Artwatch UK

Posts tagged “Velazquez

A National Gallery restoration that repudiates earlier National Gallery restorations

When major museums acquire major pictures, they invariably take additional technical and artistic possession of them through restorations. By transforming pictures’ appearances, museum staffs lay claim to an exclusive up-to-the-minute knowledge of a picture’s material and artistic traits that renders all earlier studies obsolete and activates use of the possessive “our” – as in “our Duccio” or “our Artemisia Gentileschi”. For much-criticised museums like the National Gallery and the Metropolitan Museum of Art, New York, the introduction of a well-preserved picture within a collection risks spotlighting in-house restoration damage – as might well have happened, for example, had the Met exhibited its newly acquired, fabulously well-preserved Velazquez portrait Juan de Pareja and its Perino del Vaga The Holy Family with the Infant St John the Baptist before restoring them. Today, the National Gallery seeks to counter long-standing criticisms by allowing its restorers to present their own interventions and purposes through broadcast social media. In a press release of 2 August 2019, the gallery’s Director of Collections and Research, Caroline Campbell, said of a restored panel painting:

“The National Gallery is one of just a handful of institutions across the world that is able to carry out painting conservation of this complexity. As this work has been carried out behind closed doors, this display is an opportunity to share this expertise with the public and also to celebrate our conservation skills, in a similar way to how we shared the conservation of our Artemisia Gentileschi self-portrait via a series of films.”

Such hubristic public relations manoeuvres are risky. As Michel Favre-Felix, painter and President of ARIPA (Association Internationale pour le Respect de l’Intégrité du Patrimoine Artistique), demonstrates below, restoration errors are still to be encountered among the nation’s pictures and the restorers’ own explanations leave conspicuously unaddressed questions. [M.D.]

Above, Fig. 1: Left, the National Gallery’s Artemisia Gentileschi Self-portrait as Saint Catherine of Alexandria, as presented by the Paris-based auctioneer Christophe Joron-Derem for the 19 December 2017 auction; right, as subsequently restored by the National Gallery.

Michel Favre-Felix writes:

Artemisia Gentileschi’s Self-portrait as Saint Catherine of Alexandria, which was acquired two years ago for £3.6 million (a record for the artist), has already become a new iconic painting of the National Gallery. To the appeal of a self-portrait by the most celebrated female painter of the 17th century, the picture (above, Fig. 1) adds a telling symbolic aura. Commentators have not failed to underline that this martyred saint Catherine, holding the instrument of her ordeal – the miraculously broken spiked wheel – persevering in her faith in the midst of persecution and rewarded with eternal salvation, mirrors the shattering life-story of Artemisia herself, young victim of a rape, maintaining her testimony under torture and finally triumphing in her female artist career. This emblematic portrait is the central feature in the present major exhibition of her work – the first ever in UK – “Artemisia”, National Gallery, London WC2, until 24 January 2021.

Just two weeks after announcing the purchase, in July 2018, the National Gallery began posting on YouTube the first of what became a long series of videos of the restoration in progress (see the list at the end of this article). No fewer than four of them deal with the picture’s cleaning – the need for it (which will be discussed below); the expected effects; its progress and its results.

Such a pedagogic/celebratory (to re-use Caroline Campbell’s expression) public programme is unprecedented. Hitherto, if the Gallery decided to communicate an account of one of its restorations it usually appeared in its scholarly Technical Bulletin, with a strong emphasis on the scientific analysis of the picture’s material structure and a minimal part, if any, given to the hands-on cleaning process itself. The set of YouTube videos exactly reverses that relationship.

As in political discourses, vocabulary plays a key role and carries far-reaching meanings. Old traditional terms might surface, as when the curator, Letizia Treves, observes rather innocently that ‘the picture is quite dirty’, expressing her expectations from its forthcoming cleaning (in Who was… 4:18). ‘Dirty’ is the customary loaded codeword used to justify a total varnish removal. It leaves no room for investigations or nuances: ‘dirt’ cannot reasonably be even partially kept on a painting; it must be entirely wiped out. (See Fig. 1, above, left, for the pre-restoration condition.)

