Pallava Architecture, Part I (MASI 17)

A. H. Longhurst, 1924.

Language: English
Extent: 242 passages, 23,687 words
Text: complete digital text
Source: archive.org · scan on archive.org, OCR Gemini 3.8 Flash; structure (chapters, sections) by Gemini 3.8 Flash · Memoirs of the Archaeological Survey of India (Government of India), 1924: public domain in India

c1 Front matter — 4 parts
c2 List of Plates — 129 parts
c3 Index — 105 parts

Text

Front matter

| | | | | | | | PAGES. |
| I.—History | .. | .. | .. | .. | .. | .. | 1—7 |
| II.—Religion | .. | .. | .. | .. | .. | .. | 7—8 |
| III.—Architectural Character | .. | .. | .. | .. | .. | 8—12 |
| IV.—Examples in the Mahendra Style (Cir. 610 to 640 A. D.)— | | | | | | | |
| (1) Dalavānūr | .. | .. | .. | .. | .. | .. | 12—13 |
| (2) Trichinopoly | .. | .. | .. | .. | .. | .. | 13—15 |
| (3) Mandagapattu | .. | .. | .. | .. | .. | .. | 15—16 |
| (4) Pallāvaram | .. | .. | .. | .. | .. | .. | 16 |
| (5) Mahēndravādi | .. | .. | .. | .. | .. | .. | 16—17 |
| (6) Vallam | .. | .. | .. | .. | .. | .. | 17—18 |
| (7) Mēlachēri | .. | .. | .. | .. | .. | .. | 18 |
| (8) Singavaram | .. | .. | .. | .. | .. | .. | 18—19 |
| (9) Tirukkalukkunram | .. | .. | .. | .. | .. | .. | 19—21 |
| (10) Kilmavilangai | .. | .. | .. | .. | .. | .. | 21 |
| (11) Bezwada | .. | .. | .. | .. | .. | .. | 22—24 |
| (12) Mogalrājapuram | .. | .. | .. | .. | .. | .. | 24—27 |
| (13) Undavalli | .. | .. | .. | .. | .. | .. | 27—30 |
| (14) Bhairavakonda | .. | .. | .. | .. | .. | .. | 30—37 |

c1.4

21911. 8 12 55. R 913. 03/I.D.A/long

List of Plates

c2.1

Frontispiece.—Detail of Gangādhara sculpture, Trichinopoly. Plate I (a) Dalavānūr temple—Elevation.
(b) Trichinopoly temple—Elevation.
(c) Dalavānūr temple—Plan.
(d) Trichinopoly temple—Plan. Plate II (a) Dalavānūr temple—Detail of doorkeepers.
(b) Dalavānūr temple—Detail of pillars and cornice.
(c) Trichinopoly temple—Detail of pillars.
(d) Trichinopoly temple—Detail of doorkeepers. Plate III (a) Mandagapattu temple—Front elevation.
(b) Pallāvaram temple—Elevation.
(c) Mandagapattu temple—Plan.
(d) Pallāvaram temple—Plan. Plate IV (a) Mahēndravādi temple—Elevation.
(b) Mēlachēri temple—Plan.
(c) Mahēndravādi temple—Plan.
(d) Mēlachēri temple—Section. Plate V (a) Vēdagirisvara Hill at Tirukkalukkunram.
(b) Kilmavilangai. S. Arcot District.
(c) The Orukal Mandapa at Tirukkalukkunram.
(d) Akkanna Madanna temple. Bezwada. Plate VI (a) The Orukal Mandapa at Tirukkalukkunram—Plan.
(b) Akkanna Madanna temple. Bezwada—Plan.
(c) Mogalrājapuram temple No. I—Plan.
(d) Mogalrājapuram temple No. II—Plan. Plate VII (a) Mogalrājapuram temple No. III—Plan.
(b) Mogalrājapuram temple No. IV—Plan.
(c) Mogalrājapuram temple No. V—Plan.
(d) Mogalrājapuram temple No. V—Detail A.
(e) Mogalrājapuram temple No. V—Detail B. Plate VIII (a) Mogalrājapuram temple No. III.
(b) Mogalrājapuram temple No. IV. Plate IX Undavalli temple. Guntur Dist.—Ground and first floor plans. Plate X Undavalli temple. Guntur Dist.—Second and third floor plans. Plate XI Undavalli temple. Guntur Dist.—Sections through centre and on line A. B. Plate XII Undavalli temple. Guntur Dist.—Anantasayana and Vaishnava figures in the second storey. Plate XIII Undavalli temple. Guntur Dist.—Rock-cut Memorial shrines on the hill.

c2.2

Plate XIV (a) Undavalli Rock-cut temple.
(b) Bhairavakonda. Nellore Dist.—Bhairava temple and six rock-cut shrine cells. Plate XV (a) Bhairavakonda—General view of ten rock-cut temples.
(b) Bhairavakonda temple No. I—Shrine entrance showing doorkeepers. Plate XVI (a) Bhairavakonda temple No. I—Plan.
(b) Bhairavakonda temple No. II—Plan.
(c) Bhairavakonda temple No. III—Plan.
(d) Bhairavakonda temple No. IV—Plan. Plate XVII (a) Bhairavakonda temple No. III—General view of entrance.
(b) Bhairavakonda temple No. III—Detail of doorkeeper. Plate XVIII (a) Bhairavakonda temple No. IV (lower) and No. V (upper)—General view.
(b) Bhairavakonda temple No. VII—Detail of entrance and Lion pillars. Plate XIX (a) Bhairavakonda temple No. V—Plan.
(b) Bhairavakonda temple No. VI—Plan.
(c) Bhairavakonda temple No. VII—Plan.
(d) Bhairavakonda temple No. VIII—Plan. Plate XX (a) Bhairavakonda temple No. VI—General view.
(b) Bhairavakonda temple No. VII—General view.

