Vase with Viscaria

A flower piece painted in Paris in the summer of 1886, 65 by 54 centimetres, catalogued F 324a and JH 1137. It was bought in Paris by Mohamed Mahmoud Khalil, an Egyptian politician who was twice prime minister, and it may be the first van Gogh acquired by a collector outside Europe and North America. It went to the Egyptian state with the rest of his collection and the villa in Giza opened as a museum on 23 July 1962.

It has been stolen from that collection twice. The first time it came back, from Kuwait, in circumstances that have never been explained, and with a rumour attached that has never been settled. The second time, on 21 August 2010, a thief cut it out of its frame during opening hours and walked out. Egypt's prosecutor general found that not one alarm in the museum was working and that seven of the forty-three cameras were. The culture minister announced a recovery at Cairo airport the same day and withdrew the claim the next.

Everyone calls it Poppy Flowers. The standard catalogue calls it Vase with Viscaria, because most of the flowers in it are not poppies. That is not pedantry: a picture with three names and one photograph is exactly how a stolen object disappears twice.

What it is, and what it is not called

Oil on canvas, 65 by 54 centimetres. A dark ground, a plain earthenware vase on a table, and a mass of small yellow flowers with two or three red ones low on the left. A few sprigs lie on the table on either side.

The world knows it as Poppy Flowers. The standard catalogue of the paintings knows it as Vase with Viscaria, and lists it under F 324a and JH 1137. In German it is Vase mit Pechnelken. Viscaria, pechnelke and lychnis all name the catchfly, a small campion, and that is what the yellow mass is. The red flowers low on the left are poppies. So a picture in which poppies are a minority element has been renamed after them by press coverage of its own theft.

Migrating titles are a direct cause of version confusion, and this object demonstrates the mechanism at full length. The archive files it under the catalogue title and lists the others, because a reader searching for Poppy Flowers and a reader searching for Vase with Viscaria are looking for the same missing canvas and do not know it.

The date has moved as well. Much of the press, and older reference material, gives 1887. The catalogue and the specialist literature give the summer of 1886, a few months after van Gogh arrived in Paris. This record follows the catalogue.

Why he was painting flowers at all

Van Gogh reached Paris in March 1886 and moved in with his brother Theo; from June they were at 54 rue Lepic in Montmartre. He had come from Antwerp and from the brown palette of the Dutch years, and in Paris he found Impressionism, Japanese prints and colour theory in the same eighteen months. He made more than two hundred paintings there before he left for Arles in February 1888.

The flower pieces were not decoration and they were not commissions. They were exercises, and he said so in a letter to the English painter Horace Livens, written in Paris in 1886: he had lacked the money to pay models, so instead he had made a series of colour studies of flowers, red poppies, blue cornflowers and forget-me-nots, white and pink roses, yellow chrysanthemums, looking for oppositions of blue with orange, red with green, yellow with violet, and for the broken and neutral tones that would harmonise brutal extremes.

That letter is the best available account of what this picture is for. It is a yellow mass set against a dark ground with red at its edge, made by a man teaching himself what colours do next to each other because he could not afford a sitter.

How it got to Cairo

Mohamed Mahmoud Khalil, born in 1877, was an Egyptian politician who served as prime minister twice, from 1928 to 1929 and from 1937 to 1939. He bought a palace on the bank of the Nile at Giza in 1918, and he and his wife Emiline Hector Lock assembled a collection of French nineteenth-century art in it: Monet, Renoir, Gauguin, Rodin, Ingres, and this van Gogh.

He bought the picture in Paris, which he visited regularly and where he dealt with the Impressionist trade, possibly as early as the 1920s. The exact date, the dealer and the price are not recorded in the sources consulted. What can be said, and is worth saying, is that this may be the first van Gogh bought by a collector outside Europe and North America. The market for van Gogh in the 1920s was almost entirely European and American; an Egyptian buyer in Paris in that decade was ahead of a curve that took another sixty years to bend.

