though still sloping slightly inwards, the jambs more nearly parallel, and in fact we have nearly all the features of a well-designed chaitya cave. The two pillars in front, however, as will be seen from the plan (Woodcut No. 49), are so much too large in proportion to the rest, that they are evidently stambhas, and ought to stand free instead of supporting a verandah. Their capitals (Woodcut No. 50, p. 113) are more like the Persepolitan than any others in India, and are each surmounted by horses and elephants bearing men

and women of bold and free execution. From the view (Woodcut No. 51) it will be seen how much the surface is covered with the rail decoration, a repetition on a small scale of the rails described in the last section, and which it may here be mentioned is a fair test of the age of any building. It gradually becomes less and less used after the date

of these two chaitya caves, and disappears wholly in the 4th or 5th centuries, but during that period its greater or less prevalence in any building is one of the surest indications we have of the relative age of any two examples. In this cave, as will be observed, nearly the whole of the ornamentation is made up of miniature rails, and repetitions of window fronts or façades. It has also a semicircular open-work moulding, like basket-work, which is only found in the very oldest caves, and is evidently so unsuited for stone-work that it is no wonder it was dropped very early. No example of it is known after the Christian Era. There is an inscription in this cave in an ancient form of letter, but not sufficiently distinct to fix its age absolutely without further evidence.

The third cave is the chaitya at Nassick. Its pillars internally are so nearly perpendicular that their inclination might escape detection, and the door jambs are nearly parallel.

The façade, as seen in the woodcut (No. 52, p. 115), is a very perfect and complete design, but all its details are copied from wooden forms, and nothing was executed in wood in this cave but the rafters of the roofs internally, and these have fallen down.

Outside this cave, over the doorway, there is an inscription, stating that the cave was the gift of a citizen of Nassick,[141] in the reign of King Krishna, the second of the Andrabritya kings, who reigned just before the Christian Era,[142] and inside, on the pillars, another in an older form of character, stating that it was excavated in honour of King Badrakaraka,[143] who was almost certainly the fifth king of the Sunga dynasty, and who ascended the throne about B.C. 129. It may be possible that a more critical examination of these inscriptions may render their testimony less absolute than it now appears, but, taking them in conjunction with the architecture, the age of this cave hardly seems doubtful. For myself, I see no reason for hesitating to accept B.C. 129 as the date of its inception, though its completion may be a century later, and, if this is so, it carries back the caves of Bhaja and Bedsa to a period considerably before that time, while, on the other hand, it as certainly is older than the Karli cave, which appears to come next to it in age.

Karli.

The fourth cave mentioned above, known as that at Karli, is situated on the road between Bombay and Poonah, and is the finest of all—the finest, indeed, of its class. It is certainly the largest as well as the most complete chaitya cave hitherto discovered in India, and was excavated at a time when the style was in its greatest purity. In it all the architectural defects of the previous examples are removed; the pillars of the nave are quite perpendicular. The screen is ornamented with sculpture—its first appearance apparently in such a position—and the style had reached a perfection that was never afterwards surpassed.

In this cave there is an inscription on the side of the porch, and another on the lion-pillar in front, which are certainly integral, and ascribe its excavation to the Maharaja Bhuti or Deva Bhuti,[144] who, according to the Puranas, reigned B.C. 78, and, if this is so, they fix the age of this typical example beyond all cavil.

The building, as will be seen by the annexed illustrations (Nos. 53, 54, 55), resembles, to a very great extent, an early Christian church in its arrangements: consisting of a nave and side-aisles, terminating in an apse or semidome, round which the aisle is carried. The general dimensions of the interior are 126 ft. from the entrance to the back wall, by 45 ft. 7 in. in width. The side-aisles, however, are very much narrower than in Christian churches, the central one being 25 ft. 7 in., so that the others are only 10 ft. wide, including the thickness of the pillars. As a scale for comparison, it may be mentioned that its

arrangement and dimensions are very similar to those of the choir of Norwich Cathedral, or of the Abbaye aux Hommes at Caen, omitting the outer aisles in the latter buildings. The thickness of the piers at Norwich and Caen nearly corresponds to the breadth of the aisles in the Indian temple. In height, however, Karli is very inferior, being only 42 ft. or perhaps 45 ft. from the floor to the apex, as nearly as can be ascertained.

