The Madrid Annunciation Diptych by Jan van Eyck

The Madrid Annunciation Diptych

Two small panels face one another in Jan van Eyck's Annunciation Diptych. Gabriel occupies the left wing, Mary the right. Each seems to be a pale sculpted figure standing in a niche against a glossy black background. The illusion is exact enough that a viewer might initially attend to the imagined stone before noticing the actual medium: oil paint on wood. Words across the frames give the angel's greeting and Mary's response. The two sides form one exchange, made to be viewed together: Gabriel’s announcement needs Mary’s answer, and her turn toward him needs the figure opposite. The Museo Nacional Thyssen-Bornemisza in Madrid dates it to about 1433–35; the inherited address of this page ends in “1440,” but that is not the museum's current date.^[1]

The scale is intimate. The museum records Gabriel's panel at 38.8 by 23.2 centimetres and Mary's at 39 by 24 centimetres. It places the diptych among small pictures probably made for private devotion, while leaving its first patron and exact setting unknown. The individual inventory numbers help distinguish the pair: 137.a (1933.11.1) identifies Gabriel, while 137.b (1933.11.2) identifies Mary. The borders make them appear substantial, although the pictured stone is an effect created on wood. Those choices are inseparable from the subject: a message crosses the gap between two figures, while their images seem poised on the edge of the viewer's own space.^[1]

The two wings together

A conversation across two wings

The narrative comes from Luke's account of the Annunciation. Gabriel tells Mary she will bear Jesus; Mary accepts the message. Instead of enclosing them in one continuous room, Jan sets each in a separate framed opening. The gap between the panels becomes part of the encounter. Gabriel turns toward Mary, while she faces him from the opposite side. Their paired placement lets the viewer follow the announcement from left to right and then return to Mary's answer. If either wing is reproduced alone, the figure remains remarkable, but the subject loses the physical form of its dialogue.^[1]

Gabriel is drawn through contrasts of contour and light. He has long hair and abundant drapery, with folds that break sharply around his body. One wing seems to turn outward from the dark niche, casting a shadow that makes it appear to occupy real depth. His standing base is hexagonal and projects visually past the lower edge of the inner fictive frame. These devices invite a momentary mistake: the viewer is looking at a flat painting yet feels as though a statue and its pedestal have been set into a shallow shrine. Jan uses that illusion to give the messenger a presence close to the picture's actual boundary.^[1]

Mary answers the angel in a matching niche. The museum's tactile guide describes the open book held before her and the dove of the Holy Spirit near her head. The book is a conventional part of Annunciation imagery, suggesting sacred reading interrupted by Gabriel's arrival. The dove marks divine action without making it another human figure in the conversation. Mary's placement parallels Gabriel's, but the book and dove make her side more than a mirror image of his. What is shared is the stone-like mode of representation; what differs is each figure's role in the event.^[2]

The two hexagonal bases are especially useful for understanding how the diptych works. They appear to project beyond the niches and rest over the pale lower border, as if the figures could step forward into the viewer's room. In physical fact, the pedestals, niches and figures are all painted. A small panel can therefore make a surprisingly complex claim about depth: the actual wooden object is shallow, the represented recess looks deep, and the painted base seems to cross between those levels. Looking from Gabriel to Mary repeats the trick, making the two wings feel like a single staged encounter.^[1][2]

Stone made with paint

The figure painting is a grisaille. By modulating white, black and grey, Jan evokes carved figures rather than living bodies in coloured clothes. The museum describes the niches as having a smooth black surface that reflects their outlines. Mary's reflection even suggests something of her back, as though a polished material has caught what a normal frontal view could not show. A painted reflection adds an apparent second view of a figure within the same narrow panel. It makes the fictive shrine look material and deep while showing the painter's ability to invent an optical effect.^[1]

That reflection has been interpreted as more than a display of skill. The museum connects the reflective ground with the idea of an unstained mirror, a traditional image of purity associated with Mary. The association is plausible within a work devoted to her acceptance of Gabriel's message, but the first visual fact remains the illusion of polished dark stone. The meaning emerges because Jan has made the surface convincing enough for it to be noticed. Theology and technique meet at the point where a viewer follows a pale figure into a black reflection.^[1]

“Monochrome” needs one qualification here. The principal figures mimic white stone against black, yet the original outer frames imitate reddish marble, and the inner painted architecture suggests veined pale stone. The diptych renounces the usual bright costume colours of a painted Annunciation, not every hint of hue on the object. Its restricted palette makes distinctions among materials more important: pale statue, black polished ground, redder outer border. A viewer is asked to read colour as the appearance of different stones rather than as the colour of Gabriel's and Mary's bodies.^[1][3]

