The Amaravati Collection

Over 120 carved limestone panels, railings, pillars and inscriptions from the great Buddhist monument at Amaravati in Andhra Pradesh, held by the British Museum since 1880 and displayed in a humidity-controlled room paid for by a Japanese newspaper. They were dug out by Sir Walter Elliot of the Madras Civil Service in 1845, left on open ground in Madras, shipped to London at the end of the 1850s, and moved between four unsuitable buildings before a museum would take them.

They are called the Amaravati Marbles and they are not marble. They are also called the Elliot Marbles, a nineteenth-century coinage made in deliberate imitation of the Elgin Marbles, and the man the name honours is not the first person to have taken the monument apart: that was a local landholder, thirty years earlier, whose workmen fed the stone into local house-building.

Indian calls for their return are live and unresolved, and the retention rests on British statute, not on any argument about the objects. The best collection of Amaravati sculpture anywhere is in Chennai, not in London.

The name is wrong in three different ways

Start with the label, because almost everything published about this collection carries an error inherited from it.

They are not marble. The stone is a limestone known in the trade as Palnad marble, and even the quarry attribution is not decisively proven. The word marble in the popular name describes an appearance, not a geology.

They are not the Elliot Marbles in any sense Elliot would have recognised. That name is a nineteenth-century construction, applied to the whole group he shipped to Madras in explicit emulation of the Elgin Marbles and the other Greco-Roman groupings then defining what a prestigious classical collection looked like. Attaching a British administrator's surname to a south Indian Buddhist monument was a way of making it legible to a London audience as a trophy of the same order.

And Amaravati is not the ancient name of the place. The modern village took its name from the Amaresvara temple built in the eighteenth century. The ancient settlement beside it is Dharanikota, probably a capital. Colin Mackenzie's maps of 1816, the oldest there are, label the monument simply deepaladimma, the hill of lights. The ancient inscriptions do not call it a stupa either; they call it the mahacetiya, the great sanctuary.

Four names for one thing, three of them modern impositions and one of them the site's own. This record uses the British Museum's form for the collection and records the rest.

What the objects physically are

The British Museum's group runs to over 120 pieces: sections of the railing that enclosed the monument, coping stones, crossbars, pillars and pillar fragments, drum slabs from the stone facing of the mound itself, two guardian lions, inscriptions and miscellaneous fragments from several periods. A typical drum slab is about 124 centimetres high, 86 wide and 12.5 thick.

Most of the figurative work is relief, and the most crowded scenes illustrate jataka tales, the vast body of narrative about the Buddha's previous lives. Medallions and roundels carry the finest surviving carving; many hold nothing but a single stylised lotus.

The monument these came off was enormous. Burgess's nineteenth-century estimate put the late railings at roughly three metres tall and fifty-nine metres in diameter, with 136 pillars and 348 crossbars, running some 803 feet in total. Those numbers do not reconcile with each other and the record says so rather than choosing between them. The stupa itself was in its day described as the greatest monument in Buddhist Asia and the most richly decorated stupa known.

Many of these slabs are two sculptures

This is the technical fact that changes how the whole collection should be read.

Amaravati was rebuilt. In the late phase, beginning around the middle of the first century BCE, the exterior surfaces were entirely renewed, and the sculptors did not start with fresh stone. They took slabs already carved in the earlier phase, turned them round, cut new reliefs on the previously plain backs, and remounted them on the drum with the new work facing outward.

The consequence is that the earlier carving on many pieces spent centuries facing inward against the body of the monument, where it was abraded and worn down. A drum slab in the British Museum is dated by the museum on its two faces separately, because the two faces are separated by a century or more of the same object's working life.

Burgess recorded the same economy in the fabric: dig behind the outer rail and you find broken slabs and statues jammed in as fill, and the slate paving of the processional path laid on a concrete of chips, broken slabs and pillars. The monument recycled itself. Anyone doing diligence on a piece of Amaravati sculpture should be asking which face is being sold.

The first excavation was Indian, and it was destructive

Colin Mackenzie reached the site in 1797, the first European to describe it. When he returned in 1816 the monument had been badly damaged, and not by Europeans. The powerful local zamindar Vasireddy Venkatadri Nayudu had run indiscriminate excavations, and much of the stone and brick had been carried off and reused to build local houses.

