There is no record of the portrait having been cleaned previously, which has led to the claim that it had not been touched by restorers since the seventeenth century (1).
The appearance of the painting and its varnish layers, however, indicated that it is likely to have last been cleaned in the nineteenth century.
The natural resin varnish, either dammar or mastic, had inevitably become discoloured and slightly opaque with age. The discoloration also emphasised the unevenness of its application: for example part of the veil on the right appears much brighter as a result of the thinner and less discoloured varnish in this area. A greyish layer of pigment applied over the white veil by the previous restorer, presumably to tone down the original creamy white colour, contributed further to the patchy appearance.
The old restoration of the paint losses, which are very small, often no more than a chipped corner to a raised crack, was unnecessarily extensive and had darkened. The reddish brown blotches on Margaret's check bone, for instance, could easily be mistaken for skin blemishes. Other very evident retouchings included the brown smears on the grey fur, covering small losses and open cracks in the original paint.
Furthermore, with time the surface of the painting had accumulated various accretions, mainly small but disfiguring to a work of such refinement. For example, there were several dark marks on Margaret's left eyelid.
With the help of magnification we can see that only one of these marks, in the centre of the eyelid, is a retouched paint loss. Above it is a lump of a dark brown substance (clearly sitting on the surface and crossing the old cracks in the paint). Towards the corner of the eye darkened varnish has filled a small depression in the surface.
Scattered across the paint surface were numerous spots of dark green paint, presumably splashed onto the portrait during redecoration of a room, and also occasional flecks of gold leaf. Although minor, these defects combined to give the painting a shabby appearance and made the sitter seem older and more tired than her 33 years might lead one to expect.
(1) Condition report dated 10 April 1967 signed by A. Janssens de Bistohoven, V Vermeersch and C. Leegenhoek. Dossier on the painting at the Groeningemuseum, Bruges.
Aged natural resin varnishes based on dammar or mastic can be dissolved by solvents and removed from the surface of a painting on small swabs of soft cotton wool.
In the case of Margaret Van Eyck this procedure was carried out under high magnification with the aid of a stereo binocular microscope. On the computer screen the restorer is able to refer constantly to photographs of the painting taken before cleaning and also to useful images such as X-radiographs and infrared reflectograms. Sometimes old retouchings over losses and areas of damage become insoluble to normal cleaning solvents and have to be removed with a scalpel, a slow and painstakingly process carried out under the microscope. Fortunately most of the previous restoration on the portrait had been carried out using a varnish medium and could therefore be dissolved off with the discoloured varnish layers.
This included the grey toning over the white veil and a dark chocolate brown paint that had been applied over much of the black background in order to disguise the chipped and eroded edges to some of the cracks.
As the removal of the discoloured varnish and the old restorations proceeded, the well preserved original paint surface was gradually revealed. Evidently the painting has been treated with care in the past.
Indeed it is possible that some of Van Eyck's original surface coating has survived. When cleaning a painting the restorer will regularly evaluate progress by examining the surface with an ultraviolet light source. If necessary, the cleaning can also be documented with an ultraviolet fluorescence photograph. The technique is useful because natural resins and oils exhibit strong and characteristic types of fluorescence under ultraviolet, whereas the paint layers, which contain pigment, are generally less fluorescent. The nineteenth-century varnish still present on the left side of the portrait has the strong greenish-yellow fluorescence typical of an aged natural resin such as dammar or mastic.
In the area that has been cleaned this fluorescence is no longer visible (apart from some small patches at the edges). Over Margaret's face, headdress and much of the red dress there is still a thin fluorescent layer, but of a more orange-yellow colour. In the lower right corner, however, an area of the dress appears dark and without fluorescence. In a paint sample from a loss in this area, it can be seen that there is a thin layer of an un-pigmented fluorescent substance sandwiched between the penultimate and final layers of red lake glaze. It seems possible therefore that Van Eyck applied a layer of varnish to the painting when it was almost complete in order to saturate the colours but then decided to apply a further glaze layer over the varnish in the lower right corner in order to deepen the colour of the folds (1). This glaze remained glossy and saturated and so did not need to be varnished again. The slightly orange colour of the fluorescence is characteristic of a varnish containing a resin dissolved in a drying oil, as would have been used in the fifteenth century, and so the layer may consist of the same heat-boded linseed oil with pine resin found in the more substantial surface coating on the marbled reverse (2).