Fifteen pillars on each side separate the nave from the aisles; each pillar has a tall base, an octagonal shaft, and richly ornamented capital, on which kneel two elephants, each bearing two figures, generally a man and a woman, but sometimes two females, all very much better executed than such ornaments usually are. The seven pillars behind the altar are plain octagonal piers, without either base or capital, and the four under the entrance gallery differ considerably from those at the sides. The sculptures on the capitals supply the place usually occupied by frieze and cornice in Grecian architecture; and in other examples plain painted surfaces occupy the same space. Above this springs the roof, semicircular in general section, but somewhat stilted at the sides, so as to make its height greater than the semi-diameter. It is ornamented even at this day by a series of wooden ribs, probably coeval with the excavation, which prove beyond the shadow of a doubt that the roof is not a copy of a masonry arch, but of some sort of timber construction which we cannot now very well understand.

Immediately under the semidome of the apse, and nearly where the altar stands in Christian churches, is placed the dagoba, in this instance a plain dome slightly stilted on a circular drum. As there are no ornaments on it now, and no mortices for woodwork, it probably was originally plastered and painted, or may have been adorned with hangings, which some of the sculptured representations would lead us to suppose was the usual mode of ornamenting these altars. It is surmounted by a Tee, the base of which is similar to the one shown on Woodcut No. 13, and on this still stand the remains of an umbrella in wood, very much decayed and distorted by age.

Opposite this is the entrance, consisting of three doorways, under a gallery exactly corresponding with our roodloft, one leading to the centre, and one to each of the side-aisles; and over the gallery the whole end of the hall is open as in all these chaitya halls, forming one great window, through which all the light is admitted. This great window is formed in the shape of a horseshoe, and exactly resembles those used as ornaments on the façade of this cave, as well as on those of Bhaja, Bedsa, and at Nassick described above, and which are met with everywhere at this age. Within the arch is a framework or centering of wood standing free (Woodcut No. 55). This, so far as we can judge, is, like the ribs of the interior, coeval with the building;[145] at all events, if it has been renewed, it is an exact copy of the original form, for it is found repeated in stone in all the niches of the façade, over the doorways, and generally as an ornament everywhere, and with the Buddhist “rail,” copied from Sanchi, forms the most usual ornament of the style.

The presence of the woodwork is an additional proof, if any were wanted, that there were no arches of construction in any of these Buddhist buildings. There neither were nor are any in any Indian building anterior to the Mahomedan Conquest, and very few indeed in any Hindu building afterwards.

To return, however, to Karli, the outer porch is considerably wider than the body of the building, being 52 ft. wide, and is closed in front by a screen composed of two stout octagonal pillars, without either base or capital, supporting what is now a plain mass of rock, but which was once ornamented by a wooden gallery forming the principal ornament of the façade. Above this a dwarf colonnade or attic of four columns between pilasters admitted light to the great window, and this again was surmounted by a wooden cornice or ornament of some sort, though we cannot now restore it, since only the mortices remain that attached it to the rock.

In advance of this screen stands the lion-pillar, in this instance a plain shaft with thirty-two flutes, or rather faces, surmounted by a capital not unlike that at Kesariah (Woodcut No. 6), but at Karli supporting four lions instead of one, and, for reasons given above (p. 55), they seem almost certainly to have supported a chakra or Buddhist wheel. A similar pillar probably stood on the opposite side, but it has either fallen or been taken down to make way for the little temple that now occupies its place.

The absence of the wooden ornaments of the external porch, as well as our ignorance of the mode in which this temple was finished laterally, and the porch joined to the main temple, prevents us from judging what the effect of the front would have been if belonging to a free-standing building. But the proportions of such parts as remain are so good, and the effect of the whole so pleasing, that there can be little hesitation in ascribing to such a design a tolerably high rank among architectural compositions.

Of the interior we can judge perfectly, and it certainly is as solemn and grand as any interior can well be, and the mode of lighting the most perfect—one undivided volume of light coming through a single opening overhead at a very favourable angle, and falling directly on the altar or principal object in the building, leaving the rest in comparative obscurity. The effect is considerably heightened by the closely set thick columns that divide the three aisles from one another, as they suffice to prevent the boundary walls from ever being seen, and, as there are no openings in the walls, the view between the pillars is practically unlimited.

These peculiarities are found more or less developed in all the other caves of the same class in India, varying only with the age and the gradual change that took place from the more purely wooden forms of these caves to the lithic or stone architecture of the more modern ones. This is the principal test by which their relative ages can be determined, and it proves incontestably that the Karli cave was excavated not very long after stone came to be used as a building-material in India.