Larry Keith’s expressions are purposefully different. He not only restrains himself from using the loaded and derogatory, non-scientific term ‘dirt’ to describe what is in reality a coat of old varnishes, but he takes care to amend the ambiguous twin-word of ‘cleaning’, by changing its sense, at the start of his talk (in Cleaning… 0:25): ‘Cleaning meaning the… [short pause] …reduction of the old discoloured degraded varnishes’ (“reduction” being the operative word). This singular short pause in his otherwise fluent and dynamic speech is eloquent.

A closer look shows that this change of definition has matured over several years. The cleaning of the Virgin of the Rocks, in 2009/2010, was already presented as a ‘reduction’ [Endnote 1], although this peculiar aspect went rather unnoticed at the time [2]. Earlier, when commenting on the restoration of Guido Reni’s The Adoration of the Shepherds in 2007, Larry Keith mentioned that to clean might be ‘to remove or reduce the old discoloured varnishes’ [3]. If cleaning now means a reduction rather than an elimination, this new position has generated a number of unaddressed questions.

1) First, what does this policy change reveal about the systematic total cleanings made in the past? What happened to the previous certainties on which the gallery’s conservation policy was grounded and which had served to authorise its restorations? Since the post-Second World War ‘Great Picture Cleaning Controversy’, the gallery’s conservation department maintained, against its national and international critics, that a complete removal of varnish was the only way to establish the true, objective, unfalsified state of a painting, and to recover as closely as possible its original appearance as created by the artist. This was not held to be one option among others. It was the inescapable and inevitable conclusion of methodical reasoning itself. The leading proponent of this policy, the de facto chief restorer, Helmut Ruhemann, went so far as to list nine ‘main Arguments against Part Cleaning’ in a crucial chapter of his 1968 book The Cleaning of Paintings (pp. 214-217), which had set the Gallery’s official institutional methodology for more than half a century – and still exerts an influence.

Part-cleaning was not only ruled out in theory but was held to be both unfeasible and deceiving in practice. Ruhemann’s strongest and most persuasive arguments were technical ones. Using the authority of the practitioner, he asserted that a reduction of the varnishes regularly produces an uneven result leaving disruptive and disfiguring ‘patches’ scattered all over the paint. He claimed that a half-way cleaning was arbitrary and inevitably imprecise, the restorer being ‘condemned to groping in the dark’. He stressed that, if there was some old varnish left, it would be impossible to suppress all the faulty and distorting old retouching that might lie underneath. Moreover, he added that the new retouches would never correctly match the still imperfectly cleaned paint.

This argumentation, unchallenged for decades, happens to have been refuted by Larry Keith’s recent practical demonstration. Although Keith used traditional means (no revolution in tools or solvents or monitoring is used in the Gallery) his ‘reduction’ did not generate the Ruhemann-predicted failures: it neither failed to suppress the old retouches nor to avoid uneven ‘patches’ – nor even failed to achieve perfectly matching indiscernible new retouches.

2) What is the reason for adopting partial cleaning today? On the one hand, in hindsight, we can see that the previous policy of total cleaning was based on spurious arguments but, on the other, it is striking that no revised or new justification is provided in support of the present policy.

Why is it now considered to be appropriate, required – or even essential – to keep a part of the old so-called ‘degraded and discoloured’ varnishes on this painting? Is it to serve as a guarantee for the safety of the paint and possible original glazes underneath when subjected to the cleaning with solvents? Does this last layer of old varnish bear a meaningful aesthetic and/or historic value that ought to be preserved? Does the remnant of the surface coating constitute part of the artistic authenticity of the work of art? Keith provides no indication at all. A full range of arguments in favour of part-way cleaning have been put forward elsewhere since the 1950s by connoisseurs, critics and art historians but Keith refers to none.

