To the artist or the architect the group of caves situated on the Udayagiri hill in Orissa is perhaps even more interesting than those in Behar just described, but to the archaeologist they are less so, from the difficulty of fixing their dates with the same certainty, and because their forms have not the same direct bearing on the origin or history of the great groups of caves on the western side of India. Notwithstanding this, the picturesqueness of their forms, the richness of their sculptures and architectural details, combined with their acknowledged antiquity, render them one of the most important groups of caves in India, and one that it is impossible to pass over in such a work as this, without describing them in very considerable detail.

The caves in question are all situated in a picturesque and well wooded group of hills that rise out of the level plains of the Delta of the Mahânaddi, almost like islands from the ocean. Their composition is of a coarse sandstone rock, very unusual in that neighbourhood, but which from that circumstance offered greater facilities for their excavation than the laterite rocks with which the country everywhere abounds. Their position is not marked on any of the ordinary maps of the country, but may easily be fixed, as their bearing is 17 miles slightly to the east of south from Katak, and 4 miles north-west from Bhuvaneśwar. The great Saiva temple of that city, one of the oldest and finest in India, being easily discernible from the tops of the hills in which the caves are excavated.

Besides the facilities for excavation, there were probably other motives which attracted the early Buddhist hermits to select these hills as their abode and continue to occupy them during three or four centuries at least. We may probably never be able to ascertain with accuracy what these reasons were, or how early they were 80 occupied: We know, however, that Aśoka about the year 250 B.C. selected the Aśwatama rocks, near Dhauli, about 6 miles south-east from these hills, as the spot on which to engrave one of the most complete and perfect sets of his series of edicts1 and he hardly would have chosen so remote a corner of his dominions for this purpose, had the place not possessed some previous sanctity in the eyes of his co-religionists. Unfortunately we are not able to fix with anything like certainty the site of Danta-puri, the city in which the celebrated Tooth Relic was enshrined, and where it remained till carried off to Ceylon in the beginning of the fourth century of our era.2 It certainly was not far from this, and may have been in the immediate vicinity of the caves, though the evidence, as it at present stands, seems to favour the idea that it was at Puri where the famous temple of Jagannâth now stands, some 30 miles south of the caves. The fact, however, that it is recorded by the Buddhists that the Tooth Relic was brought to this neighbourhood immediately after the cremation of his body, and the certainty of its being chosen by Aśoka B.C. 250 to record his edicts, is sufficient to show that early in the history of that religion this neighbourhood was occupied by Buddhists. There is however no record or tradition of Buddha himself ever having visited the locality, or of any event having occurred there that gave rise to the erection of any Stûpa or other monument in the neighbourhood, and even Hiuen Thsang, when passing through the country in A.D. 640, does not mention any spot as sanctified by the presence or labours of Buddha or of any of his immediate disciples.3

There are some 16 or 17 excavations of importance on the Udayagiri hill, besides numerous little rock-cut hermitages-cells in which a single ascetic could dwell and do penance. All these belong to the Buddhist religion and there is one Buddhist cave in the Khandagiri hill—the Ananta. The others there, though large and important, are much more modern and all belong to the Jaina form of faith. There is also a modern Jaina temple built by the Marâthas on the top of that hill, and I cannot help believing that Kittoe was correct when he says that there has been a large circular building on the corresponding summit of the Udayagiri rock;4 but I have not been able to ascertain for certainty whether the foundations still to be seen there are either ancient or in the form of a dâgoba.