(1) Layers of varnish interspersed with paint layers have been found on the thirteenth-century English Westminster Retable; see M.L. Sauerberg, A. Roy, M. Spring, S. Bucklow and M. Kempinski, ‘Materials and Techniques', The Westminster Retable: History, Techniquet and Conservation, eds. P. Binski and A. Massing, with M.L. Sauerberg, London-Turnhout 2009, pp. 244-6.
(2) An example of an early Netherlandish painting in National Gallery with residues of a similar varnish is The Virgin and Child Enthroned, with Four Angels (NG6282) by Quinten Massys. See J. Dunkerton, ‘The Technique and Restoration of ‘The Virgin and Child Enthroned, with Four Angels' by Quinten Massys', National Gallery Technical Bulletin, 29, 2008. pp. 60-75.
Following cleaning the small losses and areas of damage needed to be retouched so that they do not distract from the compelling image and from Van Eyck's immaculate painting technique.
The small size and extreme refinement of the painting meant that the retouching had to be very precise. It was therefore mostly carried out under the magnification of a microscope. The smallest size of brush was used to apply the retouching paints, which are manufactured especially for restorers (1). The materials used for the new restoration have to be stable, not changing colour like the old varnish and retouchings, and they must remain easily resoluble so that the painting can be safely cleaned again in the future. Carefully selected and tested modern synthetic resin paints are therefore employed.
The most damaged part of the painting was the black background, especially in the area above Margaret's head. By eliminating these losses, the velvety depth of the black is recovered and the astonishing painting of the edges of the white headdress can be appreciated without interruption.
The network of cracks which has developed with time in the ground and paint layers is a distinctive feature of the painting - a similar craquelure can be seen on some other works by Van Eyck., including A Portrait of a Man.
In some areas, however, the cracks have become widened by small flake losses from their edges and corners. The effect was to disrupt the subtleties of modelling in the white headdress and in a few places on Margaret's face and neck. These minute losses were retouched, taking great care not to eliminate the cracks completely as they are intrinsic to a painting of this age.
(1) For Gamblin Conservation Colours see conservationcolors.com
The panel consists of a single board of oak, about 8mm thick. It has a regular vertical grain. This indicates that it is quarter sawn (cut radially from the tree trunk like a slice of cake), as is usually the case with Netherlandish oak panels. Since the frame prevents access to the end grain, dendrochronology (tree-ring dating) is not possible.
The four sections of the frame, carved with the same profile on both front and back, have been assembled around the panel using simple mortise and tenon joins, with the tenons in the side pieces slotted into the mortises in the upper and lower mouldings (1).
The joins were then pegged with short wooden dowels, visible in the X-radiograph and at the upper corners of the frame where movement in the wood has caused them to push through, cracking the decorated surface. The remains of small metal pins are visible in the X-radiograph at each corner. They appear to have been hammered through from the back and are likely to have been part of a later repair and reinforcement of the frame. Two more metal pins and the circular nail heads at the centre of the upper edge are the remains of some sort of hanging device, probably not original.
The panel surfaces and frame mouldings were both coated with a chalk and glue ground in preparation for painting; originally they would have appeared as continuous. As the oak board ages and dries out, it contracts slightly, causing the ground to crack at the junction between the panel and the frame moulding. The gap at the left edge is now wide enough for the wood of the oak panel to be visible. Cracks have also developed in the frame joints but Van Eyck cleverly anticipated this by painting joins in the imaginary marble to echo the wooden construction. In the upper left corner the painted join and the crack from the real join coincide exactly.
(1) For diagrams of the construction of the portrait, see Hélène Verougstraete-Marcq and Roger Van Schoute, Cadres et supports dans la peinture flamande aux 15e et 16e siècles, Heure-le-Romain, 1989, pp. 177-8.
