The two Rathas bearing the names of Arjuna and Dharmarâja, the second and fourth in the row, are identical in so far as their architectural ordinance and general appearance is concerned, the only difference being that the first named is very much smaller than the other. They in fact form a pair and represent on a small scale the three and four storeyed Viharas of the Buddhists, in the same manner that the Ganeśa temple and Bhîma's Ratha may be taken as representations of the halls, or Śhâlâs, which were adapted for ecclesiastical purposes by the votaries of the same region from the earliest times to which we can go back.

Arjuna’s Ratha, though so very different in design, is very nearly of the same dimensions as that of Draupadî, which stands next to it. In plan it is a square measuring 11 feet 6 inches each way, or with its porch, 11 feet 6 inches by 16 feet, and its height is about 20 feet. Inside a cell has been excavated, and though only 4 feet 6 inches by 5 feet, seems to have been the cause why the Ratha is cracked from top to bottom, and a part of its finial fallen off. The roofs both of the lower and of the first storey of this little temple are ornamented with those ranges of little simulated cells which became the distinguishing characteristics of Dravidian architecture from that day to the present hour, and it is surmounted by a dome, which is an equally universal feature, though whether it is copied from an octagonal apartment, or from a Dâgoba as at Boro Buddor, is not quite clear. There is no image in the sanctuary, though the first gallery is ornamented with 12 statues, three in each face, representing either gods of the Hindu Pantheon or mortals. Some have inscriptions above them, but none of these afford any information, we cannot gather from the statues themselves.

The Ratha bearing the name of Dharmarâja is the most southerly and is the largest and finest of the group, though like everything  else about the place it is unfinished. As will be seen from the annexed woodcut, its dimensions in plan are 26 feet 9 inches by28 feet 8 inches, and its height is rather more than 35 feet. It consequently occupies more than six times the area of Arjuna’s Ratha, and is nearly twice as high, but even with these dimensions it can only be considered as a model. It would require to be magnified to twice or three times these dimensions to be a habitable building. The four upper storeys of the Undavilli cave (ante, p. 96), which resemble this Ratha more nearly than any other known building, are upwards of 50 feet in height, and they are only on the verge of habitability. The simulated cells there are still too small to be occupied by human beings.

Its general appearance will be seen from the annexed woodcut, from which it will be perceived that it is a building of four storeys arranged in a pyramidal form. The lowest storey, which was also the tallest, in the building from which this one is copied, was probably intended to be constructed in stone, as the Gopuras and temples in the south of India almost invariably are, while all the upper or pyramidal parts in them are as generally built with bricks and wood. In this instance, the upper part could only have been constructed with similar materials, and if meant to be inhabited, in wood only. The pillars of the basement all are, or were intended be slender examples of the Elephanta order, (woodcut No. 29), with yalis or conventional lions forming their base. The three upper storeys are all ornamented with those little simulated cells described above, in speaking of the Ganesa temple and Arjuna'a Ratha, and which are so universal in the , south of India—there are 16 of these on the first story, 12 on the second, and eight on the third. The front of each of these cells with their connecting links, is adorned with a representation of one of those semicircular dormer windows which are so usual in Buddhist architecture. Here each has a human head represented as if looking outwards. Behind these cells the walls are divided by slender pilasters into tall compartments, and in each of those which would have been an opening in the original building there is now placed the statue of either a deity of the Hindu pantheon or of some now undistinguishable mortal. Among the gods are found representations of Brahma, Vishṇu, and Śiva, but without any of those extravagances which afterwards deformed the imagery of the Hindu pantheon; none of the gods have more than four arms, and except for this, are scarcely distinguishable from ordinary mortals. The Ardhanâri, a favourite form of Siva, as half a male half female, occurs several times, and Vishṇu as Narasimha or the boar Avatâr. There is, however, no attempt at a bas-relief or any connected story, and unfortunately none of the inscriptions over these figures, though numerous and easily legible, do more than supply laudatory epithets to the gods over whose heads they are engraved.1 At each angle of the lower storey which was meant to be solid there are two niches, one of which contains a figure of Śiva or Ardhanâri, and another apparently a Deva, or it may be only a  mortal.