These caves were first noticed and partially described by Stirling in his admirable account of Cuttack, in the 15th volume of the Asiatic Researches published in 1824, and that was the only authority existing when I visited them in 1836. At that time, however, all the more important caves were occupied by Fâkirs and Bairagis who violently resented intrusion on their premises, and besides my time was too limited for any elaborate examination of the whole. In 1838 they were visited by Lieut. Kittoe, and his account, with the drawings that accompanied it, published in the seventh volume of Prinsep’s Journal for 1838, still remains the best account of these caves yet given to the world. His visit, however, like mine, was too hurried to enable him to make plans and draw details, while in his time, as in mine, the caves were still inhabited; otherwise with more leisure and better opportunities he would have left little to be done by his successors. Since then the caves have been photographed by Col. Dixon, Mr. Murray, and others, but without descriptions or plans, so that they are of very little use for our present purposes.5

In attempt to investigate the history of these caves, it is tantalizing to discover how narrowly we have missed finding in Orissa a chronicle of events during the whole Buddhist period as full, perhaps even more so, than those still found in Kashmir, Ceylon, or any other outlying provinces of India. It is true that the palm leaf records of the temple of Jagannâth at Puri, in which alone the fragments of this history are now to be found, date only apparently from the 10th century, and it would be idle to look in a work compiled by Brahmans at that time for any record of the acts, even perhaps of the names, of Buddhist kings of that country, still less of their building temples or excavating caves, devoted to the purposes of their—to Brahmans—accursed heresy. Notwithstanding this, if we possessed a continuous narrative of events occurring in the province we might be able to interpolate facts so as to elucidate much that is now inexplicable and mysterious.6

What these palm leaf records principally tell us is, that from a period vaguely contemporary with Buddha, i.e., from 538-421 B.C. till 414 A.D., in fact, till Yayati Kesari finally expelled the Buddhists and established the Brahmanical religion in Orissa, the country was exposed to frequent and nearly continuous invasions of Yavanas generally coming from the north west.7 Who these Yavanas were it is nearly impossible to say. The name may originally have been applied to Greeks or Romans, but it afterwards was certainly understood as designating all who, from an Indian point of view, could be considered as foreigners or outside barbarians, and so it must be understood in the present instance.

The account of these Yavana invasions in the Puri Chronicle looks at first sight so strange and improbable that one might almost be inclined to reject the whole as fabulous, were it not that the last of them, that under Rekta Bahu, which Stirling looked upon as so extraordinary and incomprehensible,8 has by the publication by Turnour of the Daladawansa9, been elevated to the dignity of an established historical fact10, and there seems no difficulty in believing that the others may be equally authenticated when more materials are accumulated for the purpose.

It is of course impossible to form an opinion as to what reliance should be placed on the facts narrated in these palm leaf records till we see what the text is, in which they are imbedded.11 All that at present can be said regarding them is that they are curiously coincident with what we know, from other sources, of the introduction of Buddhism into Orissa, and with the architectural history of the province. In the present state of our knowledge it is equally difficult to say how far we may place any dependence on the tradition that immediately after his death, the relics of his body were rescued from the funeral pyre and distributed to eight different cities in India.12 According to these accounts the left canine tooth fell to the lot of Orissa, and was received by a king named Brahmadatta, whose son named Kâśi and grandson Sunanda continued to worship and hold it in the greatest possible respect.13 These names, however, do not occur in any lists that have come down to our time, and the first, as king of Benares (Kâśi), occurs so frequently in Buddhist legends and jâtakas that no reliance can be placed in any tradition regarding him or his acts, as being authentic history. The second name looks like the name of his capital, and the third as one of the many Nandas who figure in the history of Magadha before the time of Aśoka. Be this, however, as it may, it seems tolerably certain that a tooth, supposed to be that of Buddha, was enshrined in this province in a magnificent Chaitya, in a city called Dantapura from that circumstance, before Aśoka’s time, and remained there till the beginning of the fourth century A.D., when it was conveyed to Ceylon under the circumstances narrated in the Daladawansa, and where it now remains the palladium of that island under British rule.14

What we gather, from all this practically is, that Yavanas from the north-west, probably bringing Buddhism with them, invaded Orissa before the time of Aśoka, and consequently before the first rock-cut temple was excavated. It seems also nearly certain that Orissa remained Buddhist, and the tooth relic was honoured there—intermittently it may be by the kings—but certainly by the people, down to the year 322 AD.15 when it was transferred to Ceylon, and subsequently to this, that the province remained Buddhist under the last Yavana dynasty, 328 to 474 A.D., when that religion was finally abolished by the Késari dynasty of kings.

