Pramāṇa · Search · About · API

History of Indian and Eastern Architecture v2.c25: Book VI. Northern or Indo-Aryan Style.

History of Indian and Eastern Architecture v2.c25

Language
English

Contents

Book VI. Northern or Indo-Aryan Style.

v2.c25.2

INTRODUCTORY.

v2.c25.3

CONTENTS. Introductory—Dravidian and Indo-Aryan Temples at Pattadakal—Modern Temple at Benares.

v2.c25.4

Of the three styles into which Hindû architecture naturally divides itself, the northern is found spread over a far larger portion of the country than either of the other two. It wants, however, the compactness and strongly-marked individuality of the Dravidian, and never was developed with that exuberance which characterised the southern style from the 15th to the 18th century. In many respects it resembles more the Chalukyan style, the examples being small and elegant, and found dispersed over the face of the country, where wanted, without any apparent massing together in particular spots.

v2.c25.5

Unfortunately, we have no name which would describe the style in its ethnographical and geographical relations without being open to the objection of expressing either too much or too little. In this respect the southern style is singularly fortunate: Dravidian correctly limits it to people speaking Tâmil, Telugu, or some cognate dialect; and the country where the people speaking those tongues are to be found is generally and correctly known as Drâvida-desa, or country of the Dravidians.

v2.c25.6

The term Chalukyan, applied to the second style, is not so expressive; but it is unobjectionable, as it cannot mislead any one. It is only a conventional term, derived from the principal known dynasty ruling in that country, applied to a style occupying a borderland between the other two, but a land whose boundaries cannot yet be fixed with precision. Till they are, a conventional name that does not mislead is all that can be hoped for.

v2.c25.7

If it were allowable to adopt the loose phraseology of philological ethnography, the term Aryan might be employed, as it is the name by which the people practising this style are usually known in India, and it would be particularly convenient here, as it is the correct and direct antithesis of Dravidian. It is evident, however, that any such term, if applied to architecture, ought to be descriptive of some style practised by that people, wherever they settled, all across Europe and Asia, between the shores of the Atlantic and the Bay of Bengal;¹ and it need hardly be said that no such style exists. If used in conjunction with the adjective Indian or Indo-Aryan, it becomes much less objectionable, and has the advantage of limiting its use to the people who are generally known as Aryans in India—in other words, to all those parts of the country where Sanskrit was spoken, or where the people now speak tongues so far derived from Sanskrit as to be distinguishable as offsets of that great family of languages. Its use, in this respect, has the great convenience that any ordinary ethnographical or linguistic map of India is sufficient to describe the boundaries of the style. It extends, like the so-called Aryan tongues, from the Himâlayas to the south of the Vindhya mountains. On the east, it is found prevalent in Orissa; and on the west in Mahârâshtra. Its southern boundary between these two provinces will only be known when the Nizam's territory is architecturally surveyed.

v2.c25.8

¹ In 1848 Gen. Cunningham applied the term Aryan to the architecture of Kashmîr, apparently on the strength of a pun ('Journal of the Asiatic Society of Bengal,' vol. xvii. pt. ii., 1848, p. 242). This, however, was limiting a term that belongs to two continents to an insignificant valley in one of them. It was, besides, wholly uncalled for. The term Kashmîri was amply sufficient, and all that was wanted for so strictly local a style.

v2.c25.9

Another reason why the term Aryan should be applied to the style is, that the country just described, where it prevails, is, and always has been, called Âryâvarta by the natives themselves. They consider it as the land of the pure and just—meaning thereby the Sanskrit-speaking peoples—as contradistinguished from that of the casteless Dasyus, and other tribes, who, though they may have adopted Brahmanical institutions, could not acquire their purity of race.

v2.c25.10

The great defect of the term, however, is that the people inhabiting the north of India are not Aryans in any reasonable sense of the term, whatever philologists may say to the contrary.

v2.c25.11

The Sanskrit-speaking people, who came into India 2000 or 3000 years B.C., could never have been numerically one-half of the inhabitants of the country, except, perhaps, in some such limited district as that between the Satlaj and the Jamnâ; and since the Christian Era no Âryan race has migrated eastward across the Indus, but wave after wave of peoples of Turanian race, under the names of Yavanas, Sakas, Hûnas, Tûrks, or Mongols, have poured into India. This, combined with the ascendancy of the mixed or aboriginal races during the period when Buddhism was the prevailing religion of the country, has so completely washed out Âryanism from northern India during the building ages, that there is probably no community there which could claim one-tenth of pure Âryan blood in its veins, and with nine-tenths of impurity the term is certainly a misnomer. If it were not, we would certainly find some trace of external Âryan affinities in their style; but in fact, no style is so purely local, and, if the term may be used, so aboriginal, as this. The origin of the Buddhist style is obvious and unmistakable; that of the Dravidian and Chalukyan nearly as certain, though not quite so obvious; but the origin of the northern Hindû style remains a mystery, unless, indeed, the solution suggested above (*ante*, vol. i. p. 325) be considered an explanation. It may be so, to some extent; but I confess it is to my mind neither quite satisfactory nor sufficient.

v2.c25.12

The style was adopted by the Jains, and several examples of the peculiar forms of their vimânas, or sikharas have already been given (Woodcuts Nos. 290, 299, etc.); but it still remains to be ascertained from what original form the curvilinear square tower could have arisen. There is nothing in Buddhist, or any other art, at all like it. It does not seem to have been derived from any wooden form we know, nor from any brick or stone, or tile mode of roofing found anywhere else. I have looked longer, and, perhaps, thought more, on this problem than on any other of its class connected with Indian architecture, but I have no more plausible suggestion to offer than that hinted at above. The real solution will probably be found in the accidental discovery of old temples—so old as to betray in their primitive rudeness the secret we are now guessing at in vain. Meanwhile, we probably may remain sure that it was not an imported form, but an indigenous production, and that it has no connection with the architecture of any other people outside of India.

