The Kailasa Temple at Ellora Was Cut Downward Out of One Basalt Slope, and Its Builders Removed an Estimated 150,000 to 200,000 Tons of Rock

The Kailasa temple at Ellora, cave 16 in the numbering of the site, is a complete Shiva temple that masons cut downward out of a basalt slope, leaving it standing in a pit they had dug around it. Britannica says construction began in the 8th century, in the reign of the Rashtrakuta king Krishna I, and took out 150,000 to 200,000 tons of solid rock.

It stands in Maharashtra, about 28 kilometres from Aurangabad, according to the Archaeological Survey of India’s report to UNESCO. The cliff around it holds 34 caves in all, which Ajanta and Ellora: carved from rock describes. This piece stays with cave 16.

A temple dug out of a slope

Britannica says the other temples at Ellora were first cut horizontally into the rock face, while the Kailasa complex was excavated downward from the slope and so stays largely open to sunlight. It measures some 50 metres long, 33 metres wide and 30 metres high, with four levels.

The Survey’s report measures the whole court instead. It is 82 metres by 46 metres, “entirely cut out of solid rock” with a scarp 32 metres high at the back. Inside are a shrine for the linga with a tower in the southern Dravidian style, a pillared hall (mandapa) on 16 pillars, a separate porch for Nandi, Shiva’s bull, and five detached shrines, including ones for the river goddesses Ganga, Jamuna and Saraswati. Two pillars carrying emblems stand in the courtyard, with two elephants at a slightly lower level.

The ICOMOS evaluation that went to UNESCO in 1983 says the temple transposes the models of constructed architecture into rock. Mark Cartwright of the World History Encyclopedia lists what that means in practice: bases, beams, columns, capitals, brackets and pilasters, none of them structurally required, all cut anyway.

How much rock had to go

Britannica’s figure of 150,000 to 200,000 tons is an estimate, and the trench sizes below are reconstructions. Cartwright describes the first stage as digging two massive trenches into the sloping hill, each 90 metres long, joined by a connecting trench of 53 metres, and then carving the temple from the block left in the middle.

Subhash Kak, writing in Swarajya in 2023, gives a different layout. He says it has been speculated that two trenches of about 270 by 50 by 100 feet and one of about 150 by 30 by 100 feet were cut to isolate the central mass. By his count roughly two million cubic feet of rock (about 57,000 cubic metres) had to come out before carving of the temple itself could start. He adds that a temple cut from a single rock leaves no margin for error.

The best-known account of how the work was organised was written down some seven centuries after the temple. Kak retells a story from the Katha-Kalpataru, a text of about 1500 by Krishna Yajnavalki. A queen vows to fast until she sees the tower of a temple to Shiva. Architects tell the king that a tower will take at least 16 months, and an architect named Kokasa promises it in a week. He carves the tower first, from the top, and the work proceeds downward from there. In that story he leads 7,000 craftsmen and the temple is called Manikeshvara, after the queen. It is a tradition, not a record, and Kak presents it as one.

Where the design came from

The Survey’s page on Pattadakal in Karnataka says the Virupaksha temple there, itself influenced by the Kailasanatha temple at Kanchipuram, later served as a model for Krishna I when he carved the Kailasa at Ellora. UNESCO dates the Virupaksha to about 740 and says Queen Lokamahadevi built it. Kak says it is about half the size of Kailasa. The Chalukya temples at Pattadakal are covered in Pulakeshin II and the Chalukyas.

So the Rashtrakuta masons took a southern temple form and rebuilt it at a larger scale, in living rock, with no stones to lift. Britannica notes a group of elephants struggling with lions all around the plinth. Britannica also records life-size carved elephants in the complex, and Kak adds that most deities to the left of the entrance are Shaiva and those to the right Vaishnava.