There is no evidence that this last dynasty excavated any caves, and as there are no remains of any structural buildings belonging to the Buddhist religion, in the province, our history halts here, and there is at present nothing to lead us to believe that any of the caves were excavated within even a century before 322. The architectural history of the province, in Buddhist times is consequently, it must be confessed, very incomplete, and all that remains to be done is to try and find out wh en the earliest cave was excavated, and then to trace their development, so far as it can be done, till the time when cave digging ceased to be a fashion in Orissa.

~ * ~

As just mentioned, history will hardly help in this. Such records as we have, were written, or rather compiled, by Hindus, haters of Buddhism, and not likely to mention the names of kings belonging to that sect, and still less to record any of their actions or works. Inscriptions hardly give us greater assistance. It is true about one half of the caves at Udayagiri do bear inscriptions, but none of them have dates, and none of the names found in them have yet been identified with those of any king who figures in any of our lists. What they do tell us, however is, from the form of the characters employed that all the inscribed caves are anterior to the first century B.C. Unfortunately, however, the two principal and most interesting caves, the Râni kâ Nûr and the Ganesa Gumpha, have no contemporary inscriptions, so that this class of evidence for their age, is not available. There remains consequently only the evidence of style. For that, fortunately, the materials are abundant, and the testimony is as complete as could well be expected. We have at least three monuments, whose date we may say is known with sufficient certainty for our purposes, and which, as we shall presently see, were almost as certainly contemporary with these caves.

The first of these is the rail which Aśoka (B.C. 250) is said to have erected round the Bodhi tree at Buddha Gaya. Very little of it remains, and none of it in situ, still there is enough of it existing to show exactly what the style of sculpture was at that age. Unfortunately, however, it has never been photographed, or at least no photographs of it, except of one fragment, have reached this country, and the drawings that have been published are very far from being satisfactory. The best set of drawings yet made were by Major Markham Kittoe, more than thirty years ago. They are now in the library at the India Office, but have never been published. Those in General Cunningham’s “Reports” are far from complete16, and by no means satisfactory, and the same may be said of the set engraved by Babu Rajendralâla Mitra, in his work on Buddha Gaya17, just published. Fortunately the latter does give one photograph of one gate pillar (Plate L.), but whether taken from a cast or from the stone itself is not clear. Whichever it is, it is the only really trustworthy document we have, and is quite sufficient to show how little dependence can be placed on General Cunningham’s representation of the same subject, and by implication on the drawings made by A.P. Bagchi for the Babu’s work, which are in no respect better than the General’s, if so good. It would of course be a great advantage if a few more of the sculptures had boon photographed like the pillar represented on Plate L., but it, though it stands alone, is quite sufficient to show what the style of sculpture was which prevailed in the third century B.C., when it was erected.

The Bharhut Tope, which is the second in our series, has been much more fortunate in its mode of illustration. All its sculptures have been photographed by Mr. Beglar and published with careful descriptions by its discoverer, General Cunningham.18 The date, too, has been assumed by him to be from 250 to 200 B.C. on data which are generally supposed to be sufficient for the purpose. I would suggest, however, that as this date is arrived at principally by calculating backwards at a rate of 30 years per reign from Dhanabhûti II, and as 16 years on the average is a fairer rate, it may be placed by him at least 50 years too early; the more especially as even that king’s reign is only determined from a slight variation in the form of the letters used in the inscriptions, which is by no means certain.19 On the whole, I fancy 200 to 150 B.C. is a safer date to rely upon in the present state of our knowledge. For myself I would prefer the most modern of these two dates as the most probable. It is, at all events, the one most in accordance with the character of the sculpture, which is, as nearly as may be, half way between those of the rail at Buddha Gaya, and those found on the gateways at Sanchi.20

The Sanchi Tope, which forms the third of the series, has also been illustrated with all the detail requisite for a proper understanding of its historical and artistic position. In the first place, we have General Cunningham’s work on the subject published in 1854, which is the foundation of our historical knowledge of this tope, to which may be added an extensive series of photographs by Captain Waterhouse, made in 1862. We also possess a beautiful series of drawings by Colonel Maisey; and in addition to an exhaustive transcript of its sculptures, by Lieutenant Cole21, there are also the casts he brought home, and copies of which are now in the South Kensington and Edinburgh Museums.

