This practice of setting back the upper storey may have been introduced here from the nature of the rock, and been intended to give more strength to the lower storey by relieving it from the pressure of the upper. My impression however is, that it was adopted in consequence of the Buddhist Viharas of that age—as will hereafter be explained—being, when of more than one storey in height, of a pyramidal form, each storey being consequently less in diameter than the one below it. This cave and the Vaikuntha are evidently intended to represent three sides of a structural Vihara turned inside out, to accommodate them to the nature of the material and situation in which they are excavated, all the dimensions, both in plan and section, being consequently reversed. If the wings could be wheeled back 180 degrees to first side—the principal one now standing—they would with it, form the three sides of a free standing Vihara. It is impossible to represent the fourth side or back, from its situation, in a rock-cut example. Supposing this to be the motivo of the design it appears to explain all the peculiarities of this cave. It is only necessary to assume that it is a copy of a. structural Vihara, 63 feet square at its base or lower storey, with 43 in the upper storey, and intended to have a third probably of 20 or 23 feet square. In this case the two little highly ornamented pavilions in the angles of the lower storey (shown in the plan), would represent the angle piers in which I fancy the staircases were situated in structural examples. All this, however, will be clearer when we come to describe the Raths at Mahavallipur, which are the only examples we possess showing what the external form of Viharas really was in ancient India.

The verandah in the upper storey is 63 feet in length, and opens into four cells of somewhat irregular form, by two doors in each, making eight doorways altogether. The lower verandah is only 43 feet long, and opens into three cells, the central one having three doors, the lateral ones only two each. In a structural Vihara these dimensions would of course be reversed: the upper storey being of course the smallest. Of the pillars in the upper verandah only two now remain out of nine that originally existed, and these are very much ruined, but their forms can easily be recovered from the antæ at either end. None of the pillars of the lower verandah now exist, nor can I learn if any, even of their foundations, are to be found in situ. Certain it is, however, that whether as a part of the original design, or in consequence of an accident, the roof of this lower verandah was at one time framed in wood, as shown in the diagram.1

It will be observed that the upper part of the rook forming part of the roof of the upper verandah has fallen, and carried away the pillars that at one time supported it, and the fall of such a mass may at the same time have broken through the roof of the lower verandah and caused it to be replaced in wood. Except from the form of the two antæ at either end of the range of columns, I would be inclined to believe it was originally of wooden construction; but they are so essentially lithic in their forms that the wood seems to be a later adaptation. In the earlier Vaikuntha, which, though on a smaller scale, seems to have been the model on which this one was formed, the whole is in stone, which to some extent favours the idea that this wooden verandah was a subsequent repair. In consequence, however, of its decay and destruction, which was sure to happen early in such a climate, the lower range of sculptures have from long exposure become so weather-worn as to be nearly undistinguishable. They may also have suffered from the original fall of the rock, while the upper sculptures are still partially protected by its projection, and consequently are much more perfect, and in them, as just mentioned, resides the main interest of the cave. They are in fact the most extensive series of sculptured scenes to be found in any rock-cut examples of their age. In the western caves such scenes or ornaments as are here found, were either painted on plaster or carved in wood, but on this side of India, we know from what is found at Buddha Gaya and Bharhut, the prevailing fashion in that early age was to execute these things in stone, and consequently these sculptures, even in their ruined state, are full of interest to the history of cave architecture. They are far more extensive than in any of the caves of this group previously examined, and unlike them, instead of being confined to the tympana over the doors, are placed between them, so as to form a nearly continuous frieze, merely interrupted by the semicircular heads of the doorways.

The first question that arises on examining these sculptures is, Are they Buddhist? If they are, they are in some respects unlike any others belonging to that religion we are acquainted with. We do not, of course, at that early age, expect to find any conventional representation of Buddha himself, nor even to be able to detect such scenes from his life as that represented on the Sanchi Tope.2 But there is an almost total absence of all the Buddhist symbols, or objects of worship, which we find in the Ânanta, the Jaya Vijaya, or Vaikuntha caves, and with which we have become so familiar from the sculptures at Bharhut or Sanchi. I fancy I can detect the Triśula and Shield over two doorways3, but there certainly are no dagobas, no wheels, nor are there any trees as objects of worship, and Śrî too is absent. In fact, there is nothing essentially Buddhist about the cave; but if this is so, it is equally certain that there is nothing that savours of the Brahmanical religion. There are no many-armed or many-headed figures, and no divinities of the Hindu Pantheon can be recognised in the sculptures, nor anything that can indicate that the caves were Jaina. We are consequently forced to the conclusion that they must represent scenes from the Buddhist Jâtakas, or events occurring among the local traditions of Orissa. The latter is, however, so improbable, that my conviction is that the solution will be found in the Jâtaka; but out of the 505 births therein narrated only a few have been published, and these with so many variants that it is frequently very difficult to recognise the fable, even when the name is written over it, as is so frequently the case at Bharhut, and it consequently becomes almost impossible to do so when we have no such indications to help us.4