There are caves at Ajunta and probably at Junir which are as old as the four just described, and, when the history of cave architecture comes to be written in extenso, will supply details that are wanting in the examples just quoted. Meanwhile, however, their forms are sufficient to place the history on a firm basis, and to explain the origin and early progress of the style with sufficient distinctness.

From the inscriptions and literary evidence, it seems hardly doubtful that the date of the Karli cave is about 78 B.C., and that at Nassick about 129 B.C. We have no literary authority for the date of the two earlier ones, but the archæological evidence appears irresistible. The Bhaja cave is so absolutely identical in style with the Lomas Rishi cave at Behar (Woodcut No. 43) that they must be of very nearly the same age. Their pillars and their doorways slope so nearly at the same angle, and the essential woodenness—if the expression may be used—of both is so exactly the same, that, the one being of the age of Asoka, the other cannot be far removed from the date of his reign. The Bedsa cave exhibits a degree of progress so nearly halfway between the Bhaja and Nassick examples, that it may safely be dated 150 to 200 B.C., and the whole four thus exhibit the progress of the style during nearly two centuries in the most satisfactory manner, and form a basis from which we may proceed to reason with very little hesitation or doubt.

Ajunta.[146]

There are four chaitya caves in the Ajunta series which, though not so magnificent as some of the four just mentioned, are nearly as important for the purposes of our history. The oldest there (No. 9) is the lowest down on the cliff, and is of the smallest class, being only 45 ft. by 23 ft. in width. All its woodwork has perished, though it would not be difficult to restore it from the mortices left and the representations of itself on the façade. There are several inscriptions, but they do not seem integral. They are painted on the walls, and belong, from the form of their characters, to the 2nd or 3rd century of our era, when the frescoes seem to have been renewed, so that the real tests of its age are, first, its position in the series, which make it, with its accompanying vihara (No. 12), undoubtedly the oldest there; the other test is the architecture of its façade, which so much resembles that of the Nassick chaitya (B.C. 129) that it cannot be far off in date. It may, however, be somewhat earlier, as the pillars in the interior slope inwards at a somewhat greater angle, and, in so far as that is a test of age, it indicates a greater antiquity in the Ajunta example.

The next chaitya (No. 10) is situated very near to the last, a little higher up in the rock, however, and of nearly twice its dimensions. It is 94 ft. 6 in. in depth by 41 ft. 3 in. in width internally. As may be seen from the annexed view (Woodcut No. 57), the nave is separated from the aisles by a range of plain octagonal shafts, perfectly upright, but without capitals or bases. The triforium belt is of unusual height, and was originally plastered and painted. Traces of this can still be seen, though the design cannot be made out (Woodcut No. 58). One of the most remarkable characteristics of the cave is that it shows signs of transition from wood to stone in its architectural details. The ribs of the aisle are in stone cut in rock, but copied from the wooden forms of previous examples. The vault of the nave was adorned with wooden ribs, the mortices for which are still there, and their marks can still be traced in the roof, but the wood itself is gone.

There are two inscriptions in this cave which seem to be integral, but unfortunately neither of them contain names that can be identified; but from the form of their characters a palæographist would almost certainly place them anterior to the Christian Era.[147] Taking, however, all the circumstances of the case into consideration, and so as to avoid stretching any point too far, it would, perhaps, be better to assume for the present that the cave belongs to the 1st century of our era.

The façades of both these caves are so much ruined by the rock falling away that it is impossible to assert that there was no sculpture on the lower parts. None, certainly, exists in the interior, where everything depends on painting; and it is, to say the least of it, very improbable that any figure-sculpture ever adorned the oldest, while it seems likely that even No. 10 depended wholly on conventional architectural forms for its adornment.

The next chaitya cave in this series (No. 19) is separated from these two by a very long interval of time. Unfortunately, no inscription exists upon it which would assist in assigning it any precise date; but it belongs to a group of viharas, Nos. 16 and 17, whose date, as we shall afterwards see, can be fixed with tolerable certainty as belonging to the 5th century of our era. The cave itself, as will be seen from the plan (Woodcut No. 59), is of the smallest size, nearly the same as No. 9, or 46 ft. 4 in. by 23 ft. 7 in., and its arrangements do not differ much, but its details belong to a totally different school of art. All trace of woodwork has disappeared, but wooden forms are everywhere repeated in stone, like the triglyphs and mutules of the Doric order, long after their original meaning was lost. More than this, painting in the interval had to a great extent become disused as a means of decoration, both internally and externally, and sculpture substituted for it in all monumental works; but the greatest change of all is that Buddha, in all his attitudes, is introduced everywhere. In the next woodcut (No. 60)—the view of the façade—it will be seen how completely figure-sculpture had superseded the plainer architectural forms of the earlier caves. The rail ornament, too, has entirely disappeared; the window heads have been dwarfed down to mere framings for masks; but, what is even more significant than these, is that from a pure theism or rather atheism we have passed to an overwhelming idolatry. At Karli, the eight figures that originally adorned the porch are chiefs with their wives, in pairs. All the figures of Buddha that appear there now are long subsequent additions. None but mortals were sculptured in the earlier caves, and among these mortals Sakya Muni nowhere appears. Here, on the contrary, he is Bhagavat—the Holy One—the Deity—the object of worship, and occupies a position in the front of the dagoba or altar itself (Woodcut No. 61, p. 126), surmounted by the triple umbrella and as the Numen of the place.