The whole of the three upper storeys are perfectly finished externally. But in the present state of the monument it is difficult to say how far it was intended to excavate their interiors. The upper or domical storey was probably intended to be left quite solid, like that of Bhîma's Ratha. A cell was, however, excavated to a depth of 5 feet, in the third storey, and it may have been intended to have enlarged it. A similar attempt has been made in the second storey, but carried only to the depth of 4 feet, when it was, abandoned. From there being six pilasters on the outside of the third storey, we gather that in a structural building its roof would have been supported by 36 wooden posts, and in like manner that the second storey would have had 64 supports (8 X 8), but of course some of these might have been omitted especially in the centre, in actual construction, though there probably would be no attempt to copy all these in the rock. From its extreme irregularity it is not so easy to suggest what may have been the intended arrangement of the lowest, but principal storey; but from the wider spacing of its pillars externally it is evident, that in a structural building stone, and not wood, would have been employed in its construction. From the arrangement of the exterior we gather, with almost perfect certainty, that there would have been four free standing pillars in the centre, as shown in dotted lines in the plan and section. It is not clear, however, how many of the eight piers or pillars that surrounded these four (woodcut No. 31) were free standing or attached as piers to the external walls. The four in the angles were almost certainly attached to the angle-pieces which in a structural building would have contained the staircases. Practically, therefore, this Ratha seems to have been designed to represent a building having on its lower storey 16 pillars besides piers, standing about 6 feet apart, from centre to centre, and being executed in some durable material. Above this the floors were supported by wooden posts less than half that distance apart. As before remarked, both these dimensions would require to be at least doubled to render them suitable for a habitable Vihâra. Be this as it may, there can be very little doubt that it was the intention of those who designed this Ratha to have excavated the whole of the lower storey. It is probable, however, that, warned by the fate that attended their operations in the case of Bhîma's and Arjuna's Raths, they desisted before excavating beyond a few feet on each face; and it is fortunate they did so, for had they proceeded further inwards the mass of rock they must have left above, would certainly have crushed the four slender pillars they intended to leave in the centre, and fissures, if not ruin must have been the result. It may, however, be that some social or political revolution, of which we know nothing, was the cause why this Rath was also left incomplete. It certainly was not any physical cause which led to the abandonment of the works in the caves, or on the bas-reliefs before they were completed, as no danger of crushing existed there. In the case of the raths, however, as physical causes which we can comprehend, seem amply sufficient to account for their unfinished state, it seems hardly worthwhile to speculate on one of which we know nothing. Those who first attempted to carve those rocks were certainly novices at the trade when they began them, but their experience at Arjuna's and Bhîma's Rathas must have taught them that wooden forms were not suited to monolithic masses, and that either they must desist from the undertaking, or must invent forms more appropriate to the material in which they were working.

Although these two last named Rathas are sufficiently interesting as examples of the patient labour which the Indians have at all times been prepared to spend on their religious edifices, their true value, in so far as the history of Indian architecture is concerned, lies in the fact that they are the only known specimens of a form of Buddhist architecture which prevailed in the north of India for probably 1,000 years before they were commenced, and they are the incunabula of thousands of Hindu temples which were erected in the south of India during the 1,000 years that have elapsed since they were undertaken.

To those who are thoroughly familiar with the development of Buddhist architecture during its whole course, few things seem more self-evident than that the upper storeys of these viharas were in wood or some perishable materials, like the Kyongs of Burmah at the present day, and that their forms were pyramidal. It is owing, however, to the first named cause, that there is so much difficulty in making either of these propositions clear to those who have not studied the style in all the countries where it has been practised. The originals having all perished, we are left to the careless description of unscientific writers, or to suggestions derived from conventional copies, for our knowledge of what they once were. Still there are some indications which can hardly be mistaken. There is, for instance, Fa Hian's description of the great Dakshina vihara, cut, he says, in the rock. This building had five storeys. The lower was shaped into the form of an Elephant, and had 500 stone cells in it; the second was in the form of a lion, and had 400 chambers; the third was shaped like a horse, and had 300 chambers; the fourth was in the form of an ox, and had 200 chambers; the fifth was in shape like a dove, and had 100 chambers in it.2 We know perfectly what is meant by the various storeys being said to be in the forms of these animals, because we find them, as, for instance, at Halabîd3, superimposed one over another as string courses in the basement of that and other temples in the 13th and 14th centuries. The manner in which this is done there and elsewhere makes it evident that it was a custom in earlier times to adorn the successive storeys of buildings with figures of these animals, in the order enumerated. The point that principally interests us here is the pyramidal form this vihara is said to have assumed, as indicated by the diminished number of apartments in each storey.

