Accueil Articles par thèmes Etudes Critiques Old varnishes and their preservation

by Michel FAVRE-FELIX and Paul PFISTER   (translation by Alison CLARKE) - updated 07/12/2015


The pictorial role of old varnishes
and the principle of their preservation



 Sully Wing, room 52


Oil painting differs from all other artistic techniques because the pigments
are suspended in a translucent medium of oil and/or resin, complemented by the transparency of the final varnish.
It is known that varnish is used so that the oil colours are sufficiently and evenly saturated. Without it, several of these colours — and the dark tones in particular — would be rather faded and dull. As a result, varnish acts not only as a protective coating but also as a pictorial method.

As well as its role in saturation, in this article we will discuss several additional pictorial qualities of varnish that are rarely touched upon. We will tackle the fundamental question of the ageing of varnish, and examine the multiple roles played by old varnishes in the correct perception of a painting.

Above all, it will be emphasised that the yellow colour of varnish helps to restore
a balance between colour values the relationships of which had been altered with time.
This balance is vital to insure a feeling of depth, one of the fundamental qualities of oil paintings.

We will then discuss the important role that the thickness of varnish plays in transforming the material substance of a painting into an evocative image as intended by the artist.

If a varnish has been part of the design of an oil painting from the very beginning,
the picture can only ‘work’ if its current varnish has the same qualities as the original coating that was an integral part of its artistic conception.

Over the past twenty or thirty years, the fundamental misunderstanding of oil painting has led to mistakes that cannot be considered accidental, isolated or superficial.
We will analyse these misjudgements, both technical and artistic, by considering
the mistranslation of materials, their qualities and their requirements, as well as
of paintings viewed as a whole.




When preparing to study paintings hanging together, whether in a museum or temporary exhibition, it is first of all fundamental to consider the light by which they are being displayed.
In the Louvre, as in many other museums, almost all of the rooms are lit directly from above, although this is by no means the ideal lighting for paintings. This method was developed at the end of the eighteenth century, just as the first museums were starting to appear as collections open to the general public. As a result, this became the accepted ‘museum lighting’, despite the fact that it fails to display oil paintings
to their best advantage.

louvre-atelier-davidPreviously, paintings had always been exhibited ideally with the light coming from the side of, or sometimes behind, the spectator.
Examining historical practice, it becomes clear that until the 19th century the studios of painters — including the Impressionists — would have always been lit from the side by medium or large windows. Where these windows ran from floor to ceiling on one side of the studio, they were sometimes extended by a skylight that could be covered by a blind. However, the main source of the light would still have been from the side.
Artists would therefore have elaborated their paintings and settled the final adjustments in this sideways light.

In fact, concentrated lighting from the side reveals the rich transparencies of an oil painting; conversely, overhead lighting results in a bright, diffuse haze that interferes with the transmission of the picture to the eye of the viewer. This phenomenon can be observed, and its effects slightly mitigated, by putting up a hand to shade the eyes: the tones appear with greater precision, with the darker hues and coloured shades easier to pick out.


Yellow varnishes and the effect of depth


Entering via the Lefuel Staircase (Richelieu Wing, 2nd floor), we find ourselves in
the presence of two particularly informative paintings: Frans Snyders’ The Game Stall
(fig. 1), which is still covered with an old, yellowed varnish; and The Fish Market
- or The Fish Stall (fig. 2), which is a very old copy after Snyders, possibly a studio copy, that has been fully cleaned.
Although The Game Stall is not displayed to its best advantage, as a result of the overhead lighting, it still offers a sense of space and graduated depth with a wonderful progression of layers. In contrast, The Fish Market has been reduced to a flattened,
two-dimensional image: whether designed to appear in the foreground or background of
the painting, the figures all nevertheless crowd together in the foreground.






louvre-hermitage-snydersThis lack of depth should not be attributed to the fact that it is a contemporary copy. Its faithfulness to the original, now in the Hermitage Museum, is obvious, while copies of this kind were made using the same pictorial methods as those adopted by the original artist. The effect of space was obviously well understood, with the result that it can be difficult to distinguish between these old copies and the originals. The flattening, which is a common effect, should instead be ascribed to the removal of an old varnish.