v2.c25.13

The view above proposed for the origin of the style derives considerable support from the mode in which the temples are now found distributed. There are perhaps more temples now in Orissa than in all the rest of Hindustan put together. They are very frequent in Mahârâshtra, and, if we admit the Jains who adopted this style, they are ten times more frequent in Gujarât, Rajputana and the valley of the Narbadâ than in the valley of the Ganges, or in Âryâvarta, properly so called. The first and most obvious explanation of this fact must be that the last-named country has for 600 years been occupied by a Muhammadan empire, and they, hating idolatry and idol temples, have destroyed them wherever they were so absolutely in possession of the country as to be able to do so with impunity. My impression, however, is that it does not correctly represent the whole state of the case. That the Moslims did ruthlessly destroy Jaina and Hindû temples at Ajmîr, Delhi, Kanauj, and elsewhere in northern India, is quite true, but it was, partly at least, because their columns served so admirably for the construction of their mosques. The astylar temples of the followers of Siva or Vishnu could have served principally as quarries, and stones that had been previously used in Hindû temples have not been traced to a large extent in Moslim buildings. But admitting that at Delhi or Allahâbâd, or any of their northern capitals, all Hindû buildings have been utilised, this hardly would have been supposed the case at such a provincial capital as Faizâbâd, once Ayodhyâ, the celebrated capital of Dasaratha, the father of Râma the hero of the Râmâyana, but where little besides a few pillars in Bâbar’s mosque can be discovered that belongs to any ancient building.¹

v2.c25.14

¹ ‘Gazetteer of Oudh’ (1877), vol. i. p. 7. Sâlâr Mas‘ûd Ghâzî, the nephew of Mahmûd of Ghaznî, passed through Ayodhyâ in 1033, and would hardly have failed to display his iconoclastic zeal. Gen. Cunningham attempts to identify the various mounds at this place with those described as existing in Sâketa by the Buddhist Pilgrims.—‘Ancient Geography of India,’ pp. 401 *et seqq.*; ‘Archæological Reports,’ vol. i. pp. 293 *et seqq.* The truth of the matter, however, is, that neither Fah Hian nor Hiuen Tsiang were ever near the place. The city they visited, and where the Tooth-brush-tree grew, has not been identified.

v2.c25.15

The most crucial instance, however, is the city of Benares, so long the sacred city, *par excellence*, of the Hindûs, yet, so far as is known, no vestige of an ancient Hindû temple exists within its present precincts. James Prinsep resided there for ten years, and Major Kittoe, who had a keener eye than even his great master for an architectural form, lived long there as an archæologist and architect. They drew and measured everything, yet neither of them ever thought that they had found anything that was ancient; and it was not till Messrs. Horne and Sherring² started the theory that the buildings around the Bakarîyâ Kund were ancient Buddhist or Hindû remains, that any one had discovered any traces of antiquity in that city. But the buildings about the Bakarîyâ Kund were erected by Muhammadans, and the pillars and roofing-stones, with a few possible exceptions, were carved by them for the purposes for which they were applied. They may have used the stones of deserted monasteries, or other Buddhist or Hindû buildings, in the foundations or on their terraces, or for little detached pavilions; but all the architecture, properly so called, is in a style invented, or at least introduced by the Pathâns, and brought to perfection under Akbar.

v2.c25.16

² ‘Sacred City of the Hindus,’ London, 1868, pp. 271 *et seqq.*; ‘Journal of the Asiatic Society of Bengal,’ vol. xxxiv pp. 1 *et seqq.*

v2.c25.17

That the Moslims destroyed Hindû temples all over the south of Hindustan and in their raids into the Dekhan is certain, but it was not till the time of Aurangzîb that any of their monarchs felt himself sufficiently powerful or was so bigoted as to dare the power and enmity of the Brâhmans of Benares, by erecting a mosque on the site of one of their most sacred temples as an insult and a defiance to the Hindûs. Even then, had such a temple as the great one at Bhuvaneswar existed in Benares, every stone of which, from the ground to the kalas, is covered with carving, it seems remarkable that all these carved stones should be hid away and not one now to be found. But so it appears; still we know historically that there were many temples in the city, and during the pre-Mughal period the city was often sacked, whilst the river courses have changed and probably buried what the Moslim failed to destroy.

v2.c25.18

The rock at Gwâliar was one of the earliest conquests of the Moslims, and they held it more or less directly for five centuries. They built palaces and mosques within its precincts, yet the most conspicuous objects on the hill are Hindû temples, that were erected before they obtained possession of it. In like manner Chitor was thrice besieged and thrice sacked by the Muhammadans, but numerous buildings there are comparatively intact.

v2.c25.19

The instances of early temples discovered during the last forty years, however, bears some testimony to the numbers that must have existed all over the country prior to the Musalmân conquests. These are very numerous in the west and south-west of Bengal, where the Aryan element in the population is a minimum. No temples are mentioned in the Vedas or the older Indian writings, and were not required for the simple quasi-domestic rites of their worship; and so long as they remained pure perhaps no temples were built. With the introduction of the Brahmanic ritual they became a necessity. It is to be understood then that though we may use the term Indo-Aryan as the most convenient to describe and define the limits of the northern style, the name it is intended to convey is, that the style arose in a country which they once occupied, and in which they have left a strong impress of their superior mental power and civilisation, and over which the languages spoken are of Sanskritic descent.