In reality these questions concern a majority of works because this portrait is not at all an exceptional case. It was, at the time of its acquisition, in a ‘standard’ condition that is common to so many paintings from past centuries that have been subjected to restorations: from the Gallery’s report it turns out that its surface bore the usual old retouching, and its canvas, already relined as was customary in the past, had since suffered a small tear and will be relined anew.

Acknowledging the ‘reduction’ of the varnishes as the best possible care for this painting implies/concedes that it should have been similarly prescribed and applied successfully to so many comparable paintings, affected by the same usual damages, but which were radically cleaned at the National Gallery.

3) Larry Keith never explains in his videos why he chooses to thin rather than to remove the coat of ‘degraded’ varnish, as was the rule before. He simply strives to show why the old varnish needed a treatment and to demonstrate that he achieved ‘key improvements’ on the test areas where it has been reduced.

About the state of the varnish he draws a distinction, not without reason, between two effects: ‘these old varnishes when they degrade, they turn yellow and they turn foggy…’

That is true in a general way, but it is precisely from there that reflections should begin, because while the first is the natural, predictable, regular evolution of traditional materials, the latter is an unfortunate degradation that preventive care could avoid.

Above, Fig. 2: Screen capture from the video “Cleaning…” – See the full linked-list of videos below.

On this first issue, that of yellowing, the explanations are especially puzzling:

[in Cleaning… 1:27] “You see that where the varnishes have been reduced, the overall tonalities of the picture are much less yellow. The fingers [on the left] are emerging rather pink, instead of this kind of yellow colour [on the right] and I am sure that will become more evident as we move across the picture…”

These comments are puzzling because they hardly fit with what is shown. The old varnish did not turn the skin tonalities markedly and disturbingly yellow (compare the back of the hand on the right with the old varnish on, to the ‘reduced’ one on the fingers on the left at Fig. 2 above), and it is indeed anything but ‘evident’ that it distorted the perception of the colours. It may be recalled that in December 2017, during the presentation of the painting before its auction in Paris, the expert Eric Turquin praised the ‘subtle pinks’ – in his own words – he had no trouble distinguishing in the flesh tones of the portrait with the old varnish on [4].

Above, Fig. 3: Photograph (detail) from the Hyperallergic site, 12 July 2018, showing the “Artemisia” exhibition curator, Letizia Treves, facing the self-portrait before cleaning began. Although top lighting caused a pale reflection on the canvas, lightening the dark tones, it can be seen here that Artemisia’s flesh tones are not so much yellowish as close in their pinkness to the curator’s own natural colouring.

Above, Fig. 4: The restorer Larry Keith, examining the painting before cleaning began, as shown on BBC News 6 July 2018.

The above photos published in July 2018, at the very start of the intervention, in which spectators are present confirm that the variety of colours in the painting was clearly perceptible: the shades of pink of the face, the cream tone of the headscarf, the Naples yellow of the palm leaf or the ochre of the wood read easily and naturally. One can observe that there was no oppressing monochrome veil distorting the shades of the portrait, which were quite close to the natural skin tones of the viewers, as the photographs testify (Figs. 3 and 4).

Surprisingly, if not tendentiously, Keith even evokes an ‘accumulation of varnishes’, which he ventures would result from ‘many restorations that have probably occurred’ in the past (in Cleaning… 4:35). ‘Many’ is merely hypothetical since the history of this painting is totally unknown between the years of its creation, circa 1615-1617, and the 1940s when it resurfaced, only to be quietly kept in a French family (Pes, J. 2018).

Looking at the photographs of the initial state, it is difficult to deduce a superimposition of many added layers. Fortunately, this will be checked since Keith has announced that ‘minuscule samples [will] help us understand the layers structure of the accumulation of varnishes’ (in Cleaning… 4:35). Fine. It will be of great interest for the public and the experts that the result of this investigation by the laboratory be disclosed: how many layers of old varnishes? To what total thickness? Until these results are established and cited the idea of an ‘accumulation’ of layers of varnish will remain a puzzling assumption.