We will now consider the effect produced by a yellow varnish on the sense of depth
in oil paintings.

Viewed logically, a yellow film of this kind has the effect of lessening the force of
the darker hues in a painting (because the blacks become browns) and diminishing
the power the whites (which turn yellow). The range of colour values from dark to light would therefore be reduced. As a consequence, the objects depicted should be ‘squashed’ and the image’s third dimension should be suppressed.
However, we have noticed the exact opposite: where the yellow varnish is missing,
the contrasts are indeed greater — but the image is flattened.

In fact, the argument that varnish will ‘squash’ the image relies on the simplistic hypothesis that the three-dimensional effect is more marked when the contrasts
are more intense. Yet, this is not the case when a painting is considered as a whole.
The illusion of depth does not rely on intense contrasts but on the perception of each object in its correct position within the fictional space of the painting.

In order to establish each of these elements in the right place, properly spaced,
the painter not only makes use of linear perspective but also regulates the contrast between the objects in accordance with the principles of aerial perspective. It is this adjustment of the contrast between objects, rather than the maximum contrast per se,
that determines the final depth and coherence of the whole.

This would not be a problem if this adjustment lasted as it has been achieved by
the artist. Unfortunately, with the passage of time, the oil colours beneath the varnish undergo a complex set of transformations as they age. Colours change, and pigments alter. As a result, it is wrong to claim that a restoration can restore the ‘original’ colours that the artist would have laid onto the canvas.
The different phenomena that affect hue are widely variable in nature [1]. However,
in addition to the chromatic changes, it is also important to consider changes in values. Here, it is clear that ageing works in a much less variable manner — as a general rule,
the contrasts increase.
The dark colours become murkier, even blackened, while the mid-range colours also fail to remain stable. In some cases, these latter might brighten [2] , but most often darken: in one way or another, they tend towards the extremes.
In general, the contrasts are always exaggerated to an extent that is dependent on the area and element under consideration.

This imbalance between the contrasts, which becomes apparent once the varnish is removed, acts to destroy the effect of depth as summarised below.

The greater an object’s contrast, the closer we perceive it to be. As a result, if the object’s original contrast is strengthened, it will tend to seem closer than its natural position in the painting. Similarly, if the contrast between an object and its background is heightened, it also seems to move further forward.


louvre-raphael-holy-familyIn addition, areas that have become murky black or lacklustre white tend to float like the jumbled pieces of a jigsaw, lacking any connection with the layered space of the painting. As these areas begin to float, and the figurative elements are forced out of their natural positions, the illusion of space disappears and our eyes register only the flatness of the canvas.

In The Fish Market, the strong contrast between the objects pushes them all into the foreground. We perceive the fishmonger’s face or the small fish at the back of the stall as being just as close as the seal, crab or fish on the floor in the foreground. If the yellow varnish were to be removed from The Game Stall, it is easy to see how the shaded areas would be transformed into black expanses. The stallholder would suddenly be pushed forward, seeming closer to us than the monkeys in the foreground.

The ageing of the varnish therefore plays a major role: it compensates for this dispersion and restores the value balance. Its yellow tint not only has a clear effect on the extremes of the painting, by toning down the strength of the blacks and whites, but also maintains the relative layering of the intermediate values. For excessively black or white areas, the varnish therefore works as a bridge, drawing these areas into the interior of the figurative space by once again linking them to the surrounding tones.

Eugène Delacroix noted the same effect of removing the varnish from old masters, and the bridging power of yellow varnish: "[…] paintings [with the varnish removed] seem somewhat metallic and monotone due to the uniformly dark appearance of the shaded areas […] The varnish’s yellow coloration, which develops over time and also affects the light areas, acts as a sort of bond between these lights and darks" (Letter to Dutilleux, 8th August 1858).
By re-establishing this bond, the yellowed varnish works to restore the correct sense of general space.

The restorer Sarah Walden has also described this phenomenon, giving the example of a painting undergoing cleaning. Initially covered in layers of browned varnish, the image is buried and lacks depth. Once the varnish is thinned, although still yellow, the three dimensions suddenly spring back to life. If the restorer carries on cleaning, however, hoping to enhance the depth still further, the opposite effect occurs and the image flattens once more (The Ravished Image, Weidenfeld and Nicolson, 1985, p. 137).