Jan's use of grisaille also belongs to a wider practice. A 2009 Thyssen-Bornemisza exhibition placed the diptych among monochrome painting, drawing, manuscript illumination, sculpture and textiles from the fourteenth and fifteenth centuries. That context helps explain why a painter would want to imitate sculpture with paint. The illusion makes the two figures resemble devotional images that might have occupied architectural niches, but it remains a painter's answer to the problem of how light, surface and volume can be represented. The result is neither simply a picture of a sculptural group nor a neutral reduction of colour. It is a demonstration of how painting can seem to contain another medium.^[4]

What “grisaille” means in material terms

The white-stone appearance does not mean the painter simply mixed two pigments. Conservator Noëlle Streeton’s study of Eyckian reduced-colour painting, published in the first CATS proceedings, identifies a more nuanced mixture: lead white, carbon black, and finely ground red and yellow ochres. She includes the Madrid Annunciation alongside the Ghent exterior, drawing on Emil Bosshard’s earlier examination of the diptych. Thus the painted imitation of pale stone can contain warm colours without ceasing to look like stone. “Grisaille” describes the work’s restricted visual effect more readily than a complete list of its materials.^[9]

Streeton explains why the physical behaviour of lead-white paint matters. It could supply a covering, light-toned base rather than depend on a deep succession of transparent coloured layers. The qualities needed to evoke a carved fold differ from those needed for a jewel-like blue robe. That distinction makes Madrid especially valuable within Van Eyck’s surviving work: the painter achieves substantial bodies, recesses and sharp edges with a restrained range of tones. Materials serve the illusion of weight without requiring the brilliant costume colours found in his full-colour Annunciations.^[9]

Streeton’s wider argument concerns workshop organisation. Predictable mixtures and repeatable passages might have allowed assistants to contribute to paintings bearing the master’s name. She presents this as a question and a proposed explanation, not as a newly demonstrated division of hands in Madrid. Technical likeness is not a signature, and a material recipe alone cannot identify the person who painted a particular passage. Her account nevertheless changes the useful question from “why did the painter omit colour?” to “how was this convincing appearance constructed?” The diptych can be studied as a deliberately made surface as well as an optical tour de force.^[9]

Frames that speak and deceive

The frames are a central part of the work, not an interchangeable border added by a later owner. The museum identifies the carved and painted fifteenth-century Flemish frames as original. Their outer portions imitate reddish marble; a second painted structure evokes veined pale stone, with vertical jambs and a lintel around each figure. Painted bases seem to rest on its lower edge. Real carved mouldings and false stone architecture work together, so a viewer moves across several boundaries without immediately seeing which are physical and which exist only in paint.^[3]

Watch: Jan van Eyck in context

Choose a Smarthistory close reading of the Arnolfini Portrait or the closed and opened Ghent Altarpiece.

Jan van Eyck, The Arnolfini Portrait

Smarthistory

Look closely at the room, mirror and gestures in the London double portrait.

Watch on YouTube

▶01 · PortraitThe Arnolfini PortraitSmarthistory▶02 · ExteriorThe Ghent Altarpiece, closedSmarthistory▶03 · InteriorThe Ghent Altarpiece, openSmarthistory

Words on the upper frames extend the illusion. Above Gabriel, the inscription gives the beginning of his greeting to Mary; above her, another gives the conclusion of her reply. The letters appear to belong to the fictive stone setting. Jan often made inscriptions important to the identity and meaning of his pictures, but here they also give the diptych its order. The viewer need not search the empty space between the figures for painted speech. The frames hold their words in place while the bodies seem to come forward from the niches below.^[1]

There is a deliberate difference from the Washington Annunciation. In Washington, words are suspended inside a single painted church, and Mary's response is oriented upward. In Madrid, the beginning and ending of the exchange belong to the two framed wings. The comparison does not establish a sequence of invention or prove that one composition was made for the same patron as the other. It shows how Jan could use written language to structure two otherwise distinct objects. Each gives the viewer a way to follow the biblical conversation through the painting's physical form.^[1]

The backs also matter, even though they are seldom seen in a frontal reproduction. The museum describes them as painted by Jan, somewhat deteriorated, and perhaps meant to resemble porphyry in greenish and ochre tones. “Perhaps” is essential: the proposed material effect is an interpretation of the surviving reverse, not a fully legible original label. Still, the painted backs show that the object was considered beyond its two most familiar faces. A hinged devotional work could be handled or closed; it had an outside as well as an inside. The reverse is part of the material history of the diptych, not blank space behind its celebrated figures.^[3]