4) Beside the issue of yellowing – that he admitted not to be ‘evident’ – Keith places a greater emphasize on the second, undisputable, aspect of the picture condition, that of the varnish getting foggy. This loss of its transparency is, by contrast, plainly documented.

Even during the presentation at the 2017 auction in Paris, while the subtlety of the colours was praised, the ‘dullness of the varnish’ was nonetheless underlined and attributed to the fact that the painting had remained in the same family for several generations.

The video illustrates the consequences of this phenomenon (in Cleaning… from 1:40):

“… where [the foggy varnishes] are over the darker tones, the darker tones become quite a bit lighter. You can see that here, with that sort of hazy presence. And whereas down here where I started reducing the old varnishes, you can see the darker colours are much darker and the range from light to dark is much enhanced. And I think this helps you understand how [Artemisia] has laid out the folds, and helps you understand what is in front of what.

“…I think the thing here [in the ‘reduction’ in progress] that is most significant and really very rewarding is to see now the range from light to dark, which [Artemisia] has used, and her modelling of forms, which gives this sculptural presence.”

Indeed, Artemisia’s artistic expression rests on the illusion of spatial depth and on the convincing impression of three-dimensional figures. And this pictorial achievement is only displayed when the half-tones, dark values and contrasts have their full effect, which requires a good transparency of the varnish final layer.

It is hence plainly justified to try to regain this fundamental quality. However, in the case of this painting, such faint cloudiness is a common and rather benign alteration caused by humidity (that is to say, by a lack of prevention from its keepers). Physically, this phenomenon results from the scattering of light – not exactly on the ‘varnish’s own kind of fine cracks’ as it is said rather simplistically in the video – but on a multitude of micro-fissures, much smaller than usual cracks, that have developed within the varnish film at a microscopic scale that is invisible to the naked eye.

Above, Fig. 5: Above, Fig. 5: detail of Artemisia’s arm, showing un-thinned (slightly dull) varnish on the right and thinned varnish on the left.

As can be seen on the video, the thinning of the varnish has cleared the cloudy effect and has thus enhanced saturation and contrasts [above, Fig. 5]. Yet, the cause/effect relationship is not that simple. The dissipation of the hazy opacity is the result of a specific physical process: it comes from the ‘closing’ of the micro-fissures, which is obtained through the momentary softening and swelling of the varnish film when suitable solvents are applied to it. Once the solvent has evaporated, the micro-fissures have closed and so, vanished. Since the ‘reduction’ was done with solvents, their penetration into the varnish film provoked the swelling/closing result. Thus, this was a linked side-effect and it would not have been necessary to thin the entire varnish layer for that to happen. For this kind of light haziness, a simple exposure of a varnish surface to an appropriate solvent, at much lower levels – i.e. ethanol in form of vapours – without any ‘reduction’, could have produced the same positive result (Pfister, P. 2011, Demuth, P. 2001): the saturation of colours; the in-depth setting of the figure; the sculptural modelling created by Artemisia, would all have been recovered.

Of course, when such a minimal treatment is chosen, the tonality of the varnish remains unchanged, since its thickness is undisturbed even as its transparency is regained.

Knowing this, we realize that there is confusion between the two results. In truth, a physical reduction was not essential to recover the range of values from light to dark and modelling of forms intended by the artist, which could have been achieved otherwise. Essentially, the thinning of the varnish was used principally and specifically to obtain the ‘much less yellow’ overall tone. This result is held – in Larry Keith’s account – to be such an obvious improvement as to require no further justification. Yet, it does – and we see below why it needs questioning.

5) The transparency is a basic undisputed requirement for this varnish (as for any other). But what is the justification for making it ‘much less yellow’?

When we leave aside our own era’s cultural preferences and consider the materials and varnishing practices that prevailed in the XVIIth century, we realize that the (disparaged) ‘old varnish’ found on this painting had the best chance of resembling the original finished appearance as made by Artemisia herself.

Above, Fig. 6: A section of de Mayerne’s text (Folio 151r) mentioning Artemisia Gentileschi and her varnish.

