The Concert

An unsigned Vermeer of the mid 1660s, 72.5 by 64.7 centimetres, showing three people making music in a room hung with two other pictures. It is the most valuable unrecovered stolen object in the world, taken from the Isabella Stewart Gardner Museum in Boston on 18 March 1990 by two men in police uniform who were inside the building for eighty-one minutes and left with thirteen works.

Its frame is still on the wall of the Dutch Room, empty. That is not a memorial gesture. Isabella Stewart Gardner's will forbids the trustees to change the arrangement of the museum she left, so the frame stays where it was because moving it would be a breach of trust. The museum carried no theft insurance, which means there was no settlement, no subrogation and no insurer with a claim on the picture. Title never moved.

In March 2013 the FBI announced that it had identified the thieves and would not name them. The statute of limitations on the theft itself expired in the mid 1990s, so the men who took it cannot now be prosecuted for taking it. Whoever holds it can still be charged with possessing it. Every reproduction on this page, and every reproduction anywhere, predates 1990.

The most valuable stolen object in the world is a picture of three people making music

Oil on canvas, 72.5 by 64.7 centimetres, unsigned and undated. A woman sits in profile at a harpsichord, its lid painted inside with a landscape. A man with his back to us sits between her and a second woman who stands singing, holding a sheet of music. On the carpeted table in the foreground lies a cittern, and on the tiled floor at the front of the room a viola da gamba rests untouched.

Scholars date it across a narrow band and cannot agree inside it. Albert Blankert gave 1664, Arthur Wheelock about 1665 to 1666, Walter Liedtke about 1663 to 1666, Wayne Franits about 1664 to 1665, and the curators of the 2023 Amsterdam exhibition about 1662 to 1664. There is nothing on the canvas to settle it. Vermeer dated only three pictures in his life and this is not one of them.

The room is a Vermeer room: a marble floor in black and white, a wall broken by pictures, light entering from the left. Willemijn Fock has shown that marble floors of this kind were rare in seventeenth-century Dutch houses and largely confined to the entrance halls of the very rich, so the floor is at least partly a construction. Philip Steadman noticed that the black and white marble tiles Vermeer painted are exactly twice the size of the ceramic tiles in his earlier interiors, which suggests he was working from a grid he already knew and enlarging it. In the Concert the checkerboard is broken: a white tile is periodically replaced by a black one so that groups of five black tiles form crosses.

The picture inside the picture is a brothel scene, and it belonged to his mother-in-law

On the back wall, to the right, hangs Dirck van Baburen's The Procuress of 1622. It shows a young woman playing a lute, a bearded client, and an old woman with her open palm out for payment. It is not an invention. The picture, or a version of it, belonged to Maria Thins, Vermeer's mother-in-law, in whose house on the Oude Langendijk he lived and worked; she came from Gouda, had patrician connections and collected the Utrecht painters who had gone to Rome and come back with Caravaggio's lighting.

Vermeer used the Baburen twice. Putting a paid transaction on the wall behind an unpaid one has consequences for how the room below is read. Nineteenth-century critics took the hint literally and proposed that the Concert shows a high-class brothel: the standing singer as the procuress, the seated man as the client, the harpsichordist as the girl. Modern readings are cooler. They take the Baburen as a caution set against a scene of order, a warning about where music and company can lead rather than a statement of where they have led.

Above the harpsichordist hangs a second picture, a rugged Arcadian landscape in the manner of Jacob van Ruisdael. Elise Goodman reads it as belonging to the woman beneath it, on the seventeenth-century convention that made a beautiful woman the masterpiece of nature. Arthur Wheelock reads it against her, pointing to the dead tree trunk that Ruisdael used to mean decay. Both readings are arguments by named scholars about what a viewer would have seen. Neither is a statement about what Vermeer meant, and nothing survives in which he says.

The instruments are chosen, not accumulated. This is the only harpsichord in Vermeer's work; every other keyboard he painted is a virginal. It is almost certainly a Ruckers instrument from Antwerp, the make that stood in the parlours of the rich, and it was more probably borrowed for the picture than owned. The viola da gamba is the recurring silence in his musical pictures. It appears in the Music Lesson, the Woman with a Lute, the Lady Seated at a Virginal and here, and in none of them is anybody playing it. It lies on the floor waiting for a player who is not in the room.