c2.3

PALLAVA ARCHITECTURE.

c2.4

I.—HISTORY.

c2.5

THE origin of the Pallava dynasty and that of their name has been a subject for controversy for some years, and the attempts made to throw light on it have not made the mystery less impenetrable. That the Pallavas became a great power in Southern India in the sixth and seventh centuries, and that they contributed a great deal to the growth first of Buddhism and then of Hinduism, and to South Indian architecture, sculpture and painting, are well known. But we have yet to find out who they were and whence they came.

c2.6

During the last few years, a considerable amount of material relating to the above subject has been collected and published in the various epigraphical journals and official reports but being scattered through a number of different publications the information recorded is not readily accessible to the educated public. The late Rai Bahadur V. Venkayya of the Archæological Survey of India, and Mr. G. Jouveau-Dubreuil, Professor of the Pondicherry College, have done more than any other scholars to work out a plausible history of this elusive dynasty, and the brief historical account given here is taken mainly from Mr. Venkayya's article on The Pallavas, appearing in the Annual Report of the Archæological Survey of India for 1906-07, and from Mr. Jouveau-Dubreuil's two recently published little books—The Pallavas and Pallava Antiquities. Mr. Jouveau-Dubreuil has been, perhaps, rather more successful than previous scholars in collecting material for this interesting subject, because he is the first epigraphist in India to realize the importance of studying the architecture of a monument as well as its inscriptions. His original researches in this direction are most valuable and throw quite a new light on the manner in which inscriptions should be studied in future.

c2.7

The old theory that the word Pallava is apparently the Sanskrit form of the tribal name Pahlava or Pahnava of the Purāṇas, a Pārthian tribe whose territory lay somewhere between the Indus and Persia, is no longer accepted by present day scholars, and it seems more likely that the Pallavas were a tribe, clan or caste, which was formed in the northern part of the Madras Presidency, possibly in the Veṅgi country, between the Kṛishṇā and Gōdāvari rivers. Mr. Venkayya informs us that they were Kshatriyas originally, but gradually became degraded by their omission of the sacred rites and transgressing the authority of the Brāhmaṇas. The admission that they did not conform to Brāhmaṇical practices seems to indicate that they were either foreigners, or else became Buddhists or Jains. If the tribe was formed in the Veṅgi country as suggested by the late Dr. Vincent Smith in his Early History of India, page 423, the formation must have taken place in very early times because this part of the Telugu country was included in the dominions of the Āndhras, so if the clan was actually formed there, the process must have commenced long before the Āndhra dynasty came to an end about 236 A. D. In fact, the Pallavas must have acquired a distinct political status even while the Āndhras were in power. This they appear to have done not by conquest but by contracting marriages with Āndhra princesses and thus inherited a portion of the southern part of the Āndhra country.

c2.8

We know that in the second century, the Āndhra king Pulumāvi II embellished the famous white marble stūpa at Amarāvatī on the southern bank of the Krishnā river, and that the Pallavas were the political successors of the Āndhras in that district shortly after this event. What position these early Pallavas occupied under the Āndhras and in what particular circumstances they rose into supreme power are questions which cannot be answered at present, but there is every reason to believe that future historians will eventually be able to give a fairly complete narrative of the doings of the Pallava rulers and lay open the secret of their origin.

c2.9

Referring to two Pallava copper-plate grants in the *Epigraphia Indica, Volume XV*, page 246, Mr. Krishna Sastri tells us that :—" Three, and sometimes even four, distinct periods of Pallava history are recognized, the earliest covering roughly two centuries, viz., the 3rd and the 4th and the next roughly the 5th and part of the 6th century A. D. The third or rather the third and fourth periods together, extended from the latter part of the 6th down to almost the end of the 9th century A. D., when the kingdom proper of the Pallavas, viz., the Tonda-mandalam was conquered by the Chōlas of Tanjore. The continuity of the line during these several periods has not been clearly established. The rulers of the last dynasty of Pallavas down from the time of Simhavishnu were distinguished as the first builders of lithic monuments in Southern India, the bitter opponents of the progress of the Western Chālukyas of Bādāmi in the south, and the establishers of Pallava power in the heart of the Chōla country. These facts have been practically settled and have been derived from their own copper-plates, the copper-plate records of the contemporaneous Western Chālukyas and the Pallava stone inscriptions found pretty largely in Southern India." The earliest copper-plate records of the Pallava kings are in the Prākrit language and have been assigned to a period not much later than those of the Āndhras of the 2nd and 3rd centuries of the Christian era. These are :—(1) The Mayidavōlu plates of Siva-Skandavarman, (2) The Hirehadagalli plates of the same king, and (3) the British Museum plates of Chārudēvī. The earliest record of these early Pallavas is that of Siva-Skandavarman, issued while the latter was yet a crown-prince, and is dated from his capital at Kānchī, now the modern town of Conjeeveram in the Chingelput district, and is addressed to his Viceroy at Dhannakada (Amarāvatī) in the Āndhra country. The next is a record of the same king after his accession to the throne and is also dated from Conjeeveram. It refers to the grant of a village in Sātāhani-Rattha, a territorial division located in the Bellary district. The mention of this territory in this record of about the 3rd century and of Sātavaghani-hara in an Āndhra record of the 2nd century A. D., recently discovered in the Bellary district, not only indicates the possible identification of the two territorial divisions, but infers the political succession of the Āndhras by the Pallavas of Kānchī. The British Museum plates belong to the same early period and is a record of queen Chārudēvī, the wife of the crown-prince Vijaya-Buddhavarman and mother of Buddhyankura. It comes from the Guntur district and is dated in the reign of Vijaya-Skandavarman, who was evidently the grand-father of prince Buddhyankura and the ruling sovereign at the time of the grant. The record does not inform us of the relationship between Vijaya-Skandavarman and Siva-Skandavarman. These three ancient Prākrit grants show that there was a time when the court language in Southern India was Prākrit, also that the Pallavas were the political successors of the Āndhras and had their capital at Conjeeveram and that their kingdom at that period roughly included the Tonda-mandalam and the Āndhra country right up to the river Krishnā, including the Bellary district in the west.