Throughout these supposedly informative and instructive videos it is striking that no reference is ever made to the kind of varnish that would have been used by the artist herself, or, even, to those that were common in her circle and time in Italy. This omission is hard to justify since relevant historical and technical references have survived and are accessible. For example, Turquet de Mayerne’s manuscript notebook (written between 1620 and 1646), which is the main historic testimony and source of information on the painting techniques of this period, contains a famous reference to an ‘amber varnish’ [5], ascribed to both Artemisia (active c. 1610-1653) and her father Orazio Gentileschi (active c. 1587 -1639) – see Fig. 6, above. De Mayerne specifies that this varnish had a strong reddish tone and was used by the instrument makers to varnish lutes [6].

It should be borne in mind that, at that time, in the absence of precise identification, the term ‘amber’ (otherwise called ‘c(h)arabe’) encompassed a group of resins that were close by their consistency, colour, workability and effect – and among which were chiefly the different semi-fossil resins that we now classify as copals, which range from semi-hard to hard and are easier to dissolve than true fossilized amber (Leonard et al. 2001, Holmes, M. 1999).

Furthermore, the expression amber varnish ‘coming from Venice, with which they varnish lutes’, added in the passage on Orazio (Folio 9v), most probably indicates a ready-made product. At that time in Italy many varnish formulations were no longer made in the artists workshops but prepared and sold by colours merchants. The painter Gian Battista Volpato quotes the ‘amber varnish’ as one of them [7]. De Mayerne states that a so-called ‘Oil of Amber from Venice’ (that is, a fat varnish made of ‘amber’ dissolved in a possibly larger proportion of oil), which he supposes to be the one used by Orazio, was sold in every Italian colour shop [8]. The main point is that these prepared varnishes formed a dry film that approached the legendary hardness of amber and had a similar golden-brassy colour. Some rosin (colophony) could be added, which was useful for improving the working properties of the mixtures (Leonard et al. 2001). Its marked orange hue would also increase the warm tonality of the whole – see Fig. 7 below.

Above, Fig. 7: Colophony (or rosin, resinous part remaining after the essential oil has been extracted from the balsam of Pinus maritima Lamb. by distillation.)

It is mentioned that Artemisia mixed her ‘amber’ varnish with oil and spread the blend as an intermediary layer upon the already dried parts of her work in progress, before continuing to paint (Folio 151r). This method, commonly called ‘oiling-out’, has three benefits for reworking: it brings back the initial saturation of the first colours that might have turned dull when drying; it enables a fluent application of the later colours; and, it promotes their physical adhesion to the ones beneath.

Concerning Orazio, de Mayerne notes that he used to add a drop of ‘amber’ varnish directly to his colours on his palette – especially to the ones of the flesh tones – in order to make them more ductile and quicker to dry (Folio 9v).

Was it also chosen as final varnish? The use of the same compound for mixing with colours, for intermediary ‘oiling-out’ and for final varnishing is indeed consistent with what is known of painters’ practices at the time. Examinations of paintings by Caravaggio (of whom Orazio was a disciple) have shown that remains of his final varnish – resin in oil – were similar to the ‘oiling-out’ layers found in his paint structure (Arciprete, B. 2004).On the same folio (151r) where de Mayern mentions Artemisia’s oiling-out method, he reports on another ‘charabe’ varnish, which can be used ‘for varnishing and for mixing on the palette with the colours’ [9].

Moreover, a discovery made by the Getty Conservation Institute in 2000 confirms that Orazio also adopted an ‘amber type’ varnish of his final coating. Found on one of his painting (Lot and his Daughters, J. Paul Getty Museum, Los Angeles) executed about 1622, this rare original varnish proved to be composed of Manilla copal and rosin, precisely (Leonard et al. 2001) [Figs. 7 and 8].

Above, Fig. 8: Manilla Copal (from Agathis dammara Lamb.)

