The Taking of Christ

A Caravaggio of 1602, paid for at 125 scudi by the Roman nobleman Ciriaco Mattei and entered in his bank book on 2 January 1603. It was sold out of the family in 1802 under the name of Gerard van Honthorst, went to Scotland, came to auction at Dowell's in Edinburgh in 1921 at eight guineas and again in 1922 at five, and was given in about 1930 to the Jesuit house at 35 Leeson Street in Dublin by a doctor whose husband had been shot dead by the IRA. It hung in the dining room for sixty years under a coat of dirt and brown varnish.

In August 1990 Sergio Benedetti, senior conservator at the National Gallery of Ireland, walked round to Leeson Street to look at the community's pictures and recognised it. Three years of examination and consultation followed. The painting was unveiled on 16 November 1993 and Benedetti published the case in The Burlington Magazine the same month, alongside a companion article with Francesca Cappelletti setting out the documentary evidence she and Laura Testa had found in the Mattei archive at Recanati.

At least twelve early versions of this composition exist. One of them, in Odessa, was long argued to be the original, was itself stolen in July 2008, was recovered from a car in Berlin in 2010 with the canvas shredded, and was restored between 2018 and 2021. Every plate in this record states which version it shows.

What the picture does

Oil on canvas, 135.5 by 169.5 centimetres in the National Gallery of Ireland's catalogue. Seven figures are pressed into a shallow, almost airless band across the front of the canvas. Judas has taken Christ by the arm and is completing the kiss. Christ's hands are locked together in front of him, fingers interlaced, in the only gesture of stillness in the picture. Two soldiers in black articulated armour drive in from the right, one with his gauntleted arm thrust across the whole composition. On the left a man in red flees with his mouth open and his arms up, the cloak streaming behind him. At the back right, holding up a lantern and straining to see over the soldiers, is a man whose face has been read since 1990 as Caravaggio's own.

There is no landscape, no sky and no middle distance. The garden of Gethsemane has been reduced to a black ground, and everything the viewer is given arrives on the light. That is the method: a dark ground, models posed in front of it, the picture built directly in paint without a cartoon and without preparatory drawings, guidelines scored into the wet ground with the handle of the brush. Caravaggio worked that way throughout, which is one of the reasons his pictures are so hard to copy convincingly and so easy to copy adequately.

The lantern is the argument of the picture. It is being held up by a man who wants to see, and it lights almost nothing. The betrayal happens under a light that fails.

The commission, and what 125 scudi bought

Ciriaco Mattei was one of three brothers of a Roman family with a new palace, the Palazzo Mattei di Giove, and a serious appetite for the most disturbing painter in the city. He is documented as buying three pictures from Caravaggio in a short period: The Supper at Emmaus, now in the National Gallery in London, Saint John the Baptist, now in the Pinacoteca Capitolina in Rome, and this one.

The payment for this picture appears in Ciriaco's libro di rincontro del banco, his bank reconciliation book, and it is the first documentary mention of the work in existence. It records 125 scudi paid to the artist for a painting, with its frame, of Christ taken in the garden. The entry is dated 2 January 1603, which means the picture was finished and delivered by the end of 1602.

The figure sits at the top of Caravaggio's ordinary rate for a cabinet-scale picture and it includes the frame, which is worth noticing: the frame was part of what the patron bought and part of what he paid for, and it has not survived. Set against it is a second, later payment in the same family's papers that tells you what a copy was worth. In 1626 Asdrubale Mattei paid Giovanni di Attili twelve scudi to copy this picture. The original cost slightly more than ten times the copy. That ratio is the seventeenth-century market's own estimate of the difference between a Caravaggio and a competent version of a Caravaggio, and it is a more useful number than any modern valuation of this painting, because no modern valuation exists.

How it was lost: a guidebook, a wrong name and a sale in 1802

The picture was never a secret. Gaspare Celio, a painter, wrote a guide to the pictures of modern Rome before 1620; it was published in Naples in 1638 and it records the Caravaggios in the Mattei collection at first hand. Benedetti's 1993 article opens with Celio, because Celio is the reason art historians knew a Caravaggio Taking of Christ had existed and had to be somewhere.