The sense of depth is one of the key facets—a fundamental characteristic—of paintings in oil and varnish. How can restorers fail to see how space is destroyed by the removal of varnish? How can they misconstrue the pictorial technique to the extent of destroying an artistic triumph? Is it the fault of the modern world that we no longer understand the ‘naïve’ illusion that these great artists strove towards in their art?

In room 38 (Richelieu Wing, 2nd floor), The Ray of Sunlight by Jacob van Ruisdael should act as a fabulous demonstration of the evocation of space and atmosphere. The composition of this huge, theatrical scene, lit by shafts of sunlight, leaves the viewer in no doubt as to the effect originally intended by the artist. However, the loss of the original varnish (and certainly also its glazing) leaves the painting feeling monotone and with a somewhat lunar light—in stark contrast to the painter’s conception.
If the conservators seem unable to perceive the loss of depth in this painting, this is probably because they have not had the opportunity to compare it with a similar landscape that has undergone a less thorough cleaning, such as the Ruisdael in the Kunsthaus Zürich (fig. 3)




Restoration requires points of reference: it is important that these are found in the best-preserved works, and not just in paintings from the restorer’s own museum collection.

In cases of over-cleaned paintings, an alternative solution can be to add a layer of lightly tinted varnish (as reversible as the other retouches and duly reported in the painting’s conservation file). This method, recommended by Paul Pfister, is already in use in various museums but, to date, the Louvre has refused to countenance it. In other instances, a traditional varnish can be applied thickly enough so that its ageing should be perceptible within several decades.


Progressive perception


The confusion of relative contrasts, which is heightened by the lack of an old varnish, not only impairs the rendering of space but also has a further consequence. The painter will not have concentrated solely on ensuring that the figures and objects in the painting are correctly positioned in the image’s fictional space. In addition, he will also have ensured that these various elements are ranked in their ‘order of appearance’ upon the scene.

The fact that viewing is not an instantaneous act is normally completely ignored when it comes to the perception of paintings. While painting itself is a static art, the reception of a painting has to take place in real time: neither eyes nor brain are capable of absorbing all of the components of a work all at once. This is, in practice, not a problem. Indeed,
it becomes a positive advantage in the hands of a painter who can work to develop his expression in the temporal dimension. He can lead the viewer on a step-by-step journey of discovery through multiple aspects of his creation.
‘Painting is silent poetry’, as Simonides said and Leonardo popularised. Unfolding like poetry, line by line, a painting works by taking the viewer on a journey laid out by the artist.

When we look at a well-preserved painting, we share this experience. The artist initially offers us a first impression of the painting’s main forms, striking groupings, foremost figures and key colour arrangements. However, he has also built in paths to other aspects that are not so obvious at first glance. Little by little, we begin to pick out more complex relationships such as colour tones that we had previously overlooked, or figures or objects emerging into our awareness. The whole painting begins to come together to form an entire world.
Ultimately, the artist leads the eye of the viewer on a voyage of discovery into the heart of the painting, tapping into a unique rhythm that allows us to escape the continual passage of ordinary time.

Where the state of the painting still permits it, this experience of progressive revelation goes hand in hand with our equally progressive immersion in its depth of space. This is not a coincidence: the two effects of depth and temporal discovery are both woven into the carefully adjusted contrasts throughout the painting. For this reason, the inconsistencies that appear when the old varnish is missing also affect the viewer’s temporal journey. We have often noticed this: it is hard to know where to look first when faced with a confused assortment of objects, all fighting for our attention. It seems impossible either to begin looking, or to continue to do so. Our gaze is assaulted by the minutiae of details, the 'readability' of which the restorer may have endeavoured to enhance but which in fact ensure that all of the elements of the painting are presented to us at once. The secondary elements no longer function as enhancements, as the painter had intended, but instead add their voices into a cacophony.

Our observations thus far can be summarised as follows: an old painting and an old varnish form a coherent unit or system where the various ageing processes balance and compensate for each other. Is this really so surprising? Should we not, instead, suspect that something has gone awry when faced with an old master that has been sprayed with a colourless, synthetic varnish?


