The Sevso Treasure

Most Roman silver is anonymous. This is not. Across the top of the Hunting Plate, inside a wreath and flanked by a Chi-Rho monogram, runs a two-line hexameter couplet: Hec Sevso tibi durent per saecula multa, posteris ut prosint vascula digna tuis. May these vessels endure for you, Sevso, through many ages, and serve your descendants worthily. It is a dedication, a blessing and, as it turned out, an inventory label that survived its owner by sixteen centuries and made the whole hoard identifiable as a single group.

Directly below it, in the medallion at the centre of the plate, a picnic scene is set beside water, and the water is named. The letters read PELSO. That is the Roman name for Lake Balaton, in the province of Pannonia, in what is now western Hungary. A late Roman silversmith labelled the lake on his patron's dinner service, and in doing so put a place name on an object that would spend the 1980s travelling on paperwork claiming it came from Tyre and Sidon.

Nothing else about Sevso is known. He appears in no inscription, no list of officials, no letter. The name is not common. The Chi-Rho makes him a Christian, the silver makes him very rich, and the scale of the service - fourteen vessels, several of them close to seventy centimetres across - makes him rich at the level of a senator or a provincial governor rather than a merchant. Beyond that the archive is empty, and the emptiness is itself the point: the object names its owner and the historical record cannot corroborate the name.

This is the part of the story that almost every account of the Sevso affair leaves out, and it is the part that matters most to anyone assessing how provenance failures actually occur.

In 1984 the treasure was offered for sale to the J. Paul Getty Museum in Los Angeles. Janos Gyorgy Szilagyi, then head of the Collection of Classical Antiquities at the Museum of Fine Arts in Budapest, happened to be at the Getty while it was there. He looked at the Hunting Plate and saw the word PELSO. He knew immediately what it meant, because a Hungarian classical archaeologist knows the Roman name of Lake Balaton the way a British one knows Verulamium. On returning to Budapest he told the archaeologist Andras Mocsy.

That was the moment Hungarian scholarship first learned that a major late Roman treasure with Pannonian lettering on it was circulating on the market. The Hungarian National Museum's own account of the affair states plainly what followed: no official steps were taken. Six years passed. The state made no enquiry, filed no notice, alerted no police force and told no auction house. When Sotheby's announced the sale in February 1990, Hungary's claim rested on data it had held since 1984 and done nothing with.

The Getty did not buy it either, and its reason is the second half of this record. The two facts belong together: in 1984 an institution rejected the treasure because its papers were forged, and a scholar identified its likely homeland from an inscription, and neither event produced any action for six years.

Between November 1980 and December 1981 a Guernsey company called Art Consultancy, jointly owned by the recently retired Sotheby's chairman Peter Wilson and the dealer Rainer Zietz, bought piece after piece through a chain running from the Vienna-based dealer Anton Tkalec and the Lebanese dealer Halim Korban, with the London dealer Mansur Mokhtarzade involved. The recorded prices are precise and they are worth reading in sequence: a geometric ewer at 48,000 pounds in November 1980; the Hunting Plate itself at 225,000 pounds between November 1980 and January 1981; the Achilles Plate at 525,000 pounds in May 1981; a geometric plate at 180,000 pounds in September 1981; the Hippolytus situla at 525,000 pounds in December 1981.

The silver came first. The paperwork came afterwards. In June 1981, seven months into the buying, Korban produced a Lebanese export permit, and it was applied retrospectively to material already in hand. That is the structural fact this record exists to record: the licence was not a condition of the purchase, it was a product commissioned to cover one.

In March 1984, when the treasure was offered to the J. Paul Getty Museum, the Getty examined the permit and established two things. The authorising signatures were forged, and Korban was not recognised by the Lebanese authorities as a registered dealer at all. The Getty declined.