What removed it from view was a name. In 1786 Giuseppe Vasi catalogued the painting as a Honthorst, and a Mattei family inventory of 1793 repeated the error. Gerrit van Honthorst was a Utrecht painter who had been in Rome, absorbed Caravaggio's lighting and built a career on candlelit night scenes. He was exactly the wrong answer that looks right: a picture with a lantern in it, painted in Rome, in a dark manner, by a follower.

On 27 January 1802 Duca Giuseppe Mattei sold the picture, with others, to William Hamilton Nisbet, a Scottish collector then in Rome. It went to Biel House in East Lothian and hung there for more than a century under the Dutchman's name. When the last of that line, Mrs Mary Georgiana Constance Nisbet Hamilton Ogilvy, died in 1920, the contents came to be sold.

The point to hold on to is that nothing was hidden at any stage. The picture was in a great house, in a documented collection, with a catalogue entry. It was invisible because it was labelled.

Eight guineas, then five

At Dowell's in Edinburgh on 25 June 1921 the painting came up as lot 30 and was bought by John Kent Richardson for eight guineas. Eleven months later, at Dowell's again on 3 May 1922, it was lot 480, and this time it made five guineas, to the Hon. Major Charles Noel.

Those are hammer prices at a provincial saleroom for a picture then believed to be a Honthorst. The second is lower than the first, which is what happens when a dealer buys wrong and takes the loss. Eight guineas in 1921 is a modest sum for a large seventeenth-century canvas even by the standards of a depressed post-war market for old masters; five guineas is close to the value of the stretcher and the frame.

Every discussion of this picture reaches for a modern figure and none exists, because the work has not been on the market since 1922 and is owned by a religious community that has placed it on indefinite public loan. The only prices ever recorded for it are 125 scudi in 1603, eight guineas in 1921 and five guineas in 1922. That is the entire market history of one of the most reproduced paintings of the seventeenth century.

The woman who bought it

Percival Lea-Wilson was a District Inspector of the Royal Irish Constabulary. After the surrender that ended the Easter Rising of 1916 he had charge of some two hundred and fifty republican prisoners held overnight in the grounds of the Rotunda in Dublin, and he abused and publicly humiliated them; the best-known incident is his stripping of Tom Clarke, the first signatory of the Proclamation. On 15 June 1920, aged thirty-three, he was shot dead in Gorey, County Wexford, walking home with a newspaper. A car with five men in it left the scene. The IRA claimed the killing as retribution for 1916.

His widow, Marie Lea-Wilson, commissioned a memorial window from the stained-glass artist Harry Clarke for the church at Gorey, on the subject of the martyrdom of Saint Stephen, and had it inscribed with Stephen's words: Lord, lay not this sin to their charge. She then read medicine, qualified at forty-one, and became a senior paediatrician at St Gabriel's Hospital in Dublin.

In the early 1920s, in Edinburgh, she bought a seventeenth-century picture catalogued as Gerard van Honthorst's Taking of Christ in the Garden and brought it to Ireland. The earliest Irish document connected with it is a receipt for twenty pounds paid to the Dublin cabinetmaker James Hicks for restoring its frame. In the early 1930s, in gratitude to Father Thomas Finlay of the Jesuit residence at 35 Leeson Street, who had counselled her after the killing, she gave the picture to the Jesuits.

She died in 1971. She never knew what she had given away.

Sixty years in a dining room

The Jesuits hung it in the dining room of the house of study on Leeson Street, where it stayed for about sixty years. It was not stored, not neglected in an attic and not forgotten: it was a large religious picture over a refectory table in a house full of educated men, and everyone who ate under it knew it as a Honthorst.

What sixty years of a working dining room did to it is on the record, in the conservator's own words. Benedetti described a surface obscured by a film of dust, grease and soot, and a varnish gone yellow-brown, so that the flesh of the faces and hands read as tobacco and the red of Christ's robe read as the colour of dead leaves. The soot and grease are the dining room. The picture was being smoked, gently, for six decades.