c2.10

The later Pallava records are written in Sanskrit and some of them give a genealogy of the earlier generations of the Pallava family, but they are mostly very confusing and difficult to understand. One great difficulty about fixing the dates of these records is, that so many Pallava princes had not only the same name but also the same epithets, which makes it almost impossible to know exactly which prince is referred to in the inscriptions. The age when the kings who issued the Sanskrit charters flourished is not known. But approximately, they may be assigned to about the 5th century and continued down to almost the beginning of the 7th century A. D.

c2.11

Pallava history up to this period has to be reconstructed solely from copper-plate records, as no stone monuments of the period have been discovered. Anterior to the 7th century A. D. Indian builders and sculptors appear to have been engaged exclusively in erecting monuments for the Buddhists and Jains. If any Hindu temples were erected before this period they must have been built of wood or some other perishable material, because no trace of such buildings has hitherto been discovered.

c2.12

With the beginning of the seventh century we enter a period of Pallava history for which the records are more numerous and the chronology is not altogether a field of conjecture and doubt. The earliest stone monuments of Southern India belong to this period. In fact, the history of Hindu Architecture in Southern India may be said to begin with the reign of the Pallava king Mahēndravarman I (610 to 640 A. D.). Mahēndra was the son of a Pallava king named Simhavishnu who reigned in the latter part of the sixth century. We know very little about Simhavishnu and the relationship which the Pallava kings of this series bore to the earlier dynasty is nowhere explained, though four of the latter are mentioned among the ancestors of the former in a Pallava copper-plate charter of the eighth century. The following is the pedigree of this line of Pallava kings, as given by the late Mr. Venkayya :—

c2.13

PEDIGREE OF THE SIMHAVISHNU FAMILY. (Ancestor unknown.)

c2.14

c2.15

The earliest king of this series is Simhavishnu who claims to have vanquished the Malaya, Kalabhra, Mālava, Chōla and Pāndya kings, “ the Simhala king proud of the strength of his arms ” and the Kēralas. From this it would appear that the Chōla country did not belong to the Pallavas before the reign of Simhavishnu and that it was he who conquered it. It must have been a difficult task since all the southern kings appear to have opposed him. His son and successor was Mahēndravarman I.

c2.16

The Pallavas now engage in a deadly struggle against the Chālukyas of Bādāmi in the Bombay Presidency. The causes which brought about this long war are not stated, but the hostility between the two tribes became so intense, that each looked upon the other as its natural enemy. The history of this period consists mainly of the events of this war with the Chālukyas, which lasted for nearly a century and was the ultimate cause of the decline and fall of both the Pallavas and Western Chālukyas about the middle of the eighth century. The Chālukyan king Pulikēsin II of Bādāmi, who was Mahēndra’s contemporary, ascended the throne in 609 A. D. and soon overcame the great Harsavardhana of Kanauj. He then turned his arms against the south. Pishtapura, the modern Pithāpuram in the Gōdāvari district, was the first to fall. Then Pulikēsin drove the Pallavas out of their northern province, advanced far into the Pallava country and even threatened their capital. It is stated, with regard to this event, that “ Pulikēsin subsequently caused the splendour of the lord of the Pallavas, who had opposed the rise of his power, to be obscured by the dust of his army and to vanish behind the walls of Kānchipuram.”

c2.17

This invasion of the Pallava kingdom is indirectly admitted by Mahēndravarman I, who claims to have defeated the Chālukyas at Pullalūra near Kānchipuram, when the enemy were driven back and the capital saved. However, after this event the Chālukyas permanently occupied the northern province of the Pallava dominions, and Vishnuvardhana, the younger brother of Pulikēsin II, who was probably sent out as governor of the newly acquired province, eventually established himself at Vēngi and started the Eastern Chālukya dynasty sometime before 632 A. D.

c2.18

The Telugu attributes appearing in some of the Pallava inscriptions found in cave-temples excavated during the reign of Mahēndravarman I, seem to indicate that Mahēndra ruled his empire from the Telugu country, at least at the beginning of his reign. He could not have done so for long because we know that by the middle of the seventh century, this country was in the hands of the Chālukyas. It is, therefore, likely that Pulikēsin II conquered it about 610 A. D., that is, at the very beginning of Mahēndra’s reign. It is probable that this defeat and loss of his dominions in the north led Mahēndra to extend his territory in the south. At any rate. no Pallava monument has been found anterior to the seventh century A. D. in the Tamil country.

c2.19

Like the great Asoka, Mahēndra had occasion to change his religion. He was at first a Jaina, but was afterwards converted to the cult of the linga by the saint called Appar or Tirunāvukkaraiyār, who was first persecuted and then patronised by Mahēndravarman I (Epigraphia Indica, Volume III, page 278) The two great Saiva devotees Appar and Tirunānasambandar were contemporaries of the two Pallava kings Mahēndra and his son Narasimhavarman I. Having once adopted Saivism, Mahēndra lost no time in giving a new impulse to that religion in the Tamil country by excavating a number of imperishable rock-cut Siva temples in the Kanchipuram district. Other inscriptions show that he was a poet and a musician, a soldier and a good administrator, and he also constructed several useful irrigation tanks in his kingdom. Mahendravarman I may be regarded as one of the greatest figures in the history of Tamil civilization and the founder of Hindu architecture in Southern India.

