Salvator Mundi

A walnut panel bought at an estate sale in New Orleans in April 2005 for a sum recorded on the auction house's own prices-realised sheet as 1,175 dollars, restored over roughly twelve years by Dianne Dwyer Modestini, exhibited at the National Gallery in London in 2011 as Leonardo, sold privately in 2013 twice in two days at 80 and then 127.5 million dollars, and sold at Christie's New York on 15 November 2017 at a hammer of 400 million dollars, 450,312,500 dollars with fees. It has not been publicly seen since.

The crux of the record is not the price. It is a 46-page book. In December 2019 the Louvre printed Leonard de Vinci. Le Salvator Mundi, by its own curator and its own laboratory scientists, concluding that the picture is by Leonardo. The book went briefly on sale in the museum bookshop, was withdrawn, and the museum later told a reporter that it did not exist. A national museum's attribution was therefore published and unpublished within weeks, and the painting it described was never hung.

Two further things must be held in mind on every page about this object. There are more than thirty extant paintings of the composition and at least three of them circulate under captions that would fit this one. And the picture the world knows is a restored state: photographs of the same panel before cleaning show something markedly different, and this record labels which state every plate shows.

The object, and which state of it you are looking at

A single plank of walnut, painted in oil, showing Christ half-length and frontal, the right hand raised in blessing with two fingers extended, the left holding a transparent sphere of rock crystal. The painted image measures 64.5 by 44.7 centimetres. The panel itself is 65.5 by 45.6 centimetres at the top and 65.5 by 45.1 at the bottom, a taper that Christie's published figure of 45.4 by 65.6 does not record. A check runs the entire length of the panel and passes through the head.

The reverse carries an early nineteenth-century cradle, and on that cradle, in white paint, the letters CC and the number 106. Those two marks are the reason the picture has any pre-1958 provenance at all. Robert Simon matched them to the Cook Collection inventory after the panel reached New York in 2005.

The single most important fact about every reproduction of this painting is that there are at least three different states of it in circulation and they do not look alike. There is the picture as it was photographed in the Cook Collection and again in the 2005 auction catalogue, with a grotesquely repainted face that Martin Kemp described as resembling a drug-crazed hippy. There is the cleaned state, photographed in 2006 and 2007, stripped back to what survived. And there is the restored state, the image that was reproduced worldwide in November 2017 and that almost every reader has in mind when the title is used. All three are the same object. A record that does not say which one a plate shows is not describing the object at all.

The 2005 lot, and four figures that are not the same figure

Warren and Minnie Kuntz of New Orleans bought the picture at the Cook sale in 1958. It passed to Mrs Kuntz's nephew, Basil Clovis Hendry Sr, who hung it in the stairwell of his house in Baton Rouge. He died in 2004. The preliminary appraisal of his estate described it as Continental School (19th century), Portrait of the Head of Christ, oil on panel, framed, poor condition, and assigned it a value of 750 dollars. On the appraiser's list, as reproduced by the Wall Street Journal, there is a red check beside the entry and a pencilled question mark.

Part of the estate went to Christie's in New York. This went to the St Charles Gallery, a small house in New Orleans, for its sale of 9 and 10 April 2005, as lot 664: After Leonardo da Vinci (Italian, 1452-1519), Christ Salvador Mundi, oil on cradled panel, 26 by 18 1/2 inches, presented in a fine antique gilt and gesso exhibition frame, estimate 1200 to 1800, illustrated. The misspelling of Salvator is the catalogue's own.

Four numbers attach to that sale and they are routinely used interchangeably. The estate appraisal was 750 dollars. The printed estimate was 1,200 to 1,800 dollars. The winning bid, as the buyer Alexander Parish has described it, was 1,000 dollars. The house's own prices-realised sheet gives an aggregate against lot 664 of 1,175 dollars. Only the last two are prices, and they are prices of different kinds: 1,000 dollars plus a buyer's premium of 17.5 per cent is 1,175 dollars exactly. That reconciliation is DERIVED here from the two published figures; the St Charles Gallery premium schedule for April 2005 has not been checked against a printed one, and the sheet gives an aggregate rather than a hammer and a premium separately. Note also that Wikimedia Commons, on its own copy of the catalogue page, states that the painting sold for just under 10,000 dollars. It did not.