At a future stage of our inquiries we may be able to fix more nearly the time in which this portentous change took place in Buddhist ritual. For the present it is sufficient to remark that images of Buddha, and their worship, were not known in India in the 1st century of our era, and that the revolution was complete in the 5th century.

Before leaving this cave, however, it may be well to remark on the change that had taken place in the form of the dagoba during these 500 years. If Woodcut No. 61 is compared with the dagobas in Nos. 56 and 57, it will be seen how much the low rounded form of the early examples had been conventionalised into a tall steeple-like object. The drum had become more important than the dome, and was ornamented with architectural features that have no meaning as applied. But more curious still is the form the triple umbrella had assumed. It had now become a steeple reaching to the roof of the cave, and its original form and meaning would hardly be suspected by those who were not familiar with the intermediate steps.

I am not aware of more than three umbrellas being found surmounting any dagoba in the caves, but the annexed representation of a model of one found at Sultanpore, near Jelalabad (Woodcut No. 62), probably of about the same age, has six such discs; and in Behar numerous models are found with seven, making with the base and finial nine storeys,[148] which afterwards in China became the conventional number for the nine-storeyed towers of that land.

The last chaitya at Ajunta (No. 26) is of a medium size, 66 ft. by 36, and has a long inscription, but which unfortunately contains nothing to enable us to fix its date with certainty. It is certainly more modern than the last-named, its sculptures are coarser, and their meaning more mythological. We shall probably not err in assuming that it was excavated towards the end of the 6th or beginning of the 7th century; and that the year 600 is not far from its true date. Its chief interest is in showing how nearly Buddhism was approximating to Brahmanism when the catastrophe took place which expelled the former from the country of its birth.

Ellora.

The celebrated Viswakarma cave at Ellora is a chaitya of the first class, intermediate in age between the two last-described caves at Ajunta, or it may be as modern as the last. There are unfortunately no inscriptions nor any traditions[149] that would assist in fixing its age, which must consequently depend wholly on its position in the series and its architectural peculiarities.

The dimensions of this cave are considerable, 85 ft. by 43 ft., and the inner end is entirely blocked up by the dagoba which, instead of being circular as in all the older examples, has a frontispiece attached to it larger than that in cave No. 19 at Ajunta, which, as shown in Woodcut No. 60, makes it square in front. On this addition is a figure of Buddha seated with his feet down, and surrounded by attendants and flying figures in the latest style of Buddhist art. In the roof, all the ribs and ornaments are cut in the rock, though still copied from wooden prototypes, and the triforium has sculptured figures as in Nos. 19 and 26 of Ajunta. Its most marked characteristic, however, is the façade, where for the first time we miss the great horseshoe opening, which is the most marked feature in all previous examples. We can still trace a reminiscence of it in the upper part of the window in the centre (Woodcut No. 63, p. 128); but it was evidently considered necessary, in this instance, to reduce the size of the opening, and it is easy to see why this was the case. At Bedsa, Karli, Kenheri and elsewhere, there was a verandah or porch with a screen in front of the great window, which prevented the direct rays of the sun from reaching it, and all the older caves had wooden screens, as at Karli, from which curtains could be hung so as to modify the light to any desired extent. At Ellora, no screen could ever have existed in front, and wooden additions had long ceased to be used, so that it consequently became necessary to reduce the size of the opening. In the two later chaityas at Ajunta, this is effected by simply reducing their size. At Ellora it was done by dividing it. If we had the structural examples in which this change was probably first introduced, we might trace its progress; but, as this one is the only example we have of a divided window, we must accept it as one of the latest modifications of the façades of these chaityas. Practically, it may be an improvement, as it is still sufficiently large to light the interior in a satisfactory manner; but artistically it seems rather to be regretted. There is a character and a grandeur about the older design which we miss in this more domestic-looking arrangement, though it is still a form of opening not destitute of beauty.