This tangible historical/material evidence of what was an ‘amber-like’ formulation provides a precious testimony of its visual effect on the picture: it displayed a notable warm golden tone over the parts where it was still present (see Figs. 9 and 10 below). Given that Artemisia learnt to paint in her father’s studio, it is beyond doubt that Orazio would have shared both his materials and his practices with his daughter.

Above, Fig. 9: The effect of Orazio’s varnish on the sky of his Lot and his Daughters (in The Burlington Magazine’s article, Vol CXLII n°1174, p.5.)

Above, Fig. 10: Macro-photograph of the varnished sky. Note the bright blue colour of the paint that appears in some spots where the ‘amber-like’ varnish is missing (from the same Burlington Magazine article, p.9.)

In addition to those clues, there is the certainty that Artemisia’s varnish could only have been composed with resins among those of her time (the end of 16th/ first half of 17th centuries): sandarac; oleoresins balsams from the silver fir, the larch or the spruce; colophony; mastic; copals, with or without oil [10] (Figs. 7, 8, 11 and 12). Reconstructions of historical recipes with such ingredients, prepared and applied following traditional methods are converging to show that they provided a natural warm tone – a ‘golden glow’ – that moreover increased surprisingly quickly (Favre-Félix, M. 2017, Carlyle, L. 2005, CCI 1994) (Fig. 13).

Above, Fig. 11: Strasbourg turpentine (balsam – oleoresine – of the silver fir, Abies pectinata DC. – from Kremer Pigmente.)

Above, Fig. 12: Sandarac (from Tetraclinis articulata Mast.)

Above, Fig. 13:As an example, the reconstruction of an historical recipe, using one resin and one oleoresin – a type that became increasingly prevalent from the end of the 16th and throughout the 17th centuries – showed a notable increase of its natural coloration within a short time. 

Thus, there lies a major contradiction of modern restoration: the profession asserts a strict adherence to the scientific study of the artists’ materials and techniques, but continues to ignore the technical characteristics of the varnishes that are known to have been used in those centuries. Further, while it aims to present paintings as close as possible to the artists’ conception it still declines to take into account how their paintings had once looked with their original final layer on, and it persists in eliminating the ‘yellow tone’ of any varnish encountered on old master paintings.


A last video deals with the significant choice of a frame for Artemisia’s self-portrait. The Head of Framing, Peter Schade, points out that an authentic frame from the 16th century – wood-carved, painted or gilded – will always surpass any copy of it, even those that look to be perfect reproductions. He makes the following crucial remark [Choosing… 8:45]: “We always carry the baggage of modernity, of our time… And that gets always in some way transferred into reproduction frames. Usually, we don’t see it now but you can look it back at the history of frame reproductions, in the gallery as well, and [see that with] most reproduction frames, after twenty, thirty years they don’t match up to originals.”

Larry Keith had then to admit – albeit in carefully chosen words – that the same rule of unwilling, modern distortion applies to restoration:

“[9:12] …It is the same thing about how we… decisions we make about restoration itself, you know. We think we try to be… I guess what we can say now, is that we are very transparent about the decision-making process but it’s definitely an interpretation all the way down the line”.

Restoration being a contemporary “interpretation” of the work of the past, transparency is essential, and transparency implies clear explanations for the present and for the older interventions. But, strikingly, Larry Keith has not explained in any way the main justification for reducing rather than eliminating surviving varnishes. With regard to the use of retouching – e.g. in the reconstruction of the cropped top of the crown – his presentation and discussions are fair (see Reconstructing… and Retouching…) But on matters of cleaning and varnish this essay’s conspicuous technical, aesthetic and historical documentary omissions testify to an enduring institutional avoidance of transparency on the most vital artistic questions of art conservation at the National Gallery.

Above, Fig. 14: Artemisia’s Self-portrait, left, as in 2017 at auction; centre, same state but as provided by the National Gallery to the press in July 2018, before cleaning; and, right, as at the end of 2018 at the National Gallery, after cleaning and restoration.

Michel Favre-Felix, 9 October 2020.