From all these data the date of this monument has been ascertained with sufficient precision for our present purposes at least. The southern gateway, which is the earliest, seems to have been erected by a king who reigned between the 10th and the 28th year of the Christian era, and the other three gateways during the remaining three-quarters of that century.22

There is still a fourth building equally important for the general history of architecture in India, though not bearing so directly as that of the caves in Orissa as the other three. The principal sculptures of the tope at Amrâvatî were executed during the course of the fourth century of our era23, and are perhaps the most beautiful and perfect Buddhist sculptures yet found in India, and as such full of interest for the history of the Art. It cannot, however, be said that any of the sculptures in the caves at Udayagiri are so modern as they are, but this being so, marks at all events the limit beyond which the Orissan caves cannot be said to extend. On the other hand, with our imperfect knowledge of the Buddha Gaya rails it is not easy to determine whether any of these caves are really so old as the time of Aśoka. From a comparison of their details we may, however, feel certain that some of these caves are certainly contemporary with the rail at Bharhut, others with the gateways at Sanchi. Although we cannot fix the limit either way with absolute certainty, we may feel confident that all those which are most interesting from an architectural point of view, were excavated during the three and a half centuries which elapsed between the years 250 D.C. and 100 A.D. Some of the smaller and ruder examples may be earlier, but none of them have any characteristics which would lead us to assign them to a more modern epoch than that just quoted.

  • 1. J.A.S.B., vol. XII, p. 436, for Kittoe’s plates and description of the locality.
  • 2. J.R.A.S., vol. III, new series, pp. 149 et seq.
  • 3. Julien, vol. I. 184; III. 88.
  • 4. J.A.S.B., xii. p. 438. In a private letter from Mr. Phillips, the joint magistrate of the district, he informs me “there are the remains of some building above the Rani ka nour, i.e., on the top of the Udayagiri.” It probably would require excavation to ascertain its character.
  • 5. Some 10 years ago an opportunity occurred, which had it been availed of, would have gone far to remedy the deficiency of former explorers, and to supply an exhaustive account of these caves. In 1868-89 Babu Rajendralâla Mitra conducted an expedition for that purpose, accompanied by a staff of draughtsmen and students in the school of art at Calcutta, who were to be employed in making drawings and casts of the sculptures. Their labours, however, were almost exclusively directed to the temples at Bhuvaneswar, he himself making only personal notes of the caves. In consequence of this, mainly, if not wholly, in consequence of reclamations, made by me on the subject, a second expedition was sent down by the Bengal Government in the cold weather of 1870-71. This was conducted by Mr. C.C. Locke of the Government school of art, and resulted in his bringing back plans of all the principal caves and casts of all the more important sculptures. These were placed in Babu Rajendralâla’s hands for publication, which, however, he has not yet found it convenient to carry into effect, but meanwhile I have received photographs from the casts, and plans of the caves from Mr. Locke, and these form the basis of all our real knowledge of the subject, and what is most relied upon in the following descriptions. (Two of the plans were published in my History of Indian Architecture, woodcuts 70 and 72, and five of the casts in my Tree and Serpent Worship, Plate C., published in 1873).