v2.c25.20

If this reservation is always borne in mind, I know of no term that more conveniently expresses the characteristics of this style, and it is consequently proposed to adopt it in the following pages as the name of the style that prevailed among the Hindûs in northern India, between the Himâlaya and Vindhya mountains, and even much further south, from the 7th century to the present day.

v2.c25.21

309. Dravidian and Indo-Aryan Temples at Pattadakal. (From a Photograph.)

v2.c25.22

The general appearance of the northern temples, and the points of difference between them and those of the south, will be appreciated from the above woodcut (No. 309), representing two very ancient temples, built in juxtaposition at Pattadakal, in Bijâpûr district. That on the left is a complete specimen of Dravidian architecture (*ante*, vol. i. p. 355). There is the same pyramidal form, the same distinction of storeys, the same cells on each, as we find at Mâmallapuram (Woodcut No. 185), at Tanjor (Woodcut No. 213), or at Madurâ (Woodcut No. 195). The right-hand temple—that of Galaganâth, to the north-west of Sangamesvar's—is Indo-Aryan of somewhat later date, and in which, on the contrary, the outline of the pyramid is curvilinear; no trace of division of storeys is observable, no reminiscence of habitations and no pillars or pilasters anywhere. Even in its modern form (Woodcut No. 310), it still retains the same characteristics, and all the lines of the pyramid or sikhara are curvilinear, the base polygonal. No trace of utilitarianism is visible anywhere. If Woodcut No. 310 is compared with that at vol. i. page 339 (Woodcut No. 195), the two styles will be exhibited in their most modern garbs, when, after more than 1000 years' practice, they have receded furthest from the forms in which we first meet them. Yet the Madras temple retains the memory of its storeys and its cells. The Bengal example recalls nothing known in civil or domestic architecture.

v2.c25.23

Neither the pyramid nor the tumulus affords any suggestion as to the origin of the form, nor does the tower, either square or circular; nor does any form of civil or domestic

v2.c25.24

310. Modern Temple at Benares.

v2.c25.25

311. Diagram Plan of Hindû Temple.

v2.c25.26

architecture. It does not seem to be derived from any of these, and, whether we consider it as beautiful or otherwise, it seems certainly to have been invented principally at least for æsthetic purposes, and to have retained that impress from the earliest till the present day.

v2.c25.27

The plan of a northern temple is always a square internally, and generally the same form is retained in the exterior; but very rarely, if ever, without some addition. In some instances it is only a thin parallel projection, as at A in the diagram (No. 311). Sometimes it has two such slices added, as at B; but in the oldest examples these are only half the thickness shown here. From this they proceeded to three projections, as at C, the oldest examples being the thinnest. In more modern times the thickness of the projections became equal to their distance from each other, as at D; so that the temple became in plan practically a square, the sides of which were parallel to the diagonal of the original square or to the line E F G. Even, however, when this was the case, the cell always retained its original form and direction, and the entrance and windows kept their position on what had thus practically become the angles of the building. This is the case with the temple at Benares, shown in Woodcut No. 310, and generally also with the Jaina temples, and especially the case with the temple on the Takht-i-Sulaimân in Kashmîr. Although the depth and width of these offsets vary considerably even in the same design, the original square is never lost sight of; the four central angles, as at F, being always larger and more strongly accentuated than the others, and their line is always carried through to the summit of the pyramid.

v2.c25.28

It will be observed that by this process we have arrived at the same form or plan for a solid building that was attained by the arrangement of pillars described vol. i. page 317. In fact, the two forms were elaborated simultaneously, and were afterwards constantly used together. My impression is, that the pillared arrangement is the oldest, and led to the deepening of the additions to the solid square till the two became identical in plan. Whether this were so or not, it is one of the most distinguishing features of northern Hindû architecture.

v2.c25.29

In the very centre of India, at Amarakantak, near a place marked Ajmîrgadh on the map, is a sacred tank, from which it is said that the Sôn flows to the north, the Mahânadî to Katak in the Bay of Bengal, and the Narbadâ to the Indian Ocean. All these rivers have their sources in the hill. The spot has always been held sacred, and is surrounded by temples, two or three of them—as far as can be gathered from the imperfect accounts available—of considerable age.¹ On the south and east of this hill extends the great and fertile table-land of Chhattîsgarh. This is now, and has always been, so far as our knowledge extends, one of the principal seats of the native tribes. If that country and the surrounding districts were carefully surveyed, we might find temples, some of which would add very materially to our knowledge of the history of this style.²

v2.c25.30

¹ Mr. Beglar in 'Archæological Survey of India Reports,' vol. vii. pp. 227ff. and plates 20, 21; but the account and drawings are very defective.

v2.c25.31

² Conf. Cousens, 'Lists of Antiquarian Remains in the Central Provinces and Berar,' Calcutta, 1897, which indicate some promising remains.

v2.c25.32

ORISSA.

v2.c25.33

History—Temples at Bhuvaneswar, Kanârak, Puri, Jâjpur, and Katak.