Khalil died in 1953 and his wife in 1960. The collection passed to the Egyptian state and the palace opened as the Mohamed Mahmoud Khalil Museum on 23 July 1962, dedicated to them both.

The first theft, and the rumour it left behind

The museum did not stay a museum. From 1971 President Anwar Sadat used the palace for executive offices, and it was not returned to museum use until 1993, reopening to the public in 1995. So for twenty-two years the Khalil collection was not in the Khalil house, and the first theft of this picture falls inside that period.

The date of that theft is not settled in the sources. Martin Bailey, writing in The Art Newspaper in 2021, gives 1978 and says the picture was recovered two years later after surfacing in Kuwait. A contemporaneous account gives 1978 and the same two-year gap, adding that Egypt's interior minister said three Egyptians had been arrested and had told the police where the canvas was hidden. Wikipedia gives 4 June 1977 and a recovery a decade later. This record does not average them. Two independent accounts say 1978 and two years; one tertiary source says 1977 and ten. The disagreement is recorded as a disagreement.

What every account agrees on is that the circumstances of the recovery were never fully explained. And there was a rumour, current at the time, that the painting had been copied during the two years it was gone, and that what the museum received back was one of those copies rather than the original van Gogh.

That rumour has never been tested, and it cannot be tested now. Any examination that could have settled it would have had to happen between the recovery and 21 August 2010, and there is no published record that one did. A picture that a specialist already doubted, that had been out of institutional hands for two years, and that then vanished again before anybody looked at it properly, is about as compromised an evidentiary position as an object can be in.

21 August 2010

On a Saturday, during opening hours, in a museum on a main road in Giza, somebody cut the canvas out of its frame with a box cutter and left with it. Reports of the time describe a couch being dragged under the picture to stand on.

Egypt's prosecutor general, Abdel-Meguid Mahmoud, established the security position and it is worth reading slowly. Not one alarm in the entire museum was functional. Of forty-three surveillance cameras, seven were working. There were not enough guards to make a proper round at closing time. A security official said the cameras had not worked for a long time and neither had the alarm system.

The prosecutor had already warned the institution. In 2009 nine paintings had been stolen from another Cairo museum with comparably poor security and were found dumped ten days later. Nothing was done. After the van Gogh went, Mahmoud barred fifteen Egyptian officials from leaving the country, among them the director of the Khalil museum and the head of the fine arts department at the Ministry of Culture.

Then came the part that turned a theft into a scandal. The minister of culture, Farouk Hosni, announced that suspects had been detained at Cairo airport and the painting recovered. Within a day he retracted it, saying he had been given false information. The retraction is the moment the case went from a security failure to an institutional one, because it established that the ministry was announcing outcomes it had not verified.

The trial

In October 2010 an Egyptian court convicted eleven employees of the Ministry of Culture of negligence and professional misconduct, among them the deputy minister of culture, Mohsen Shaalan. They were sentenced to three years and released on bail of around 1,750 dollars. Shaalan subsequently served a term of a year, ending in 2013.

Very few art thefts anywhere produce criminal convictions of the custodians rather than the thieves. This one did, and that is the reason it belongs in a provenance archive rather than only in a crime file. The finding of the Egyptian court was not that a clever criminal beat a good system; it was that there was no system, and that the people paid to run one were culpable for its absence.

Nobody has been convicted of taking the painting. The Egyptian businessman Naguib Sawiris offered a reward, reported at 175,000 dollars or one million Egyptian pounds, for information leading to its return. It has not been claimed.

Fifty-five million dollars is not a valuation

Every report of this theft carries a figure. It is usually 55 million dollars and sometimes 50 million. Neither is a price, an appraisal or an insurance valuation. They are press estimates that entered circulation in the first forty-eight hours after the theft and have been repeated ever since.

This picture has never been at auction. It was bought privately in Paris by a private collector at an unrecorded price, probably in the 1920s, and it has been state property in Egypt since the 1950s. There is no hammer price, no premium, no estimate and no published probate or insurance figure anywhere in its history. The only honest statement about its market value is that no market has ever tested it.