What the restorer has published about the extent of the restoration

Dianne Dwyer Modestini has written her own account and it is more candid than the press coverage that quotes her. Robert Simon brought the panel to her apartment in 2005. Her husband, the restorer and connoisseur Mario Modestini, was ninety-eight and dying. She removed the recent varnish and repaints in a matter of hours while Simon watched, and was not immediately impressed. Later that afternoon she carried the panel to her husband. He held it, looked at it in silence for a long time, and said that it was by a very great artist whose name he did not know, a generation after Leonardo, and slightly larger than life. Mario Modestini died on 28 January 2006.

The panel went to a specialist for structural work before painting could begin. The background had been repainted a muddy brown in the nineteenth century. Beneath that lay passages of deep black and vestiges of a bright green, and beneath those the evidence that the original black had for the most part perished when the panel was, in her word, brutally scraped, probably in the early nineteenth century. The gouges had been filled with gesso and painted over. In 2011, after comparing the blacks in the freshly cleaned London Virgin of the Rocks, she obtained Simon's permission to remove the brown and retouched the losses to match what remained of the original black.

She also records three specific interventions that a buyer's diligence file would want in writing. She suppressed the pentimento of the blessing thumb, which the scraping of the background had exposed, in consultation with Simon and with Luke Syson, the curator of the London exhibition. She added missing glazes to the abraded orb. She retouched losses in the embroidery of the stole. Before the 2017 sale she unframed the picture again, reworked the folds of the shoulder because the black underpaint had asserted itself, corrected a retouch above the lip with which she had never been satisfied, and applied the final varnish.

The head is the damaged part. When she showed the cleaned-state photograph at the National Gallery technical symposium in January 2012, she records that the condition of the head created a certain amount of consternation. The collected papers were published in 2014, which means the cleaned-state image has been available to anyone who cared to consult that volume since three years before the sale.

The pentimenti, which are the technical case

A pentimento is a change of mind made during painting and covered over. In a composition surviving in many versions, pentimenti matter because a copyist has no reason to make them: they are the argument that a given panel is the prime version rather than a repetition.

Modestini records several. The thumb of the blessing hand was moved, and the earlier, fully modelled position was exposed by the historic scraping of the background. The whole hand was shifted slightly lower and towards the edge of the panel; part of the first position of the palm survives in the crevice between thumb and forefinger. The new outer contour of the third finger was incised into already dry paint, and black lines, some laid in with a liquid medium and some in a dry one such as black chalk, run along the inner contours of the third and fourth fingers. The stole was repositioned late, slightly higher in its first state and at a marginally different angle, with pleats of blue drapery from the earlier design visible under the embroidered threads.

She is equally careful about what the evidence does not establish. The upper part of the blessing hand, above the shoulder line, sits over a thin initial black wash, whereas below the shoulder the area is held in reserve. That is consistent with an early idea in which the hand sat lower, as it does in Marco d'Oggiono's version in the Galleria Borghese and Giampietrino's in the Pushkin Museum. But she notes that Leonardo's heavy use of black in his mixtures, the lead white preparation and the extreme thinness of his paint layers make such a change effectively invisible to X-ray and infrared, and that macro XRF scanning could settle it. Macro XRF scanning requires access to the painting.

Who signed off, and who changed their mind

The attribution was built in stages and the record of who agreed to what, and when, is unusually well documented because the restorer wrote it down.

Carmen Bambach of the Metropolitan Museum saw the picture once in Modestini's studio at the Conservation Center. Asked what she thought it was, and told that Simon and Modestini believed it to be Leonardo, she replied that it was not by Boltraffio. She has since attributed it to Boltraffio with retouches by Leonardo. Modestini's published response is direct: Bambach has said she followed the restoration, and that studio visit was the only time she saw the painting.

