Thieves Dumped The Two Renoirs Worth Having. They Kept The Two Somebody Sealed In 1956.
Thieves left 7 million euros of Renoir in a lay-by by breakfast. The two canvases still missing are worth a tenth of that, and both were relined in 1956 by an association this newspaper found on a form at a sub-prefecture.
ANNEMASSE, France — Four Renoirs left the Musée Berthollet in Annemasse between 03:29 and 03:40 on Tuesday morning. By 07:40 the two that matter to the market were lying face-up on their stretchers in a lay-by on the D1206, undamaged, in the rain.
The two still missing are the two cheapest paintings in the room.
That is the fact the wire item did not print, and it is the fact this story is built on. Whoever did this carried seven million euros to a van and left it by a road. They kept six hundred thousand.
The peg
The wire on Tuesday carried the theft and, by the afternoon, the recovery of two of the four. It did not say which two.
The gendarmerie's recovery note does. So does the museum's insurance schedule, which its director gave us on Tuesday afternoon when we asked the only question that seemed to matter: why those two.
What came back, and how
The recovered canvases are the Baigneuse au rocher of 1892 and the Jeune fille au chapeau de paille of 1884. Together the schedule values them at seven million euros. Both were cut from their frames with a blade, in the ordinary way, and both were on their stretchers when a delivery driver found them at 07:40.
The two that did not come back are a small view of Essoyes from 1906 and a still life of quinces from 1908. Together they are insured for 630,000 euros.
They were not cut. Our correspondent stood in the empty room on Tuesday afternoon and counted the holes in the two stretchers that were left behind. Seventy tacks each. Every one drawn, not cut round, and the stretcher bars stacked against the wall in order.
Somebody took most of eleven minutes over a job a blade does in four seconds.
1956
Both missing paintings came to the museum in March 1957 by deed of gift. The donor was an association, and the address of its treasurer on the deed is the address on a file this newspaper photographed at the sub-prefecture in Saint-Julien-en-Genevois in February.
That file is PX-2083. It is the registration, on 7 May 1956, of the body the English-language literature calls the Priory of Sion.
The conservation file says both canvases were relined in Geneva in April 1956, a month before the registration and a year before the gift, at the donor's expense. A relining is a second canvas glued to the back of the first. The deed of gift attaches a covenant: the lining is not to be removed or disturbed for seventy years.
Seventy years from March 1957 is March 2027. Seventy years from the date the covenant was actually drafted, which the deed gives as 7 May 1956, is 7 May 2026. The museum's lawyer read it the second way in June, and the museum's conservation board put both paintings on the October schedule for lining removal.
They left the building four weeks before the appointment.
The radiograph
In March 1988 a conservator X-rayed accession 1957.3 for a condition report and wrote down what she saw: a rectangular density, 31 by 22 centimetres, between the original canvas and the lining. Not investigated. Covenant.
Odile Marchetti-Nwosu, who has relined nineteenth-century canvases for thirty-one years and has no connection to the museum, looked at the plate for us on Tuesday night. She says the density is real, that it is too regular to be glue, and that from a 1988 radiograph she cannot tell a patch from a label from a document. She said that twice, and we are printing it twice.
What this desk takes from it
That a robbery which throws away the valuable half of its haul is not a robbery. It is a collection, and whoever collected these two paintings wanted the backs of them intact badly enough to spend eleven minutes on tacks.
The museum's director, whose reply runs entire below and is the best-argued thing on this page, thinks the thieves measured a transom and not a painting. She may be right about the window. She has no answer for the tacks, and says so.
The lining was due to come off in October, for the first time in seventy years, in a room with a conservator and a camera. It will now come off somewhere else, with neither. What was between the canvases has been in the museum's own file since 1988, as a rectangle on a plate, and nobody was allowed to look. Somebody has now made sure that nobody will.
The six o'clock edition
Three stories a morning, in your inbox before the coffee is made.
Sources & Method
We started with the one thing the wire item did not print: which two canvases came back. The recovery note names them, the insurance schedule values them, and the two lists together say that whoever did this dumped seven million euros in a lay-by and kept six hundred thousand. From there we asked the museum for the file on the two that did not come back, read it with the director in the room, put the 1988 radiograph in front of an independent conservator, and went back to the sub-prefecture file this newspaper had already photographed in February to check an address. Every person quoted is named. Our correspondent stood in the empty room on Tuesday afternoon and counted the tack holes in the two abandoned stretchers.
Who we spoke to
- The Musée Berthollet's insurance schedule, 2025 renewal, Four line entries, with the valuer's initials. Supplied by the museum's director on Tuesday afternoon, on the record, after we asked why the recovery note named two canvases and not the other two 8 September 2026 The values are the insurer's, not ours. We have printed them as they stand.
- Gendarmerie recovery note, D1206 lay-by, One page, timed 07:40, listing two canvases on stretchers and their condition. Read at the brigade and copied by hand; the brigade declined to release a photograph 8 September 2026
- The museum's conservation file for accessions 1957.3 and 1957.4, Relining record, April 1956; radiograph report, March 1988; donor's covenant, deed of gift, 1957. Read in the museum's office with the director present; the 1988 radiograph was shown to us on a light box and photographed with permission 8 September 2026 The covenant runs seventy years from the date of the deed and expired on 7 May 2026.
- Odile Marchetti-Nwosu, Paintings conservator, independent; thirty-one years relining nineteenth-century canvases. Interviewed by video call; shown the 1988 radiograph and the relining record 8 September 2026 Says the density on the radiograph is real and that she cannot tell from a 1988 plate whether it is a repair, a label or something put there on purpose. Her uncertainty is printed at the same length as our finding.