Owing to the sloping nature of the ground in which it is excavated this cave possesses a forecourt of considerable extent and of great elegance of design, which gives its façade an importance it is not entitled to from any intrinsic merit of its own.

Kenheri.

One of the best known and most frequently described chaityas in India is that on the island of Salsette, in Bombay Harbour, known as the great Kenheri cave. In dimensions it belongs to the first rank, being 88 ft. 6 in. by 39 ft. 10 in., and it has the advantage that its date is now almost absolutely fixed. In the verandah there is an inscription recording that the celebrated Buddhaghosha dedicated one of the middle-sized statues in the porch to the honour of the lord Bhagawan,[150] and in the same porch another inscription records the execution of the great statues of Buddha by “Gotamiputra’s imperial descendant Sri Yadnya Sat Karni.”[151] Now we know that the first-named, Buddhaghosha, went on his mission to Ceylon, B.C. 410,[152] and he is not known ever to have returned to India; and Yadnya Sri has always been assumed to have lived 408-428, generally it must be confessed on the mistaken etymology of confounding his name with that of Yuegai of the Chinese. That, however, is apparently only a translation of the “Moon beloved king,” and more applicable, consequently, to Chandra Sri or Chandragupta, who was his contemporary. The true basis for the determination of his date is the Puranic chronology, which, for this period seems indisputable.[153] Be all this as it may, the conjunction of these two names here in this cave settles their date, and settles also the age of the cave as belonging to the early years of the 5th century, at the time when Fa Hian was travelling in India.

This being so, one would naturally expect that the architecture of the cave should exhibit some stage of progress intermediate between cave(No. 10and cave No. 19 of Ajunta, but nothing of the sort is apparent here; the Kenheri cave is a literal copy of the great cave at Karli, but in so inferior a style of art that, when I first saw it, I was inclined to ascribe it to an age of Buddhist decrepitude, when the traditions of true art had passed away, and men were trying by spasmodic efforts to revive a dead art. This being now proved not to be the case, the architecture of this cave can only be looked upon as an exceptional anomaly, the principles of whose design are unlike anything else to be found in India, emanating probably from some individual caprice, the origin of which we may probably never now be able to recover.

Internally the roof was ornamented with timber rafters, and though these have fallen away, the wooden pins by which they were fastened to the rock still remain; and the screen in front has all the mortices and other indications, as at Karli, proving that it was intended to be covered with wooden galleries and framework. What is still more curious, the figures of chiefs with their wives, which adorn the front of the screen at Karli, are here repeated literally, but copied so badly as not at first sight to be easily recognisable. This is the more strange as it occurred at an age when their place was reserved for figures of Buddha, and when, at Karli itself, they were cutting away the old sculptures and old inscriptions, to introduce figures of Buddha, either seated cross-legged, or borne on the lotus, supported by Naga figures at its base.[154]

In front of this cave is a dwarf rail which, with the knowledge we now have, would in itself be almost sufficient to settle the age, in spite of these anomalies (Woodcut No. 64). Unfortunately it is so weather-worn that it is difficult to make out all its details; but comparing it with the Gautamiputra rail (Woodcut No. 32) and the Amravati rail (Woodcut No. 36), it will be seen that it contains all those complications that were introduced in the 3rd and 4th centuries, but which were discontinued in the 5th and 6th, when the rail in any shape fell into disuse as an architectural ornament.[155]

The evidence in fact seems complete that this cave was excavated in the early years of the 5th century; but, admitting this, it remains an anomaly, the like of which only occurs once again so far as I know in the history of Indian architecture, and that in a vihara at Nassick of the same age, to be described hereafter.

Dhumnar.

About half way between Kotah and Ujjain, in Rajputana, there exists a series of caves at a place called Dhumnar which are of considerable extent, but the interest that might be felt in them is considerably diminished by their being cut in a coarse laterite conglomerate, so coarse that all the finer architectural details had to be worked out in plaster, and that, having perished with time, only their plans and outlines can now be recognised. Among the sixty or seventy excavations here found one is a chaitya of some extent, and presenting peculiarities of plan not found elsewhere. It is practically a chaitya cella situated in the midst of a vihara (Woodcut No. 65). The cell in which the dagoba is situated is only 35 ft. by 13 ft. 6 in., but to this must be added the porch, or ante-chapel, extending 25 ft. further, making the whole 60 ft. On two sides, and on half the third, it is surrounded by an open verandah leading to the cells. The third side never was finished, but in two of the side cells are smaller dagobas—the whole making a confused mass of chambers and chaityas 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 dates for determining the age of this cave, but like all of the series it is late, probably between the years 500 and 600 A.D., or even later, and its great interest is that, on comparing it with the chaitya and vihara at Bhaja or Bedsa (Woodcuts Nos. 46 and 49), we are enabled to realise the progress and changes that took place in designing these monuments during the seven or eight centuries that elapsed between them.

