Girl with a Pearl Earring

The best-known head in European painting, and one of the shortest documented ownership histories attached to a famous picture. The Mauritshuis prints a chain reaching back to Vermeer's lifetime in which every link before 1881 carries a question mark, and this record keeps those question marks rather than smoothing them out. The painting enters the documentary record in 1881, at the Braams sale in The Hague, where it was bought for two guilders with a thirty cent mark-up by Arnoldus Andries des Tombe, after the cultural official Victor de Stuers agreed not to bid against him. Des Tombe left it to the Mauritshuis in a secret will; it entered the collection in 1903 as inv. 670.

It is a tronie, not a portrait: a study of a head in exotic costume, made for the open market rather than for a sitter. No identification of the girl has ever been established and the category means there may never have been one to establish. The long-standing association with a tronie in the Dissius sale of 16 May 1696 remains an argument, not a fact, and cannot be narrowed even to which of the three candidate lots is meant; the Metropolitan Museum's Study of a Young Woman is a competing claimant for the same entries.

What is known in detail is the paint. The 2018 project The Girl in the Spotlight examined the picture non-invasively for two weeks in front of the museum public and published its findings across nine open-access papers. The background is a faded green glaze of indigo and weld over charcoal black, and the imaging found diagonal modelling in it that the researchers read as folded fabric, very probably a curtain. The eyelashes everyone believed absent are there. The ultramarine came from lapis that had been heated before grinding; the lead white came from Derbyshire. The pearl is two strokes of paint over a finished neck, with no contour and no hook, and it is too large to have been a real one.

It is not a portrait, and the question everyone asks about it is the wrong question

Oil on canvas, 44.5 by 39 centimetres, signed in the upper left IVMeer with the first three letters in ligature. A young woman turns over her shoulder against a dark ground, wearing a turban wound from two pieces of cloth, one blue and one yellow, and an oversized pearl drop. The Mauritshuis dates it 1665 and holds it as inv. 670.

The picture is a tronie, not a portrait. That distinction is the whole of the difference between what this record can state and what it cannot. A tronie is a bust-length study of a head representing a type or a character, made for the open market rather than for a sitter who commissioned it. Rembrandt popularised the form in Dutch painting around 1630 and made dozens, frequently using himself. Seventeenth-century Dutch girls did not wear turbans; the headdress is a costume property that signals an exotic or biblical figure, not a wardrobe.

There would probably have been a model in the room. The point of the exercise was not to record who she was. The conservation literature makes the same point from the paint: the face is idealised and deliberately indistinct, the bridge of the nose has no contour and its tip dissolves into the cheek, and the transitions were blended with a dry brush precisely so that no sharp line would fix a specific person. Asking who she is asks the picture to be something it was not built to be.

Two guilders, plus thirty cents

The painting entered public knowledge in 1881, at the Braams sale at the Venduhuis der Notarissen in the Nobelstraat, The Hague. The exact day and month are not recorded. On the viewing day it caught the eye of Victor de Stuers, the most influential cultural official in the Netherlands, who was there with his friend and neighbour, the collector Arnoldus Andries des Tombe.

Two versions of what happened next survive and the Mauritshuis prints both rather than choosing. In one, De Stuers recognised a Vermeer under the neglect. In the other, the picture was too dirty to be judged at all and the authorship only emerged later, when cleaning uncovered the signature. What is not in dispute is the arrangement between the two men: they agreed not to bid against each other, and Des Tombe bought it alone.

The hammer price was two guilders. The buyer's mark-up was thirty cents. The total was two guilders and thirty cents, and the Mauritshuis prints the figure in its provenance as f2,30. Almost every account of this sale gives a single number and does not say which of the two it is; the two components are both documented and both belong in the record. The painting's condition at the moment of sale was described as a sorry state of neglect.

Des Tombe hung it at Parkstraat 26 in The Hague, where his collection was open to visitors, and lent it to the Mauritshuis in the year he bought it. The future director of the museum, Abraham Bredius, saw it there in 1885 and wrote that Vermeer overshadowed all the rest. When Des Tombe died on 16 December 1902 he was found to have made a secret will leaving twelve paintings to the Mauritshuis. This was one of them, and it entered the collection in 1903.

The two centuries nobody can account for

Between the studio in the 1660s and the saleroom in 1881 there is no owner this record will assert.

The Mauritshuis publishes a chain reaching back to Vermeer's lifetime, and it is worth reading exactly how the museum writes it. Every link before 1881 carries a question mark: Pieter Claesz van Ruijven of Delft before 1674, then his widow Maria de Knuijt to 1681, then their daughter Magdalena van Ruijven with Jacob Dissius, then Dissius alone with his father Abraham to 1695, then the Dissius sale in Amsterdam on 16 May 1696. The museum is not being coy. Those question marks are the finding.

The reconstruction rests on a general fact rather than a specific document. Van Ruijven and De Knuijt owned roughly twenty-one pictures by Vermeer, more than half his surviving output, and for about thirteen years from 1657 the artist worked almost exclusively for them and their circle. Maria de Knuijt's will of 1665 left Vermeer five hundred guilders should she die first, a bequest from a patron to a painter with no known parallel in seventeenth-century Dutch practice. If a Vermeer of the mid-1660s had an owner at all, it was probably one of these two. That is a probability about a body of work, not a document about this canvas.

