is either bulbous, flat or elongated in shape, is nearly superimposed on a drum having also stunted or column-like elongated forms. Formerly they were all covered with Stucco. Compared with the monolithic Stupas at Bhaja, they appear absolutely bare; even the essential rail ornament is absent from the rim of the drum, and as far as it can be guessed, this device and the favourite Caitya windows do not occur on the Harmika, disfigured as it is. The dagobas on the West Hill are comparatively better. Some of them have moulded bases, plain plasters and cornice round the drum, also a series of inverted slabs on the relic casket, in the conventional way. Others are faced with brick or made wholly of brick. "The crowning Stupa rested on a square platform, on which rested the low rock-cut cylinder which formed the lower part of the dome, the upper part being completed in brick. The complete dome must have been a low curve of less than semi-circle.........almost wholly of brick." Structural Buildings for Residential Purposes.--On the eastern end of the top of the highest terrace, the remains of a structural rectangular Caitya hall made of brick and terminating in an apse, has been excavated, Like the one at Ter, in the Nizam's Dominions, the Caityas are too small to have space or necessity for pillars. It is the main structure round which all other constructions grew up. It has been divided into two compartments by means of the usual partition wall near the apsidal end, into an ante-camber and a shrine The Caitya is replaced inside the shrine by a rectangular stone pedestal with a cavity on top, probably meant for an image. There p. 737 is a large stone-paved brick hall, faced with pilasters just in front of the Caitya, but at a lower level. "'And inside these walls and placed at right angles to them at the same level are the remains of the partition and outer walls of a continuous row of cells and shrines standing on the north, east and south sides.'' Another peculiar feature is to be met with at the entrance of the hall, which is flanked on either side by apsidal brick structures, with their entrances facing the central passage. The chambers which occupied this position, if at all, in the Western caves, were square and never apsidal. Remnants of other continuous rows of cells have also been dug out around the three sides of the raised Caitya terrace and at the same level with it, while an outer detached row stands parallel to those to the north.(1) One of the range of hills, in the vicinity of the village of Ramtirtham, in the Vizagapatam district, is also studded with the extensive ruins of a Buddhist monastety. Like Sankaramm, apart from the foundations of a large brick stupa, the most interesting buildings here are the structural Caityas so rare in India. On the Gurubhaktakondu Hill, there are remains of an apsidal brick Caitya hall, with a stone dagoba resting on a double pedestal. There is a wall across the chord of the hall. The absence of pillars was perhaps a common characteristic of the stuctural Caityas. In agreement with the Caitya at Chezarla, it has brick pilasters "with moulded bases and capitals, and at the base of each, fragments remain of three crouchant lions.'' The semi-circular slabs at the foot of the flight of stone stairs at once recalls the beautiful 'moonstones' of Ceylonese architecture. Near by it, at a lower level, is the site of a brick Vihara, the roof of which was supported by six rows of six piers each--square in section but near the top octagonal. This exceptional arrangement has made it impossible for a quadrangular space to be provided in the middle, in imitation of the Western rock-cut Viharas. The foundations of other Viharas do not show remains of columns.(2) The Date of the Ruins.--Regarding the remains at Sankaram, Mr. Alexander Rea in his Report says, "The sculptures in all the ____________________ 1. Most of the data utilised here are borrowed from Mr. Rea's Report. 2. Rea, "Buddhist Monasteries on the Gurubhaktakonda and Durgakonda Hills at Ramatirtham." Arch. Surv. Ann. Rep., 1910-11, p. 78-81. p. 738 caves and on their facades generally are crude and primitive in design and have none of the finished technique so strikingly observable at places like Amaravati, where the highest phases of the sculptor's art are so lavishly represented. The crudeness may point in either of two ways. It may either represent a very early period of undeveloped workmanship or a later decadence. The Buddhists did not survive sufficiently long after Amaravati epoch for any such decadence to have strikingly manifested itself. The inference is therefore that the period represented by these sculptures is earlier than Amaravati or possibly prior to the first century. The earliest of the remains here or the monoliths probably belong to the period of Asoka himself. Though the sites founded by him are historically and traditionally described as numerous in Southern India, no traces of any of them have hitherto been found."