Udayagiri and Khandagiri Near Bhubaneswar Hold 33 Rock-Cut Caves and a 17-Line Inscription That Records the First 13 Years of King Kharavela of Kalinga

Two low sandstone hills near Bhubaneswar, the capital of Odisha, hold 33 rock-cut caves, according to a Ministry of Culture release of 10 May 2015 that calls them one of the earliest groups of Jain rock-cut architecture in eastern India. Over the mouth of one of them, the Hathigumpha, runs a 17-line inscription in Brahmi script that follows King Kharavela of Kalinga from his boyhood to the thirteenth year of his reign.

The Archaeological Survey of India (ASI) maintains the site. Documentation sheets for the individual caves, posted on the website of the Indira Gandhi National Centre for the Arts and naming the ASI’s Bhubaneswar Circle as custodian, date most of them to the first century BCE and to the Mahameghavahana line of the Chedi dynasty, Kharavela’s family. The state profile in Odisha at a glance places the caves in the setting of today’s state.

Thirty-three caves on two hills

The cave numbers on the sheets run to 18 on Udayagiri and 15 on Khandagiri, which adds up to the ministry’s 33. Udayagiri means hill of sunrise and Khandagiri means broken hill, according to Puratattva, a website that documents Indian monuments. The release gives their older names as the Kumari and Kumara hills, the names Kharavela’s inscription itself uses.

Most of the caves are small cells, often with a benched verandah in front, cut for monks who sheltered there through the rains. Puratattva calls Udayagiri one of the earliest rock-cut excavations in India. A few of the cells carry fine carving, and the Rani Gumpha and Ganesha Gumpha are the best known.

Andrew Stirling brought the hills to notice, and the release gives 1825 as the year. Puratattva records a local belief he heard about their origin: that the hills were once part of the Himalayas and home to sages who cut the caves, and that Hanuman, carrying the mountain to build the bridge for Rama, let it fall here. The region’s best-known later monument, the Konark Sun Temple, has its own article.

What the inscription says, year by year

The Hathigumpha text opens with a salutation to the Arhats and the Siddhas, the Jain exemplars. It names Kharavela as Mahameghavahana and overlord of Kalinga. In the translation by N. K. Sahu that Puratattva reproduces, he plays boyhood games for fifteen years, trains in writing, coinage, arithmetic and law, serves nine years as crown prince, and is crowned after completing his twenty-fourth year.

The record then proceeds by regnal year. The first year goes on repairs: the gates, ramparts and buildings of the fort of Kalinganagara, which a storm had damaged, plus gardens and steps down to the tanks. Without regard for a king called Satakarni, Kharavela sends a large army west as far as the Krishna river in the second year. Being versed in the Gandharva Veda, the traditional lore of music, he fills the city with festivals, acrobatics, dance and song in the third.

The fifth year brings water. Kharavela has an aqueduct dug by King Nanda “three hundred years before” brought into the city. The Odisha Review notes an earlier dispute over whether the figure should be read as 103 or 300 and describes it as now over. Taxes for town and country are remitted in the sixth year, and the Great Victory Palace goes up in the ninth, at a cost of thirty-eight hundred thousand coins in the translation, or 3.8 million.

By the twelfth year the king of Magadha, named Bahasatimita, bows at Kharavela’s feet, and an image of a Jina, a Jain teacher, that King Nanda had carried off comes back to Kalinga. The thirteenth year is the one that concerns the caves: Kharavela has cells cut on Kumari hill for the rest of Arhats. At the wish of the Queen of Simhapatha, the king builds an edifice of 3.5 million stone slabs, and he revives the Tauryatrika, which the Odisha Review glosses as dance, song and concert and the text says had lapsed under the Mauryas. The closing lines call him a worshipper of all religious orders and a repairer of the shrines of the gods.

The tuput article on Ashoka’s edicts covers the Mauryan conquest of Kalinga and Dhauli, which it places about seven miles south of Bhubaneswar.

Reading the inscription took more than a century

James Prinsep published the first translation in the Journal of the Asiatic Society of Bengal in 1837. He worked from a lithograph that Colonel Mackenzie had made in 1820 at Stirling’s request, according to Puratattva’s chronology of readings. Prinsep wrote that the record was “perhaps the most curious that has yet been discovered to us”.

Bhagwanlal Indraji presented a fuller reading to an international oriental congress in 1885 and, Puratattva says, was the first to read the king’s name correctly as Kharavela. K. P. Jayaswal published revised readings from 1917 to 1928, and he and R. D. Banerji issued a joint reading in Epigraphia Indica in 1930. D. C. Sircar followed in 1942 and N. K. Sahu in 1963. The Odisha Review says that once the text could be read, by 1885, it confirmed Kalinga, already known from Ashoka’s edicts, as a kingdom with an overlord and a capital of its own.

The ASI built a masonry shed over the inscription in 1902 to shield it from the weather, according to the documentation sheet for the Hathigumpha. The sheet adds that the cave began as a natural rock shelter that people enlarged, about 16 metres by 9 metres. Its rear wall and ceiling carry engraved and painted motifs of antelope, human figures, footmarks and serpentine lines, which the sheet assigns to the Mesolithic.

Kharavela’s capital, Kalinganagara, has a probable site. B. B. Lal of the ASI excavated the fortified town of Sisupalgarh in 1948, and Puratattva reports that most scholars now accept it as Kalinganagara. The fort covers about 1.2 square kilometres and had eight gateways, two on each side. B. M. Barua had earlier proposed Mukhalingam in Ganjam district, and Jayaswal proposed Tosali.