This has to be stated because the standard account of Amaravati as colonial looting is true and incomplete. By the time Elliot arrived in 1845, the mound had already been quarried by its own landholder for construction material for a generation. The stone was a resource before it was a heritage, to the people who lived on top of it.

What follows from that is uncomfortable in both directions. The British removals took what an Indian landholder's works had left. And the reason anything at all survives to be argued over is that Mackenzie recorded a plan in 1816 of a monument already being destroyed.

Elliot, 1845, and thirteen years on the grass

Sir Walter Elliot of the Madras Civil Service excavated near the west gate of the railing in 1845 and removed a large number of sculptures to Madras. His notes and sketches of the excavation were extensive and most of them were lost on the way back to England, which is a permanent loss of context: the findspots of the pieces he took are largely unrecoverable.

In Madras the sculptures were placed on the greens in front of the college. After enquiries from the Court of Directors of the East India Company they were shifted to a slightly less exposed position in front of the museum. That is the whole of their care for thirteen years: outdoors, on grass, in a tropical climate.

By 1853 the curator of the Madras museum, Dr Edward Balfour of the Madras Medical Service, was reporting to London that the sculptures were deteriorating where they stood, and pressing for them to be moved. To let London decide whether they were artistically important enough to be worth shipping, Balfour commissioned a set of drawings from the Indian artist P. Murugasa Moodaliar in 1855; they are now in the British Library. Dissatisfied with their accuracy, he then had Dr A. J. Scott photograph the sculptures and sent those too. In 1858 Linnaeus Tripe made a further series, preserved in the British Library as Photographs of the Elliot Marbles, shelfmark Photo 958.

Those photographs matter more than their status as illustrations suggests. They record the condition of the stones before the sea voyage and before another century of handling, and they are the reason a scholar can say what has been lost since. They also record the name of an Indian draughtsman in a chain of custody that otherwise names only Europeans.

Twenty years in the wrong buildings

The sculptures were shipped to London on the orders of the Court of Directors at the end of the 1850s. The two standard accounts give 1858 and 1859 for the sailing, and this record does not arbitrate between them.

Their destination was the East India Company's own India Museum, and they arrived just as the Company was being dissolved. So they went into storage at Beale's Wharf in Southwark, and then to Fife House in Whitehall, a private house once occupied by the second Earl Fife. The Indian collections there opened to the public in July 1861. Conditions at Fife House were poor and the Amaravati pieces were damaged by the weather.

That damage produced the intervention that eventually saved them. James Fergusson, who published a book on Amaravati in 1868 and made the first detailed study of the sculpture, appears to have raised his concerns with Augustus Wollaston Franks at the British Museum.

The lease on Fife House expired in 1869 and the India Museum went briefly back to the India Office, where there was so little room that the Amaravati sculpture was sent to the India Office Stores in Lambeth. Meanwhile John Forbes Watson, who had run the India Museum since 1858, fought to keep it whole and lobbied for an Imperial Museum for India and the Colonies. He lost. In 1879 Sir Louis Mallet asked George Birdwood to plan the dispersal of the India Museum, and in 1880 the institution was dissolved and its collections transferred.

The British Museum acquired this material in the break-up of a rival museum, not by purchase and not by gift from India. That distinction is routinely blurred and it matters for any claim: the transfer of 1880 was an internal reorganisation of British imperial institutions.

A stairwell, a war, a basement, and finally a room

Once at the British Museum the sculptures were arranged on the main stairwell, and stayed there for decades until they were removed for safekeeping during the Second World War. They came back out in the 1950s, and London's air at that time - the era of the coal smogs - attacked them again. In 1959 they were moved into an air-conditioned basement, where they remained for over thirty years.

The first proper display came in 1992, when the Sir Joseph Hotung Gallery opened as Room 33 and the Amaravati collection was given the adjoining Room 33a, with controlled humidity and air conditioning, showing about seventy objects including a reconstructed section of the railing and a selection of drum slabs. The room is named after the Asahi Shimbun, the Japanese publishing house that paid for it.