    Through the kindness of his friend, Mr. Arthur Grote, late B.C.S., I have been permitted to see the corrected proofs of the first 56 pages or the 2nd volume of Babu Rajendras’ Antiquities of Orissa, which contains his account of these caves, with the accompanying illustrations, but under a pledge that I would not make any quotations from them, as it is possible the Babu may yet see fit to cancel them, or at all events modify them to some extent before publication. This, for his own sake, I trust he will do, for as they now stand they will do him no credit either as an archaeologist or a controversialist, and he will eventually be forced to retract nearly all he has said in the latter capacity. So far as I am capable of forming an opinion on the subject, the conclusions he arrives at as to the age of the caves are entirely erroneous, and he does not pretend that his explanations of the sculptures are derived either from local traditions, or Buddhist literature, merely that they are evolved from his own inner consciousness. Others may form a different opinion from that I have arrived at regarding his interpretation of the scenes depicted in them; to me they appear only as an idle waste of misplaced ingenuity and hardly worthy of serious consideration. —J.F.

  • 6. These chronicles were very largely employed by Stirling in his History of Orissa and Cuttack, in the 15th volume of the Asiatic Researches, and still more extensively by Mr. Hunter in his Orissa, published in 1872, vol. i. pp. 198 et seq. They were also further investigated by a Calcutta Brahman Bhawanicharan Bandopadhyaya, in a work he published in Bengali, in 1848, entitled Purushottama Chandrika, which was very largely utilised by W. W. Hunter in his last work on Orissa, vol. i. p. 198 et seq.
  • 7. The following chronological account of Yavana invasions is abstracted from Mr. Hunter’s Orissa, vol. ii. p. 184 of the Appendix :

    B.C.538-421. Bajra Deva. —In his reign Orissa was invaded by Yavanas from Marwar, from Delhi, and from Babul Deś, the last supposed to be Iran (Persia) and Cabul. According to the palm leaf chronicle the invaders were repulsed.

    B.C. 421-306. Narsingh Deva. —Another chief from the far north invaded the country during this reign, but he was defeated and the Orissa prince reduced a great part of the Delhi kingdom.

    306-184. Mankrishna Deva. —Yavanas from Kashmir invaded the country, but were driven back after many battles.

    184-57. Bhoj Deva. —A great prince who drove back a Yavana invasion, and is said to have subdued all India.

    Here follows the usual account of Vicramâditya and Śalivâhana, and we hear no more of the Yavanas till-

    A.D. 319-323. Sobhan Deva. —During this reign of four years, the maritime invasion and conquest of Orissa by the Yavanas under Rekta Bahu, the Red-armed, took place. The king fled with the sacred image of Jagannâth (the Brabmanical synonym for the tooth relic), and with those of his brother and sister Balbhadra and Subhadra, and buried them in a cave at Sonpur. The lawful prince perished in the jungles, and the Yavanas ruled in his stead.

    323-328. Chandra Deva, who, however was only a nominal king, as the Yavanas were completely masters of the country. They put him to death 328 A.D.

    328-474. Yavana occupation of Orissa 146 years. According to Stirling these Yavanas were Buddhists.

    474-526. Yayati Kesari expelled the Yavanas and founded the Kesari or Lion dynasty. This prince brought back the image of Jagannâth to Puri, and commenced building the Temple City to Śiva at Bhuvaneswar.

    After this we hear no more of Yavanas or Buddhists in Orissa. The Brahmanical religion was firmly established there, and was not afterwards disturbed till the invasion of the Mahomedan Yavanas from Delhi, repeated the old story in 1510.A.D.