The Lowa Maha Paya or great brazen monastery at Anuradhapura, is said, in the Mahawanso4to have been originally nine storeys in height, but after being utterly destroyed by Mahasena in 285 A.D., to have been re-erected by his son, but this time with only five storeys instead of nine. The forest of stone pillars, each about 12 feet in height, which once supported it, still remain, measuring in plan 250 feet each way, but no remains are found, among them, either of the primitive monastery destroyed by Mahasena, nor of the subsequent erection, which was allowed to go to decay when the city was deserted. This in itself is almost sufficient to prove that the materials of which the superstructure was formed were of a very perishable nature.

It is in Burmah, however, that we see the system carried out to its fullest extent; but even there it is now only a reflex of what it was in earlier times. There, however, all the Kyongs or Viharas, though generally supported, like the Lowa Maha Paya, on stone posts, have their superstructures, which are three, five, and nine storeys in height, constructed in wood, and all assume the pyramidal form. These differ, of course, from the earlier forms, but not more so than might be expected from their great difference of age. Perhaps, however, the best illustration, for those who know how to interpret it, is the temple of Boro Buddor, in Java. It is a nine-storeyed Vihara, converted from a residence for monks, into a temple for the reception of Buddhist images, and the display of Buddhist sculptures.5 It is nearly of the same age, perhaps slightly more modern than these Mahâvallipur Raths, and is a perfectly parallel example. In India it is an example of an earlier form, invented for utilitarian purposes, conventionalised into a temple for the worship of the divinities of a hostile religion. In Java of one as completely diverted from its original purpose, though for the glorification of that religion for which the Viharas were originally invented.

It was evidently owing to the perishable nature of the materials with which they were constructed that no remains of any of these many-storeyed Viharas of the Buddhists is now to be found in India. The foundations of several were excavated at Sarnath, near Benares. That one explored by Lieutenant Kittoe, and afterwards by M. Thomas6, was apparently only of one storey, the cells surrounding an open court; and the same seems to have been the case with another discovered in cutting through a mound in making the railway near Sultangunge7, on the Ganges; and it is a question how far these cloister courts—if they may be so called—were the models for some at least of the rock-cut Viharas in the west. Others, however, have been excavated by General Cunningham8, which were evidently the foundations of taller buildings, such as those described by the Chinese pilgrims, and more resembling the Mahâvallipur Raths in design. An opportunity occurred of ascertaining what their forms were when Mr. Broadley was authorised by the Bengal Government to employ 1,000 labourers to excavate what he supposed to be the Baladitya monastery at Nalanda.9 He published a plan of this, said to be the result of his excavations, in a pamphlet in 1872, and a restored elevation of the building in the Journal of the Bengal Asiatic Society for the same year (Vol. XLI); but in neither case is it possible to make out what he found, or what he invented, and his text is so confused and illogical that it is impossible from it to make the one agree with the other, or to feel sure of any of the results he attained. So far as can be made out, it was a five storeyed vihara, measuring about 80 feet square according to the text, though the scale attached to the plan makes it more than 100 feet, and the two lower storeys averaging about 12 feet each, were found to be nearly entire, the height of the ruins still standing being on the different sides 30 or 40 feet. There was a portico on the east with 12 pillars, which led to a cell 22 feet square, in which was found a headless statue of Buddha 4 feet in height. The second storey, 63 feet square, was set back 8 or 9 feet from the lower one, and the whole may have made up five storeys, with a height of about 70 feet, assuming the proportions to have been about those of the Dharmarâja Ratha just described. The upper storey may, however, have assumed a more spire-like form, as was the case in Burmah, und made up the total height of 100 feet, though this is still far from the height of 200 or 300 feet, which Hiuen Thsang ascribes to the building he saw.10

From a photograph it appears that the base, for a height of  about 5 feet, was adorned with courses of brickwork richly moulded, and above that with a range of niches 3 feet 3 inches in height, between pilasters 4 feet 6 high. These bore a cornice in moulded brick, with stucco ornaments making up altogether about 12 feet. Above this the whole exterior of the building seems to have been made up of wooden galleries attached to a plain central core of brickwork, in four or five offsets. It is now of course idle to speculate on what the appearance of these galleries may have been, and for our present purposes it is not of much consequence, inasmuch as an inscription found in its entrance, states that it was erected by Mahipâla, the third king of the Pâla dynasty, who, according to General Cunningham, reigned in Bengal from 1015 to 1040 A.D.11

It might at one time have been open to doubt whether this inscription was integral, and whether consequently the building was really erected by Mahipâla. The style of the architecture, however, and all the details of its ornamentation, as shown in the photograph, set that question quite at rest. The whole is comparatively modern, and  must have been erected during the reign of some king who was contemporary with that dynasty of Burmese kings who built and  ruled in Pagan between the years 850 and 1284.12 This being so, although a more complete knowledge of this building would be of the utmost importance in a general history of Indian architecture, it is evident from its date, that its peculiarities can have only a very indirect and retrospective bearing on an investigation into the form of its rock-cut temples.