Varnish thickness, transparency and the coating of brushstrokes


The purpose of varnish is not just to establish depth from the very beginning, and later to ‘restore’ this effect of depth by yellowing and compensating for the growing discordance of values over time. In addition, the thickness of the varnish allows it
to play a decisive optical role that cannot be replicated by a thin varnish.
Painters from Van Eyck until the 19th century used a fairly thick, substantial varnish
to fulfil three fundamental aesthetic functions: coating the brushstrokes, creating
a transparent layer that gives light access to the deepest layers and forming an intangible boundary between the natural world and the ideal world of the painting.


Transparency and the coating of brushstrokes


By coating the brushstrokes, the thickness of traditional varnishes gives a sense of elegance to the painting and envelops its materials in mystery. In this way, painters aimed to smooth their workaday brushstrokes close to invisibility. An illuminating comparison can be made between two Vigée Le Brun portraits in room 52 (Sully Wing, 2nd floor).





In her portrait (fig. 4), which had its varnish carefully thinned a long time ago,
Madame Rousseau’s skin still features a tender bloom because of the blending effect of the old varnish on the brushstrokes.

The painting of Madame Molé-Raymond (fig. 5), meanwhile, was completely stripped of its varnish (‘heavily thinned’) in 1994. It now features only a thin, colourless finish, which no longer conveys any sense of the subject’s skin. Instead, it resembles a palette daubed with brushstrokes: the marks of the paintbrush, such as the streaks on her cheeks, stand out vividly, and everywhere the eye is assaulted by accumulations of paint. The picture seems to have been painted in acrylic rather than oils, despite the fact that these two paintings by the same artist are both on wooden panels and feature the same artistic techniques. It is simply that the old varnish is missing from one painting and not the other.


louvre-van-dyckIt is a mistake to believe that, because of his habitual use of vigorous brushstrokes, Frans Hals would have wanted these marks to have been rendered quite as obvious as cleaning is doing today. Hals varnished his paintings and the evidence points to the use of a very substantial oil varnish. The audacious style
of his brushwork was not diminished by the varnish — as is nowadays believed — but was instead enriched by the softening effect. All of the aesthetic sophistication of the seventeenth century is bound up in this union between vigour and illusionism. The extremely well-preserved portrait of The Gypsy Girl acts as a fine example of this vigorous yet softened style (room 28, Richelieu Wing, 2nd floor).

It is worth celebrating the fact that such a significant number of the Louvre’s Dutch paintings still feature this style thanks to the efforts of Jacques Foucart, their conservator until 2005. Further examples include the Rembrandts (from Bathsheba [3] to the Carcass of Beef) and Van Dyck’s portrait of Charles I, which display the effects of the varnish thinning, carried out in the 1950s. Particularly laudable is the care taken to preserve the beautiful appearance of a rare female nude by Van Dyck, Venus at the Forge of Vulcan (room 26, Richelieu Wing, 2nd floor).


A subtle boundary


This problem merits further discussion: paintings that have been completely stripped of their varnish and then covered again with a thin, colourless layer convey an overt sense of materiality. The physical painting — a flat canvas covered with paint — takes precedence over the palpable representation of figures and space. Cesare Brandi called this ‘the material babble’, which acts to shout down the image’s mysterious presence or ‘epiphany’.
According to Alberti, painting was originally intended as an open window through which the historia — the convincing and harmonious representation of a subject — can be viewed. Traditional varnish acts as a subtle but effective boundary through which the subject becomes the representation of an ideal world. The removal of this boundary reduces the painting to its purely material state.

When seeking to explain this phenomenon, it is worth reiterating that the coloured layer ‘discovered’ during deep cleaning is very far from having remained intact.
As a straightforward example, let us consider a painting classically executed in linseed oil and pigment. As the oil dries — undergoes gradual polymerisation — a linoxyn-dense epidermis forms on the surface of the colour. This oily exudate, which can sometimes be slightly opaque due to traces of calcium, can also itself turn yellowish. The varnish, which is layered on top, acts to seal this skin.