Kholvi.

Not far from Dhumnar is another series of caves not so extensive, but interesting as being probably the most modern group of Buddhist caves in India. No very complete account of them has yet been published,[156] but enough is known to enable us to feel sure how modern they are. One, called Arjun’s House, is a highly ornamented dagoba, originally apparently some 20 ft. in height, but the upper part being in masonry has fallen away. Inside this is a cell open to the front, in which is a cross-legged seated figure of Buddha, showing an approach to the Hindu mode of treating images in their temples, which looks as if Buddhism was on the verge of disappearing.

The same arrangement is repeated in the only excavation here which can be called a chaitya hall. It is only 26 ft. by 13 ft. internally; but the whole of the dagoba, which is 8 ft. in diameter, has been hollowed out to make a cell, in which an image of Buddha is enshrined. The dagobas, in fact, here—there are three standing by themselves—have become temples, and only distinguishable from those of the Hindus by their circular forms.[157]

It is probably hardly necessary to say more on this subject now, as most of the questions, both of art and chronology, will be again touched upon in the next chapter when describing the viharas which were attached to the chaityas, and were, in fact, parts of the same establishments. As mere residences, the viharas may be deficient in that dignity and unity which characterises the chaityas, but their number and variety make up to a great extent for their other deficiencies; and altogether their description forms one of the most interesting chapters in our history.

CHAPTER VI.

VIHARAS,[158] OR MONASTERIES.

CONTENTS.

Structural Viharas—Bengal and Western Vihara Caves—Nassick, Ajunta, Bagh, Dhumnar, Kholvi, and Ellora Viharas—Circular Cave at Junir.

Structural Viharas.

We are almost more dependent on rock-cut examples for our knowledge of the Viharas or monasteries of the Buddhists than we are for that of their Chaityas or churches: a circumstance more to be regretted in this instance than in the other. In a chaitya hall the interior is naturally the principal object, and where the art of the architect would be principally lavished. Next would come the façade. The sides and apse are comparatively insignificant and incapable of ornament. The façades and the interior can be as well expressed in the rock as when standing free; but the case is different with the viharas. A court or hall surrounded with cells is not an imposing architectural object. Where the court has galleries two or three storeys in height, and the pillars that support these are richly carved, it may attain an amount of picturesqueness we find in our old hostelries, or of that class of beauty that prevails in the courts of Spanish monasteries.[159] Such was, I believe, the form many of the Indian structural viharas may have taken, but which could hardly be repeated in the rock; and, unless some representations are discovered among the paintings or sculptures, we shall probably never know, though we may guess, what the original appearances may have been.

There was, however, I believe, another form of Vihara even less capable of being repeated in the rock. It was pyramidal, and is the original of all the temples of southern India. Take, for instance, a description of one mentioned both by Fa Hian and Hiouen Thsang,[160] though neither of them, it must be confessed, ever saw it, which accounts in part for some absurdities in the description:—“The building,” says Fa Hian, “has altogether five storeys. The lowest is shaped into the form of an elephant, and has 500 stone cells in it; the second is in the form of a lion, and has 400 chambers; the third is shaped like a horse, and has 300 chambers; the fourth is in the form of an ox, and has 200 chambers; and the fifth is in the shape of a dove, and has 100 chambers in it”—and the account given of it by Hiouen Thsang is practically the same.[161] At first sight this looks wild enough; but if we substitute the assertion that the several storeys were adorned with elephants, lions, horses, &c., we get a mode of decoration which began at Karli, where a great range of elephants adorn the lower storey, and was continued with variations to Hullabîd, where, as we shall see further on, all these five animals are, in the 13th century, superimposed upon one another exactly as here recounted.