Nicholas Penny, shortly before becoming director of the National Gallery, agreed with Simon that it could be Leonardo. Three infrared scans were made at the Metropolitan Museum with an InGaAs scanner, of the head, the blessing hand, and the mantle, stole and orb. Penny then invited five scholars to see it in the National Gallery conservation studio: Bambach; David Alan Brown of the National Gallery of Art in Washington; Maria Teresa Fiorio of the Universita Statale in Milan; Martin Kemp of Oxford; and Pietro Marani of the Politecnico in Milan. No formal statement was issued. Kemp and Brown accepted it. Marani endorsed it, later asked to borrow it for an exhibition, and has since expressed reservations. Fiorio was struck by the well-preserved passages but held that the ruinous state of the head made a firm conclusion impossible. Bambach appeared to accept it and later recanted.

Luke Syson, then curator of Italian paintings at the National Gallery, included it in Leonardo da Vinci: Painter at the Court of Milan, which opened in November 2011. That decision is the hinge of the whole affair, and Modestini notes what it cost the museum in principle: museums as a rule decline to exhibit paintings that are on the market.

Against the attribution: Charles Hope has dismissed it, describing the picture as a ruin with the face much restored to make it reminiscent of the Mona Lisa. Frank Zollner has emphasised the extent of restoration and read the sfumato as closer to the manner of a talented pupil working in the 1520s. Matthew Landrus gives it to Bernardino Luini with a Leonardo contribution of between five and twenty per cent. Jacques Franck has called it a good studio work with a little Leonardo at best, and later gave it jointly to Salai and Boltraffio. Michael Daley of ArtWatch UK notes that there is no documentary record within the Leonardo literature of the artist having been involved in such a project at all.

This record does not adjudicate. It notes only that every position on the list was formed either before the final varnish went on or without access to the object, and that no one has had access since 2017.

The book the Louvre printed and then did not have

In 2018 the painting was shipped to Paris and examined by the Centre de recherche et de restauration des musees de France. In December 2019, two months after the Louvre's Leonardo retrospective opened without it, the museum published a 46-page cardboard-covered booklet, Leonard de Vinci. Le Salvator Mundi, printed in Turin, priced at eight euros, written by the Louvre's curator of Italian painting Vincent Delieuvin with Myriam Eveno and Elisabeth Ravaud of the C2RMF laboratory.

Its preface, by the Louvre's president Jean-Luc Martinez, states that the results of the historical and scientific study allow the attribution of the work to Leonardo da Vinci to be confirmed. The laboratory text states that the examination of the Salvator Mundi seems to the authors to demonstrate that the painting was indeed executed by Leonardo. Delieuvin concludes that the evidence invites one to privilege the idea of a work that is entirely autograph. The findings recorded include a walnut support consistent with Leonardo's Lombard practice, an almost imperceptible underdrawing very close to that of the Mona Lisa and the Saint John the Baptist, glazing consistent with the mature period between 1500 and 1510, and the late introduction of the blessing hand into the composition.

The book was on sale in the Louvre bookshop in December 2019, buried in a pile of exhibition catalogues, and was withdrawn. Dianne Modestini found a copy by chance and scanned it. When The Art Newspaper asked the museum about it, a spokesperson said that no such publication exists, while acknowledging that the book was a project in case the Louvre got the chance to present the painting.

That sequence is the reason this object earns a record rather than a news story. An auction house attribution is a commercial opinion and everyone treats it as one. A national museum's attribution, written by its own curator and its own laboratory, printed, priced and put on a shelf, is a different kind of document. This one existed for a few weeks in a bookshop and then, officially, had never existed. The painting it authenticated has not been hung since.

Three incompatible accounts of an empty wall

The Louvre's Leonard de Vinci exhibition ran from 24 October 2019 to 24 February 2020 and did not include the Salvator Mundi. Three explanations are in public circulation and they cannot all be true.

The first is the museum's own book, which says the picture is entirely by Leonardo. On that account there was no attribution obstacle at all.