Chronology.

v2.c25.s1.1

| Anantavarman Chodagangadeva | . *cir.* 1078 | Narasimhadeva I. | . *cir.* 1238-1264 |
| Narasimhadeva II. | . „ 1277-1305 |
| Kâmârnava | . „ 1144-1155 | Sulaimân, King of Bengal | |
| Anîyanga Bhîmadeva II. | . „ 1190-1198 | conquered Orissa | . „ 1568 |

v2.c25.s1.2

THE two provinces of India, where the Indo-Aryan style can be studied with the greatest advantage, are Dhârwâr on the west, and Orissa on the east coast. The former has the advantage of being mixed up with the Dravidian style, so as to admit of synonyms and contrasts that are singularly interesting, both from an ethnological and historical point of view. In Orissa, on the contrary, the style is perfectly pure, being unmixed with any other, and thus forms one of the most compact and homogeneous architectural groups in India, and as such of more than usual interest, and it is consequently in this province that the style can be studied to the greatest advantage.

v2.c25.s1.3

One of the most marked and striking peculiarities of Orissan architecture is the distinct and almost absolute contrast it presents to the style of the Dravidian at the southern end of the peninsula. The curved outline of the towers or vimânas has already been remarked upon, but, besides this, no Orissan towers present the smallest trace of any storeyed or even step-like arrangement, which is so universal further south, and the crowning member is never a dome, nor a reminiscence of one. Even more remarkable than this, is the fact that the Orissan style is almost entirely astylar. In some of the more modern examples, as for instance in the porches added to the temples at Bhuvaneswar and Puri in the 12th and 14th centuries, we do find pillars, but it is probably correct to state that, among the 100 or 150 original shrines at Bhuvaneswar, scarcely a pillar is to be found.¹ This is the more remarkable because, within sight of that capital, the caves in the Udayagiri (*ante*, p. 13) are adorned with pillars to such an extent as to show that their forms must have been usual and well known in the province before any of the temples were constructed. When we recollect that no great temple in the south was considered complete without its "hall of 1000 columns," and many besides this had hundreds dispersed about the place, and used for every conceivable purpose, the contrast is more striking, and shows what a complete barrier the Chalukyas interposed between the two races on this side of India, though not on the other. As a rule, every Orissan temple consists of two apartments, similar in plan, as shown in the diagram (Woodcut No. 184). The inner one is generally a cube, surmounted by a tower, here called Barâ-deül, or Dewal, corresponding with the vimâna of the south, and in it the image or images of the gods are enshrined; in front of this is a porch or antarâla, called Jaga-mohan, generally square in plan or approaching it, and surmounted by a pyramidal roof of varying pitch. The peculiarities are illustrated in the diagram (Woodcut No. 184) just referred to, which purports to be an elevation of the celebrated Black Pagoda at Kanârak. It is only, however, an eye-sketch, and cannot be depended upon for minute detail and correctness, but it is sufficient to explain the meaning of the text. Sometimes one or two more porches (mandapas) were added in front of this one, called the Nâta-mandir or dancing-hall—corresponding to the Sabhâ-mandapa in a Gujarât temple—and the Bhoga-mandir or refectory, but these, in almost every instance, are afterthoughts, and not parts of the original design. Be this as it may, in every instance in Orissa the tower with its porch forms the temple. If enclosed in a wall, they are always to be seen outside. There are gateways, it is true, but they are always subordinate, and there are none of those accretions of enclosures and gopurams that form so marked a characteristic of the southern style. There generally are other shrines within the enclosures of the great temples, but they are always kept subordinate, and the temple itself towers over everything to even a greater extent than that at Tanjor (Woodcut No. 213), giving a unity and purpose to the whole design, so frequently wanting in the south.

v2.c25.s1.4

¹ The Bhogamandapas of the Lingarâja, Jagannâth and Yamesvara temples have each four pillars supporting their roofs; but these mandapas are of later dates than the temples themselves.

v2.c25.s1.5

Other contrasts will come out as we proceed, but, in the meanwhile, few examples bring out more clearly the vast importance of ethnography as applied to architecture. That two peoples, inhabiting practically the same country, and worshipping the same gods under the guidance of the same Brahmanical priesthood, should have adopted and adhered to two such dissimilar styles for their sacred buildings, shows as clearly as anything can well do how much race has to do with these matters, and how little we can understand the causes of such contrasts, unless we take affinities or differences of race into consideration.

v2.c25.s1.6

About eighty years ago Mr. Andrew Stirling published an ‘Account of Orissa proper or Cuttack,’ giving a dynastic list of the rulers from B.C. 3101 to the beginning of last century, with dates and notable events of their reigns.¹ This was drawn up from native records belonging to the temple of Jagannâth at Purî; and, dismissing the early portions as manifestly fabulous, the record, from the accession of Yayâti Kesarî towards the end of the 5th century, was accepted as tolerably trustworthy, and was revised and published with a fuller list of the names by Sir W. W. Hunter in his ‘Orissa.’² But, like other native histories, it cannot stand examination, and must be discarded as worthless previous to the 12th century, and very inaccurate even for the last four or five centuries. Here, as elsewhere, we can hope for trustworthy historical information only from the steady pursuit of epigraphical research, which as yet has yielded but four or five names of a Somavamsi or Lunar dynasty that ruled before the 12th century; and their inscriptions are unfortunately dated only in regnal years, and must be relegated, on epigraphical grounds, to about the 11th century.³ We have thus, as yet, but little help from historical sources. It is true that the dates of two of its temples have been approximately ascertained. The great one at Bhuvaneswar is said to have been erected about A.D. 640—but possibly later—and that at Purî between A.D. 1080 and 1140, nearly the first and the last of the series. My impression is that in the later direction it can hardly be extended beyond the year 1260, but within these limits it seems possible to arrange the sequence of all the temples in the province without much difficulty, and to ascertain their dates with some degree of approximate certainty.

v2.c25.s1.7

¹ ‘Asiatic Researches,’ vol. xv. (1825), pp. 163ff.; also ‘History of the Râjas of Orissa, from the reign of Râja Yudhistira. Translated from the Vamsâvalî,’ in ‘Jour. Asiat. Soc. Bengal,’ vol. vi. (1837), pp. 756-766.

v2.c25.s1.8

² ‘Orissa’ (ed. 1872) vol. i. pp. 198f. and vol. ii. pp. 183-191; also in Sewell’s ‘Lists of Inscriptions and Dynasties of Southern India,’ pp. 204-209. An out-line of the history of Orissa is given in the new ‘Gazetteer of Purî,’ chap. ii.