What happened next is the reason this case is not merely a story about a bad document. In 1985 Peter Mimpriss, the solicitor acting for both Wilson and the Marquess of Northampton, approached a Lebanese businessman named Ramiz Rizk to obtain replacement documentation, at a cost recorded as 628,000 pounds in expenses including taxes. In 1987, when Northampton bought four further pieces from Korban and Tkalec for a combined 8.7 million dollars, export permits were obtained through Rizk for a further 460,000 dollars. Three sets of Lebanese papers, one of them proven forged by a museum before the second was commissioned, and a running total in the low seven figures spent on nothing but provenance. Not on the silver. On the story about the silver.

Spencer Compton, seventh Marquess of Northampton, did not buy the treasure in one transaction and did not own all of it from the start. Mimpriss, who acted for both parties, suggested in November 1981 that Northampton might invest. The Abraham Trust was formed on 15 September 1982 and took ownership of the eight pieces Wilson and Zietz had already acquired, together with two more that Northampton bought for 500,000 pounds.

In 1987, acting on his own, Northampton bought four further pieces from Korban and Tkalec for 8.7 million dollars. The Marquess of Northampton 1987 Settlement Trust became the sole owner of the whole group, and it is that trust, not the man, that was the defendant in New York.

The distinction is not pedantry. It explains why the treasure could later be split: after the failed 2006 attempt to sell, the hoard was divided into two parts, one of which was in the hands of the heirs of Peter Wilson and one of which remained with Northampton. Hungary bought them separately, from different holders, three years apart. Accounts that describe Hungary as buying the treasure back from the Marquess of Northampton describe only the second half.

Spencer Douglas David Compton, born 2 April 1946, became seventh Marquess of Northampton in 1978 on the death of his father. The family holds two seats, Castle Ashby in Northamptonshire and Compton Wynyates in Warwickshire, and his fortune was assessed in 2017 at around 110 million pounds. He was, in other words, exactly the kind of private buyer a dealer with an unprovenanced Roman service would want to find: solvent enough to hold an illiquid object indefinitely, advised by a City firm, and represented by a solicitor who was already acting for the seller.

One transaction in his record deserves setting beside the Sevso chronology, because the two touch. In 1985 he sold Mantegna's Adoration of the Magi at Christie's in London, and the buyer was the J. Paul Getty Museum. The reported figure is 10.5 million dollars, described as a then record auction price for the artist; the source consulted does not say whether that is the hammer or the price with premium, and this record will not guess.

The sequence matters. In March 1984 the Getty examined the Sevso export licences, found the signatures forged and the named dealer unregistered, and declined the treasure. In 1985 the same museum paid eight figures at auction for a picture from the same owner. In the same year, that owner's solicitor was commissioning a second set of Lebanese papers at 628,000 pounds. A museum's diligence operates object by object, and there is nothing improper in buying a well-documented Mantegna from a man whose Roman silver you have just refused. But a collector reading a provenance chain should register how narrow the refusal was: the Getty rejected the paperwork, not the seller, and it told nobody.

On 10 February 1990 Sotheby's announced in New York that it would sell the treasure, and held a private view at its New York headquarters. The consignor was the consortium headed by Northampton. The published estimate was 40 to 70 million dollars in some accounts and 50 to 100 million in others; both figures are estimates and neither is a price. Sotheby's, to its credit, notified twenty-nine countries whose territory had lain inside the fourth-century Roman Empire, together with UNESCO, Interpol, IFAR and ICOM.

The notification produced the opposite of a clean sale. On 15 February 1990, five days after the announcement, Lebanon filed suit in the New York State Supreme Court alleging that the export documentation was forged. A temporary restraining order followed on 20 February. By March the treasure was impounded. It never reached an auction room.

Every figure attached to the Sevso Treasure in general circulation is therefore one of three things and none of them is a hammer price: a 1980s private purchase price paid by dealers or by the trust, a 1990 pre-sale estimate for a sale that did not happen, or a twenty-first century payment by the Hungarian state under a private agreement. This record keeps them apart, because conflating them is how a withdrawn lot acquires an imaginary auction record.