Though the result of this Nâlanda investigation is certainly a disappointment, there still remains the celebrated temple of the Bôdhidruma at Buddha Gaya, which might at first sight be expected to throw considerable light on the subject. It is a nine-storeyed Vihara, and 80 far as is known the only one that ever was erected, wholly with permanent materials, by Buddhists in India, or at all events is the only one of which any remains now exist, and had it consequently been built by natives, it could hardly have failed to be of extreme interest. It is evidently, however, of a foreign design, as there is nothing in the same style in India, either before or after it, and nothing indeed at all like it, except a little temple dedicated to Târâ13 Bodhisattwa, close alongside of it, and part in fact of the same design. When, however, the thing is looked into a little more closely, it is evident that it does not require the Burmese inscriptions found on the spot14 to convince anyone at all familiar with the architecture of the East that the building now standing there was built by the Burmese in the 13th or 14th century of our era.15 It need hardly be added, if this is so, that all the controversies that have recently raged about the age and form of the arches which were introduced into its construction, fall to the ground with the foundations on which they rested.16 If the Nâlanda monasteries could be restored they would no doubt show a much greater affinity to those of Mahâvallipur than this one does; but its style having been elaborated in a foreign country, and under foreign local influences, we ought hardly to be surprised at it having assumed so, totally different an appearance during the seven centuries that elapsed between their erection.17 Had it been erected by Indians it probably would have taken much more of the form of the Tanjore pagoda, and the numberless examples of the Dravidian style to be found in the south f India. As it is, it is nearly a counterpart of the Bôdhidruma temple at Pagan18, erected by King Jaya Sinha between the years 1204 and 1227.19 The Burmese temple is, it must be confessed, a little broader in its base than that at Buddha Gaya, and its pyramid a little less steep, but this may have arisen from the architect in India being limited to the dimensions of the temple that existed there when Hiuen Thsang visited the place, and which he described as 20 paces—say 50 feet—square, which is very nearly that of the present temple. Its height, too, is nearly the same as of that seen by the Chinese pilgrim, 160 to 170 feet, but how that was made up it is extremely difficult to say.20Neither the Mahâvallipur Raths, nor any other authority we have, give us a hint of now, at that age, a building 50 feet square could have been designed so as to extend to between three and four times the height of its diameter.

As these Behar examples fail us so entirely it is very difficult to ascertain what other materials may exist in India to enable us to restore the external appearance of the tall Viharas of the Buddhists with anything like certainty. If it is decided that no structural remains exist, it only makes these Mahâvallipur Raths the more valuable in the eyes of the antiquary. They certainly approach in appearance more nearly to, what the ancient buildings were, from which they are copied, than anything else that has yet been discovered.