c2.20

It would appear that Mahendra got his taste for rock-cut temples from the Buddhists who lived in the Telugu country at the beginning of his reign, when he himself appears to have lived for a while in that part of his dominions. The rock-cut temples at Bezwada, Mogalrajapuram, and Undavalli, all belong to this same period and are similar in style to some of the rock-cut temples in the Tamil country, which we know for certain were excavated by Mahendra's orders. The Bezwada temples possess no ancient inscriptions, but their architectural style proclaims them to be Pallava monuments of the seventh century, and it is just possible that they may have been started by Mahendra himself before he was forced to retire to the south, as mentioned above.

c2.21

Mr. Jouveau-Dubreuil is of opinion that the Bezwada and Undavalli temples were excavated by the Vishnukundins (The Pallavas, pages 33, 35), who seem to have ruled on the banks of the Krishna and Godavari towards the end of the sixth century, and at a time when the Pallavas reigned over the adjacent districts of Nellore and Guntur. Thus they were neighbours, and it is possible that the Pallavas may have learned the art of excavating temples out of the rock from the Vishnukundins, but there is no proof at present that this was the case, and until we know more about the Vishnukundins and their monuments, we may assume, on purely architectural grounds, that the Bezwada cave-temples are Pallava monuments of the early part of the seventh century.

c2.22

Mahendravarman I was succeeded by his son Narasimhavarman I in the second quarter of the seventh century. Narasimhavarman I was surnamed Mamalla or Mahamalla and he founded the seacoast town of Mamallapuram, now the modern village of Mahabalipuram in the Chingleput district, and popularly known as the Seven Pagodas. He retrieved the fortunes of the family by repeatedly defeating the Cholas, Keralas, Kalabhras and Pandyas. He made war on the Chalukyas and actually captured Vatapi (Badami), their capital. This claim of his is established by an inscription found at Badami in the Bombay Presidency in which he bears the title Mahamalla (Indian Antiquary, Volume IX, page 100). The capture of Badami by the Pallavas has been assigned to the year 642 A. D. The Pallavas probably held the territory during the time of Narasimhavarman I, his son Mahendravarman II, and during the early part of the reign of the latter's son and successor, Paramesvaravarman I. The last named king is said to have defeated the Chalukyan king Vikramaditya I at Peruvalanallur and put his army to flight. On the other hand, Vikramaditya I claims to have 'received by surrender the town of Kanchipuram after defeating the lord of the Pallavas who had been the cause of the humiliation and destruction of his family.' Apparently as a result of this victory the Kurnool district, which originally belonged to the Pallavas, passed into the hands of the Chalukyas. Two copper-plate records of Vikramaditya and two of his son Vinayaditya have been discovered in this district. Paramesvaravarman's son and successor was Narasimhavarman II, surnamed Rājasimha, i.e., ‘Lion of Kings.’ Rājasimha was an ardent devotee of Siva and seems to have spent most of his time in building Siva temples and bestowing gifts on the Brāhmans. Previous to Rājasimha’s reign, all Pallava monuments were monolithic rock-cut memorials but those erected by Rājasimha were structural buildings built of stone and brick. He thus introduced a new style of Pallava architecture which may be called the style of Rājasimha, so as to distinguish it from the earlier rock-cut styles of Mahendra and Māmamalla.

c2.23

Rājasimha built the central shrine of the Kailāsanātha temple at Conjeeveram, and his son Mahendravarman III seems to have completed the work. Paramēsvara-varman II, another son of Rājasimha, constructed the Vaikuntha Perumāl temple at Conjeeveram. The death of the latter is mentioned in an inscription engraved on the wall of the verandah running round the central shrine of this building (South Indian Inscriptions, Volume II, page 344). The events which took place after his death are depicted in bas-relief panels on the walls of the same verandah. A dispute about the succession to the throne after his death seems to have occurred and Nandivarman, a cousin of Paramēsvaravarman II, was chosen by the subjects as their king. It is clear that at this period internal dissensions had set in in the Pallava family. The Chālukyas of Bādāmi were not slow to take advantage of this unsettled state of the Pallava country and Vikramāditya II, grandson of Vinayāditya, having ‘resolved to uproot completely his natural enemy, the Pallava, who had robbed of their splendour the previous kings born from his race, reached with great speed Tonda-mandalam, attacked at the head of battle and put to flight the Pallava called Nandipōtavarman (Nandivarman), who had come to meet him.’ The city of Kāñchīpuram was captured by the enemy.

c2.24

The late Mr. Venkayya commenting on this defeat of the Pallavas says: ‘The procedure adopted by Vikramāditya after the capture of the Pallava capital shows that the frequent wars waged in India by ancient kings against one another did not much affect the country or the peaceful inhabitants. In fact, the atrocities of later Indian warfare were unknown in early times. It is said of Vikramāditya that though he took Kāñchīpuram, he did not destroy it and that, ‘having made the twice-born, the distressed and the helpless rejoice by continual gifts, he acquired great merit by granting heaps of gold to the Kailāsanātha and other temples in the city.’

c2.25

This crushing defeat of the Pallavas by the Chālukyas seems to have dealt the death blow to the sovereignty of the former. Thus the history of the Pallavas emphasises the oft-repeated lesson of history that, when internal dissensions set in, the decline and disappearance of a dynasty are only a matter of time. Strangely enough, their rivals, the Western Chālukyas also ceased to be the ruling power in the west about the same time as Pallava ascendancy came to an end in the south.