Count the sequence: outdoors on a Madras green, a Southwark wharf, a damp Whitehall house, a Lambeth store, a London stairwell, wartime storage, a stairwell again, a basement for thirty years, and then a gallery. The objects have spent far more of their time in London out of sight and in unsuitable conditions than on considered display, and the argument that a British museum was the safe destination has to be measured against that record rather than against the room as it is now.

The claim, and the statute that answers it

Indian calls for the return of the Amaravati sculptures are active. In October 2021 the Deccan Chronicle reported the historian Dr Jasti Veeranjineyulu writing to the Indian Prime Minister after visiting the British Museum, saying that priceless objects had been taken from Amaravati during British rule and were being shown in gallery 33a. No response from the central government is recorded in that report.

The obstacle is not evidentiary. Nobody disputes where these objects came from, who removed them or when. The obstacle is statutory: the British Museum's ability to transfer objects out of its collection is constrained by its governing legislation, the British Museum Act 1963, and by domestic British law generally, and that is the ground on which the legal literature engages. Francesca Zepeda's article The Stupa Amaravati of India: Where British Common Law Controls the Repatriation of Cultural Property, in the Loyola University Chicago International Law Review in 2024, examines exactly that: the statutory basis of the museum's ownership, the international conventions binding the United Kingdom, and the museum's standing after recent theft allegations, and proposes reconciling the contradiction between British and international law.

The 1970 UNESCO and 1995 UNIDROIT instruments do not reach a removal of the 1840s in any event. What is left is politics and statute, and the statute currently says no.

The best collection is in Chennai

The largest and best collection of Amaravati sculpture is in the Government Museum, Chennai, supplemented by later finds and by the friezes from Goli. There is a substantial collection at the site museum at Amaravati itself, and smaller groups in museums around the world. The British Museum's group is important and it is not the principal one.

Nearly all of the London material comes out of Elliot's shipment, which means the London and Chennai collections are two halves of a single removal to Madras, split by a decision taken in the 1850s about what was worth putting on a ship.

The Chennai collection has its own conservation problem: pieces not in air-conditioned storage have been reported as showing damage from the atmosphere and from salt, and the museum's plans for a climate-controlled gallery have not been realised. That is the honest state of the comparison. Neither city has solved the problem, and only one of them has had the resources of a national museum applied to it since 1880.

The monument is not dead

Amaravati is often written about as a ruin and a source of museum objects. It is also a living site.

The Chinese pilgrim Xuanzang came in 640 CE, stayed some time and studied there, and wrote an enthusiastic account of its monasteries. The place was still being named in Sri Lanka and Tibet as a centre of esoteric Buddhism in the fourteenth century, and a fourteenth-century Sri Lankan inscription records repairs made to the stupa. It is bound into the Vajrayana teaching of the Kalachakra, which is still practised in Tibetan Buddhism, and in 2006 the Dalai Lama conducted a Kalachakra initiation at the site before more than a hundred thousand pilgrims.

That is the fact that distinguishes an Amaravati claim from an antiquarian one. The railings in Room 33a were not detached from a dead religion. They were detached from a sanctuary that people were still travelling to in this century.

What it is worth, and the market that should not exist

There is no sale, no auction record, no valuation and no insurance figure in the public domain for the British Museum's Amaravati collection. It arrived by an administrative transfer in 1880, no consideration passed, and it cannot be sold.

The market fact that matters here is a negative one and it is genuinely useful. The site was quarried by its landholder, excavated without adequate recording by Elliot, further dug by Robert Sewell under James Burgess in the 1880s with drawings but not to a standard anyone would now accept, and re-excavated since. Loose Amaravati stone therefore exists, and the recording is poor enough that a piece with no documented findspot cannot easily be shown to be either legitimate or not. Anything offered as Amaravati sculpture without a pre-1970 published photograph and a named collection should be treated as a title problem.

The published photographic record is the strongest diligence tool that exists for this material: Murugasa Moodaliar's drawings of 1855, Scott's photographs, Tripe's 1858 series in the British Library under Photo 958, Fergusson 1868, Barrett's British Museum catalogue of 1954, the 1992 catalogue, Shimada 2013, and the World Corpus of Amaravati Sculpture run with the Archaeological Survey of India. A piece that appears in none of them is a piece with a question attached.