Small scale and probable devotion

The museum associates the diptych with private devotion because its two wings are compact and its subject invites close attention. A person could approach each face at a distance that makes the inscriptions, shadows and reflections meaningful. That is a reasonable account of likely use, though no checked document names an original owner, chapel or room. The picture gives many clues about how it might have been viewed; it does not give a complete history of who first held or commissioned it.^[1]

The experience of looking may have involved more than finding Gabriel and Mary. A viewer could begin with the figures, notice that their apparent stone bodies cast convincing shadows, then move outward to the bases and inward again to the black reflections. Reading the words across the top would return the eye to the story. The original frame pulls that movement together: it protects the panels physically, orders the conversation and supplies the false architecture in which the figures stand. The eye repeatedly returns to the transition between the body and its support. At that boundary, a ridge of cloth can seem to continue into the space in front of the niche.^[1][3]

There are limits to reconstructing a first viewing. The museum's statement that the work was probably for private devotion is a typological judgement, not an inventory entry documenting a specific household. We also cannot assume the diptych was always displayed open or at a fixed height. Its backs and hinged format invite questions about closing and handling, but the checked sources do not establish a daily ritual. What can be said securely is that the small size and paired arrangement made a close, changing encounter possible. The work's physical features support that inference without turning it into an invented scene of use.^[1][3]

Date, authorship and comparison

The current collection record names Jan van Eyck and dates the diptych to about 1433–35. The museum traces its modern recognition to Max J. Friedländer’s 1934 publication and describes the attribution as unquestioned since then, while acknowledging debate over its date. That statement represents the owning museum’s account; a contrary assessment survives in another institutional catalogue, discussed below. An earlier Thyssen-Bornemisza exhibition labelled it about 1435–40, and the museum's VERONA technical-project account called it a work of the middle to later 1430s. These are not three separate diptychs. They are different datings of the same object, made at different moments of study. The current collection date belongs first in the identification. The different labels show how dating can change without a change of object or a newly discovered date inscription.^[1][4][5]

The comparisons behind the dating are revealing. The museum sets the diptych beside the 1437 Dresden triptych, whose outside wings also show a separated Annunciation in grisaille. It also mentions the 1437 Saint Barbara and the 1439 Madonna at the Fountain when explaining why a place in Jan's chronology has been discussed. Resemblance in an approach to pale figures, delicate modelling or small devotional format can help situate an undated work. It cannot convert the Madrid painting's present estimate into a date written on its frame. In particular, the old site address ending in “1440” is not a reliable substitute for the museum's c. 1433–35 catalogue date.^[1]

The diptych may also have had a later influence. The museum reports Frédéric Elsig's comparison with an Annunciation by the French painter Nicolas Froment on the reverse of his Burning Bush altarpiece in Aix-en-Provence. That is an attributed scholarly proposal based on visual parallels, not a documented loan, encounter or act of copying. It is useful because it shows why Jan's stone-like Annunciation mattered beyond its own private scale: another artist could adapt the idea of isolated, sculptural figures to a different painted structure. The possibility belongs to the picture's afterlife, while the evidence for a direct transmission remains limited.^[1]

Its recorded collecting history is much shorter than its artistic afterlife. The museum says the diptych was formerly in a French private collection and was acquired in 1933. Friedländer published the attribution to Jan the following year. Those facts explain when the two panels entered the modern scholarly record, but they do not identify the fifteenth-century commissioner or trace every owner between Jan's workshop and the twentieth century. The preservation of the original frames is unusually informative; it does not close the documentary gap. A reader can distinguish the work's convincing physical wholeness from the incomplete written account of where it travelled before reaching Madrid.^[1][2]

A disagreement in the French catalogue record

The Louvre’s French-language entry for its separate diptych of Saint John the Baptist and the Virgin and Child contains an unexpected qualification. In comparing that work with other grisaille pictures, it calls Dresden an original by Van Eyck but describes the Madrid Annunciation as the work of a faithful follower. This conflicts with the Thyssen’s direct attribution and its account of unchallenged acceptance. The Louvre entry, updated in March 2026, also cautions that its content may not reflect the latest state of knowledge. It supplies no new technical demonstration for reassigning Madrid.^[11]

The discrepancy should remain visible without turning a comparative phrase into a settled reattribution. This page follows the owning museum’s Jan van Eyck designation and current date, while acknowledging the divergent institutional wording. Nor should the Louvre diptych’s own ownership, altered hinges or later dating be transferred to Madrid. They are different objects with related devices. Their comparison shows how apparently similar simulated statues can generate separate arguments about authorship, chronology and the survival of Eyckian designs.^[1][11]