c2.26

Nandivarman is said to have reigned for about fifty years, among his later successors being Aparājita, who vanquished the Pāndya king, Varaguna II, at the battle of Sri Purambiya, but was himself overcome by the Chōla king Aditya I, about the end of the ninth century. From that time Pallava supremacy finally passed away and was transferred to the Chōlas, who brought all the southern kingdoms under their control during the tenth and eleventh centuries.

c2.27

At this period, Pallava architecture ceases to exist as a separate style and merges into that of the Chōlas.

c2.28

II.—RELIGION.

c2.29

The numerous important Buddhist remains at Amarāvatī, Guntapalli, Anakapalli, Rāmatirtham and other places in the northern part of the Madras Presidency, and the Jaina monuments in the southern districts, show that these two creeds flourished side by side in the seventh century A. D. and were in fact. the prevailing religions of the country at that period. Although it is probable that some of the Pallava kings were either Jainas or Buddhists, such names as Siva-Skandavarman and Skandavarman borne by some of the earlier kings, appear to show that some of them were Saivas at a very early date. The fact that they had the bull for their crest and the club (khatvānga) for their banner corroborates this. On the other hand some of the kings who issued the Sanskrit charters were probably adherents of the Vaishnava faith as indicated by such names as Simhavishnu, Vishnugōpa, etc. The fact that some of the Pallava Siva temples contain images of Vishnu along with the lingas shows that both deities were worshipped with equal veneration and that the Pallavas must have been wonderfully free from religious bigotry.

c2.30

After Mahēndravarman I became converted to Saivism by the saint Appar as related above, he excavated a number of cave-temples and dedicated all of them, except one, to Siva. The one exception is the temple at Mahēndravādi which was dedicated to Vishnu. Although apparently at first a Jaina, it is stated that after he became converted to the cult of the linga, Mahēndra demolished the Jaina monastery at Pataliputtiram, a seat of Jaina learning in the South Arcot district, and built a Siva temple on the spot. As a Jaina he is also said to have persecuted the Saivas and the saint Appar in particular. However, be this as it may, he evidently did not persecute the Vaishnavas nor the Buddhists; for the Chinese pilgrim Hiuen Tsiang who visited Kānchīpuram about the end of Mahēndra's reign, says with regard to the religious condition of the Pallava country—"There are some hundreds of Buddhist monasteries and ten thousand priests. They all study the teaching of the Sthavira School belonging to the Great Vehicle. There are some eighty Dēva temples and many heretics called Nirgranthas." Even at the present day, there are sufficient Buddhist antiquities surviving at Conjeeveram to testify to the truth of Hiuen Tsiang's statement.

c2.31

Mahēndra's son and successor, Narasimhavarman I, surnamed Māamalla, was just as zealous a devotee of Siva as his father had been before him. This is exemplified by the wonderful group of temples at the Seven Pagodas, most of which are dedicated to Siva and were excavated during Māamalla's reign.

c2.32

Narasimhavarman II, surnamed Rājasimha, was equally devout and built a number of structural temples in honour of Siva at Māhabalipuram, Conjeeveram and Panamalai in the South Arcot district. In fact Rājasimha appears to have devoted his whole life to temple building and bestowing gifts upon the Brāhmans.

c2.33

The result of Buddhism flourishing alongside this revival of Saivism, was that early Hindu architecture became strongly influenced by the older art of the Buddhists.

c2.34

III.—ARCHITECTURAL CHARACTER.

c2.35

The architectural character of early Pallava monuments is simple and severe but indicative of great strength and durability. Being excavated in the natural rock, the temples of the earlier period have but one external facade which is in the face of the rock, and therefore, the architecture is mainly internal. In most styles of architecture, we can generally distinguish three distinct periods, an early period, an intermediate period and a late period. Since the style of one period naturally merges into that of the next, the line of demarcation is not always well defined in the intermediate period, but the difference between the early and late periods is always apparent. The classification of Pallava monuments is a comparatively simple matter, as the differentiation is unusually well marked in each period. In the first place, all the earlier Pallava monuments are rock-cut cave-temples with only one external facade cut in the face of the rock. Secondly, monuments of the intermediate period, although all are monolithic, comprise free-standing rock-cut temples commonly known as rathas as well as cave-temples. The latter possess certain features which are not found in the cave-temples of the earlier period. Thirdly, monuments belonging to the late period are structural buildings built of stone and brick.

c2.36

All of the earlier Pallava cave-temples were excavated during the reign of Mahēndravarman I, in the early part of the seventh century and may, therefore, be said to belong to the Mahēndra period. Those belonging to the intermediate period were mostly cut out of the rock in the reign of Māmalia during the latter half of the seventh century and, therefore, belong to the Māmalia period. The structural monuments of the later period were started by Rājasimha at the beginning of the eighth century and thus belong to the style of Rājasimha. There is also a still later style than that of Rājasimha, dating from about 800 to 900 A. D., which may be called the style of Nandivarman. There are very few temples in this last-named style as at this period we find Pallava architecture losing its identity and merging into the Chōla style.

c2.37

We may, therefore, adopt the following classification, and divide the history of Pallava architecture into four styles :—

c2.38

(1) Mahēndra style, 610 to 640 A. D.
(2) Māmalia style, 640 to 674 A. D.
(3) Rājasimha style, 674 to 800 A. D.
(4) Nandivarman style, 800 to 900 A. D.