Examination by infrared reflectography has revealed a detailed underdrawing for Margaret's head and headdress, but with only a simple outlining for the lower part of the portrait, the design of which seems to have been more fully developed at a later stage. The underdrawing has been executed with a brush and a liquid medium. The lines are sometimes relatively broad and free, for example those that indicate the crimped edges of her headdress, including the marks which suggest that originally the edge of the fabric was to echo the curve of her right cheek.
The shadowed side of her face is modelled with very fine closely hatched lines of exceptional delicacy, almost as detailed as the shading on Van Eyck's portrait study in metalpoint on paper of a Cardinal, usually identified as Niccolò Albergati (Kupferstich-Kabinett, Staatliche Kunstsammlungen, Dresden). Van Eyck seems to have made this very elaborate drawing, complete with colour notes, in order to produce the painted portrait (Kunsthistorisches Museum, Vienna) when he no longer had access to the sitter (1). This would not, of course, have been the case when executing a portrait of his wife.
As well as the possible change to the edge of the white fabric, there are several differences between the underdrawing and the final painting. The diagonal lines sketched across her ear imply that they may have been intended to be covered, the ‘horns' of coiled hair were originally shorter and Margaret's eyes, especially her left eye, were drawn in a higher position.
Both in its technique and in the alterations made during painting, the underdrawing in this portrait is strikingly similar to that of the considerably smaller scale head of Giovanni(?) Arnolfini's wife in The Arnolfini Portrait. In this instance, however, the eyes have been moved up (2).
(1) See T. Ketelsen, I. Reiceh, O. Simon, and S. Merkel, ‘New information on Jan van Eyck's portrait drawing in Dresden', The Burlington Magazine, CXLVII, 2005, pp. 170-5.
(2) For the underdrawing on the Arnolfini Portrait see R. Billinge and L. Campbell, ‘The Infra-Red Reflectogram of Jan van Eycks's "Portrait of Giovanni (?) Arnolfini and his Wife"', National Gallery Technical Bulletin, 16, 1995, pp. 47-60 and also R. Billinge, ‘Examining Jan van Eyck's Underdrawings' in S. Foister, S. Jones and D. Cool (eds), Investigating Jan van Eyck, Turnhout 2000, pp. 80-96.
The precision of the underdrawing revealed by infrared reflectography in Margaret's face and the perfectly calculated geometry of the final design might lead one to suppose that Van Eyck had worked out every detail of the portrait in advance. Examination of the X-radiograph, in conjunction with the infrared image, reveals that in fact he made several adjustments in the course of painting.
Since X-rays are transmitted through the whole object, not just the paint surface, an image of the wood of the panel and the marbling on the reverse is also recorded. The X-ray opacity of the dense vermilion spatters means that they appear as grey splotches across the radiograph, in places reducing its legibility.
The X-ray image confirms that Van Eyck had begun to apply the white paint of Margaret's headdress before he decided that the horns of her hair needed to be made longer than in the underdrawing. The exceptional X-ray opacity of the lead white paint of the headdress around her right cheek may simply be the result of careful painting up to this important contour, or there may be an extra thickness of paint because he began to paint the headdress following the first sketched position of the crinkled edge of the fabric and then cancelled it.
The quickly brushed strokes of white paint in the area of Margaret's ear appear to confirm that originally they were to be covered, as indicated in the underdrawing. It is difficult to imagine how this might have worked if the horns of the hair were to remain exposed.
Several alternations were made to the arrangement of the folds and edges of the headdress. On the left side of the portrait the white paint extended further to the left, as did the red paint of her sleeve.
Since the red pigments are almost transparent to X-rays and in this area there is notable interference from the decorated reverse, this change can only be seen by close examination of the painting surface, in particular along the edges of the cracks in the black paint of the background. On the right side a hidden triangular corner of fabric appears between the present hanging folds.Several alterations were made to the arrangement of the folds and edges of the headdress. On the left side of the portrait the white paint extended further to the left, as did the red paint of her sleeve.
Its elimination was important for the jagged pattern formed by the creamy white shapes as they intersect with the deep red of the dress. Equally important for the design is the painting out of the fine white linen at Margaret's neckline, even though on the evidence of other similarly-costumed women in Netherlandish paintings of this period it seems to have been the custom to show the linen worn between the skin and the fur lining of the dress. If left visible this extra area of white might have spoilt the effect of the headdress.