Lebanon claimed first, on the strength of the export licences, and abandoned the claim on the eve of trial once those licences were shown to be forged. The irony is exact: the forged document that was manufactured to give the treasure a Lebanese origin was the only basis of Lebanon's own claim, and its exposure destroyed the claim it was invented to support.

Yugoslavia, and after 1991 Croatia as successor state, claimed on the basis that the silver had been found near Barbariga in Istria, not far from Pula, and smuggled out in the late 1970s, citing soil analysis. Hungary claimed from 1991 that it had been found near Polgardi, some sixteen kilometres east of Lake Balaton.

The trial ran for seven weeks from September 1993. The jury found that neither Hungary nor Croatia had proved that the treasure was found within its borders, and possession stayed with the Northampton trust. Two things about that verdict are consistently misreported. First, the court did not rule that Northampton had good title; it ruled that the claimants had not established theirs, which is a different finding with different consequences. Second, and more seriously, the Hungarian National Museum records that the comparative silver and soil samples, the results of the Hungarian police investigation and the Koszarhegy silver stand were not put before the jury: the court did not permit it, on procedural grounds. The Appellate Division affirmed the dismissal in 1994.

A jury verdict that a claimant has failed to prove a fact is not a finding that the fact is false. It is a finding about the state of the record in that courtroom on that day. Where the strongest part of a claimant's record was excluded before it reached the jury, the verdict tells a provenance researcher considerably less than its citation suggests.

In December 1980 the body of a twenty-four-year-old Hungarian named Jozsef Sumegh was found hanging in a wine cellar near Polgardi. His compulsory military service was ending. He collected coins and antiquities, and he was said locally to have found a large silver treasure before he was called up. The death was recorded as suicide. It was later reinvestigated as murder, and the criminal file has never been closed.

What connects the death to the silver is not rumour but a hole in the ground. During the investigation, traces of a filled-in pit were noticed in the floor of the wine cellar where the body was found. The pit was later excavated by archaeological method. The copper cauldron in which the Sevso silver was buried - the cauldron that came back to Hungary with the first tranche in 2014 - fits the pit. The Hungarian National Museum states that the fit has since been verified by an authentication excavation.

The cauldron is the single most important object in the group for provenance purposes and it is the one nobody photographs. It is not decorated, it carries no inscription, and it would not be worth recording as an object of art. It is worth recording as evidence, because a vessel that matches the void it was lifted out of is a better provenance link than any signature.

In 1878, fragments of a four-legged folding silver stand, a quadripus, were found on Szarhegy, in the administrative area of the village of Koszarhegy, a few kilometres from Lake Balaton and close to Polgardi. It went into the Hungarian National Museum, where it has been ever since, and it was regarded as a unique and slightly inexplicable object: a piece of late Roman display silver of a scale and quality with no local comparison.

During the New York proceedings in 1993 Hungarian researchers made the connection. The stand matches the Sevso vessels in the style and subject matter of its ornament and in its scale, and it was found in the district that the plate's own inscription points to. Read together with the cauldron and the pit, the quadripus is the strongest circumstantial argument that the treasure is Pannonian. Zsolt Mrav published the case at length in 2012.

Excavations were conducted at the stand's find-spot in both the nineteenth and twentieth centuries, and no further fragments of it have ever been found. That absence is part of the evidence too, and this record does not resolve it: the museum's own position, published on its Seuso research site, is that all of this points to a Hungarian find-spot only indirectly, and that the question will be settled if at all by the scientific programme begun in 2014 and by the police investigation, not by anything already known.

Under pressure from Yugoslavia, the Metropolitan Police in London opened a criminal investigation in early 1990. By July the police considered the evidence to suggest a conspiracy to defraud. Northampton was interviewed in October 1990 and was treated as a victim rather than a suspect, cleared of criminal intent. Attention moved to Mimpriss. In 1991 the Foreign Office and the Crown Prosecution Service recommended that the investigation be wound down for want of evidence, and it was, against Northampton's own wishes.