  • 1. Trans. R. A. S., vol. ii. Plates XVI. and XVII.; see also Carr’s compilation, p. 224, for Dr. Burnell’s transliteration, unfortunately without translation.
  • 2. Beal's Fa Hian, pp. 139, 140.
  • 3. History of Indian Architecture, p. 402, woodcut 226.
  • 4. Mahawanso, p. 163. See also Hist. of Indian Arch., p. 195.
  • 5. History of Indian Architecture, p. 643 et seq., woodcut 363.
  • 6. J. A. S. B., vol. xxiii. pp. 469 et seq. General Cunningham's Report, vol. i. Plate XXXII.
  • 7. J. A. S. B., vol. xxiii. pp. 360 et seq.
  • 8. Reports, vol. i. Plate XXXIII. pp. 120 et seq.
  • 9. I am unable to ascertain how far these excavations are coincident with those of Captain Marshall in 1871. The latter are described by General Cunningham, vol. i. of his Reports, p. 33, but he does not, so far as I am aware, allude to Mr. Broadley's either in this or a subsequent Report, in his third volume published in 1874, and the dimensions he quotes in describing this Vihara by no means agree with those given by Mr. Broadley. I have since the above was in type, received from Mr. Beglar, a photograph of the part uncovered by Mr. Broadley, but unfortunately taken from so low a point of view, as hardly to assist in understanding the form of the building. It is, however, sufficient to show how utterly worthless Mr. Broadley's drawings are, and to enable us to ascertain the date of the building with very tolerable certainty.
  • 10. Julien's Translation, vol. i. p. 160; vol. iii. p. 50. If the latter dimension is assumed as the correct one, as the Chinese foot is nearly 13 English inches, the Vihara must have been as high as the cross on the dome of St. Paul's.
  • 11. Reports, vol. iii. p. 134.
  • 12. Yule's Mission to Ava, pp. 32 et seq.; Crawfurd, pp. 111 et seq. of vol. i., 8vo. edition. It may be observed, there is a discrepancy of from 10 to 14 years in die dates or the kings' reigns quoted by Crawfurd and Burney, and those employed at the present day. This arises, as Sir Arthur Phayre informs me, from the Burmese having recently revised their chronology, with the aid of inscriptions and other data hitherto neglected, and adopted revised lists, in many instances showing differences from the old ones to the extent just stated.
  • 13. Hiuen Thsang, vol. iii. p. 51. The modern Hindus have converted this info Târâ Devî,an idea adopted by Rajendralâla Mitra, Buddha Gaya, p. 136, Plate XX., Fig. 1. Târâ is one of the favourite Saktîs of the modern Buddhists in Nepal. She is a Mahâyâna divinity associated with the Bôdhisatywas, and figures in the Nâsik, Elurâ, and Aurangâbâd caves.
  • 14. These inscriptions are given at full length, and with all the necessary details and translations in Rajendralâla Mitra's Buddha Gaya, p. 206 et seq.
  • 15. There is some little difficulty about the exact date of these inscriptions. According to Sir Arthur Phayre, who is probably the best authority on the subject, there are two dates. The first records the repairs or rebuilding of the temple by a Burmese king, A.D. 1106. The second its final completion and dedication by a king of Arakan, 1299 A.D., 193 years afterwards, during the reign of Nasiru'd-din, Sultan of Bengal. It is impossible now to discriminate between the parts that may belong to each of these two dates, or whether any parts of the older erection may be incorporated in the present building, but it seems quite certain that all its architectural features belong to the two centuries that elapsed between them. See Sir A. Phayre's paper, J. A. S. B., vol. xxxvii. p. 97.
  • 16. Mr. Beglar, General Cunuingham's assistant, has recently sent me home an account of certain arches of construction, which he has found inserted sporadically into certain brick buildings in Bengal. So far as I can make out from his photographs, all the temples or Topes in which these are found belong to the age of the Pàla dynasty, and are consequently posterior to the beginning of the 9th century. Some of them considerably more modern. This is only what might be expected, as we know from Yule's Mission to Ava, Plate 9, and other authorities, that arches, round, pointed, and flat, were currently used in the brick buildings at Pagan, between 850 and 1284 A.D., and this being so, it always appeared a mystery to me that none were found in contemporary buildings in Bengal. One advantage of Mr. Beglar's discoveries is, that they tend to show that there was a considerable intercourse between Bengal and Burmah in these ages. This, however, has always been suspected though difficult to prove, and every step in that direction is consequently welcome, besides removing to a great extent, any difficulty that, might be felt in believing that the Buddha Gaya temple was actually erected by the Burmese.
  • 17. In his work on Buddha Gaya, Babu Rajendralâla Mitra adduces the form of the temple at Konch (Plate XVIII.) in support of his theory of the Buddha Gaya temple. It would, however, be difficult to find two building so essentially different as these are. That at Konch is a curvilinear spire of the Northern Aryan or Bengal style; that at Buddha Gaya is a straight lined many-storeyed pyramid, deriving its form from those of the ancient Buddhist Viharas. The only advantage that can be derived from their juxtaposition is to prove that they were built by different people ; at distant times, and for dissimilar purposes; there is absolutely no connexion between them.
  • 18. In a private letter to me Sir Arthur Phayre says that when he first saw the Buddha Gaya temple, he at once came to the conclusion, from the style of its masonry and whole appearance, that it must have been erected by the Burmese, and no one probably is a better judge and more competent than he is to give an opinion on the subject. —J.F.
  • 19. Crawfurd’s Embassy to Ava, vol. i., p. 117, 8vo. edition.
  • 20. In his work on Buddha Gaya, at pp.204 et seq., Babu RajendralâIa proves beyond all cavil, that the famous inscription which Sir Charles Wilkins published in the first vol. of the Asiatic Researches is a manifest forgery. The fable, consequently, that this tower was erected by the Brahman Amara, one of the jewels of the court of Vicramâditya in the sixth century, is shown to have no foundation in fact, and must be relegated to the company of many others which have been invented to account for the exceptional appearance of this celebrated tower. It is curious, however, that the Babu does not see how completely his learning upset his own theories of the history of the temple.