The opposite woodcut (No. 66), taken from one of the raths at Mahavellipore, probably correctly represents such a structure, and I believe also the form of a great many ancient viharas in India. The diagram (No. 67) is intended to explain what probably were the internal arrangements of such a structure. As far as it can be understood from the rock-cut examples we have, the centre was occupied by halls of varying dimensions according to height, supported by wooden posts above the ground-floor, and used as the common day-rooms of the monks. The sleeping-cells (Woodcuts Nos. 68, 69) were apparently on the terraces, and may have been such as are frequently represented in the bas-reliefs at Bharhut and elsewhere. Alternately they seem to have been square and oblong, and with smaller apartments between. Of course we must not take too literally a representation of a monastery, carried out solidly in the rock for a different purpose, as an absolutely correct representation of its original. The importance, however, of this form, as explaining the peculiarities of subsequent Buddhist and Dravidian architecture, is so great that it is well worth quoting here, though this will be more evident in the sequel than it can be at present. In construction the breadth, in a structural building, would probably have been greater in proportion to the height than in this example, but that is of little consequence for our present purposes.

It is, of course, always difficult, sometimes impossible, to realise the form of buildings from verbal descriptions only, and the Chinese Pilgrims were not adepts at architectural definitions. Still Hiouen Thsang’s description of the great Nalanda monastery is important, and so germane to our present subject that it cannot well be passed over.

This celebrated monastery, which was the Monte Cassino of India for the first five centuries of our era, was situated thirty-four miles south of Patna, and seven miles north of the old capital of Rajagriha. If not founded under the auspices of the celebrated Nagárjuna in the 1st century, he at all events resided there, introducing the Mahayana or great translation, and making it the seat of that school for Central India. After his time six successive kings had built as many viharas on this spot, when one of them surrounded the whole with a high wall, which can still be traced, measuring 1600 ft. north and south, by 400 ft., and enclosing eight separate courts. Externally to this enclosure were numerous stupas or tower-like viharas, ten or twelve of which are easily recognised, and have been identified, with more or less certainty, by General Cunningham, from the Pilgrim’s description.[162] The general appearance of the place may be gathered from the following:—“In the different courts the houses of the monks were each four storeys in height. The pavilions had pillars ornamented with dragons, and had beams resplendent with all the colours of the rainbow—rafters richly carved—columns ornamented with jade, painted red and richly chiselled, and balustrades of carved open work. The lintels of the doors were decorated with elegance, and the roofs covered with glazed tiles of brilliant colours, which multiplied themselves by reflection, and varied the effect at every moment in a thousand manners.” Or as he enthusiastically sums up:—“The Sangharamas of India are counted by thousands, but there are none equal to this in majesty or richness, or the height of their construction.”[163]

From what we know of the effects of Burmese monasteries at the present day this is probably no exaggeration; and with its groves of Mango-trees, and its immense tanks, which still remain, it must have been, as he says, “an enchanting abode.” Here there resided in his time—within and without the walls—10,000 priests and neophytes, and religion and philosophy were taught from a hundred chairs, and here consequently our Pilgrim sojourned for five years, imbibing the doctrines of the Law of Buddha. What Cluny and Clairvaux were to France in the Middle Ages, Nalanda was to Central India, the depository of all true learning, and the foundation from which it spread over all the other lands of the faithful; but still, as in all instances connected with that strange parallelism which existed between the two religions, the Buddhists kept five centuries in advance of the Christians in the invention and use of all the ceremonies and forms common to both religions.

It would indeed be satisfactory if the architecture of this celebrated monastery could be restored and its arrangements made clear. Something has been done by Cunningham[164] towards this, and excavations have been made by Mr. Broadley and Captain Marshall. The former it is feared has destroyed more than he has restored, and his drawings are so imperfect as to be utterly unintelligible. The latter has not yet published his discoveries. Nothing, however, would probably better repay a systematic exploration than this celebrated spot, if undertaken by some one accustomed to such researches, and capable of making detailed architectural drawings of what is found.

If, however, it should turn out, as hinted above, that the whole of the superstructure of these viharas was in wood, either fire or natural decay may have made such havoc among all that remains of them, as to leave little to reward the labours of the explorer. What has been done in this direction certainly affords no great encouragement to hope for much. At Sultangunge, near Monghyr, a large vihara was cut through by the railway, but except one remarkable bronze statue of Buddha[165] nothing was found of importance. The monastery apparently consisted of two large courtyards surrounded by cells. What was found, however, could only have been the foundations, as there were no doorways to the apartments or means of communication between each other or with the exterior.[166]

The vihara excavated by Captain Kittoe and Mr. Thomas, at Sarnath, seems certainly to have been destroyed by fire. All that remained was a series of some twenty cells and four larger halls surrounding a pillared court 50 ft. square. On one side were three cells evidently forming a sanctuary, as is frequently found in the later rock-cut examples.[167]

The excavations conducted by General Cunningham, at the same place, are hardly more satisfactory in their result. The two buildings he explored seem to bear the relation to one another of a vihara 60 ft. square over all, and a temple of little more than half these dimensions with a projecting porch on each face.[168] Only the foundation of these buildings now remains, and nothing to indicate how they were originally finished.