However, this original exudate rarely survives the repeated application of solvents.
Not only are our modern varnishes too thin to coat the brushstrokes, as the original varnishes would have done, they also fail to compensate for the loss of this linoxyn layer.

A rare example of a painting that has retained the majority of its original structure is Poussin’s 1657/58 Santa Francesca Romana (room 14, Richelieu Wing, 2nd floor) (fig. 6). From what we know of the painting’s history, it appears that it was not completely stripped of its varnish until it was cleaned in 1999, when the varnishes were removed but the top skin of the painting was not entirely eliminated [4].




The delicacy of the painting’s colours, such as the reddish glazes that still tint the angel’s green robes, or the soft blue of the drapery of the prone figure, should be compared with the current colours of Poussin’s other works in the same room. The crudity of these later paintings is a symptom of the flayed surfaces on which the colours stand out vividly and incongruously.
When approaching a Poussin, restorers should remember that the painter began by making miniature models out of wax. These were dressed and placed on a board inside a box to create a little theatrical scene in relief. The aim should therefore be to recreate this feeling of looking at a scene inside a theatre, with all its attendant mystery and illusion, while at the same time maintaining the distance fitting for a member of the audience.


The evolution of French museums


Although in theory French museums are committed to the doctrine of varnish thinning as formulated in the 1950s, over the years they have leaned towards the practice of removal known as ‘advanced thinning’. Proof of this is provided by Domenichino’s
The Virgin and Child with Saint Francis, a small painting on copper that was rediscovered in June 2004. The official label next to the work tells us that the ‘varnish was thinned’ during its restoration in 2006. However, the Louvre’s curator agreed that the varnish
had been completely removed.
It is absolutely vital for these constant euphemisms to be abandoned: without honest language, any dialogue between curator and restorer is doomed to failure from the start

The cleanings at the Louvre over the last two decades were not carried out for reasons of preservation, as the paintings were in no danger. Instead, they were carried out for aesthetic purposes. One might therefore expect some level of aesthetic justification, whether in terms of the creation of the work (what materials were used?) or in terms of the artist’s expression (what effect was intended?). However, instead of such arguments, the only motivations to be found in the dossiers are to get rid of the yellow tone of an old varnish (termed ‘oxidised’) and, therefore, to achieve ‘readability’ for paintings.


The illusion of readability


‘Readability’ is a deceptive notion, as a picture is not designed to be read [5].

Nevertheless, if the word is to be understood in a certain sense — as ‘highlighting artistic expression’, for example — it is important to note that paintings restored in this fashion have not been made more ‘readable’, but in fact less so.

louvre-huysmans-1For example, in Herds in a Hilly Landscape by Cornelis Huysmans (Lefuel Staircase), all of the vegetation has darkened over time. The sky has remained light, and has doubtless been over-cleaned. The removal of the varnish has not revealed any additional elements; it has done nothing but exaggerate the brutal disconnect between two overly contrasting areas.

Canvases prepared with red-brown grounds pose even more frequent problems.
As they age, these coloured grounds end up becoming visible beneath the colours and stifling the mid-hues of the painting by erasing invading its tones. It is absurd to believe that removing the varnish would lead to an improved ‘readability’ for this — extremely common — kind of painting (as in the example below).



With the varnish gone, the painting simply appears to be made up of dark, monotone areas, from which emerge disassociated elements in colours that have often been attacked by solvent leaching.

In the examples that we have seen, the removal of the old varnishes simply makes the artist’s original message unreadable. The gradual discovery intended by the painter becomes a confused tangle, with any perception of depth reduced to a sensation of flatness. Instead of bringing readability to the tangible interpretation of subjects such as the fresh skin of the sitter, cleaning instead forces us to ‘read’ brushstrokes and paint marks.


The obsession with yellowing


The current ideas on restoration can be boiled down to an obsessive struggle against
the yellowing of varnishes. It is constantly said, in the name of science, that varnishes have ‘oxidised’ — as if this is some kind of sickness. Oxidation is not an illness, however: it is no more than a natural, evolutionary process that also affects the oil in the painting to much the same extent. The changes and deterioration in the colours themselves are, as we have seen above, pathological, complex and uneven, and cannot be anticipated by artists. On the other hand, varnish is a simple material whose transformation would have been familiar to artists. The yellowing process was regarded as obvious, natural and unsurprising.