In March 1991 Northampton sued Mimpriss and the law firm Allen and Overy for fraud, deceit and misrepresentation over the acquisitions between 1981 and 1990, on the ground that the treasure had been rendered unsaleable by the forged licences and the publicity. The action was settled out of court in 1999, with Northampton reported to have received 15 million pounds.

Take the money in this case as a whole. The dealers were paid. The producers of the export documents were paid twice, in 1985 and 1987. The buyer recovered a reported 15 million pounds in damages from his own solicitors. The state that says the treasure was dug out of its soil recovered nothing in court and then paid 43 million euros to get the objects. Nobody was convicted of anything, and the man who is thought to have found the silver was buried in 1980.

After the trial the objects were legally in the possession of a trust and commercially worthless, because no museum could buy them and no private collector could ever show them. In October 2006 the hoard was brought back together and shown for a few days to an invited audience at Bonhams in London. It was a viewing, not a sale, and it failed. The group was then divided in two. In 2012 a Hungarian private individual attempted to buy the treasure and failed.

The Hungarian state then negotiated directly, and it negotiated with the Wilson side first. The half that included the Hunting Plate and the copper cauldron was held not by the Marquess of Northampton but by a trust benefiting the two sons of Peter Wilson, the Sotheby's chairman who had started the buying in 1980. That tranche came to Hungary in spring 2014, was unveiled in the Parliament building by Viktor Orban with Jozsef Palinkas of the Hungarian Academy of Sciences, and moved to the Hungarian National Museum in July of that year. The second and, on the museum's own assessment, more richly decorated half was Northampton's, and it followed in summer 2017, announced on 12 July.

Now the money, and this is the part that no single account gets right. The sums are 15 million euros for the first tranche and 28 million euros for the second, 43 million euros in all. They have been published under three different legal descriptions. The Hungarian National Museum calls them a compensation fee, paid in harmony with the international agreements on illegal art trade. Artnet News, reporting The Art Newspaper in April 2014, called the first payment a retention fee and gave it in dollars, at 20.7 million. Most of the press has simply called both of them purchases. The Hungarian government's own framing, in the account of the 2017 tranche, is that two family foundations received compensation for relinquishing the treasure.

These are not synonyms. A purchase price concedes that the seller had title to sell. Compensation concedes nothing about title and characterises the payment as the price of ending a dispute. A retention fee is different again and implies payment for possession rather than for ownership. Hungary has never conceded that the holders owned the silver, having spent a seven-week trial arguing that it did, and the vocabulary it uses is the residue of that position. A reader who sees 43 million euros described as the price Hungary paid for its own national treasure is reading a characterisation that the paying state itself declines.

Viktor Orban, presenting the second tranche in Parliament, used the language of family property rather than acquisition: the government believed the place for the family silver was in Hungary, and it had done its duty and got it back.

Fifteen objects are in Budapest: fourteen silver vessels and the cauldron. Persistent reports, repeated in the specialist literature on the case, hold that the known group is only part of what came out of the ground, with perhaps eight further vessels and as many as 185 spoons unaccounted for. No court and no excavation has confirmed or dismissed this. If the reports are right, a second Sevso Treasure exists somewhere and has never surfaced.

The find-spot has never been established judicially. The New York jury did not find that the silver came from Hungary; it found that Hungary had not proved it, on a record from which Hungary's physical evidence had been excluded. The Croatian claim was dismissed on the same basis and has never been withdrawn as a matter of scholarship. The Hungarian criminal investigation into Jozsef Sumegh's death remains open. The scientific programme the museum began in 2014 - archaeological, art-historical and archaeometric - is the only route by which the question is now likely to be answered.

The treasure is in the Hungarian National Museum, complete as far as anyone knows it, bought back by a state that a court had told, twice, that it had not proved its case. That is the settlement: not a judgment, not a restitution and not a repatriation in any legal sense, but a private purchase which the buyer insists on calling something else.