In a monograph of the caves in Katak, it might be expedient to describe the sculptures of the Râni kâ Nûr in detail, but even then it would hardly be possible to render their story intelligible to others without publishing at the same time the photographs from the casts made from them by Mr. Locke in 1871-2. These have been entrusted to Babu RajendraIâla Mitra for publication5, and when given to the world it may be worthwhile to go more carefully into the subject. At present, it may be sufficient to indicate their general character.

The frieze occupying the upper part of the verandah of the upper storey is divided by the heads of the eight doorways into seven complete and separate bassi rilievi with two half ones at the ends. The latter, which are about the best protected from the weather, are occupied by two running figures with their faces turned towards the centre; the one on the left bearing a tray, apparently with offerings, while the corresponding figure at the other end carries a wreath, such as that which forms the frieze of the outer rail at Amrâvati6, only of course on a much smaller scale.

The first bas-relief between the doors represents three very small elephants issuing from a natural rocky cavern, apparently to attack a man (query, giant), who is defending himself with an enormous club, worthy of Hercules. On his right hand in front of him is a Yakkhinî, known by her curly locks, standing on end, and behind him are a number of females either seeking shelter in various attitudes of consternation, or by their gestures offering to assist in repelling the attack. If this is meant for history, it probably represents some episode in the story of the conquest of Ceylon by Vijaya, which is a very favourite subject with Buddhist artists, and where elephants with Yakkhos and Yakkhinîs always perform important parts. It is one too of the most likely subjects to be depicted in these caves, as it is always from this country of Kalinga that the conquest of that island is said to have originated.7 But it may be some Jâtaka to whose interpretation we have no clue, and regarding which it is consequently idle to speculate.

The second bas-relief (Plate I., fig. 4) is certainly the most interesting of the series, not only because it is one of the best preserved, but also because it is repeated without any variation in the incidents, though in a very different style of sculpture, in the Ganeśa cave, to be next described. This bas-relief contains eight figures, four males and four females, in four groups. The first represents a man apparently asleep in the doorway of a hut, and a woman sitting by him watching. In front of these is a woman leading a man by the hand apparently to introduce him to the first pair. Beyond these, on the right, a man and woman are engaged in mortal combat with swords of different shapes, but both bearing shields of very unusual form, which I have never seen elsewhere. Beyond these, on the extreme right, a man is carrying off in his arms an Amazonian female, who still carries her shield on her arm, though she has dropped her sword, and is pointing with the finger of her right hand to the still fighting pair. Here again the first suggestion is Ceylon, for nowhere else, that I know of, at least, do Amazons figure in Buddhist tradition. But they are represented as defending Ceylon against the invasion from Kalinga in the great fresco in Cave XVII. at Ajaṇṭâ, engraved by Mrs. Speir in her Ancient Life in India, and repeated further on in a woodcut in the second part. It is by no means impossible that this bas-relief may represent an episode in that apocryphal campaign. It may, however, from its being repeated twice in two different caves, be some local legend, and if so the key will probably be found in the palm leaf records of the province, whenever they are looked into for that purpose, which has not hitherto been done. If not found there, or in Ceylonese tradition, I am afraid the solution may be difficult. It does not look like a Jâtaka. At least there is no man in any of these four groups whom we can fancy could have been Buddha in any former birth. But nothing is so difficult as to interpret a Jâtaka without a hint from some external source.