v2.c25.s1.9

³ Among four successive rulers, of which we have inscriptions, tentatively placed in the 11th century, the third, Yayâti Mahâ-Sivagupta, may possibly be the Yayâti Kesarî, which the Vamsâvalî makes the founder of the dynasty in 474-526, or five centuries before his probable date.—Dr J. F. Fleet in ‘Epigraphia Indica,’ vol. iii. pp. 323-359.

v2.c25.s1.10

With the exception of the great temple of Jagannâth at Purî, the buildings described in this chapter were mostly erected before the commencement of the 12th century, when we find Anantavarman Chodagangadeva (1078-1142) of the Eastern Ganga dynasty recording that he replaced the fallen lord of Orissa in his kingdom.¹ About this period the Ganga-vansa dynasty succeeded, the second of whom was the builder of the great Purî temple—or at least completed it—for its erection is ascribed to his father, Chodaganga, thirty years earlier. They were nominally Saivas, but patronised also the Vaishnavas, whilst the preceding dynasty seem to have been devoted Saivas.

v2.c25.s1.11

¹ 'Indian Antiquary,' vol. xviii. p. 171; 'Jour. Asiat. Soc. Bengal,' vol. lxxii. pp. 101ff.

v2.c25.s1.12

Owing to its remoteness from the seats of Muhammadan power, Orissa almost entirely escaped the ravages which devastated the principal Hindû cities in the earlier and more intolerant age of their power. The first serious invasion of Orissa was only made about 1510 by 'Alâu-d-Dîn Hasain Shâh, King of Bengal, whose army sacked Katak and plundered Purî, but was driven back; and it was not till 1567-1568 that Sulaimân Khân Karârâni, the Afghân Viceroy of Bengal, finally defeated the Orissa king at Jâjpur. Soon after it was annexed by Akbar, and after four more years of contests it became a province of his empire in 1578, after which further outrages were hardly to be feared.

v2.c25.s1.13

At Jâjpur the Muhammadans had already wreaked their vengeance on all that was Hindû; but elsewhere the monuments were left more nearly intact than any other group in the north of India. Neither at Bhuvaneswar nor at Purî or Kanârak are marked traces of their violence. In later times the Orissa remains have suffered from the sordid proceedings of the Public Works Department, which destroyed the fort at Bârbâti and other public buildings, to mend roads or to save some money in erecting a lighthouse at False Point. Further injury has been done by the antiquarian zeal of the officers who removed some of the best statues of the Râjarânî temple,² and by the vandals who pulled down the Navagraha sculpture from the Kanârak temple. Lastly, and worst of all, by the Archæological Survey, a few years ago, which caused the interior of the mandap of this famous monument to be completely filled up with stones and sand and so "shut up for ever."³

v2.c25.s1.14

² Râjendralâl Mitra's 'Antiquities of Orissa,' vol. ii. p. 90.

v2.c25.s1.15

³ This was conceived to be the only way of preventing the roof from falling in.—Mr. Marshall's 'Annual Report,' 1903-1904,' p. 48.

v2.c25.s1.16

Besides their immunity from the ordinary causes of destruction of Hindû buildings, the Orissa group forms in itself one of the most complete and interesting in all India. The Khajurâho group is nearly as extensive and magnificent, but they were all erected within the limits of about a century, 950 to 1050,¹ so that little sequence can be traced among them. There are also temples in the Kanarese districts more magnificent than any in Orissa, and extending through a long series of years; but they are scattered over a wide extent of country, and are consequently

v2.c25.s1.17

¹ Cunningham, 'Archæological Survey Reports,' vol. ii. p. 416; *infra*, p. 141.

v2.c25.s1.18

312. Temple of Parasurâmeswar. (From a Photograph.)

v2.c25.s1.19

varied by local peculiarities of style. It therefore requires more knowledge and experience to classify them than it does those in this province. Altogether there is not, perhaps, any group which, if properly investigated, would add more to our knowledge of Indian architecture, and give it more precision, than the Bhuvaneswar temples.²

v2.c25.s1.20

² The late Râjendralâl Mitra, who was sent with the expedition organised by the Bengal Government in 1868 to survey the antiquities of Orissa, most unfortunately had no knowledge whatever of architectural surveying or draughtsmanship; nor had he any acquaintance with Indian styles to guide him in determining the periods to which different buildings belonged. Even his vaunted acquaintance with epigraphy was superficial and inexact; and the two folio volumes he prepared at public expense, added little, if anything, to our knowledge.—'Indian Antiquary,' vol. ix. pp. 113f. and 142f.; Fergusson's 'Archæology in India,' pp. 48ff.

v2.c25.s1.21

The oldest temple in the town of Bhuvaneswar is probably that called Parasurâmeswar (Woodcut No. 312), which, from the termination of the name, as well as the Linga in the cella, and subjects portrayed in the three principal niches of the tower, mark it as a Saiva shrine. It may belong to the 7th century, though it may be as late as the 8th. Its style is certainly different from the other early temples here, and more like what we find at other places outside the province. It is not large, being only 20 ft. square¹ at its base; but its sculptures are cut with a delicacy seldom surpassed, and there is an appropriateness about the ornaments greater than is seen in most of the temples.

v2.c25.s1.22

¹ This dimension is from Bâbu Râjendralâl's 'Orissa Antiquities,' vol. i. p. 41, but I don't like it. [Judging from a photograph—the estimated height being about 43 ft.—this dimension seems to be *at least* 27 ft.]

v2.c25.s1.23

The temple itself is apparently 42 ft. in height, and from the summit to the base it is covered with sculptures of the most elaborate character,² but still without detracting from the simplicity and vigour of its outline.