c2.39

This division of Pallava architecture applies chiefly to the country lying immediately round Kānchīpuram, i.e., Tonda-mandalam. Further south, between Tanjore and Pudukkottai for example, Pallava architecture was influenced by the neighbouring cultures of the Pandyas and the Chōlas. However, as all the most important groups of Pallava monuments are situated in the districts around Kānchīpuram, the classification of style defined above will serve all practical purposes in studying the art and architecture of the Pallavas.

c2.40

(1) Mahendra style.—All monuments in this style are subterranean rock-cut excavations usually known as cave-temples. They have but one external facade which is in the face of the rock. On plan, the temple consists of a rectangular pillared hall with a small square shrine chamber excavated in one of the side walls. As a rule, the shrine faces the east or the west, usually the latter. With the exception of the Vishnu temple at Mahendravadi, all are dedicated to Siva and originally contained stone lingas or images of that deity. The lingas were large and cylindrical in shape, highly polished and mounted on yoni pedestals of the usual kind. The shrine chamber is free from all ornament within, but usually has a dvarapala, or doorkeeper, carved in high relief standing on each side of the entrance into the sanctum. The external ends of the facade are also often provided with figures of doorkeepers, one at each end protecting the outer entrance. Some of these figures are very large and have a characteristic pose which is somewhat different to that of the doorkeepers of the later periods. They generally stand facing the spectator, in rather an aggressive attitude with one hand resting on a huge club. Others have one hand raised to the head in the act of adoration.

c2.41

A very characteristic feature of temples in this style is the type of pillar found in them. These are about two feet square in section and about seven feet in height. The upper and lower portions are cubical, while the middle portion of the shaft has the angles bevelled off, which makes the middle third octagonal in section. Sometimes the cubical portions are decorated with a conventional lotus flower design similar to the lotus medallions appearing on the stone rails of the Amaravati stupa. The capitals of the pillars are simple corbels or brackets supporting the architrave above. The lower or underside of the bracket is rounded upwards and sometimes decorated with horizontal rows of roll ornament. Each pillar has a corresponding side pilaster.

c2.42

The cornice of the facade and that over the shrine doorway when decorated, usually takes the form of a heavy projecting convex moulding decorated with the Buddhist gable-window ornament. This ornament simulates a miniature barrel-vaulted roof decorated with little horse-shoe-shaped gable-windows. Human heads are portrayed peering through the windows and the gables have large flat-headed finials shaped like garden spades. [Plate I (a).]

c2.43

Another Buddhist feature that is sometimes met with in early Pallava monuments is the Buddhist rail ornament. Floriated toranas or arches, are also sometimes depicted in bas-relief, spanning an entrance or as an ornament over a niche containing an image. [Plate II (b).]

c2.44

The floor of the temple is usually raised a few feet above the natural ground-level and approached by a flight of rock-cut steps. In many cases the steps have been left unfinished, indicating that this work was always left to the last. When inscriptions occur in these temples, they are generally to be found engraved on the pillars or along the architrave of the facade.

c2.45

(2) Mamalla style.—The monuments in this style are of three kinds :—Cave-temples, monolithic free-standing temples commonly known as rathas and rock sculptures. The

c2.46

cave temples are excavated in the face of the rock in the same manner as those belonging to the earlier period but their facades are usually more ornamental and contain pillars of a different variety. On plan, the interiors are much the same as those of the cave-temples in the Mahendra style, with similar small square shrine chambers cut in the back wall, which are usually free from ornament within. The pillared hall in front of the shrine often contains large panels cut in the side walls filled with sculptural figures in high relief. These are usually of great beauty and executed with remarkable skill. The interior too, of the more ornamental cave-temples, is provided with cornice and plinth mouldings. The cornice is decorated with the Buddhist gable-window ornament often with a frieze of sacred geese below it and the plinth with two or three flat horizontal bands or mouldings running all round the base of the hall.

c2.47

The most striking feature about the Māmalla style is the curious shape of the pillars. The square heavy pillars with corbel capitals of the earlier period are replaced by pillars of a more elegant shape and better proportion. The base of the pillar is carved in the shape of a conventional lion sitting very erect and carrying the shaft of the column on the top of its head. The shaft is still octagonal in section but of better proportion and crowned with a bulbous capital having a flat abacus. Between the capital and the architrave is usually a double bracket supporting the cornice. This latter feature is obviously a stone copy of a wooden model.

c2.48

The so-called rathas are rock-cut models of structural temples chiefly of timber construction. They vary in style and on plan but are decorated in the same manner as the cave-temples of this period. They will be described in detail later on.

c2.49

The rock sculptures are large bas-relief sculptural scenes carved on the face of a cliff standing in the open air and are different to those within the temples. There are only three or four examples of this type of monument, all of which are situated at the Seven Pagodas.

c2.50

The earliest sculptural representations of Indian deities are usually portrayed as ordinary mortals with only one pair of arms. As a rule, additional heads and limbs denote a later period, but there are exceptions to this rule, as we find images of Siva with four arms at the very beginning of Hindu art in the seventh century. The big image of Siva in the Trichinopoly cave-temple is an example of this. We know from the inscription in this temple that this image was carved at the same time as the rest of the temple and is not a later addition. But images of the minor deities and figures of doorkeepers are usually portrayed with only one pair of arms. In later times they are portrayed with four arms. Sacred symbols, such as the sankha (conch) and chakra (discus) are represented in early Indian art without flames of fire issuing from their sides. In later art (after 800 A. D.) these symbols are decorated with flames of fire. Figures of the gods and ordinary mortals portrayed in early Indian art are of much better proportion and more dignified and lifelike than those of the later period. The early Indian sculptors took Nature as their model and were not handicapped in expressing their ideals by a set of conventional rules of art like those laid down in the Silpa Sastras, which were strictly followed to the detriment of Indian art by the sculptors and builders of later times.