Perhaps the most significant alteration of all is the introduction of Margaret's hands, crossed and lightly resting against the folds of her skirt. The long heavy sleeves make them appear undersized, as does their cutting by frame edge. How Van Eyck originally envisaged the pose of Margaret's arms is not clear. Both infrared and X-ray images suggest a triangular gap between sleeve and waistband, in which case the arms would have been held further apart and the hands perhaps dropped out of the picture. The decision to include them was made at the underpainting stage, when they appear to have been outlined in black over the first layers of red paint and then painted with the same flesh tints as for the face. The increase in transparency with age of the upper paint layers means that the red underlayers are now visible, especially in the thinly painted further hand and its fur cuff.
One detail which might be thought to be a pentimento, the extension of the frilled edge of the headdress over the black background at the upper edge, is probably not one; is just as likely to be a deliberate effect to demonstrate the gauzy fineness of the linen fabric.
Margaret is expensively dressed in a red wool gown, lined with grey fur, probably squirrel. The rich red colour, beautifully preserved, is built up with several layers of paint. Over the white chalk ground is a thin layer of the opaque red pigment, vermilion, bound in a medium of heat-bodied linseed oil (identified by GC-MS). This seems to have been a flat, unmodelled layer, brushed in broadly, the sweeping brush marks clearly visible in the X-radiograph, especially in the area of her hands, which were added over the red underpaint.
In places the red underpaint encroached onto the mouldings of the frame (now visible in small areas of damage).
Over this red underlayer, the folds of the dress were modelled with layers of red lake glaze. Unfortunately, the tiny paint sample, taken from the damaged lower edge, split at the interface between the underlayer and the subsequent paint layers.
The lowest layer of the upper cross-section is in fact a grey paint consisting of bone black and lead white. Since the sample point is at the conjunction between image and frame this is probably an overlap with the greyish marbled frame, confirmation that the frame decoration was applied concurrently with the painting of the portrait. Three layers of red glaze can be seen in the cross-section, most clearly when viewed in ultraviolet light. There may be a small amount of lead white in the first glaze layer which fluoresces more than the other red layers. The strongest fluorescence is from a layer of varnish between the final and penultimate glazes, perhaps applied to saturate the colours before the final touches. Some of this varnish may survive as a surface coating elsewhere on the paint surface.
Analysis by HPLC reveals that the paint of the glazes contains two separate dyestuffs. The principal component is kermes from the scale insect, Kermes vermilio Planchon, the most expensive red dyestuff, but some madder, the dyestuff from the plant Rubia tinctorum L., is also present. Its distinctive fluorescence in ultraviolet can be seen in particles in the cross-section. Although GC-MS analysis confirmed that the binding medium of the glaze layers is the same heat-bodied linseed oil as in the underpaint, some protein was detected by FTIR. This indicates that one of the dyestuffs used to make the red lake pigment, most probably the madder, was sourced from shearings of a dyed red woollen textile (1).
To darken the shadows Van Eyck added a little bone black and also some natural ultramarine. The deepest folds were reinforced with final touches of ultramarine. Unfortunately the rich purple blue pigment is now degraded and blanched, a common phenomenon in paintings by Van Eyck (including A Man in a Turban NG222) and so the original effect is lost. Under magnification it can be seen that Van Eyck blotted the sticky paint with his fingers or the palm of his hand in order to ensure even application of the glazes (2).
Embedded in the paint of the fur where it meets the edge of the veil on the right side of her neck is an S-shaped red fibre. Is this a fibre from Margaret's red wool dress, shed as she leant over to inspect her husband's portrayal of her in all her finery?
(1) For sources and manufacture of pigments from dyestuffs see J. Kirby, ‘The Identification of Red Lake Pigment Dyestuffs and a Discussion of their Use', National Gallery Technical Bulletin, 17, 1996, pp. 56-80.