v2.c25.s1.24

² G. Le Bon, 'Les Monuments de l'Inde,' p. 70 and figs. 49, 50.

v2.c25.s1.25

If I am correct in assigning so early a date to the tower of this temple, it is evident that the porch must be a subsequent addition, because it fits badly to the tower. It may, however, be that if this is really the oldest temple of its class in Orissa, its design may be copied from a foreign example, and borrowed, with all its peculiarities, from a style practised elsewhere. Be that as it may, it is interesting as showing the mode by which light was sometimes introduced into the porches of these temples between the ends of the beams of the stone roof. As the sloping roofing-stones project considerably beyond the openings, a subdued light is introduced, without either the direct rays of the sun, or the rain being able to penetrate.³

v2.c25.s1.26

³ This temple has of late been in the hands of the official, and we learn that it "has been thoroughly restored; the whole roof of the mandapa was dismantled and rebuilt."—'Archæological Survey Annual Report, 1902-03,' p. 46.

v2.c25.s1.27

The temple of Mukteswar (Woodcut No. 313) is very similar in general design to that of Parasurâmeswar, but even richer and more varied in detail, and its porch partakes more of the regular Orissan type. It has no pillars internally, and the roof externally exhibits at least the germ of what we find in the porches of the great temple at Bhuvaneswar and the Black Pagoda. Its dimensions are somewhat less than those of the last temple described, but in its class it may be considered the gem of Orissan architecture.⁴

v2.c25.s1.28

⁴ This temple is surrounded by "a number of small shrines, and close to it also stands the temple of Siddhesvara, a larger structure. All the shrines are inferior works of art, but nevertheless they have all been restored."—Dr. T. Bloch, *ibid.* p. 46. No survey is mentioned as having been made.

v2.c25.s1.29

Besides these, there are several other temples which, from the style of their architecture, I would feel inclined to place as

v2.c25.s1.30

313. Temple of Mukteswar. (From a Photograph.)

v2.c25.s1.31

earlier than the great temple. One is known as Sâri Dewal, near the great temple, and another, a very complete and beautiful example, is called Maitreswar, which is almost a duplicate, on a small scale, of the great temple, except that it has no repetition of itself on itself. As above pointed out, almost all the ornaments on the façades of Buddhist temples are repetitions of themselves; but the Hindûs do not seem to have adopted this system so early, and the extent to which it is carried is generally a fair test of the age of Hindû temples. In the great Pagoda there are eight copies of itself on each face, and in the Râjarânî the system is carried so far as almost to obliterate the original form of the temple.

v2.c25.s1.32

314. Plan of Great Temple at Bhuvaneswar. Compiled partly from Plan in Bâbu Râjendralâl Mitra's work, but corrected from Photographs. Scale 50 ft. to 1 in.

Great Temple of Bhuvaneswar.

v2.c25.s2.1

The great temple of Bhuvaneswar, known as the Lingarâja, is one of the landmarks in the style. It is traditionally ascribed to a Lalâtendra Kesarî, who is said to have ruled in the 7th century; though this is mere fable, the temple may tentatively be ascribed to about the 9th or 10th century; but be this as it may, taking it all in all, it is perhaps the finest example of a purely Hindû temple in India.

v2.c25.s2.2

Though not a building of the largest class, the dimensions of this temple in plan are, so far as I can make out, far from contemptible. The whole length is about 210 ft., with a breadth varying from 60 ft. to 75 ft. The original temple, however, like almost all those in Orissa, consisted only of a vimâna, or Baradewal, and a porch or Jagamohan, shaded darker in the plan (Woodcut No. 314), and they extend only to 160 ft. The Nâta- and Bhoga-mandaps, shaded lighter, were added possibly about the 12th century or even later. Though several temples have all these four apartments, so far as I can make out, none were originally erected with them. The true Orissan temple is like that represented in Woodcut No. 184, a building with two apartments only, and these astylar, or practically so: the pillars were only introduced in the comparatively modern additions.

v2.c25.s2.3

315. View of Great or Lingarâja Temple, Bhuvaneswar. (From a Photograph.)

v2.c25.s2.4

The outline of this temple in elevation is not, at first sight, pleasing to the European eye; but when once the eye is accustomed to it, it has a singularly solemn and pleasing aspect. It is a solid, and would be a plain square tower, but for the slight curve at the top, which takes off the hardness of the outline and introduces pleasingly the circular crowning object (Woodcut No. 315). As compared with that at Tanjor (Woodcut No. 213), it certainly is by far the finer design of the two. In plan the southern example is the larger, being 82 ft. square. This one is only about 66 ft.¹ from angle to angle, though it is 75 ft. across the central projection. Their height is nearly the same, both of them being over 180 ft., but the upper part of the northern tower is so much more solid, that the cubic contents of the two are probably not very different. Besides, however, greater beauty in form, the northern example excels the other immeasurably in the fact that it is wholly in stone from the base to the apex, and—what, unfortunately, no woodcut can show—every inch of the surface is covered with carving in the most elaborate manner. It is not only the divisions of the courses, the roll-mouldings on the angles, or the breaks on the face of the tower: these are sufficient to relieve its flatness, and with any other people they would be deemed sufficient; but every individual stone in the tower has a pattern carved upon it, not so as to break its outline, but sufficient to relieve any idea of monotony. It is, perhaps, not an exaggeration to say that if it would take a sum—say a lakh of rupees or pounds—to erect such a building as this, it would take three lakhs to carve it as this one is carved. Whether such an outlay is judicious or not, is another question. Most people would be of opinion that a building four times as large would produce a greater and more imposing architectural effect; but this is not the way a Hindû ever looked at the matter. Infinite labour bestowed on every detail was the mode in which he thought he could render his temple most worthy of the deity; and, whether he was right or wrong, the effect of the whole is certainly marvellously beautiful. It is not, however, in those parts of the building shown in the woodcut that the greatest amount of carving or design was bestowed, but in the perpendicular parts seen from the courtyard (Woodcut No. 316). There the sculpture is of a very high order and great beauty of design. This, however, ought not to surprise us when we recollect that at Amarâvatî, on the banks of the Krishnâ, not far from the southern boundary of this kingdom, there stood a temple more delicate and elaborate in its carvings than any other building in India,¹ and that this temple had been finished probably eight centuries before this one was erected; and though the history of art in India is now written in decay, its growth and vitality had, in earlier times, been vigorous.