Even a certain level of yellowing has now come to be perceived as an unacceptable fault, although this had previously been seen, with reference to the pioneering cleaning doctrine developed by René Huyghe at the Louvre in the 1950s, as attractive and important for the perception of the work. For example, Huyghe warned against the deep cleaning of the seventeenth-century paintings, arguing that "the blues of Poussin, Claude, Le Sueur and others have become somewhat aggressively glaring over time and so stand out from the other colours, upsetting their original balance".


louvre-lesueurThe Louvre’s seventeenth-century French paintings demonstrate the wholesale abandonment of these principles. Notable is the series of 21 paintings in the St Bruno cycle, painted by Le Sueur between 1645 and 1648 and stripped of their varnish between 1981 and 1993 (room 24, Sully Wing, 2nd floor).
These pictures are the perfect example of the imbalance, such as the striking discord between whites and blues, that Huyghe wanted to avoid. It took only one generation, over the course of the 1980s, for Huyghe's careful argument on the limits of cleaning to go up in smoke.

The large altarpieces in room 29 (Sully Wing, 2nd floor) have undergone such serious thinning of their varnish, both recently and long ago, that these oil paintings now look more like frescoes. Examples include Laurent de La Hyre’s Christ Appearing to the Three Marys and Noël Coypel’s St James the Elder. This is an aberration because, although the seventeenth century valued clarity in artistic expression, painters could not or did not want to surpass the methods inherent in the oils technique.


louvre-vernis-embusIn some cases, paintings that have been weakened by the application of solvents become porous. In Sébastien Bourdon’s
The Deposition from the Cross, for example, the thin, synthetic varnish has been entirely absorbed in these parched sections to create matte, murky areas.


This obsession with yellowing equally applies to revarnishing, where a sprayer is used to apply an ultra-thin layer of synthetic resins. In this field, as with cleaning, the cautious reputation enjoyed by French museums is no longer at all justified: for some time now, the use of acrylic resins and particularly ketones (polycyclohexanones) has been widespread. The presence of these chemicals is fairly easy to detect on several canvases in the Louvre — including Poussin’s masterpieces Spring, Summer and Autumn — due to their lifeless, plastic appearance.

Because the molecules in these resins are too large, they cannot be used to achieve the correct saturation of colours. These were at first championed as they would not yellow and would be easy to remove. Actually, they might become even more difficult to remove than natural mastic varnishes, sometimes requiring the use of strong, penetrating solvents.

The most egregious example of modern times is the Van Eyck brothers’ masterwork known as the Ghent Altarpiece. This is in urgent need of restoration, currently ongoing, to remove the synthetic varnish applied in 1950's and later to the panels of this treasure of world heritage. These ketone resins are now threatening the painting, and get harder and harder to dissolve the older they become.


A note on colour perception


We still need to consider the main accusation constantly levelled against the yellow tone of varnish: that it denatures the cool colours and, in particular, turns blues green. This prejudice withers in the face of evidence, however, as can be seen from paintings such as Vouet’s Polyhymnia, Muse of Eloquence (Fig.7) (room 30, Sully Wing, 2nd floor). Here, the even blond varnish acts as no impediment to the full enjoyment of the picture’s coloured tones, which range from ultramarine to pinks, ochres and off-white.




The same experience holds true in front of Chardin’s masterpieces The Attributes of Civilian Music and The Attributes of Military Music (Fig. 8) (room 49, Sully Wing, 2nd floor). These two paintings are associated with a famous remark by Diderot, who, on coming to admire the newly finished works at the Salon of 1767, warned visitors not to be disconcerted by the "somewhat raw and harsh glare of the fresh colours" that would be "extinguished" by time. He advised the viewer to "revisit these works once Time has painted them". When the two paintings entered the Louvre from a private collection, their varnish had been preserved in an even, transparent state without additional thinning. Its golden tone helps to highlight all of the subtle differences between the coloured tones: the blue has not turned green, while the pinks, reds, brownish reds, vermillion and carmine have all kept their hues without any confusion.