(2) Many painters blotted glazes and other paint layers with their fingers or with rags. For some other examples see J Dunkerton and R. Billinge, Beyond the Naked Eye, National Gallery, London 2005, pp. 48, 58-9. For blotting of lake glazes over the whole area of a red drapery, in this instance with a fabric pad, see J. Dunkerton, ‘The Technique and Restoration of ‘The Virgin and Child Enthroned, with Four Angels' by Quinten Massys', National Gallery Technical Bulletin, 29, 2008. pp. 60-75.
The old tradition that Jan van Eyck was the inventor of oil painting has long been disproved. Drying oils had been used as binders for pigments to make paints for centuries before van Eyck took up his brushes. By chance a hair shed from a paint brush used for his portrait of his wife remains embedded in the black paint at the left edge. Its length (about 6 mm) and colour indicate that it is probably from a fine brush made from the tail fur of a squirrel or a member of the weasel family, much like a modern sable brush. Heat-bodied linseed oil was identified in all the paint samples that were analysed from the painting and the front frame mouldings. Heat-bodied oils have been partially pre-polymerised (polymerisation is the process by which it sets or ‘dries') by gentle heating. Oils could also be placed in a dish exposed to sunlight (1).
Paints made with heat-bodied oils form a softly rounded impasto when they contain pigments such as lead white and smooth glossy glazes with translucent pigments such as red lakes and copper greens.
Margaret's complicated white headdress provided a superb vehicle for van Eyck's descriptive powers and close examination demonstrates the speed and virtuosity with which he manipulated his simple painting materials. These headdresses, fashionable in Bruges at the time, consisted of a single length of fine linen with ruffled edges introduced as part of the weaving process. The fabric was repeatedly folded back on itself until the edges formed a dense frill. In Margaret's case seven layers of fabric can be counted, while Giovanni Arnolfini's wife, whose head is dressed in similar fashion, has only five (2).
The precision and confidence with which van Eyck's brush followed the twists and turns of the edges of the ruffles, sometimes picking them out with short dots of dense white, is astonishing.
At the right side of the painting he suggested a single thickness of fine linen by bringing touches of white over the black background with deft flicks of the brush. In other places he defined the little pleats of the ruffles by scraping into the soft white paint with a tool, perhaps the end of the handle of his paintbrush. In the X-radiograph these scraped out lines show as darker than the rest of the white paint.
Along the top of the headdress van Eyck has taken a small dry brush and dragged the wet white paint up into the black in tiny vertical strokes. These may have been to indicate the softness of the linen but they also soften the contour, making it vibrate.
He has used the same short wet-in-wet strokes to blur the edge of the upper lid of Margaret's right eye.
The painting of the soft fur gives the impression of precision but in fact the individual hairs of the pelt are defined only at the edges where the fur overlaps her dress and neckline.
Margaret's rich green sash or girdle is depicted as a silk damask and its simple pattern suggests the possibility that she may have woven it herself. It would have been woven as a single strip and not cut from a larger piece of fabric. Small tablet looms for weaving ribbons and girdles are recorded as having been owned by gentlewomen in the Renaissance (3). Van Eyck has described the belt with his usual confident skill, the rapid flick of his brush evident at the points of the zigzags. It is completed with a rich glaze of copper green pigment, almost certainly in the same heat-bodied linseed oil used for the rest of the painting. Here it has resulted in a smooth glossy surface appropriate for the silk that it represents. This contrasts with the blotting of the final glazes on the scarlet dress, which both ensures an even application and suggests the different texture of the matt woollen fabric.
(1) For paint media in early Netherlandish and German painting, including heat-bodied oil, see ‘Methods and materials of Northern European painting', National Gallery Technical Bulletin, 18, 1997, pp. 6-55, esp. pp. 40-3.
(2) See the catalogue entry in L. Campbell, M. Falomir, J. Fletcher and L. Syson, Renaissance Faces. Van Eyck to Titian, Exh. Cat., National Gallery. London 2008, p. 180 and S. M Newton and M .M. Giza, ‘Frilled Edges'. Textile History, 14 (2), 1083, pp. 141-52.
(3) Information supplied by Lisa Monnas.