A chain that runs through six salerooms

The picture was very probably one of the twenty-one Vermeers that came to auction in Amsterdam on 16 May 1696 from the estate of Jacob Abrahamsz Dissius, who had married into the family of Pieter van Ruijven, the Delft patron who bought a large share of Vermeer's output. The twenty-one made about 1,500 guilders between them, roughly seventy guilders each. That is the whole of the artist's market at the moment his name meant nothing.

Then it moves. Amsterdam again in 1780, in the Strantwijk sale, lot 150, bought by A. Delfos for the man the catalogue calls the Heer van Vlaardingen, who was Diederik van Leyden. Paris in 1804, in van Leyden's own sale run by Paillet, lot 62, bought in by Paillet. London in 1835 at Foster's, lot 127, in an anonymous sale. London again on 2 April 1860 at Christie's, in the Admiral Lysaght sale, lot 49, to a buyer recorded only as Toothe. Paris on 1 April 1869, in the Demidoff sale, lot 14.

Six salerooms in a hundred and seventy-three years, and in the sources consulted here not one of those sales has a recorded price. That is not unusual for the period and it is worth stating plainly rather than filling in. What the sequence does show is a picture circulating steadily among dealers and collectors as a good Dutch interior by nobody in particular, at prices nobody thought worth writing down.

The man who invented Vermeer bought it a month before he died

The buyer at the Demidoff sale of 1 April 1869 was, on the evidence of the catalogue annotations, Etienne Joseph Theophile Thore, who wrote as Bruger and is known as Thore-Burger. He was a republican journalist who had been exiled after 1849, and he had spent the 1860s assembling out of scattered museum labels and saleroom notices the artist that art history now calls Vermeer. His catalogue raisonne appeared in the Gazette des Beaux-Arts in 1866 and attributed something over seventy pictures to a painter who had been effectively invisible for a century and a half.

He was wrong about most of them. Of the seventy-odd he gathered, about thirty-four are accepted today. That is the standard fate of a first catalogue and it does not diminish the achievement: before Thore-Burger there was no Vermeer to be wrong about.

He owned the Concert for at most a month. Thore-Burger died in April 1869, weeks after the Demidoff sale. About a hundred and twenty pictures went to Apolline Lacroix, a family friend, and her family kept them together until December 1892, when a selection was put up at the Hotel Drouot in Paris. So the picture that would become the most valuable stolen object in the world spent twenty-three years in a private house in France as part of the estate of the critic who had rescued its author from oblivion, and then was sold off.

The handkerchief, and the two national museums that stopped bidding

The Thore-Burger sale was held at the Hotel Drouot on 5 December 1892. The Concert was lot 31. Isabella Stewart Gardner, a Boston collector then in her early fifties, was in the room but did not bid. Convention did not have women bidding in Paris salerooms, and in any case a known American buyer in the room raises the price. She signalled with her handkerchief to Fernand Robert, her regular agent in Paris, and Robert bid.

The underbidders were the Louvre and the National Gallery in London. Both had sent representatives. The accounts agree that the two agents worked out that they were bidding against each other rather than against a private buyer, and withdrew. Robert took the lot.

Two figures circulate for what she paid, and they are not in conflict. The Gardner Museum's own records give 29,000 francs; other accounts give 31,175 francs. The second is the first plus exactly 7.5 per cent, which is the buyer's charge levied at the Hotel Drouot in that period, so the likeliest reading is that 29,000 francs is the price at which the lot was knocked down and 31,175 francs is what Robert actually paid out. This record treats 29,000 as the hammer and 31,175 as the total, and flags the reading as an inference from the arithmetic rather than a fact read off the sale record. Dollar equivalents of 5,000 and 6,000 both appear in the literature. They are later conversions by writers, not the price of anything.

An underbidder establishes the depth of a price, and here the underbidders were two national collections. In 1892 the Louvre and the National Gallery both thought this picture worth competing for, and both were willing to be outbid at a level a private American buyer could reach. That is the market position of a Vermeer four decades before Vermeer became unbuyable.

Eighty-one minutes

Fenway Court opened on 1 January 1903 and the Concert hung in the Dutch Room on the second floor, catalogued P21W27. It stayed there for eighty-seven years.

Shortly after midnight on 18 March 1990, in the hours after St Patrick's Day, two men in Boston police uniform presented themselves at the museum's side entrance on Palace Road and said they were responding to a disturbance. Rick Abath, a twenty-three-year-old night watchman, let them in. They told him to step away from the desk, which took him away from the only alarm button in the building. He and a second guard, Randy Hestand, aged twenty-five and rarely on nights, were handcuffed, wrapped in duct tape and left in the basement.