What technical study can establish

In 2015, the VERONA project came to Madrid to document the diptych. The collaboration between the museum and the Royal Institute for Cultural Heritage used close photography under ordinary and raking light, ultraviolet fluorescence, infrared imaging and reflectography. The museum's project page makes photographs, infrared views and X-radiographs available as evidence of the examination. Such records allow conservators to study paint, drawing and structure beyond what a gallery visitor sees. These images are a research resource, rather than a published list of conclusions. A dark mark in an infrared image needs interpretation in relation to the visible surface and the imaging method; it is not automatically a changed outline or a second artist’s intervention.^[5]

The Belgian institute’s Dutch-language account explains the importance of a consistent procedure. VERONA used the same equipment, lighting and resolution to document dispersed works, making comparisons possible across museums. The campaign also photographed reverses and recorded objects from different angles. For Madrid, that approach suits a picture whose effect depends on the relationship between panel, surround and apparent projection. Ordinary frontal photographs answer one set of questions; views of construction and surface answer others. The institute’s account makes the evidence accessible without claiming that every question of authorship or workshop practice has been resolved.^[10]

The work's original frame presents a practical reason for care. The museum's preventive-conservation account illustrates a climate case made for this diptych and explains that such protection can accommodate an original frame without altering it. This is a modern part of the object's history, not a clue about Jan's workshop. It does show how the surviving wood and painted surfaces are treated as one vulnerable ensemble. The frame carries inscriptions, false stone, mouldings and a painted reverse; protecting it preserves evidence that a reproduction of the two central figures alone would omit.^[3][6]

One of several Annunciations

Jan's Madrid diptych is distinct from the Washington panel and the Annunciation on the exterior of the Ghent Altarpiece. Washington gives Gabriel and Mary a richly coloured church interior, with biblical images on walls and floor. Ghent distributes their encounter across shutters that belong to a much larger opening and closing programme. Madrid presents two apparently sculpted figures in a portable pair whose inscriptions are held by its frames. The same Gospel episode can therefore be read in three different physical languages. In Madrid there is little narrative setting to distract from the meeting. Empty black areas isolate the figures, and the two lit bodies direct attention toward each other. The small dove, open book and framed words carry the event without a furnished room or receding church.^[1][7][8]

The Madrid work rewards slow looking because nearly every apparent material asks to be checked. The figures are painted like sculpture, the black setting like polished stone, the borders like two kinds of marble, and the letters like writing cut into that stone. At the same time, the carved and painted frames really survive from the fifteenth century. The diptych's power lies in that interplay between what exists and what appears to exist. Gabriel and Mary can be followed as a sacred conversation, while the object's making remains visible in every attempted crossing between painting, sculpture, frame and reflected image.^[1][3][5]

Related reading

The Annunciation in Washington

Ghent Altarpiece

The Arnolfini Portrait

References

1. Mar Borobia, “The Annunciation Diptych,” Museo Nacional Thyssen-Bornemisza, Madrid, collection record 137.b. View Source.
2. Museo Nacional Thyssen-Bornemisza, “Tocar para ver,” tactile description of The Annunciation Diptych. View Source.
3. Museo Nacional Thyssen-Bornemisza, “Exceptional picture frames,” Jan van Eyck section. View Source.
4. Museo Thyssen-Bornemisza, “Jan van Eyck: Grisailles,” exhibition site, 2009–10. View Source.
5. Museo Thyssen-Bornemisza and VERONA project, “Technical and scientific documentation of the Annunciation Diptych,” 2015 project account. View Source.
6. Museo Nacional Thyssen-Bornemisza, “Preventive Conservation,” climate-case account. View Source.
7. National Gallery of Art, Washington, “The Annunciation,” collection record 1937.1.39. View Source.

8. Saint Bavo Cathedral, “The Ghent Altarpiece,” history and programme. View Source.

9. Noëlle L. W. Streeton, “Emulating Van Eyck: The Significance of Grisaille,” European Paintings 15th–18th Century: Copying, Replicating and Emulating, CATS Proceedings I, 2012 conference (published 2014), pp. 29–35, especially pp. 31–32 and notes 17, 25. View Source.

10. Royal Institute for Cultural Heritage, Brussels (KIK-IRPA), “VERONA – Van Eyck Research in OpeN Access,” Dutch project account, standardised photography and comparative research. View Source.
11. Musée du Louvre, “Diptyque avec saint Jean Baptiste et la Vierge à l’Enfant,” RF 1938 22, French collection record, comparative attribution wording; updated 6 March 2026, with catalogue currency caveat. View Source.