That is also what saved it. A dirty picture of a subject nobody was looking for, under the name of a minor Dutchman, in a religious house that was never going to sell it, is about the safest place a Caravaggio could be.

August 1990: what Benedetti actually did

The identification has been told as a moment of recognition and it partly was. In August 1990 Sergio Benedetti, senior conservator at the National Gallery of Ireland and a specialist in early Italian painting, walked round to Leeson Street to look at the community's pictures at their invitation. He saw a composition he knew from copies and photographs, in a canvas whose quality he could read through the varnish, and he thought it was the lost Mattei original.

Then came three years of work, which is the part that matters and the part the story usually drops. The gallery undertook the research, the technical analysis and the consultation with international specialists. The picture was cleaned. Benedetti wrote it up and published it in The Burlington Magazine in November 1993, volume 135, number 1088, under the title Caravaggio's 'Taking of Christ', a Masterpiece Rediscovered.

In the same issue, immediately after it, ran a second article: The Documentary Evidence of the Early History of Caravaggio's 'Taking of Christ', by Francesca Cappelletti and Benedetti. Cappelletti and Laura Testa, then graduate students at the University of Rome, had gone through the Mattei family archive in a cellar in Recanati in the Marche and found the payment of 125 scudi in Ciriaco's bank book. Their work was independent of Benedetti's and it landed at the same time.

This is why the identification held. A connoisseur's recognition and a laboratory report will always be arguable. A recognition, a laboratory report and a bank entry naming the artist, the subject and the sum, published together in a peer-reviewed journal within three years, are something else. The picture was shown to the public on 16 November 1993 in an exhibition called Caravaggio: The Master Revealed, and put on indefinite loan to the National Gallery of Ireland by the Jesuits.

The other versions, and why they matter

At least twelve early versions of this composition are known. Some are workshop or contemporary copies, one of which is documented: in 1626 Asdrubale Mattei commissioned Giovanni di Attili to copy the picture for twelve scudi, and that copy is the reason there is a Mattei-provenance version at all. Others are later. The existence of the group is why the original could go missing in plain sight: for most of the twentieth century, anyone wanting to see the Taking of Christ could see a version of it.

Two of the versions have their own histories and both belong here.

The first is in the Odesa Museum of Western and Eastern Art in Ukraine. It surfaced at a Paris auction in 1868 among the property of the Russian collector Alexander Petrovich Basilevsky, catalogued under Caravaggio's given name. Basilevsky gave it to the Grand Duke Vladimir Alexandrovich Romanov in 1870; it was shown at the Academy of Fine Arts in St Petersburg and transferred to Odessa in 1901. It was hidden in a cathedral during the Second World War, returned to Soviet hands, and X-rayed at the Grabar Institute in Moscow in 1951. Those X-rays are still the centre of the argument about it.

The second is the fragment from the Ruffo collection, a cut-down early version whose central group survives. Its status is unsettled.

The Odessa picture was stolen too

In July 2008 the Odessa version was taken from the museum. The thieves cut it roughly from its frame and shredded the canvas in the process. The museum had no alarm connection and one guard on duty.

The recovery ran through an undercover operation led by Major Denis Formichev of the Ukrainian organised crime police, working with German and Italian counterparts, with Formichev himself posing as a dealer. The painting was found in Berlin in 2010. Formichev's account is of a car boot being opened and a torch shone on to the damage. The recovery was announced jointly by Angela Merkel and Viktor Yanukovych.

Then it stalled. The canvas was taken to the National Art Restoration Centre in Kyiv and immediately impounded as evidence in the criminal case. Five men were indicted and none was convicted; Formichev has said publicly that politics and corruption are the reason. The museum's director spent years warning that the picture had gone more than five years without conservation and that the consequences would be irreversible. Restoration finally began in March 2018, after intervention by the Italian embassy, and ran to 2021.

The restored picture was shown in Kyiv in 2026, in the exhibition In the Light of Caravaggio at the National Scientific Restoration Centre and at the Saint Sophia of Kyiv National Sanctuary, its first public appearance in eighteen years. A ruling of the Podil District Court of Kyiv of 28 August 2025 ordering its return to the Odesa museum came into force after the prosecutor's appeal was dismissed. As of August 2026 the picture is in Kyiv and the war is the reason it has not gone home.