What the 1696 sale can prove, and what it cannot

On 16 May 1696, twenty-one paintings by Vermeer were sold at the Oude Heeren Logement in Amsterdam from the Dissius estate, for a total of 1,503 guilders and 10 stuivers. It is the single most important document in the history of Vermeer provenance; roughly fifteen surviving pictures are identified through its catalogue descriptions, and most Vermeer ownership histories begin there.

Three of its lots are heads. Lot 38 is described as a tronie in antique dress, uncommonly artful, and fetched 36 guilders. Lot 39 is another by the same hand, 17 guilders. Lot 40 is a companion to it, 17 guilders. No dimensions, no sitter, no distinguishing detail. The Mauritshuis records this painting as lot 38, 39 or 40, without choosing, and Vermeer scholarship has never been able to choose either.

This is why the identification is graded disputed here and why it must stay that way. Two other tronies compete for the same three entries, chiefly the Study of a Young Woman in the Metropolitan Museum, which is close to the same size, was probably painted within a year or two, and is signed IVMeer in the same corner of the canvas. Any argument that puts this painting in lot 38 has to displace that one, and the catalogue wording supports neither.

The prices are worth holding on to for a different reason. A tronie sat low in the seventeenth-century hierarchy of subjects, not far above landscape. In the same sale, comparable canvases by Vermeer went for very much more: the sums attached to lots 38, 39 and 40 are the prices of a category, not of a masterpiece. A Rembrandt tronie in the same room went for seven guilders.

One published figure conflicts. The Essential Vermeer catalogue entry gives 46 guilders for the tronie in antique dress; the Mauritshuis provenance and the same site's own account of the Dissius auction both give 36. The lower figure is recorded here and the discrepancy is flagged.

The Girl in the Spotlight: what was actually done in 2018

For two weeks, from Monday 26 February to Sunday 11 March 2018, the painting was examined twenty-four hours a day inside a glass enclosure in the Golden Room of the Mauritshuis, in front of the public. The project was called The Girl in the Spotlight, it was led by the Mauritshuis paintings conservator Abbie Vandivere, and it was carried out within the framework of the Netherlands Institute for Conservation, Art and Science, with scientists from the University of Antwerp, the National Gallery of Art in Washington, Delft University of Technology, the Rijksmuseum, the Cultural Heritage Agency of the Netherlands and others.

Every method used on the painting itself was non-invasive. Nothing touched it. The list was: technical photography in visible, ultraviolet, raking and infrared light, X-radiography, reflectance and fluorescence imaging spectroscopy, fibre optic reflectance spectroscopy, multispectral infrared reflectography, macroscopic X-ray fluorescence scanning, macroscopic X-ray powder diffraction, multi-scale optical coherence tomography, 3D scanning, 3D digital microscopy, and computer-assisted thread-level canvas analysis. Separately, the microscopic samples taken during the 1994 restoration were re-analysed with electron microscopy, focused ion beam scanning transmission electron microscopy, chromatography, synchrotron X-ray techniques at Grenoble and Hamburg, and lead isotope analysis. Four new samples were taken from areas the 1994 campaign had missed, including the skin.

The results were published across nine papers in a dedicated collection in the journal Heritage Science in 2019 and 2020, all open access. The claim made for the outcome, in the museum's own press language before the work began, was that the painting would afterwards be one of the best documented works of art in the world. That claim held.

What 2018 established about the background

The background is the headline, and it is also the place where the popular account and the published account diverge.

What the paint is, is settled. The dark space behind the figure is a black underlayer of charcoal, over which Vermeer laid a translucent glaze coloured with two organic dyes, indigo and weld, with chalk as the substrate and a copper compound added as a drier. It was originally a deep, even, glossy green. Both organic colourants have faded, which is why the area now reads as a mottled dark grey and why the signature is barely legible to the naked eye.

What was new in 2018 is what the imaging found inside that degraded surface. The X-ray fluorescence maps for lead, copper, potassium and calcium show diagonal lines in the upper right and lower right corners that resemble folds in fabric, with a narrow dark line beside a wider lead-rich band reading as shadow beside a highlight. Optical coherence tomography confirms the glaze is thicker along that line, and the infrared images show that Vermeer had already varied the thickness of the black underlayer in the same places. The modelling was therefore part of his original conception, not a later addition, and it stops behind the figure; the left side is a flat expanse with no diagonals at all.

The published wording is careful and this record follows it. The paper's own abstract says the imaging revealed the suggestion of folded fabric (possibly a curtain), and its conclusion says the folds suggest a draped fabric, possibly a curtain, that puts the Girl into a defined space. The press coverage of 2020 turned this into the flat statement that she stood in front of a green curtain. The evidence is diagonal modelling in a degraded green glaze, consistent with drapery and comparable to the curtain the Metropolitan Museum found by infrared in Vermeer's Mistress and Maid in the Frick. No earlier composition was found beneath the background of this painting. Folded fabric, very probably a curtain, painted green: that is what was established.