When Kharavela reigned

The inscription gives no date that scholars agree on, so the caves are dated by comparison. Puratattva lists the proposals by scholar. R. L. Mitra put Kharavela in the fourth century BCE, J. F. Fleet and Heinrich Lüders in the third, and Indraji, Sten Konow, Jayaswal, Banerji and K. C. Panigrahi in the second. R. P. Chanda, H. C. Raychaudhuri, N. N. Ghosh and Sircar chose the first century BCE, and B. M. Barua the first century CE. Puratattva calls the first century BCE the latest accepted date. Odisha Tourism’s page for the caves says the second century BCE.

Indraji had read a Maurya-era year, 165, into the inscription and placed the king in the second century BCE. Jayaswal and Banerji later wrote that the line carries no date. Nabin Kumar Sahu, writing in the Odisha Historical Research Journal, weighed the script, language and art and dated the inscription to the later part of the first century BCE, with Kharavela born about 64 BCE and consecrated about 40 BCE. The Odisha Review article reaches the same birth year by counting 300 years from the reign of King Nanda. Janice Leoshko, as Puratattva reports her, suggests the caves were cut together around 50 BCE.

The Rani Gumpha, a two-storey cave with a long carved frieze

The largest cave on Udayagiri, the Rani Gumpha or Queen’s Cave, wraps around three sides of a quadrangle. The ASI sheet gives its footprint as 25.75 by 24.50 metres and its height as about 8 metres. The upper storey is set back from the lower one, which leaves the roof of the lower verandah as an open terrace. Puratattva says the cave took its name from a belief that the queen of the Somavamsi king Lalatendu Kesari lived in it.

Arched doorways in the lower storey frame nine sculpted panels. Puratattva describes a double-storey building, royal processions under an umbrella, elephant riders, and a king receiving figures who kneel before him. Debala Mitra’s 1960 ASI guidebook, as Puratattva cites it, calls a headband on one kneeling figure a Greek feature. Sahu reads the sequence as Kharavela’s campaign and the kneeling scene as the surrender of the Magadha king Bahasatimita. A reading by Mitra takes it as a Yavana, or Greek, king submitting, and O’Malley, another scholar Puratattva cites, read the panels as a procession for a saint. Puratattva adds that no panel matches a single episode the inscription lists.

Another relief shows a dance, with musicians playing the mridangam, a drum called the dhakka, a harp and a flute. One of the guards on the upper storey is a foreigner in boots and a kilt. Dhiren Dash, quoted by Puratattva, argues that the cave measures exactly 64 cubits and was built as a playhouse of the kind the Natya Shastra, the ancient treatise on drama, specifies. The article on the classical dances of India lists Odisha’s own form, Odissi, among the eight.

Later hands in the same caves

Sahu reads inscriptions in the Manchapuri cave as naming the chief queen of Kharavela on the upper storey, and a Maharaja Kudepa and a Prince Vadukha on the lower, whom he takes to be Kharavela’s son and grandson. The Odisha Review article calls Kharavela the principal builder of 117 caves, a figure above the 33 now listed, and attributes later work to the Bhauma king Santikara, in 891 CE, and to the Somavamsi king Lalatendu Kesari, who ruled from 1040 to 1065 CE.

The Ganesha Gumpha, Cave 10, has two cells and a verandah. The ASI sheet says a Ganesha image was added to the rear wall of the right cell later, and that a five-line inscription of the eighth or ninth century CE records a donation by a physician named Bhimata in the reign of Santikara. The sheet concludes that the cave was in use from Kharavela’s time to at least the Somavamsi period.

On Khandagiri, the Ananta Gumpha is named for twin serpents on its door arches, and the ASI sheet dates it to the first century BCE. The Barabhuji Gumpha cell holds 25 Tirthankara figures, with Parsvanatha twice, each above his Sasanadevi, an attendant goddess. The sheet dates the images to the tenth or eleventh century CE and says local people now worship them as Hindu deities. The Odisha Review counts 24 figures and says the Sasanadevis came to be worshipped as Barabhuji Durga. The Lalatendu Kesari Gumpha was cut in the first century BCE, the sheet says, and carries eleventh-century sculpture.

What the documentation sheets record on condition

The sheets use the old spelling Orissa and so probably predate 2011. They list weathering and vandalism as threats at several caves. They record quarrying that took the floor of the Alakapuri Gumpha down by 2.2 metres, which that sheet places in the early medieval period, and the floor of the Lalatendu Kesari Gumpha down by 5.6 metres. The Barabhuji sheet says that vermilion, oil and water applied to the images risk the carving, and that a concrete roof and a new Shiva temple have been built in front of the cave. The Rani Gumpha sheet says that lighting installed by the state tourism department was unscientific.

The same documents record the ASI’s work. Pillars were re-erected at the Rani Gumpha, Alakapuri, Manchapuri and Navamuni, the Ananta Gumpha verandah was paved and given glass doors, and in 1958 the ASI exposed an apsidal laterite structure on Udayagiri, whose plan the sheet likens to a Buddhist chaitya hall. Khandagiri was declared protected in 1912, Puratattva reports, citing a 2015 chapter by Umakanta Mishra. Hindu and Jain stakeholders contested its custody until 1947 and after, according to the same source.

A site in use

Odisha Tourism lists visiting hours of 8:00 am to 5:00 pm, entry fees under ASI norms, and October to March as the best season. Puratattva calls the hill equally sacred to Hindus and Jains and one of the most important pilgrimage sites for Jains in Odisha. It reports that the Khandagiri Mela is held every year on Magha Saptami, when the presiding deity Ananta Kesari Dev is brought in procession from the village of Jagamara.