The thieves were in the building for eighty-one minutes. They took thirteen objects: this picture, Rembrandt's The Storm on the Sea of Galilee and A Lady and Gentleman in Black, a Rembrandt self-portrait etching, Govaert Flinck's Landscape with an Obelisk, Manet's Chez Tortoni, five works on paper by Degas, a bronze gu beaker of the Shang dynasty and a French imperial eagle finial. The two large Rembrandts were cut out of their frames. Nothing has ever been recovered.

One detail from that night has never been explained. The museum's motion detectors recorded no one in the Blue Room, where the Manet was taken from a frame beside the security desk; the only movement the system logged in that room all night was Abath on his rounds before the thieves arrived. Abath maintained to the end of his life that he had done nothing wrong, and he was never charged. He died at his home in Brattleboro, Vermont, in February 2024, aged fifty-seven.

What the FBI said in 2013, and what it did not say

On 18 March 2013, the twenty-third anniversary of the theft, the FBI held a press conference in Boston. Richard DesLauriers, the Special Agent in Charge, was joined by the United States Attorney Carmen Ortiz, the case agent Geoff Kelly and the museum's security director Anthony Amore.

The Bureau said it had identified the men who entered the museum, and that it had done so on the basis of evidence rather than inference. It said the thieves were associated with a criminal organisation based in the mid-Atlantic states and New England. It said that in the years after the theft the works were moved to Connecticut and to the Philadelphia area, and that around a decade earlier some of them had been offered for sale in Philadelphia.

Then it stopped. The FBI declined to name the men, declined to name the organisation, and said the investigation was active. And it said, in the same breath as the announcement that the thieves were known, that it did not know where the pictures were. The trail it had reconstructed ran out at the attempted Philadelphia sale.

The announcement is therefore two things at once, and the coverage generally reported only the first. It is a claim to have solved the identity question. It is also, read carefully, an admission that solving the identity question did not recover a single object, and that the Bureau had no current location for anything. Thirteen years on from that press conference, in 2026, that remains the position.

Why nobody can be prosecuted for taking it

The federal statute of limitations on the theft ran out in the mid 1990s. Carmen Ortiz put the period at six years from the crime; the museum's own chronology and most reporting date the expiry to 1995. Either way, by the time the FBI announced in 2013 that it knew who had done it, the men who did it could not be charged with doing it.

That is not the end of the criminal exposure, and the distinction matters to anyone holding one of these objects. Possession of stolen property and concealment are continuing offences, and the clock on them runs from the possession, not from 1990. Ortiz was explicit that charges remained available against anyone currently holding the works or anyone who had misled investigators. She was equally explicit that she would not offer blanket immunity in advance, but that immunity was a strong possibility for someone who came forward and produced the objects.

The practical shape of the case is therefore unusual. The prosecution of the theft is dead. The prosecution of the custody is alive, and is the only lever the government has. Everything the FBI has said publicly since 2013 has been addressed not to the thieves but to whoever is sitting on the pictures now, or to whoever knows who is.

The reward is a reward. It is not a valuation

Three different kinds of number attach to this picture and they are routinely printed as if they were one.

The first is the reward. The museum's trustees offered one million dollars within three days of the theft, raised it to five million in 1997, doubled it to ten million in May 2017 with a year-end expiry, and in January 2018 extended it indefinitely. It is payable for information leading to the recovery of all thirteen works in good condition. It is the largest reward ever offered by a private institution, and it is a bounty on an outcome, not an opinion about what anything is worth.

The second is the aggregate estimate for the thirteen objects, which is a press and law-enforcement figure and has drifted upward with retelling: about 200 million dollars in the FBI's early statements, 500 million by the 2000s, and 600 million in the mouths of some dealers by the end of that decade. Nothing was appraised to produce these.

The third is the figure attached to this picture alone, commonly given as 250 million dollars and traceable to the accounts by Ulrich Boser in 2009 and Stephen Kurkjian in 2015. It is an author's estimate of what a Vermeer would fetch if a Vermeer could be sold, in a market where no Vermeer has come up at auction in living memory. It is not a hammer price, not a premium, not an insurance valuation and not a probate figure.

The insurance point is the one that most reporting misses. The Gardner did not carry theft cover. The museum has said its works were insured for damage, restoration and conservation but not against theft, on the reasoning that premiums for a collection of this value would have been prohibitive and that no payout could have replaced what was lost, since the will forbids buying anything. So there was no settlement in 1990, no insurer stepped into the museum's shoes, and no subrogated claim exists. Legal title to the Concert has never left the Isabella Stewart Gardner Museum, which is why the museum, and not a syndicate, is the party offering the reward.