A colour measurement instrument would obviously reveal the presence of yellowing,
but here the human eye has an advantage over technology as the yellow tone can be cancelled out from the overall colour perception. The brain scales down the appearance of yellowing to allow the nuances between colours to spring quickly back into view.
This adjustment of the eye takes place even more quickly and instinctively when a painting is lit from the side with a warmish natural light. This is not surprising, given that paintings made in previous centuries were designed to be viewed under exactly these lighting conditions.


Authenticity, original and historical varnishes


Now that this journey is complete, we can suggest a general course of action:


For all paintings that were originally designed with a varnish but have lost their original coating, an old varnish, featuring natural resins and the attendant yellowing, should be considered authentic and should be preserved.
This rounds off a pictorial structure that, without such a varnish, no longer works, or even works against the artistic intentions of the artist.
At the very least, the thickness of the varnish must match the characteristics of the original varnish, while it must also be sufficiently transparent to function on an optical level.
‘Cleaning’ programmes have been rolled out without proof of their legitimacy.
In contrast, however, the removal of an old varnish must not be considered without proof that it is totally incompatible both with the varnish applied by the artist and with the pictorial structure of the work.


The public believe that museums study the materials used by artists and endeavour to reconstruct the type of finish that would have originally been chosen. This is not the case. How would the public react if they were to be told that restoration instead strives to impose a modern, industrial finish upon old paintings?
However, museums are reluctant to supply full explanations of their restoration work, let alone exact information. For example, when interviewed, the conservator responsible for the controversial restoration of The Marriage at Cana was at pains to reassure the public that Veronese’s masterpiece had benefitted from a revarnish "using natural resins, as it would previously have been varnished". This was entirely untrue, as at the end of the restoration six months prior the painting had been sprayed with three layers of a synthetic ketone varnish.


louvre-copal-colophaneAs for the study of varnishes used by artists in the past, this has long been ignored by museums and labelled useless or hypothetical. However, this position can no longer be justified: in addition to the wide range of knowledge already available from historical sources such as recipes or written testimonies, new types of scientific analysis are now revealing the composition and characteristics of varnishes, such as the resins they contain.


In the case of Orazio Gentileschi (1562-1639), several years ago at the Getty Museum an original varnish was identified on his Lot and his daughters executed in around 1622. It consisted of colophony (or rosin, resin extracted from the pine) and Manila copal (sometimes bracketed with other closely related resins under the general term ‘amber’) [6]. This final, golden layer also corresponds to the warm coloured "amber varnish from Venice" that Turquet de Mayerne, a contemporary of Gentileschi, describes him using [see Nuances 36-37 : Les vernis originaux dans la peinture italienne -14eme-17eme ].

Given this crucial technical information, what can we deduce about the two Gentileschi paintings owned by the Louvre, which were painted between 1624 and 1628 and are currently on display in the Grande Galerie?



The Rest on the Flight to Egypt (fig. 9) still features a faintly golden natural varnish. Although unlikely to be original, this varnish is very similar to the original coating identified by the Getty. Its presence is therefore fully justified because of the aesthetic authenticity that it brings to the work, and specific arguments in technical art history support its conservation.


louvre-gentileschi-c In contrast, extremely powerful solvents were used to strip the varnish from Gentileschi’s other painting, Public Happiness Triumphs over Danger, in 1978. The work was then sprayed with a synthetic varnish. Neither historical nor aesthetic justifications can be found to defend this treatment, or its appalling results. Nevertheless, this type of systematic, ungrounded ‘restoration’ has been imposed on almost all of the Italian paintings of this period, including those by Guercino, Guido Reni, Pietro da Cortona, Albano, Carracci — who share, along with Gentileschi, the characteristics of Counter-Reformation art — and of others Italian painters of the fifteenth and sixteenth centuries, in spite of their various aesthetic conception.


The study of original varnishes, thanks to newly discovered analysis techniques, is just getting off the ground. The discovery of many of these original varnishes, even just as traces, still in situ on old master painters has already — finally — supplied incontrovertible evidence that calls into question museum practice and the received ideas of modern restoration.


louvre-caravaggio-1To contemporary tastes used to bright,
pure colours, Caravaggio’s masterpiece
The Death of the Virgin
at the Louvre can seem somewhat dampened by its current ‘uniformly yellowed’ varnish. However, this corresponds directly to the type of varnish used by Caravaggio — the thick, naturally coloured vernice grosso. Traces of this have been found on several of his works and a direct account of its use by him was recorded by a contemporary [7].