The third compartment I have very little doubt contains a representation of one of the various editions of the Mṛiga or Deer Jâtaka; not exactly that narrated by Hiuen Thsang8, nor exactly that represented at Bharhut9, but having so many features in common with both, that it seems hardly doubtful the story is the same. The principal figure in the bas-relief is undoubtedly a king, from the umbrella borne behind him and the train of attendants that follow him. That he is king of Benares is also probable, from his likeness to the king represented at Sanchi in the Sama Jâtaka.10 The winged deer is almost certainly the king of the herd, who was afterwards born as Buddha, but whether the second person represented is the king repeated, or some other person,—as would appear to be the case at Bharhut,—I am unable to guess. The deer at his feet is probably the doe who admitted that her turn to be sacrificed had come, but pleaded that she ought to be spared in consequence of the unborn fawn she bore in her womb, whose time had not yet arrived. I am unable to suggest who the woman in the tree may be. I know of no Dêvatas or female tree divinities elsewhere, though there may have been such in Orissa.

The fourth, which is the central compartment, is the only one in which anything like worship can be traced, but at its right hand corner, though much injured, I think we can detect something like a miniature dagoba or relic casket with someone praying towards it, and above a priest or someone seated in the cross-legged attitude afterwards adopted in the statues of Buddha. To the left of these is a figure in an attitude sometimes found at Amrâvatî, bearing a relic.11 It is difficult to say who the great man or woman is who is seated further to the left and surrounded with attendants. He or she is evidently the person in whose honour the puja or worship in the right hand corner is being performed, but who these may be must be left for future investigation.

The next compartment is so completely destroyed that no cast was taken of it, and its subject cannot of course be ascertained. The following one, however, containing three couples with possibly a fourth—for the right-hand end is very much ruined—at once calls to mind the scenes depicted at Sanchi on Plate XXXVIII. of Tree and Serpent Worship. The first pair are seated on a couch, the gentleman with his arms round the lady's waist, and a wine bottle on the ground in front of them. In the second group the lady is seated on the gentleman's knee, and there is a table with refreshments before them. The third it is difficult to describe, and the fourth is too nearly obliterated—if it ever existed—for anything to be made out regarding it.12

The seventh bas-relief is partially destroyed and was not cast.

As it at present stands, the evidence derived from these bas-reliefs is too indistinct to admit of any theory being formed of much value regarding their import. It looks, however, as if the first, the third, the fifth and seventh were Jâtakas, while the even numbers—the remaining four—represented local legends or scenes in the domestic life of the excavators of the cave.

Several of the reliefs on the front of the lower storey were cast by Mr. Locke, but they are so fragmentary and so ruined by exposure to the weather, that no continuous group can be formed out of any of them, nor can any connected story be discerned either of a legendary or religious character. Whether on the spot in the varying lights of the day, anything could be made out of them it is impossible to say, but neither the photographs nor the casts give much hope of this being done. They seem to represent men and women following their usual avocations or amusements, and certainly nothing can be discerned in them that illustrates either the religion of Buddha, or the history of the country.13

This fortunately cannot be said of the sculptures on the right-hand wing, where they are perfectly well protected from the weather by a verandah 8 feet in depth. This leads through three doors into an apartment measuring 7 feet by 20, on the front of which there is consequently space for two full and two half compartments, which are filled with sculptures. In the left-hand half division, a man and his wife are seen approaching the centre with their hands joined in the attitude of prayer. Behind them is a dwarf, and before them a woman bearing offerings. In the corresponding compartment at the other end of the verandah, three women—one may be a man—and a child are seen bearing what may also be offerings. The left-hand full compartment is occupied by a woman dancing under a canopy borne by four pillars, to the accompaniment of four musicians, one playing on a flute14, another on a harp, a third on a drum, and a fourth apparently on a Vina or some guitarlike instrument. In the other full division are three women, either sitting on a bench with their legs crossed in front, or dancing. My impression is that the latter is the true interpretation of the scene, from two women in precisely similar attitudes being represented as Boro Buddor, in Java15, but there so much better executed that there is no mistake as to their action. Whether, however, these women represent the audience, or are actually taking part in the performance, it is quite certain that the sculptures on this façade are of a wholly domestic character, and represent a Nâch and that only. As such, they would be quite as appropriate to a Queen's palace—as this cave is called—as to the abode of cœnobite Priests, to which purpose it is generally supposed to have been appropriated.