We may eventually hit on some representation which may enable us to form definite ideas on this subject, but till we do this we probably must be content with the interiors as seen in the rock-cut examples.

Bengal Caves.

None of the Behar caves can, properly speaking, be called viharas, in the sense in which the word is generally used, except perhaps the Son Bhandar, which, as before mentioned, General Cunningham identifies with the Sattapanni cave, in front of which the first convocation was held 543 B.C. It is a plain rectangular excavation, 33 ft. 9 in. long by 17 ft. wide, and 11 ft. 7 in. to the springing of the curved roof.[169] It has one door and one window, but both, like the rest of the cave, without mouldings or any architectural features that would assist in determining its age. The jambs of the doorway slope slightly inwards, but not sufficiently to give an idea of great antiquity. In front there was a wooden verandah, the mortice holes for which are still visible in the front wall.

The other caves, at Barabar and Nagárjuna, if not exactly chaityas in the sense in which that term is applied to the western caves, were at least oratories, places of prayer and worship, rather than residences. One Arhat or ascetic may have resided in them, but for the purpose of performing the necessary services. There are no separate cells in them, nor any division that can be considered as separating the ceremonial from the domestic uses of the cave, and they must consequently, for the present at least, be classed as chaityas rather than viharas.

The case is widely different when we turn to the caves in Orissa, which are among the most interesting, though at the same time the most anomalous, of all the caves in India. They are situated in two isolated hills of sandstone rock, about twenty miles from Cuttack and five from Bhuvaneswar. The oldest are in the hill called Udayagiri; the more modern in that portion designated Khandagiri. They became Jaina about the 10th or 11th century, and the last-named hill is crowned by a Jaina temple, erected by the Maharattas in the end of the last century.

What we know of the age of the older caves is principally derived from a long inscription on the front of the oldest, known as the Hathi Gumpha, or Elephant Cave.[170] From it we learn that it was engraved by a king called Aira, who ascended the throne of Kalinga in his twenty-fourth year, and spread his power by conquest over neighbouring rajas. He seems at first to have vacillated between the Brahmanical and Buddhist faiths, but finally to have adopted the latter and distributed infinite alms. Among other good works, he is said “to have constructed subterranean chambers—caves containing a chaitya temple, and pillars.”

Palæographically, the forms of the letters used in this inscription are identical with those used by Asoka in the copy of his edicts on the Aswatama rock close by, and that recently found at Aska, near the northern corner of the Chilkya lake. The first presumption, therefore, is that they may be of about the same date. This is justified by the mention of Nanda in the past tense, while there seems no reason for doubting that he was one of the kings of that name who immediately preceded the revolution that placed Chandragupta on the throne. Beside these, there are other indications in this inscription which seem to make it almost certain that Aira was contemporary with the great Mauryan dynasty of Magadha; but whether he preceded or followed Asoka is not quite so clear. Still it appears unlikely that Asoka would have been allowed to set up two copies of his edicts in the dominions of such powerful kings as Aira and his father seem to have been, and as unlikely that Aira should make such a record without some allusion to the previously promulgated edicts, had they then existed. On the whole, I am inclined to believe that Aira lived before Asoka, and, if so, that this is the oldest inscription yet found in India. Be this as it may, the cave in which it is found is certainly the oldest here. It is a great natural cavern, the brow of which has been smoothed to admit of this inscription, but all the rest remains nearly in a state of nature. Close to it is a small cave, the whole “fronton” of which over the doorway is occupied by a great three-headed Naga, and may be as old as the Hathi cave. The inscription on it merely says that it is the unequalled chamber of Chulakarma, who seems also to have excavated another cave, here called the Pawan Gabha,[171] or Purification Cave.

Besides these, and smaller caves to be noticed hereafter, the great interest of the Udayagiri caves centres in two—the so-called Ganesa cave, and that called the Raj Rani, or Rani Hanspur, from a tradition—Hindu—that it was excavated by the Rani of Lelat Indra Kesari, the celebrated builder of the Bhuvaneswar temple in the 7th century.