The empty frame is a legal object

Visitors are told that the Gardner leaves the frames empty as a gesture of hope, or of mourning, or as a standing provocation to whoever has the pictures. Those readings are all available. None of them is the reason.

Isabella Stewart Gardner died in 1924 and her will requires that the collection be kept for the education and enjoyment of the public forever, arranged as she left it. Nothing may be moved. Nothing may be sold. Nothing may be added. The trustees do not have the discretion to rehang the Dutch Room around a gap, because rehanging the Dutch Room is precisely what they are forbidden to do. The empty frames are what compliance looks like.

How hard that constraint bites can be measured by what it took to build an extension. When the museum wanted to add its new wing it had to go to court in 2008 for permission, and its own leadership acknowledged afterwards that the addition was technically a breach of the will and was made because there was no alternative. A museum that needs a judge's leave to put up a building is not a museum that can quietly take a frame off a wall.

So the frame in the Dutch Room is not a curatorial decision that could be revisited. It is the operation of a testamentary instrument written in 1924 on an object stolen in 1990, and it will hold whether or not the picture ever comes back.

There is no photograph of this painting as it is now

Every reproduction of the Concert in existence was made before 18 March 1990. That includes every plate on this page. It is a limitation that has to be stated on any record of a missing object, and it has consequences that go beyond illustration.

The picture has not been examined since 1990. There is no infrared reflectogram made with modern equipment, no X-radiograph a student can consult, no pigment analysis from the current generation of instruments, no IIIF manifest, no technical bulletin. The entire physical study of this painting was completed under twentieth-century conditions and cannot be extended. Where a comparable Vermeer in a museum gains a new layer of evidence every decade, this one is frozen at the state of knowledge of a generation ago.

The condition question is worse. Whatever the painting looked like in March 1990, that is not what it looks like now. It has spent thirty-six years somewhere unknown, out of climate control, in the hands of people with an interest in concealment rather than conservation. The two Rembrandts were cut from their stretchers, which tells us something about the handling. The Concert was removed from its frame. Any recovery would be followed immediately by a conservation report, and nobody can say now what that report would contain.

The oldest reproduction reproduced here is a heliogravure of about 1887 to 1892, made in Paris by the plate-maker Simonet after the painting while it still hung in the Lacroix house, and printed as a book plate. It is now in the Rijksmuseum. It was made to sell the idea of Vermeer to a European readership, in the years immediately before the picture crossed the Atlantic. It is also, at 136 by 116 millimetres, an object in its own right with its own inventory number, and it will outlast the argument about where the painting is.

The painter as an authenticator

One episode in Vermeer's life belongs in a provenance archive more than in a monograph. In the early 1670s the Amsterdam dealer Gerrit van Uylenburgh sold a group of Italian paintings to Frederick William, the Elector of Brandenburg, who was then told by his adviser Hendrick Fromantiou that they were copies and sent twelve of them back. Uylenburgh assembled a panel of thirty-five painters to give a counter-opinion. Vermeer was one of them, and he travelled to The Hague to pronounce on whether the pictures were what they were said to be.

He was, in other words, hired as an expert in version confusion: the question of whether a canvas is the original or a copy after it, argued in front of a client who had paid for one and suspected he had the other. Three centuries later the same question is the reason this publication exists. It is also, in a colder sense, the reason the Concert is unsellable. A stolen Vermeer cannot be authenticated in public, cannot be exhibited, cannot be catalogued and cannot be insured, which is why the FBI's working theory has always been that these objects have been used as collateral and currency inside criminal networks rather than sold to a collector.

Where the case stands in 2026

The ten million dollar reward remains open with no expiry. The FBI investigation remains open, and the Bureau has not added materially to what it said in 2013. Rick Abath, the guard who opened the door, died in 2024 without changing his account. Most of the men whose names have circulated as suspects, among them Bobby Donati, who was murdered in 1991, and George Reissfelder, who died the same year, are dead. Robert Gentile, the Connecticut figure the Bureau pressed hardest in the 2010s, denied knowing anything to the end.

Nothing about the legal position has changed. Title is the museum's. The theft cannot be prosecuted. Possession can. The frame is on the wall because a will says so.

A reader who wants to know what this object is worth should note that the only number anyone has ever actually paid for it is 29,000 francs, at the Hotel Drouot, on 5 December 1892.