In the inscription on the frame it is Margaret herself who announces ‘My husband Jan completed me on 15 June 1439' when she was 33 years old. Although no beauty by modern tastes, she comes across as a forceful personality. Her maiden name is not known, but she seems to have had a relatively high social status, probably coming from a similar rank as Jan. Her clothes are expensive, and certainly not those of the wife of a craftsman painter. She had several children and outlived her husband by fifteen years or more, probably continuing to run his workshop after his death.
Van Eyck used a number of devices to create this impression of a living, speaking woman. As with many of his portraits, the head is given emphasis by making it over large in proportion to the body. Although his masterly understanding of the planes of Margaret's face is demonstrated by the shading of her left temple, cheek, and jaw, and above all in the placing and depiction of her ear, her nose is presented slightly too far in profile for the rest of her face. Van Eyck did this in other portraits as well, drawing attention to an easily recognisable feature, but perhaps also giving an impression of mobility in the sitter as the viewer unconsciously accommodates the shifting viewpoint.
Much of the sense of character comes from her watchful blue-grey eyes, the graduation to hazel towards the pupil precisely observed. When they are examined separately it becomes apparent her right iris has been painted larger than her left. This is not a defect of hers or of her husband's brush control; rather it is an indication of van Eyck's readiness to distort literal observation to enhance the sense of a living being.
One small imperfection that he did faithfully record is in her left eyebrow. At first glance it might appear that the interruption in the eyebrow is the result of damage to the paint surface. Under magnification, however, we can see that he has deliberately left a break in the painting of fine red-brown hairs. Margaret must have had a small scar or perhaps she had the somewhat patchy eyebrows typical of a redhead. The gingery hair coiled around the horns may well have been false, but presumably matched her real hair colour.
Now that the painting is free of the discoloured yellow varnish, her cool pink and typically Northern complexion is evident. The area of pink skin on the side of her imposing nose is painted with hatched brushstrokes that define its shape. The slight flush around the right side of her chin, on the other hand, is added with small pointilliste flecks of colour.
When magnified her thin-lipped mouth becomes almost sensuous as a result of the streak of deep pink where her lips meet.
(1) See the catalogue entry in L. Campbell, M. Falomir, J. Fletcher and L. Syson, Renaissance Faces. Van Eyck to Titian, Exh. Cat., National Gallery. London 2008, p. 180 and also p. 43.
(2) For distortion to emphasise likeness in Renaissance portraiture see L. Campbell, Renaissance Portraits, New Haven and London 1990, pp. 9-12.
Jan van Eyck both signed and dated the portrait and identified the sitter with the inscriptions on bands along the upper and lower edges of the frame. The Latin inscription on the upper moulding reads: ‘CO(N)IU(N)X M(EU)S IOH(ANN)ES ME (COM)PLEVIT A(N)NO . 1439˚ . 15˚ . IUNII' (My husband Jan completed me on 15 June 1439) (1).
It continues on the lower moulding with ‘[A]ETAS MEA TRIGINTA TRIU[M] AN[N]ORUM . AΛL (can't find correct character) ICH CAN' (My age being 33 years. As I can). The same pun on the word Ich and van Eyck's name appears on the possible self-portrait in the National Gallery.
The way the light catches the right ends of the inscription panels indicates that we are to understand them as set into the marbled frame mouldings. The illumination of the letters and numerals also implies that they are indented and perhaps gilt. It is not clear what the brown background to the letters is meant to represent. Apart from the glint on the letters there is no suggestion of the reflective surface of a metal. The facetted rhomboid dots at either end of the strips look more like punctuation points (and are lit as such) rather than pins securing metal plates. Perhaps the inscriptions are to be thought of as strips of tooled leather glued into a shallow recess in the marble.
The mouldings of the frame on the front face of the portrait were painted a warm yellow-grey and streaked with a translucent brownish black to represent the veins of the marble. In the recent treatment the frame was cleaned lightly, removing mainly surface dirt and smears of beeswax that had been applied to consolidate the frame. It retains various surface coatings (in part possibly original), but also the inevitable build of encrusted dirt and minor damage from handling to which frames are vulnerable.