The second is Antoine Vitkine's documentary The Savior for Sale, broadcast in April 2021, which reports through anonymous French officials that the C2RMF examination concluded Leonardo had only contributed to the picture and that its authenticity could not be confirmed; that the owner was told; that Crown Prince Mohammed bin Salman then demanded it hang beside the Mona Lisa and be presented as one hundred per cent Leonardo with no qualifying text; and that President Macron refused in September 2019 on the ground that agreeing would amount to laundering a work at 450 million dollars. The film places the episode inside the French-Saudi negotiations over the Al-Ula cultural agreement.

The third, reported separately, is that the sticking point was placement rather than attribution: that the Saudi side required the picture to hang next to the Mona Lisa and that this was impossible for reasons of crowd management and security.

Iconotheca does not choose between them. What is documented is that the loan did not happen, that the museum printed a book saying the picture is autograph, and that the museum subsequently denied the book's existence. Any of the three accounts leaves that pair of facts standing.

Hammer against premium, on every sale in the chain

Every headline figure attached to this object is premium-inclusive or private, and they are not comparable to one another.

2005. The prices-realised sheet gives an aggregate of 1,175 dollars against lot 664. The buyer has described his winning bid as 1,000 dollars. A buyer's premium of 17.5 per cent on 1,000 dollars produces 1,175 dollars exactly. That schedule is DERIVED from the two figures and has not been checked against a printed St Charles Gallery schedule for April 2005.

2013. Neither of the two private sales carries a premium at all, which is why they are the cleanest numbers in the chain. Sotheby's Private Sales approached the three owners in April 2013 on behalf of a client. Yves Bouvier bought at 80 million dollars and sold to Dmitry Rybolovlev the following day at 127.5 million. Sotheby's fee on the transaction has been reported at 3 million dollars. The 47.5 million difference is not a market movement; it is the subject of the litigation Rybolovlev brought against Bouvier in several jurisdictions.

2017. Christie's New York, 15 November 2017, in the Post-War and Contemporary Art evening sale, with an estimate reported in the region of 100 million dollars. Bidding ran, in Scott Reyburn's first-hand account, more than eighteen minutes between clients on telephones, and closed on a jump bid of 400 million dollars. That 400 million is the hammer. The published total was 450,312,500 dollars. The premium is therefore 50,312,500 dollars, or 12.578 per cent of hammer.

That percentage is not a rate; it is the blended result of a stepped schedule. It reconciles exactly, to the dollar, against 25 per cent on the first 250,000 dollars, 20 per cent on the portion from 250,000 to 4,000,000 dollars, and 12.5 per cent above 4,000,000 dollars: 62,500 plus 750,000 plus 49,500,000 equals 50,312,500. That schedule is DERIVED from the arithmetic and has not been checked against a printed Christie's premium chart for November 2017. What can be stated without derivation is that the hammer was 400 million and the total 450,312,500, that the difference of 50,312,500 is the fee, and that a reader who compares the 2017 total against the 2013 private figure of 127.5 million is comparing a fee-inclusive auction total against a net private price.

Thirty versions, and why a caption is not evidence

More than thirty extant paintings derive from this composition or from the related Christ the Redeemer. Only three reproduce the details visible in Wenceslaus Hollar's etching of 1650: this panel, the version formerly with the Marquis de Ganay, and the version in San Domenico Maggiore in Naples. Both of those repeat the elaborate knotwork of the stole and give the orb as translucent, and both differ from this panel in other respects.

Two of the best-known versions have had their traditional attributions destroyed by the wood rather than by the eye. The panel at the Wilanow Palace Museum in Warsaw, accession Wil.1016, is catalogued as Cesare da Sesto and dated 1516 to 1517; dendrochronology carried out in 2005 dated the felling of its oak to about 1586 in the Seine basin, and Cesare da Sesto died in 1523. The panel at the Detroit Institute of Arts is on three horizontally joined boards of Baltic oak felled after 1569, and Giampietrino, to whom it has been given, died in 1549. Neither attribution can stand, and neither has anything to do with connoisseurship.