(1) Firstly, we cannot concur with the view of Mr. Rea, that the Buddhists and their art "did not survive sufficiently long after Amaravati epoch for any such decadence to have strikingly manifested itself." The statement falls to the ground in the face of the discovery of Buddha and Bodhisattva images at Amaravati, Jaggayyapeta and other places belonging to the 6th and 7th centuries A.C. and betraying obvious signs of degeneration in technique and treatment. Again we can hardly ignore the invaluable testimony of the famous Chinese traveller Huen-tsang, who passing through the countries of Kalinga, Kosala, Andhra and Dhanyakataka in the 7th century noticed stupas and numerous Sangharamas peopled by hundreds of Buddhist priests. Secondly, the architecture of the caves, monoliths and other structures, itself does not warrant us to accept the conclusion of Mr. Rea. If the rail pattern, Caitya window motif, sloping door jambs, wooden ribs of the barrel vaulted roof and the wooden screen and the purlins in front of the Caitya hall are indicative of an early age, surely all these features are prominent by their absence at Sankaram. The occurrence of the miniature Caitya windows over the door and some of the windows and the so-called "horse-shoe arch" over the entrance of the vestibule, closely resembling the facade of the Lomasa Rsi cave, in some of the Buddhist caves at Guntupalle, Godavari district, in the heart of the Kalinga country, is sufficient proof of their pre-Christian age. But we search in vain for these typical details here, ________________________ 1. Rea, Arch. Sur. Ann. Report, 1907-08, foot-note, p.158. p. 739 The facade is extremely simple in design, the ceiling is entirely plain and the pillars themselves have not any affinity with the earlier types we are familiar with. On the contrary, many points of similarity can be detected between them and the later cave pillars. Indeed, none of the architectonic features at Sankaram is reminiscent of wooden construction--they are purely lithic in design and conception and indicate a late period when the transition from the wooden to the stone construction has been complete. Then again the arrangement of the pillars, the square plan of the Caitya Chamber and the situation of the Caitya itself, in the centre of the hall and on a pedestal, are unique in the history of Buddhist architecture in India proper, The presence of the stupas with square bases, is noticed first in the caves at Kholvi in Rajputana and also at Dhamnar. About the former's date, Mr. Fargusson is of opinion that "they are probably the most modern group of Buddhist caves in India."(1) As regards the monolithic dagobas, they cannot reasonably be assigned to a period earlier than Amaravati, much less to the age of Asoka--for the outlines of all the stupas, large or small, is flat and stunted. We come across the bulbous domes for the first time at Amaravati and it is an admitted fact that stunted domes resting upon elongated pedestal is a later development. If we also take into account the structural building at Sankaram, the peculiar combination and arrangement of rows and cells with and around a Caitya, primarily appears in the caves at Dhamnar, about which Fergusson remarks "..........,.the whole making a confused mass of chambers and caityas in which all the original parts are confounded and all the primitive simplicity of design and arrangement is lost, to such an extent that without previous knowledge they would hardly be recognisable.........There are no exact date for determining the age of this cave but like all of these series, it is late, probably between A.D. 600-700."(2) So in the absence of any good photographs of the sculptures we have been compelled to take recourse to architecture, and in the light of the above facts, it may not be quite correct to maintain that the period represented by the ruins at Sankaram "is earlier than Amaravati or probably prior to the first century A.D." It is very probable, ____________________ 1. Fergusson, History of Indian and Eastern Arch., vol.I, p. 166. 2. Ibid. p. 740 on the other hand, that they belong to a much later period. The crudeness of the sculptures, of which Mr. Rea speaks, is the natural concomitant of a decayed art, when Buddhism was apparently in its last gasps in Southern India.