Examination of the edges of open cracks at the joins, and also a paint sample, suggest that originally Van Eyck may have planned a more colourful marbled effect since the lowest layer is a bright green colour containing verdigris mixed with a lead-tin yellow and a little black and white. Over it lies the warm yellow grey of the present decorative scheme. The medium of both layers is the same heat-bodied linseed oil used for the portrait itself. Another possibility is that the green underlayer may have been applied to impart a cooler tonality to the superimposed layers, or it could have been simply to use up surplus green paint from the painting of Margaret's sash - not sampled because of its perfect condition - but almost certainly painted with the same pigments. To neaten the junction between frame and painting at the lower edge, and perhaps also to suggest the light catching the rim, Van Eyck brought the frame paint slightly over onto the painted surface, finishing it with a fine line of dark green.
The paint on the frame mouldings on the reverse is not original. In some areas where the later layers have flaked away, glimpses of the original decoration of black, possibly veined or spotted with red, are visible. This seems to have become damaged at a relatively early date and so the mouldings were covered with a new layer of white ground, much thicker than the original one and several layers of paint applied.
The lowest is black in imitation of the original, but over this is a sequence of cream and yellow brown layers. The tendency for the new ground to blister and flake away from the remains of the original paint and ground means that there are extensive opaque brown retouchings. Even these retouchings are very hard and insoluble in normal cleaning solvents, an indication of the age of the repainting of the frame. The removal of the brown layers would be laborious and very difficult. Given the likelihood that little original decoration has survived, it was decided that this repaint should be left in place.
(1) Transcription and translation from catalogue entry in L. Campbell, M. Falomir, J. Fletcher and L. Syson, Renaissance Faces. Van Eyck to Titian, Exh. Cat., National Gallery. London 2008, p. 180. The date is read as 17 June in A. Janssens de Bisthoven, M. Baes-Dondeyne and D. de Vos, Les Primitifs Flamands I. Corpus. Musée Communal des Beaux-Arts (Musée Groeninge) Bruges, 3rd edition, Brussels 1983, Vol I, p. 179.
Cleaning of the marbled decoration on the reverse involved no more than the removal of a thick coating of wax (applied probably at some time in the last century) and a great deal of embedded grey dirt. The surface underneath was found to be rich and glossy, and with no need for a new varnish. It is exceptionally well preserved.
Although the pattern does not resemble that of any known marble the illusion of polished stone is remarkable and has been achieved by a complex technique.
Over a white chalk ground, similar in appearance in the cross-section to that on the front (although the paint surface is notably less cracked) is a layer of black pigment, probably a carbon black as in the black background of the portrait.
Paint was then spattered over the black, first of all a porphyry colour mixed from red lake, vermilion, chalk (more translucent than the usual lead white) and a little black, as in the cross-section, followed by scattered spots of a greyish white and then finally the brilliant red of vermilion. FTIR analysis of the binding medium of the spatters indicates that it contains a protein, either glue, egg or possibly casein.
This is not surprising as oil paint is too thick and sticky for effective spattering. When Italian painters such as Mantegna executed spattered marbling using an egg medium craters and air bubbles from the foaming paint are usually visible under magnification (1). The absence of such craters here suggests that a glue medium may have been used. Between the porphyry-coloured and vermilion layers in the cross-section there is a translucent layer of unpigmented material which fluoresces strongly under ultraviolet. This intermediate application of varnish or oil medium contributes to the sense of depth in the layers of spattered paint.
Analysis of the smooth glossy surface coating indicates that is contains mainly heat-bodied linseed oil (the medium of the portrait itself) with a small amount of pine resin and other minor components, yet to be identified. It is slightly discoloured and clearly very old, having developed its own network of cracks and the ultraviolet image confirms that it is interrupted only in the areas where there are small scratches and flake losses, which show as black. Varnish recipes from the fifteenth century indicate that they were based on resins dissolved in drying oils rather than the volatile solvents of more modern varnishes. It seems likely, therefore, that this is Van Eyck's original surface coating.
(1) See J Dunkerton and R. Billinge, Beyond the Naked Eye, National Gallery, London 2005, pp. 10-13.
Text: Jill Dunkerton (National Gallery, London)
Your browser does not meet the minimum requirements to view this website. The browsers below are compatible. If you do not have one of these browsers, click on the icon to download the preferred browser.