v2.c25.s2.5

¹ This and the dimensions in plan generally are taken from a table in Bâbu Râjendralâl's ‘Antiquities of Orissa,’ vol. i. p. 41. I am afraid they are only round numbers, but they suffice for comparison. They are certainly incorrect. In the table the tower is described as 66 ft. by 60, while all the photographs prove that it is undoubtedly square. In the plan (vol. ii. pl. 48) the sides are represented as 66 by 54 ft. from angle to angle, and the internal dimensions are given in the table as 42 square. In the plan they are 44 by 46, and approach so nearly to the exterior, that if the tower had been built, as represented in his plan, it would not have stood for an hour. In figure 314 the internal dimension is reduced to 40 ft. with the larger external one of 65 ft. The Bhoga-mandapa is said in the text (p. 72) to be 56 ft. square; by scale it is 63 by 70. The Nâta-mandir is said to be 52 ft. square, and scales 58 by 61. The Jagamohan in the text is said to measure 65 ft. by 45; on the plan it measures 70 by 50. Making these and other adjustments from the plan, it reduces the total length to about 210 ft., instead of the 290 of the plan. This is confirmed by Mr Atkinson’s plan (pl. xxviii.). In like manner the temple of Bhagavatî (pl. xlviii.) is represented as 160 ft. in length, while Mr. Atkinson makes it only 110.—‘Archæology in India,’ pp. 49, 50.

v2.c25.s2.6

¹ 'Tree and Serpent Worship,' plates 48-98; 'Amarâvatî and Jaggayyapeta Buddhist Stûpas' (1887).

v2.c25.s2.7

316. Lower part of Great Tower at Bhuvaneswar. (From a Photograph.)

v2.c25.s2.8

Attached to the Jagamohan of this temple is a Nâta-mandir, or dancing-hall, whose date is, traditionally assigned to about the year 1100: but this is perhaps too early, as there are inscriptions of the 12th and 13th centuries on the doorway of the temple porch, and they are probably earlier than the Nâta-mandir. But even then it enables us to measure the extent of this decay with some degree of certainty. It is elegant, of course, for art had not yet perished among the Hindûs, but it differs from the style of the porch to which it is attached more than the leanest example of Tudor art differs from the vigour and grace of the buildings of the early Edwards. All that power of expression is gone which enabled the early architects to make small things look gigantic from the exuberance of labour bestowed upon them. A glance at the Nâta-mandir is sufficient for the mastery of its details. A week's study of the Jagamohan would every hour reveal new beauties.

v2.c25.s2.9

The last woodcut may convey some idea of the extent to which the older parts were elaborated: but even the photograph hardly enables any one not familiar with the style to realise how exquisite the combination of solidity of mass with exuberance of ornament really is.

v2.c25.s2.10

317. Plan of Râjarânî Temple. (Compiled from a Plan by Bâbu Râjendralâl and corrected from Photographs.) Scale 50 ft. to 1 in.

v2.c25.s2.11

During the five centuries which elapsed between the erection of these two porches, Bhuvaneswar was adorned with some hundreds of temples, some dozen of which have been photographed, but hardly in sufficient detail to enable the student to classify them according to their dates.¹ On the spot it probably would be easy for any one trained to this class of study, and it would be a great gain if it were done. The group nearest in richness and interest is that at Khajurâho, mentioned above (p. 49); but that group belongs to an age just subsequent² to that of the Bhuvaneswar group, and only enables us to see that some of the most elaborate of the Katak temples may extend to the year 1000 or thereabouts. It is to this date that I would ascribe the erection of the Râjarânî temple. The names of the more notable, of which I have photographs, with their approximate dates, are given in the list at the end of this chapter; but I refrain from burdening the text with their names, as I despair, by any reasonable number of woodcuts, of illustrating their marvellous details in anything like a satisfactory manner.

v2.c25.s2.12

¹ Photographs have enabled me to supply to some extent the deficiency of my local knowledge; but unless photographs are taken by a scientific man for scientific purposes, they do not supply the place of local experience; but a full architectural survey also is much desiderated.

v2.c25.s2.13

² Cunningham's 'Reports,' vol. ii. p. 416.

v2.c25.s2.14

The Râjarânî temple, as will be seen from the woodcut (No. 317), is small; but the plan is arranged so as to give great variety and play of light and shade, and as the details are of the most exquisite beauty, it is one of the gems of Orissan art.¹ The following woodcut (No. 318), without attempting to illustrate the art, is quoted as characteristic of the emblems of the period. Below the pillar are three kneeling elephants, over which domineer three lions or leogriffs.