Two paintings of the same composition, both stolen. One was recovered in a Berlin car park after two years and has been on an operating table ever since. The other, in Dublin, has never been stolen at all, and its worst injury was sixty years of dinner.

Is the Dublin picture the original? The case against

This record does not treat the attribution as closed, for the reason that the archive exists: correcting version confusion is the whole job, and a record that reports only the winning side of a live dispute is doing the opposite.

The Ukrainian case for Odessa is made most fully by the art historian Natalia Chechykova and rests on the 1951 Grabar X-rays. She argues they show changes and second thoughts between a first lay-in and the final composition, notably in the position of the hand of the lantern-bearing figure and in the expression given to Christ, and that a copyist does not think again about a composition he is copying. Benedetti answered this directly. He pointed to the documented 1626 copy commission and wrote that the Odessa picture, although faithful in design, reveals its status as a copy by the evident weaknesses of its execution.

A separate line of attack comes from the Italian side. Francesco Petrucci, curator of the Ariccia version, has argued from technical examination that the Ariccia canvas shows pentimenti and period pigments consistent with Caravaggio, and that one passage in the Dublin picture reads as a copyist's error rather than an artist's variation: a pronounced scar on the left hand with which Judas grips Christ, clearly present in the Ariccia and Pitti versions, is distorted in Dublin. On that reading the Dublin canvas would revert to something like the Honthorst label it wore for two centuries.

The National Gallery of Ireland's position, put by its curator Aoife Brady, is that pentimenti alone do not settle anything, because copyists also change their minds and correct mistakes, and that the specific kinds of revision associated with Caravaggio's practice are not what the Odessa evidence shows. She has also made the point that the argument does not diminish what Odessa has.

Weighing it honestly: the Dublin attribution is the only one supported by a payment record that names the artist, the subject and the sum, published alongside the technical case in a refereed journal, and it is accepted by the overwhelming majority of Caravaggio scholarship. The dissent is real, it is published, and it is named here. It has not moved the field.

The man with the lantern

The figure at the back right, holding the lantern up and craning to see, is generally identified as a self-portrait. It is the only face in the picture that is not doing anything to Christ, and it is the only one that is looking rather than acting.

Read against what is documented about the painter, that placement is not sentimental. Caravaggio's own life is one of the best-recorded of any artist of the period, because so much of the record is police record. On 29 May 1606 he killed Ranuccio Tomassoni in a fight in Rome and was sentenced to beheading; he spent the last four years of his life in Naples, Malta and Sicily working for a pardon that never came, and he died at Porto Ercole on 18 July 1610, aged thirty-eight. This picture was painted four years before the killing, in the middle of the best-paid and most productive stretch of his Roman career, for a patron who kept buying.

The subject was not chosen by the painter. Ciriaco Mattei ordered a Christ taken in the garden and paid for it with its frame. What the painter did with it was to remove the setting, remove the divine light, and put himself in it as a man holding up a lamp that does not reach.

Where it stands in 2026

The picture is in Dublin, in the National Gallery of Ireland, as loan number L.14702. It is not the gallery's property. It belongs to the Jesuit Community at Leeson Street, who placed it on indefinite loan to the Irish public in 1993 and whose credit line acknowledges Marie Lea-Wilson by name. It has been lent out repeatedly since the rediscovery, to London, Boston, Washington, Milan, Amsterdam and Rome among others.

The Odessa version is in Kyiv, restored, exhibited in 2026 for the first time since it was cut out of its frame, and under a court order to go back to a museum in a city under bombardment. The Ruffo fragment is in Italy. The Ariccia version is the basis of a published dissent from the Dublin attribution. The 1626 copy commissioned by Asdrubale Mattei is documented in the family papers.

Marie Lea-Wilson's twenty pounds to a Dublin framemaker is the last money anyone is recorded as having spent on this object. The frame Ciriaco Mattei paid for in 1603 is gone.