What 2018 established about everything else

The eyelashes. The Girl has been described in the literature as having no eyebrows or lashes, and one writer proposed alopecia. She has lashes. Those on her proper left eye are visible at high magnification as tiny black hatched lines. Those on her proper right eye, the one against the background, were visualised for the first time in the macro X-ray fluorescence map for iron: they were painted thinly in an iron-containing brown paint on top of the finished background, and appear as a negative image in the potassium, calcium and copper maps.

The ultramarine. The headscarf is natural ultramarine, ground from lapis lazuli that in the seventeenth century came from what is now Afghanistan, and it is of high quality: large particle size, strong colour, few accessory minerals. Synchrotron sulphur K-edge absorption spectroscopy indicated that the lazurite came, at least in part, from lapis that had been heated before grinding, a treatment some period recipes prescribe because it deepens the blue and makes the rock easier to mill. Ultramarine also turns up mixed into the yellow jacket and in trace amounts in the pink flesh paint. It does not appear in the surviving inventories of the Delft shops that sold painters' materials.

The lead white. Lead isotope analysis of eight samples traced the ore to the Peak District of Derbyshire. Vermeer used at least two grades of it, deliberately, with the finest particles reserved for the thin light scumble over the reddish-brown underpaint in the shadowed side of the face.

The changes. He shifted the ear upwards and repainted the earlobe, moved the shadow along the jawline to suit the new position, straightened the line between cheek and headscarf at the final stage, and softened the back of the neck. Each eye began as a small black dot placed at an early stage; the pupils moved.

The contour that is not there. Along the left side of the figure Vermeer left a gap one to two millimetres wide between the background and the face, jacket and headscarf, in places showing brown underlayer. It is deliberate, it softens the transition, and it produces the impression of light circulating round the side of her head.

There is no pearl

The object the painting is named for is not depicted. It is not modelled, it has no contour, and it hangs from nothing.

The Mauritshuis puts it plainly: Vermeer painted it with two strokes of white, one at the bottom to catch the reflection of the collar and a thick dab at the top, and there is nothing more, not even a silver hook. The technical examination confirms the construction from the paint: a scumble creates the soft counter-reflection from the white collar, a teardrop-shaped highlight was applied in thick impasto, and both sit on top of the already-finished skin tone of the neck. The pearl was added last, onto a completed passage, as two marks.

Nor was there necessarily a pearl in the room. It is too large to be a real pearl of that period. The museum's own catalogue text proposes a glass drop varnished to a matte sheen, or simply Vermeer's invention. Imitation pearls were a substantial industry by the seventeenth century, and pearls both real and false were fashionable between about 1650 and 1680.

One further loss should be recorded against the impasto. Historic linings flattened the raised paint in several areas of this picture, most conspicuously in the pearl itself. The highlight that carries the whole illusion is not as high as Vermeer left it.

What has been done to it since

Five interventions are documented and each left a mark. In 1881, immediately after the sale, a restorer named Van de Haeghen in Antwerp lined the painting with a starch-based adhesive and probably consolidated the front with animal glue; the resulting exposure to moisture shrank the canvas and lifted the edges of the cracks, dislodging fragments of paint. In 1915 Derix de Wild regenerated the top varnish layer, almost certainly by the Pettenkofer method using alcohol vapour and copaiba balsam, and repeated the treatment in 1922; the copaiba was later detected by chromatography and has visibly softened the paint in one place, the highlight on her forehead. In 1960 Jan Cornelis Traas relined the canvas with wax-resin, removed most of the old varnish, revarnished and retouched.

The 1994 treatment, carried out at the Mauritshuis under Jorgen Wadum alongside View of Delft and watched by the public through a glass wall and a skylight, was aesthetic rather than structural. Discoloured varnish and retouchings were removed or adjusted with solvent mixtures, losses were filled with chalk and polyvinyl alcohol, an isolating dammar varnish was applied, retouching was done in watercolour and then in dry pigment in polyvinyl acetate, and the final dammar varnish was given a light stabiliser to slow yellowing. It was published as Vermeer Illuminated, and it was one of the first restorations ever staged in front of a museum audience, a precedent the 2018 examination consciously repeated.

The painting is currently in a stable state of preservation and there are no plans to treat it.

Where it is, as of this record

Mauritshuis, The Hague, inv. 670, on view in Room 15.

It is about to move, briefly. The Mauritshuis closes to the public from Monday 24 August to Sunday 20 September 2026 for building alterations, and during that closure the painting will be shown at the Nakanoshima Museum of Art in Osaka, in a presentation organised by the Asahi Shimbun. The museum will use the proceeds to fund the alterations and the refurbishment of the building on Korte Vijverberg that it took over on 1 December 2025, where an education centre is planned for 2028.

The loan matters to a provenance record because this object effectively does not travel. Its last loan was to the Rijksmuseum's Vermeer retrospective in 2023, and before that its only substantial absence was the 2012 to 2014 world tour, mounted for the same reason: to pay for building work. The museum's director has said publicly that the Osaka showing may be the last time the painting leaves the country. Anyone checking this record against a location should check the date on it.