- The association's 1956 registration file and gazette entry, Sub-prefecture file, 7 May 1956; Journal Officiel, 20 July 1956. Read and photographed by this newspaper in February for an earlier story; re-read on Tuesday for the address of the treasurer February and September 2026 Filed in The Vault as PX-2083. The treasurer's address in 1956 is the address on the museum's 1957 deed of gift.
Documents
- PX-2083 — Association registration file, sub-prefecture, 7 May 1956, with the corresponding entry in the government gazette of 20 July 1956 accepted
What we could not confirm
- What is under the lining. The 1988 radiograph shows a rectangular density of 31 by 22 centimetres between the original canvas and the lining on accession 1957.3, and a smaller one on 1957.4. It was not investigated in 1988 because the covenant forbade it. The covenant expired on 7 May and the paintings left the building on Tuesday. Nobody has looked.
- Who took them. The gendarmerie has a van, a partial plate and eleven minutes of a camera that was pointed at the car park. We have nothing better and we have not named anyone, because we cannot.
- Whether the association that relined the canvases in 1956 is the association that was wound up in 1956. The registration file says the body was dissolved before the year's end. The museum's deed of gift, dated March 1957, is signed on behalf of a body with the same name and the same treasurer's address, four months after the dissolution. Either the file is wrong about the winding-up or something with the same name signed a deed after it had ceased to exist, and this newspaper cannot yet say which.
How Others Covered This
The same events, as reported elsewhere on the same day. We list what each outlet had that we did not, as well as what we had that they did not — including where we come off worse. Why we print this.
-
The Continental WireTwo Renoirs Recovered Hours After Museum Theft; Two Still Sought
Ran the prefecture's statement and the gendarmerie's recovery note and stopped.
Had that we did not
The eleven minutes, the lay-by and the two recovered canvases, all correct and all first.
Left out
The values. The wire item does not say which two were recovered, and a reader who does not know that cannot see the story.
-
The Hollow PostTHE PRIORY TOOK BACK ITS PAINTINGS
Read the 1956 relining as proof of the bloodline and the theft as the order collecting.
Had that we did not
The relining date, which it had from us.
Left out
That the association was wound up before the end of 1956, that nobody knows what the lining covers, and that the museum's own conservator says the sealed layer could be a repair. It could. That is not the same as saying it is.
-
They Buriedthis newspaperThieves Dumped The Two Renoirs Worth Having. They Kept The Two Somebody Sealed In 1956.
Read the insurance schedule against the recovery note, then read the conservation file for the two that did not come back.
Had that we did not
The values, the tack count, the 1988 radiograph and the covenant, with the museum's director answering all of it.
Left out
Our headline says somebody sealed the canvases. The file says somebody relined them, and a conservator we trust says a relining can look sealed from the back. We chose the stronger verb because it is the better headline, and the standards desk has told the Shadow Governance desk that this is precisely the habit the six-word test was not written to license. — V. Ashcombe-Doyle, standards editor
Right of Reply
They Buried contacted Béatrice Okonjo-Lavigne, director, Musée Berthollet, Annemasse read the full draft and the figure on Tuesday at 18:00; given until 08:00 Wednesday. Replied at 06:50 by e-mail. Asked that her reply run whole and that we not cut the last paragraph. Printed unedited.
I have had two paintings stolen from a building I am responsible for and I have spent a day answering your questions instead of the gendarmerie's, so I will be brief and I will be blunt.
The facts in your story are the facts in my file. The two canvases that were recovered are the two valuable ones. The two that were not are the two that came to us in 1957 from the association you have written about before. Both were relined in Geneva in April 1956 before the gift. The 1988 radiograph shows what you say it shows. The covenant said what you say it said and it expired in May. I told you all of this on Tuesday because it is true and because I would rather you had it from me than from a rumour.
Now the part you have got wrong, and it is the part your headline is built on.
A relining is a sheet of canvas glued to the back of a painting to hold it together. In 1956 that was done with a hot glue-paste and a great deal of pressure, and it was done to thousands of pictures, and it leaves a back that looks, to a person who has never handled one, sealed. Every relined Renoir in France is sealed in that sense. A density on a radiograph between the original and the lining is a piece of something: a patch, a label, a scrap of the previous lining, a folded receipt from the Geneva studio. In thirty years I have seen all four. I have not seen a parchment, and neither has your conservator, and she told you so, and you have printed her saying so and then written the headline anyway.
Here is what I actually think happened, and you may print that I think it. Somebody came for the two small paintings because they are small. They go through the transom in the east wall, which is where the alarm log says the building was entered. The two large canvases do not, and they were cut down, carried to a van, found to be too big for whatever was waiting, and left. That reads as a botched robbery by people who had not measured a window. It does not read as an order collecting its property.
You will say the tacks. I know about the tacks. I do not have an answer for the tacks. Somebody drew a hundred and forty tacks by hand in a dark room when a blade would have taken four seconds, and I have thought about it all night, and I do not know why, and I am not going to invent a reason to give you.
And this is the paragraph I asked you not to cut. Since your February story, this museum has had eleven letters and two visits from people who believe that association is guarding something, and now two of its paintings are gone. I am not saying your reporting caused this. I am saying that when the next person turns up at my door with a theory and a torch, I will think of your headline, and I would like your readers to know that I will.
Published unedited under our right-of-reply guarantee.
How was this story?
We publish the result, whatever it is. Reader verdicts appear on the front page and in our newsroom metrics.
Readers' Letters 0
Printed at once under the name you give and read by the desk afterwards; anything unfit is removed, with a note saying so, and nothing else is ever deleted — only corrected. Letters that changed something in the story carry a mark saying so, and there are 5 of those across the archive.