Likewise, it is very likely that Rembrandt used a mastic resin varnish — possibly mixed with a resinous balsam, Venitian turpentine — this can be used as a reference point for judging the conservation of such paintings [8].
In the Louvre collections, there are still Rembrandt's paintings with mastic resin varnishes (or dammar, which is similar) that were perfectly thinned in the 1950s. These varnishes would fairly correspond to the artist’s practice and there is no reason why they should be touched.


All the same, not all old varnishes are in a satisfactory condition, and treatment is required for their preservation. A lack of transparency is notable on the lower half of The Plague at Jaffa by Gros, for example. A greyish haze is also present on Georges de La Tour’s St Sebastian Tended by St Irene.

The true aim of a conservation project should be to bring transparency to a varnish, and to treat it without resorting to removal. There are solutions that keep this aim firmly in mind — as opposed to the unjustified cleaning programmes put into practice to date.



Michel Favre-Félix and Paul Pfister

Translated by Alison Clarke


for the translation © Alison Clarke




[1] In terms of blue alone, smalt can either remain cobalt blue or transform into greyish-brown; lapis lazuli can either whiten or maintain its dark blue radiance; and azurite has a general tendency to darken.

[2] Either due to the discolouration of their pigment, or because the bright sections have often been excessively scoured duringold cleanings that have wiped away the painter’s careful graduations in tone.

[3] Since we first published this study in French (2007), the Louvre has planned a restoration of Rembrandt’s Bathsheba. At the very beginning of the cleaning tests (March 2014) we have been given the chance by the Head of the painting department, during a meeting attended by the curators, the restorer and a member of the C2RMF, to present our arguments defending the pictorial role of the old varnish still present on this painting and the aesthetic adverse consequences that might come from any pronounced cleaning.

Eventually, the Louvre’s intervention was restricted to the lightest surface cleaning, preserving - in our opinion - both the overall unity of the painting and the sensitive perception of this outstanding nude. The result can be appreciated on the Bathsheba, back in the museum’s room since June 2015.

[4] The haste of the 1999 restoration must still be condemned, however, as there is every reason to believe that the original varnish was still in place. It was removed before the results of the analyses showed — too late — evidence of ingredients typically in use in Poussin’s era (mastic resin, colophony and sandarac).

[5] On the French term lisibilité — which should be translated as ‘readability’, rather than simply ‘legibility’, in order to encompass the additional dimension of ‘comprehension’ — see Michel Favre-Félix, ‘Ambiguïtés, erreurs et conséquences : Rendre l’œuvre lisible’, Revue CeROArt [Conservation, exhibition, Restoration of works of Art.], 3-2009. URL: . See also Salvador Muñoz Viñas, Contemporary Theory of Conservation, pp. 99-101, Elsevier Butterworth-Heinemann, 2005.

[6] see Mark Leonard, Narayan Khandekar and Dawson W.Carr,'Amber varnish' and Orazio Gentileshi's 'Lot and his daughters', in The Burlington Magazine, Vol CXLIII, number 1174, January 2001, pp. 4-10.

[7] For the traces found on Carravagio’s works, see Roberta Roani, in Riflessioni sulle “Regole per comprare, collocare e conservare le pitture” di Guilio Mancini, p. 31, Edifir Edizioni, Firenze, 2005.

[8] Samuel Dirksz van Hoogstraten, who was one of Rembrandt’s pupils, notes in his book Inleyding tot de Hooge Schoole der Schilderkonst (Introduction to the Academy of Painting — Rotterdam 1678) on the varnish in current use in his circle: “Our varnish, consisting of turpentine [balsam from pinaceae, usually Venitian turpentine], spirit of turpentine, and pulverised mastic dissolved, is sufficiently convenient for our works” (Inleyding …, p.223). It would have been even more coloured than pure mastic, given the original orange hue of the Venitian turpentine, which both increase over time.




Achat publications

Mentions légales