Ravana at the foot of the mountain

The temple is named for Mount Kailasa in the Himalayas, where Shiva resides. The relief that gives the name a picture is on the south face, in the lower storey of the hall. Sahapedia’s description, by the art-history researcher Nikita Rathore, has Ravana kneeling with his back to the viewer, many-armed, trying to lift the mountain. Above him a four-armed Shiva sits and reassures a frightened Parvati with a hand on her shoulder. Monkeys and Shiva’s attendants fill the base.

In the Ramayana’s account, as Kak summarises it, Ravana tries to uproot Kailasa and Shiva presses the mountain down with his toe. Ravana, the villain who was a scholar tells the episode, and What the Ramayana is gives the epic around it. UNESCO calls the carving especially noteworthy. Britannica calls it “a composition of considerable grace and charm”. The Survey’s report notes that the same scene appears in Cave 29, the Dumar Lena.

Paintings also survive. UNESCO says remains of paintings from different periods are preserved on the ceilings of the front mandapa, and the Survey’s report says there are two periods of painting in the hall.

Who ordered it, and how sure scholars are

The proof for the patron is a copper plate. It was issued by Karka II of the Gujarat branch of the Rashtrakutas in Shaka 734, which Sahapedia gives as 812 to 813 CE, a generation after Krishna I. It describes a Shiva temple on the hill at Elapura, the old name of Ellora.

In 1883 R. G. Bhandarkar published a reading of the verses in the Indian Antiquary. Earlier translators, Bühler and Fleet, had taken them to describe a fort. Bhandarkar argued they describe a temple that Krishnaraja had built, and that Elapura becomes Elur in the vernacular. His translation has the architect “suddenly struck with astonishment”, saying “Oh, how was it that I built it!” Bhandarkar’s conclusion was that “very likely it was the Kailasa itself”.

E. Hultzsch, editing the Kadaba plates of the Rashtrakuta king Prabhutavarsha for Epigraphia Indica, agreed that a splendid Shiva temple at Elapura was meant, and noted that nearly a sixth of that inscription is spent describing a temple. He also pointed out that none of the temples at Ellora bears the name Krishnesvara or anything like it, which he thought left the temple’s name in doubt.

Later scholars disagree on how much belongs to whom. According to Sahapedia’s summary, James Burgess dated the temple to Dantidurga, Krishna I’s predecessor, because it resembles the Virupaksha. M. K. Dhavalikar argues from the Baroda plate that Krishna I was the patron. K. V. Soundararajan suggests the upper shrine scheme and the main temple up to the front hall were built under Krishna I, with other parts added later. The Archaeological Survey, as Sahapedia reports it, holds that work began under Dantidurga, that most of it was done under Krishna I, and that it took 200 years and ten generations. ICOMOS wrote that the temple was most likely undertaken in Krishna I’s reign. A Dantidurga inscription survives in Cave 15, the Dashavatara cave, according to Cartwright.

Even the dates of the king vary by source. Britannica gives Krishna I as c. 756 to 773, Sahapedia 757 to 772, the Survey 757 to 783, and Bhandarkar, in 1883, 753 to 778.

The 1983 listing and what came after

UNESCO inscribed the Ellora Caves in 1983 under criteria (i), (iii) and (vi), and the Survey’s report gives the date as 9 December 1983. UNESCO’s text says the Brahmanical caves, 13 to 29, including Kailasa, were excavated between the 7th and 10th centuries, and that cave 16 “marks the culmination of rock-cut architecture in India”. The ICOMOS evaluation of May 1983 says that if one considers only the work of excavating the rock, the temple is “a technological exploit without equal”.

Britannica says some features have been damaged over the centuries, such as a rock-hewn footbridge that once joined two upper thresholds. UNESCO lists seepage and cracking in the caves, visitor management and the capacity of conservation staff among the potential threats to the property.

Later builders copied the temple. The Survey’s report says the Indra Sabha, a two-storeyed Jain shrine at Ellora, is an attempt to copy the Kailasa cave, and ICOMOS saw echoes of it in cave 32.