  • 8. Asiatic Researches, vol. xv. p. 263.
  • 9. J.A.S.B., vol. vi. p. 856 et seq.
  • 10. Journal R. A. S., New Series, vol. iii. p. 149 et seq.
  • 11. A golden opportunity for effecting this was presented by Babu Rajendralâla’s mission to Katak in 1868-69. As a Brahman he had access to the temples and their treasures to an extent that could not be afforded to any Yavana inquirer, and indeed be seems to have intended to have transcribed and translated them (Hunter’s Orissa, vol. i. p. 198, note), but his ambition to be considered an archæologist of the European type, led him to neglect a task for which he was pre-eminently fitted, and to waste his time instead, in inventing improbable myths to explain the sculptures in the caves.
  • 12. Journal Asiatic Soc. Of Bengal, vol. vii.; p. 1014; Foë Kouë Ki, 240.
  • 13. Turnsour’s account of the Daladawansa, J.A.S.B., vol. vi., p. 856 et seq.
  • 14. I have already detailed so fully the circumstances under which the transfer took place in a paper on the Amrâvatî tope, which I read to the Asiatic Society in 1847 J.R.A.S., vol. iii. N.S., pp. 132 et seq., that I may be excused repeating what I then said. The particulars will also be found, Tree and Serpent Worship, pp. 173 et seq.
  • 15. There is a discrepancy here of about 10 years between the dates in the Orissan chronicles and those derived from the Mahâwanso according to Turrnour. On the whole I am inclined, from various collateral pieces of evidence, to place most reliance on that derived from the Puri chronicles.
  • 16. Reports, vol. i. Plates VIII. to XI.; vol. iii. Plates XXVI. to XXX.
  • 17. Buddha Gaya, Plates XXXIII. to XXXVIII.
  • 18. The Stupa of Bharhut, by General A. Cunningham, London, 1879.
  • 19. The Stupa of Bharhut, pp. 15 and 16.
  • 20. From the great similarity that exists between the alphabetical characters found at Bharhut, and those employed by Aśoka in his numerous inscriptions, General Cunningham was no doubt perfectly justified in assuming that the stupa’s age could not be far distant from that of his reign. At the same time, however, almost as if to show how little reliance can he placed on Palæographic evidence alone, where extreme precision is aimed at, and no other data are available, he quotes an inscription found at Mathura recording some gifts of a king of the same name, whom he calls Dhanabhûti II, and joins the two together in his genealogical list, with only one name, that of Vâdha Pâla, between them. (Stupa at Bharhut, p. 16.)

    When General Cunningham first published this Mathura inscription (Reports, III, p. 36, Plate XVI) he placed it in chronological series, between one dated Samvat 39 and another dated Samvat 135, and from the form of its characters he was no doubt correct in so doing, more especially as in Plate XIV of the same volume, he quotes another inscription or Huviskha dated Samvat 39, where the alphabet used is very little, if at all earlier. If the Samvat referred to in these inscriptions was that (If Vikramâditya, as the General assumes, this would place this second Dhanabhûti about A.D. 50 or 60. But as it seems certain this era was not invented at that time, it must be Salta, and accordingly he could not have reigned before the end of the second century of our era, and his connexion with the Bharhut Stûpa is out of the question.

    Another point that makes the more modern date extremely probable, is that the sculpture on the Mathura pillar represents the flight of the prince, Siddhârtha, with the Gandharvas holding up the feet of his horse in order that their noise might not awaken the sleeping guards (Stupa at Bharhut, p. 16). As General Cunningham knows, and admits, no representations of Buddha, are found either at Bharhut or Sanchi (Stupa at Bharhut, p. 107), and this legend, though one of the most common among the Gandhara sculptures, does not occur in India, so far as is at present known, before the time of the Tope at Amrâvatî in the fourth century (Tree and Serpent Worship, Plate LIX, fig. 1.), and consequently this sculpture cannot certainly be earlier than the second century A.D., and may be much more modern. It is just possible, no doubt, that it may not be integral, but may have been added afterwards when the larger rails were inserted, which cut through the inscription. This, however, is hardly probable, but until this is explained all the evidence, as it now stands, tends to prove that this Mathura inscription is much more likely to be 200 years after Christ instead of 200 before that era, as General Cunningham seems inclined to make it.

  • 21. All these have been utilised, and form the first 45 plates of my Tree and Serpent Worship, published in 1873, second edition.
  • 22. Tree and Serpent Worship, p. 99.
  • 23. Tree and Serpent Worship, Plates XLVI to C. (For dates see p. 178.) probably from about A.D. 322-380