What can be said instead is what a comparable object does when it is present and can be examined. The Wadsworth Atheneum in Hartford owns a van Gogh flower piece of the same Paris summer, Vase with Poppies, F 279, 56 by 46.5 centimetres, bequeathed in 1957. Its attribution was questioned by Bogomila Welsh-Ovcharov in 1976 and again by Walter Feilchenfeldt around 1990, and the museum put it into storage. In 2019, with new imaging equipment, the conservation laboratory found an earlier painting beneath the surface, apparently a self-portrait, of which the outline of an ear was legible. The picture went to the Van Gogh Museum in Amsterdam, was confirmed as autograph, and went back on display in April 2019.

That is what a doubted van Gogh in a museum gets: thirty years in storage and then an answer. The Cairo picture will not get one.

Why the attribution matters more than usual here

Van Gogh has a forgery problem older than most collectors realise, and it is baked into the catalogue that gives this painting its number. In 1928 the Berlin dealer Otto Wacker was found to be selling van Goghs that his brother Leonhard, a painter and restorer, had almost certainly made. The story attached to them was a Russian owner who had moved them out through Switzerland, which conveniently explained why none of them had papers. Grete Ring and Walter Feilchenfeldt spotted the fakes in January 1928 when the last of the pictures were delivered for an exhibition.

Jacob-Baart de la Faille published his catalogue raisonne of van Gogh the same year. It is the source of the F numbers still used, including the F 324a on this record. In November 1928 he had to publish a supplement identifying the suspect pictures as forgeries. The National Gallery in Berlin found a resin in them used to speed drying that appeared in no genuine van Gogh, and the canvases were not French. Wacker was tried in 1932, convicted, and given nineteen months and a fine of thirty thousand Reichsmarks.

The point for this record is not the scandal. It is that van Gogh attribution has always been done with laboratories as well as eyes, that de la Faille himself had to correct his catalogue within a year of publishing it, and that the corpus has been argued over continuously ever since. Louis van Tilborgh accepts this picture. Ronald Pickvance did not. In any other case that would be a file waiting for an infrared reflectogram. Here the file is shut, because there is nothing to put under the lamp.

No photograph of this painting is any good

This is the plainest way to say it: there is no high-quality reproduction of this object in existence that this record could find under any licence.

What survives is a handful of small plates from printed books and press files, most of them well under a thousand pixels on the long side, all made before August 2010 and most of them a good deal earlier. The colour differs between them. In 2013 a touring exhibition of reproductions in Budapest showed a digital reconstruction of the picture's original colours; a photograph of that panel is on this page, and it is the largest image available, and it is a photograph of a modern reconstruction rather than of the painting.

The consequence is that nobody can now compare a candidate canvas against a good reference image. If the picture surfaced tomorrow, the identification would rest on small, colour-unreliable plates and on whatever the Khalil museum holds in its own files. That is a bad position for an object with a live authenticity question and a twenty-year-old rumour about a substituted copy attached to it.

Compare the Gardner Vermeer, missing since 1990, which is reproduced in high resolution from pre-theft photography and is instantly recognisable. Illustration quality is not a cosmetic matter for a stolen object. It is the recovery evidence.

Where it stands in 2026

The museum was closed after the theft and reopened in 2021. The van Gogh was not in it. There has been no trace of the picture since it went out of the door in August 2010, and the specialist view is that the chances of recovery after this long are slim.

The convictions stand against the officials. Nobody has been convicted of the theft. The Sawiris reward has not been claimed. The picture is the only van Gogh the museum had, and it may have been the first van Gogh bought by anyone outside Europe and North America.

Two thefts, one recovery, one rumour of substitution, one unresolved attribution and no usable photograph. The second theft of this object was not a break-in. It was the consequence of a decision, taken over years, not to fix the alarms.