Besides the bassi rilievi just described, there are throughout these caves a number of single figures in alto rilievo. They are generally life-size and placed at either end of the verandahs of the caves, as dwârpâlas or sentinels. They are generally dressed in the ordinary native costume, and of no especial interest; but in this cave there are two which are exceptional, and when properly investigated may prove of the utmost value for the history of these caves. These two are situated at the north end of the upper verandah of this cave. The first is of a singularly Bacchic character, and is generally described as a woman riding astride on a lion, and is certainly so represented in Captain Kittoe’s drawing.16 From Captain Murray's photograph, however, the stout figure of the rider appears to me very much more like the Silenus brought from Mathura and now in the Calcutta Museum, and the animal is as likely to be a tiger as a lion.17 It is, however, too much mutilated to feel sure what it may represent.

Behind this group stands a warrior in a Yavana costume, (woodcut No. 22), which, so far as I know, is quite unique in these caves though something very like it occurs at Sanchi.18 There, as here, the dress consists of a short tunic or kilt reaching to the knee, with a scarf thrown over the left shoulder and knotted on the right. On his left side hangs a short sword of curiously Roman type, and on his feet he wears short boots, or hose reaching to the calf of the leg, whether they are bound like sandals as at Sanchi is not quite clear, but the whole costume is as nearly that of a Scotch Highlander of the present day as it is possible to conceive. Those wearing this costume at Sanchi are known from their instruments of music and other peculiarities to be foreigners, though whence they came is not clear, and this one, we may safely assert, is not an Indian, and his costume not such as was adapted to the climate, or ever worn by the people; nor is it found in any of the bas-reliefs just described. Bearing in mind what we learn from the palm leaf records of the Yavana invasions of Orissa, there seems little doubt that these two figures do represent foreigners from the north-west, or at least a tradition of their presence here. In the present state of our knowledge, however, it is impossible to form even a plausible theory as to who they were, nor to guess at what time they may have been present in this country, beyond what we gather from the age of the caves in which they are represented.

• 1. The diagram is compiled by me, from Mr. Locke’s two plans and the photographs, and must not therefore be considered as quite correct, though sufficiently so to explain the text.—J. F.
• 2. Tree and Serpent Worship, Plate XXXIII.
• 3. The casts made by Mr. Locke are generally divided at the apex or the arch over the doorways, where these emblems are usually found. I cannot, therefore, feel certain that what I have taken for the Triśula and Shield emblems may not be, after all, mere architectural ornaments.
• 4. I have shown the photographs from the caste of the bas-reliefs to Messrs. Fausböll, Rhys Davids, Sénart, and Feer, who are perhaps the four persons who at the present day are most competent to give an opinion on such a subject, but none of them have been able to offer any plausible suggestions on this subject.
• 5. As the plates of this work have been complete for several years, and the text printed, it is much to be regretted that the Government did not entrust their publication to Mr. Locke or someone else, so that the public might have the advantage of the information obtained at their expense. I am afraid there is very little chance of their being published by the Babu within any reasonable time.
• 6. Tree and Serpent Worship, Plates LVI. and XCII. Y 182
• 7. Turnour's Mahawanso, chap. vi. p. 43, et seq.
• 8. Translated by Julien, vol. ii. p. 355.
• 9. The Stupa at Bharhut, Plate XXV. Fig. 2.
• 10. Tree and Serpent Worship, Plate XXXVI. Fig. 1.
• 11. Ibid., Plate LI., Fig. 1.
• 12. A similar scene occurs at Buddha Gaya. See Cunningham’s Reports, vol. i., Plate X., Fig. 33. Rajendralâla’s Buddha Gaya, Plate XXXIV., Fig. 3. It is moat unmistakably a love scene.
• 13. They have all been lithographed for Babu RajendraIâla's second volume, so that when that is published the public will have an opportunity of judging how far this account of them is correct.
• 14. This, as in all the ancient sculptures in India, is the “Flauto Traverso," supposed to be invented in Italy in the 13th or 14th century.
• 15. Boro Buddor, 4 vols. folio, published by the Dutch Government at Batavia, vol. i. Plate CX., Fig. 189.
• 16. J. A. S. Bengal, vol. vii. Plate XLI.
• 17. There is a second figure of Silenus presented to the Calcutta Museum by Col. Stacey, brought also from Mathura, with female attendants, the whole of which, with the trees behind, was certainly sculptured in India about the period to which I assign this cave. There is also the patera brought by Dr. Lord from Budakshan, now in the Indian Museum, representing Silenus in a chariot, drawn by panthers, also of Indian workmanship.
• 18. Tree and Serpent Worship, Plate XXVIII. Fig. 1.