v2.c25.s2.15

¹ Both the temples of Muktesvara and Râjarânî have been restored by Dr. Bloch, the Archæological surveyor, who gives photographs of them "before, and after repairs," but from so very different points of view, that it is not clear what is the extent of these repairs;—but in the Reports, this method of photographing from different points "before and after" meddling with the buildings, is remarkably frequent. No mention seems to be made of securing correct plans of the temples, which might readily have been made whilst the works were going on. The work done is thus described: "The temples generally were fairly intact, but a number of stones had become either loose or unsafe, in the roof of the mandapa and the upper parts of the spire. These had to be dismantled and built up again, using as far as possible the ancient materials. Carvings, when broken and lost, were replaced by new ones, and the work of the modern stonemason does not fall much behind the old work, except that modern restorations of human or animal figures are less graceful than their older models. Only such carvings have been replaced by new ones of which the original pattern was available."—'Archæol. Survey Annual Report, 1902-03,' pp. 45-46. It is pitiable to think of the barbarity of 20th century imitations, or supposed—but very inferior—imitations, being inserted in these venerable structures.

v2.c25.s2.16

318. Doorway in Râjarânî Temple. (From a Photograph.)

v2.c25.s2.17

Above this a Nâgnî, or female Nâga, with her seven-headed snake-hood, adorns the upper part of the pillar. They are to be found, generally in great numbers, in almost all the temples of the province. Over the doorway are the Navagraha, or nine planets, which are almost more universal at the Lingarâja temple.

v2.c25.s2.18

Throughout the province, from the time we first meet it, about the 7th century, if so early, till it dies out about A.D. 1300, the style seems to be singularly uniform in its features, and it requires considerable familiarity with it to detect its gradual progress towards decay. Notwithstanding this, it is easy to perceive that there are two styles of architecture in Orissa, which ran side by side with one another during the whole course. The first is represented by the temples of Parasurâmeswar and Mukteswar (Woodcuts No. 312, 313); the second by the great temple (Woodcut No. 315). They are not antagonistic, but sister styles, and seem certainly to have had at least partially different origins. We can find affinities with that of the Mukteswar group in Dhârwâr and most parts of northern India: but I know of nothing exactly like the great temple anywhere else. It seems to be quite indigenous, and if not the most beautiful, it is the simplest and most majestic of the Indo-Aryan styles. And I cannot help suspecting a wooden origin for it—the courses look so much more like carved logs of wood laid one upon another than courses of masonry, and the mode and extent to which they are carved certainly savours of the same material. There is a mosque built of Deodar pine in Kashmîr, to be referred to thereafter, which certainly seems to favour this idea; but till we find some older temples than any yet discovered in Orissa this must remain in doubt. Meanwhile, it may be well to point out that the majority of the older temples in Orissa follow the type of the great temple, and the rest that of Parasurâmswar; but the two get confounded together in the 9th and 10th centuries, and are mixed together into what may almost be called a new style in the Râjarânî and temples of the 11th and 12th centuries.

Kanârak.

v2.c25.s3.1

With, perhaps, the single exception of the temple of Jagannâth at Purî, there is no temple in India better known, and about which more has been written than the so-called Black Pagoda at Kanârak, 19 miles north-east from Purî; nor is there any one whose date and dedication is better known, since the literature on the subject can here be depended upon. Stirling's statement that the present edifice was built by the Râja Narasingh-deva I., who ruled from about 1238 to 1264, is supported by copperplate inscriptions.¹ Complete as this evidence appears, one is almost tempted to question it, for the simple reason that it seems improbable—after the erection of so inferior a specimen of the art as the temple of Purî (*cir.* A.D. 1100) appears to be—the style could have reverted to anything so beautiful as this. In general design and detail it is so similar to the Jagamohan of the great temple at Bhuvaneswar that at first sight I should be inclined to place it in the same century;

v2.c25.s3.2

¹ 'Asiatic Researches,' vol. xv. p. 327; and 'Journal Asiatic Society of Bengal,' vol. lxxii. (1903), part i. p. 124. Râjendralâl Mitra, ascribed it to Narasingha II., who began to reign, A.D. 1278. 'Antiq. of Orissa,' vol. ii. p. 156.

v2.c25.s3.3

still the details of the tower exhibit a progress towards modern forms which is unmistakable.¹ Abul Fazl after describing the temple with considerable detail and circumstantiality and ascribing it to Râja Narasingh-deva I., adds that "it is said to be a work of 730 years' antiquity."² In other words, it was erected about A.D. 860, or just about 400 years before Narasingh's date which must arise from an error in the hundreds figure. Narasingh-deva must, however, have employed architects of very different tastes and abilities to those engaged a century earlier in erecting the Purî temple.

v2.c25.s3.4

¹ When I visited Orissa in 1837 and sketched this temple, a great part of the tower was still standing. See 'Picturesque Illustrations of Indian Architecture,' plate iii. It has since fallen entirely.

v2.c25.s3.5

² 'Ayeen Akbery,' Gladwin's translation, vol. ii. p. 16. Jarrett's version (vol. ii. pp. 128-129) reads :—" It is said that somewhat over 730 years ago Râja Narsing Deo completed this stupendous fabric and left this mighty memorial to posterity."

v2.c25.s3.6

Another point of interest connected with this temple is, that all authors, apparently following Abul Fazl, agree that it was, like the temple of Mârtând, in Kashmîr (*ante*, vol. i., p. 259), dedicated to the sun.³ Sun-worship, we know, was prevalent in various parts of India, previous to the 12th century, but it seems to have become merged in the Vishnu cult—Sûrya-Nârâyana being regarded as a form of Vishnu. In the west of India there are remains of quite a number of sun-temples of about the eleventh century,⁴ and probably others will be found in Central India and elsewhere, when looked for.

v2.c25.s3.7

³ Arka is a name of the sun as the "lightner"; the place is mentioned as Arka-kshetra or Padma-kshetra.

v2.c25.s3.8

⁴ 'Archæological Survey of Western India,' vol. ix. pp. 73, 74.

First 90 of 435 passages. Read on

Linked texts (43)