c2.51

One of the most interesting points about monuments in the Māmalla style is, that the architectural details such as the pillars, architraves, cornices and the roofs of the rathas, show unmistakably that these monuments are stone models of structural buildings mainly of timber construction. The brackets, joists, rafters, and cross beams that would be required in a half-timbered construction are all laboriously reproduced in stone without the slightest structural advantage being gained. This clearly shows that ordinarily, the masons who produced these rock-cut temples were accustomed to erect structural buildings of half-timbered construction. The pillars and frame-work being of timber with the interstices filled with brick and plaster, while the ornamentation was in stucco and wood-carving. It was no doubt the perishable nature of such structures which led the masons to reproduce these buildings in stone when ordered to erect everlasting memorials in honour of their religion and to the greatness of their kings. It is only the religious monuments which are executed in stone. No remains of palaces, public buildings or domestic architecture have been discovered and yet we know that such buildings must have existed at this period.

c2.52

Examples of the Māmalla style are only found at the Seven Pagodas. It will be remembered that it was Māmalla who founded Māmallapuram, naming the town after himself. Although most of the monuments there may be said to be in the Māmalla style, it is quite certain that all of them could not possibly have been excavated during his life-time, as the work must have taken over a century and even then was never finished. Therefore, we may presume that Māmalla's son and successor Mahendravarman II, and also the latter's son Paramēśvaravarman I, carried on the work begun by Māmalla, keeping more or less to the original style. Being rock-cut monuments their orientation was somewhat a matter of chance and not of choice perhaps, but most of them face the east or the west, usually the latter. A few were dedicated to Vishnu but most of them are dedicated to Siva and his consort Pārvatī, the latter sometimes in the form of Dūrga. Several of the temples were dedicated to Siva in the form of Sōmaskanda and these have a panel or niche carved in the back or west wall of the sanctum immediately behind the linga and facing the east. This panel contains a bas-relief sculpture representing Siva and Pārvatī seated on a pedestal with the child Skānda between them. Sometimes Brahma and Vishnu are depicted at the back of the panel as attendants on Siva.

c2.53

As a rule, the interior of the shrine of a Hindu temple is not decorated. It is only in these Pallava temples dedicated to Siva in the form of Sōmaskanda where this very unusual feature is found. Another uncommon feature about these Sōmaskanda temples is the style of the linga enshrined within them. Generally, the lingas of this period, like those of the Mahendra period, are carved out of the local granite and cylindrical in shape, the shaft having a smooth polished surface. But the lingas in the Sōmaskanda temples are carved out of black basalt, specially imported from a distance for the purpose. The shaft of the linga is cut into eight or sixteen vertical facets which are generally slightly fluted and terminate in the centre of the crown of the linga. The fluting is highly polished. Some of these temples contain three lingas of this kind, each in a separate shrine, the one in the central shrine usually being the biggest. Perhaps they are meant to represent Siva, Pārvatī and Skānda, as depicted in the panel.

c2.54

on the back wall described above. In all probability, the rock-cut Sōmaskanda temples at Māmallapuram were excavated at the end of the Māmalla period, possibly by Paramēśvaravarman I, because in the next period, we find the latter's son and successor, Rājasimha, introducing the same uncommon features in his structural Siva temples. With the close of Paramēśvaravarman's reign we come to the end of the history of rock-cut architecture in Southern India.

c2.55

(3) *Rājasimha style.*—With the reign of this king we start a new epoch in the history of South Indian architecture. From this period onwards structural buildings alone are met with. Temples in the style of Rājasimha are built of stone with sometimes a brick superstructure covered in plaster and decorated in stucco. On plan, the shrine is a small square cella surrounded by a circumambulatory passage and faces the east. All Rājasimha temples are dedicated to Siva, presumably in the form of Sōmaskanda, since they all possess fluted basalt lingas and have the Sōmaskanda panel carved on the back wall of the shrine.

c2.56

Externally, a lofty tower rising in tiers which diminish in size as they approach the summit, is built over the central shrine, in front of which is a small porch which leads into a large pillared hall or mandapa. Built up against the external walls of the central shrine, are usually three or more small attendant shrines containing fluted lingas. The bases of the pillars and the angles of the building are decorated with conventional lions mainly executed in stucco. The Kailāsanātha temple at Conjeeveram is one of the best examples of this style of Pallava architecture.

c2.57

(4) *Nandivarman style.*—This style flourished during the second half of the ninth century and is a development of the apsidal-ended temple of the Māmalla period and similar in style to the Sahadēva temple at the Seven Pagodas. It is the intermediate style between the Rājasimha period and that of the Early Chōla. The lingas are cylindrical and generally smaller than those of the Māmalla period. The doorkeepers have four arms. The gable-window ornament is different in style to that of the earlier period. The pillars and angles of the building have no conventional lions. There is no Sōmaskanda panel in the sanctuary. The niches in the external walls of the sanctuary are generally filled with stone images carved in high relief or in the round. The upper portion of the temple is built of brick with the ornamentation in plaster.

c2.58

IV.—EXAMPLES IN THE MAHENDRA STYLE.

c2.59

(CIR. 610 TO 640 A. D.)