The practical consequence for anyone assembling a file on this object is severe. Wikimedia Commons hosts a photograph of a version whose file page simultaneously names the Museo Diocesano in Naples as its location, describes it as formerly in the Marquis de Ganay collection, and cross-references a separate Ganay file. Those are two different paintings in two different countries, and the Naples picture is in the treasury of the Basilica of San Domenico Maggiore rather than the Museo Diocesano. Iconotheca has not published that plate, because publishing it would require asserting which of the two pictures it shows. That is the version-safety problem in its ordinary form: not a forger, not a fraud, just a caption written quickly by someone who had two entries open.

The money in period terms, and who the finders were

The 2005 purchase is often reported as though a masterpiece cost the price of a meal. It did not quite. One thousand dollars in New Orleans in April 2005 was a fortnight's wage for a median American worker, and the buyers were not passers-by. Alexander Parish is a professional sleeper hunter who buys on his own eye at regional sales; Robert Simon is a dealer and an art historian of Italian Renaissance painting with a doctorate. They then spent years and their own money on structural work, on restoration, on infrared scanning at two museums and on the scholarly campaign that produced the 2011 exhibition. Warren Adelson, a dealer in American pictures, joined the partnership and took charge of selling it.

The selling campaign failed before it succeeded. The picture was offered to the Dallas Museum of Art after the London exhibition. Modestini went to Dallas to unframe it and spoke to trustees alongside Luke Syson; her impression was that the money was there and the appetite was not, and that Ross Perot expressed active distaste for the picture. Negotiations are said to have run nine months and produced nothing. By 2013 the painting had, in her account, been shopped around to the point of being burned, market slang for unsaleable, and what remained was a buyer from the world of freeports and shell companies. She calls that the great tragedy in the saga of the painting.

The 1958 sale is the other number worth holding. On 25 June 1958, 136 paintings from the Cook Collection were sold at Sotheby's. This one made 45 pounds, about 120 dollars. Kenneth Clark, Ellis Waterhouse and the Florentine dealer Gianni Salocchi were in the room. The post-sale list recorded the buyer as Kuntz and nobody asked who that was for sixty years.

What the chain does not contain

The provenance offered in 2017 ran back to the English royal collection: a Salvator Mundi by Leonardo recorded in the possessions of Charles I, valued at thirty pounds in the Commonwealth inventory, etched by Hollar in 1650 with an inscription naming Leonardo as its author, and traceable thereafter through the Restoration and the Buckingham House sale of 1763.

Set against that, Modestini's account of the 1913 Cook catalogue is flat: the picture had been bought from Sir Charles Robinson in 1900 for 120 pounds as a Luini, and there was no earlier provenance. Borenius and Cook described it as a free copy after the Salvator Mundi by Boltraffio then with Signor G. B. Vittadini at Arcore, and Herbert Cook added that he preferred to call it a parallel work by some contemporary painter of Leonardo's school.

The two accounts are not in direct contradiction, because a chain can be reconstructed after 1913. They are, however, evidence of different weight. Everything before 1900 rests on identifying this panel with a picture in a documentary record, and identification of that kind is an argument, not a receipt. This record grades the pre-1900 links disputed and says why, rather than printing a smooth line from Charles I to Christie's.

Where the painting is

In December 2017 the Department of Culture and Tourism of Abu Dhabi announced that the picture would go to the Louvre Abu Dhabi, and released a photograph of it. In September 2018 the display was postponed indefinitely, with no reason given. In 2018 the panel was in Paris for examination. In 2019 it was not in the Louvre's exhibition. Since then it has been reported hanging on the yacht Serene, then removed to Saudi storage while the vessel was refitted, and latterly held against the opening of a museum at Al-Ula.

None of that is documented in the sense this archive uses the word. There is no accession number, no loan agreement, no exhibition label, no conservation report and no insurance schedule in the public domain. The buyer named in the room was Prince Badr bin Abdullah al-Saud; the Saudi embassy said he was acting as an agent for the Abu Dhabi department; other reporting names Crown Prince Mohammed bin Salman as the owner. Iconotheca records the location as unlocated, which is the honest term for an object whose ownership is asserted by three parties and confirmed by none.

The last authorised photograph of the picture was posted to a government social media account in September 2018. The most expensive object ever sold at auction has been out of sight for longer than the twelve years it spent being restored.