c2.60

(1) *Dalavānūr.*—The village of Dalavānūr is at a distance of ten miles by road to the south-east of Gingee in the South Arcot district. It is noteworthy as containing one of the best rock-cut Pallava temples in the district. The temple is dedicated to Siva and is excavated in the southern face of a small granite hill lying to the north of the village and is known locally as the Pancha Pāndava Malai. The temple is no longer in use as a place of worship, although it still contains a stone linga. The facade faces the south but the little shrine containing the linga faces the east. The shrine chamber is 7' 10" long, 8' 6" wide and 6' 10" in height and the walls are free from ornamentation. It contains a cylindrical granite linga, 2' 5" in height and measuring 4' 6"

c2.61

in circumference. The linga is fixed in a yoni pedestal of the usual type which measures 6' 2" in length. Down the front of the linga two narrow vertical lines are cut in the stone. Both the linga and yoni pedestal are detached stones and are not monolithic like the rest of the temple and its ornamentation. Since the facade and entrance hall face the south [Plate I (a) ] and in order that the linga should face the east, the little shrine chamber has been excavated in the west wall of the main hall with its doorway facing the east [Plate I (c) ]. On each side of this doorway are two doorkeepers or dvāra-pālas in high relief, one on each side of the entrance. They are standing figures facing the spectator, with lofty headdresses and each is shown with one hand raised to the head in the act of adoration [Plate II (a) ]. In front of the shrine are two pillars about two feet square in section with corbel capitals, rising seven feet in height from the floor and forming a little porch leading into the sanctuary. The bracket capitals support cross beams carved out of the natural rock forming the roof over the main hall. The latter, including the little porch just mentioned, is a rectangular chamber measuring 21' 10" long, 19' wide and is 8' 10" in height. In the centre of the facade are two handsome pillars with corresponding pilasters at each end of the facade dividing the latter into three openings of equal size for admitting light and air into the interior of the temple. The central opening forms the entrance and has an unfinished flight of steps leading up to it. The floor of the hall is about four feet above the level of the natural ground which tends to heighten the effect of the facade. On each side of the two pilasters at the ends of the facade and facing the south, are two large crudely carved figures about six feet in height representing doorkeepers standing in large panels similar in shape to the three openings in the facade. The one on the left or western end of the facade is depicted with the left hand raised in adoration, and is a similar figure to the one on the left side of the shrine doorway mentioned above. The one on the eastern side of the facade is depicted resting the right hand on a huge club.

c2.62

On the outer surface of the pilaster at the left end of the facade is an inscription written in the Sanskrit language in Grantha-Pallava letters, from which we learn that the temple was excavated by Mahendra's orders (vide Epigraphia Indica, Part V, Volume XII, January 1914).

c2.63

The cubical portions of the two pillars in the centre of the facade are decorated with conventional lotus flower medallions similar to those on the rails of the Amarāvatī stūpa. Above the main entrance and springing from the two corbel capitals is a bas-relief representation of a makara-torana, in which the two ends or volutes of the torana are shown emerging from the mouths of makaras. Above the torana running the entire length of the facade, is a heavy projecting convex cornice decorated with the Buddhist gable-window ornament. With the exception of the linga and the figures of the doorkeepers, the entire ornamentation is Buddhist in character.

c2.64

(2) Trichinopoly.—Half way up the famous Trichinopoly rock and excavated in its southern face, is a fine example of a Pallava temple in the Mahendra style. It is a particularly interesting example because it contains some very valuable inscriptions which prove beyond all doubt that it was excavated by Mahendra's orders. The pillars of the facade contain a number of inscriptions written in the Grantha-Pallava alphabet (vide South Indian Inscriptions, Volume I, Nos. 33 and 34, pages 28 to 30).

c2.65

This alphabet is very archaic and similar in style to the one employed in the inscriptions in the Pallāvaram temple. Engraved on the pilaster to the right of the sanctuary and at a spot which appears to have been selected for the principal inscription, is the following name “ Sri Mahēndra-Vikrama, ” exactly as in the Pallāvaram inscription. The pillars at the other end of the hall contain a number of names and birudas, among them, “ Gunabhara, ” who is said to have had the temple excavated and placed within it a stone linga and a portrait image of himself. The inscription also records that the beautiful large sculptural panel facing the shrine entrance, representing Siva in the form of Gangādhāra, was also executed at the same time and is not a later addition.

c2.66

It is now generally accepted that “ Gunabhara ” was a surname of Mahēndravarman I. One inscription (No. 33 above) states that “ Now that Gunabhara is a worshipper of the linga, let the knowledge that he has turned back from hostile conduct be spread far and wide by this linga.” This evidently refers to Mahēndra’s former persecution of the saint Appar, who afterwards converted him to the cult of the linga as already related above (vide Epigraphia Indica, Volume III, page 277).

c2.67.1

The facade faces the south and is simple and severe in character [Plate I (b)]. It has four cubical pillars dividing the facade into five openings of equal size, with a second row of pillars inside the temple and in line with those of the facade. In style, the pillars are similar to those at Dalavānūr, but they have no bas-relief representation of a torana over the entrance and there is no ornamental cornice above the architrave of the facade. The cubical portions of the pillars are decorated with lotus flower medallions and floriated ornament similar in design to that appearing on the stone rails of the Amarāvati stūpa. The underside of the corbel capitals is decorated with roll ornament [Plate II (c)]. On plan, the temple consists of a rock-cut hall measuring 30' 0" in length, 15' 0" in width and 9' 0" in height. Cut into the east wall is a small sanctuary 7' 10" square and 7' 0" in height. The shrine faces the west. Cut in the centre of the floor of the sanctuary is a socket-hole, 2 feet square, for the reception of the linga, by the side of it is a second socket-hole 1 foot square, the latter being no doubt for the reception of the portrait image of Mahēndra mentioned in the inscription. The shrine is now empty. At one time, the temple was used as a powder magazine, presumably by the French, and the openings between the pillars of the facade were closed with brick walls, which have since been removed.

First 71 of 247 passages. Read on

Pallava Architecture, Part I (MASI 17)

Contents

Text

Front matter

List of Plates

Reported in