1) Starting the restoration of Artemisia Gentileschi’s ‘Self Portrait’ – Now entitled: “The art restoration plan for Artemisia Gentileschi’s ‘Self Portrait'” – as posted on the 20th of July 2018.

2) Cleaning Artemisia Gentileschi’s ‘Self Portrait’ – as posted on the 27th of July 2018.

3) ‘It’s such a 17th century thing to do’ | Cleaning Artemisia Gentileschi’s ‘Self Portrait’ – as posted on the 3rd of August 2018.

4) Who was Artemisia Gentileschi? – as posted on the 20th of August 2018.

5) Finishing the cleaning | Cleaning Artemisia’s ‘Self Portrait’ – as posted on the 30th of August 2018.

6) Repairing a 17th century canvas – as posted on the 10th of September 2018.

7) Applying the moisture treatment – as posted on the 21st of September 2018.

8) Finishing the relining – as posted on the 2nd of October 2018.

9) Reconstructing the unusual composition of Artemisia’s ‘Self Portrait’ – as posted on the 9th of October 2018.

10) Retouching a 17th century painting – as posted on the 13th of November 2018.

11) Choosing a frame – as posted on the 26th of November 2018.

12) Framing Artemisia – as posted on the 14th of December 2018.


[1] “Indeed not all the old varnish was removed – it was simply reduced to a level which helps us to fully appreciate the painting.” (Larry Keith – Restoring Leonardo, National Gallery website.)

[2] “By removing the ugly varnish…” Jonathan Jones commenting on this cleaning in The Guardian, 13 July 2010. When reviewing the National Gallery’s restoration of its Leonardo Virgin of the Rocks, Jones expressed delight that the painting had been “freed from an amber prison”.

[3] National Gallery Podcast: Restoring Reni’s ‘Adoration of the Shepherds’, 1 :48.

[4] ARTECENTRO – Artemisia Gentileschi, Sainte Catherine d’Alexandrie, vente le 19 décembre 2017. Time 2: 47.

[5] ‘Vernix d’Ambre venant de Venise’ (Sloane MS 2052 Folio 9v).

[6] ‘Ce Vernix est fort rouge & est celuy des faiseurs de Luths’ (Sloane MS 2052 Folio 150 v).

[7] ‘quella d’ambra si compra, quella di mastice la facio io’ (Merrrifield, M.P, p. 743).

[8] ‘Chés touts les vendeurs de couleurs en Italie on vend une huile espaisse, qu’ils appellent Huile d’Ambre de Venise […] Je croy que c’est ceste huyle dont m’a parlé & se sert Gentileschij ’ (Sloane MS 2052 folio 146v).

[9] ‘Et pour vernir: & pour mesler sur la palette avec les couleurs’ (Sloane MS 2052 folio 151r).

[10] Such a choice of resins for varnishes is also noted by Van Dyck, at the same period, on a folio of a sketchbook: fir balsam, colophony, unspecified ‘vernizia’ and amber varnish (Kirby, J. 1999, p. 13).


Arciprete, B. (2004), ‘Il restauro’, La Flagellazione di Caravaggio, il Restauro, Electa Napoli.

Carlyle, L. (2005) ‘Representing authentic surfaces for oil paintings: experiments with 18th and 19th-century varnish recipes’, Art of the Past, Sources and Reconstructions. Proceedings of the first symposium of the Art Technological Source Research study group. Archetype Publications.

CCI (1994) Varnishes: Authenticity and Permanence Workshop, Canadian Conservation Institute, (Reviewed by Neil Cockerline).

Christiansen, K., Mann, J. W. (2001) Orazio and Artemisia Gentileschi – New York Metropolitan Museum of Art, Yale University Press, New Haven.

De Mayerne, T. Turquet, (1620) Pictoria Sculptoria & quae subalternarum artium, British Library, Sloane MS 2052. Trancription in Berger, E. (1901) Quellen für Maltechnik Während der Renaissance und Deren Folgezeit (XVI.-XVIII. Jahrhundert), München.

Demuth, P. (2001) ‘Regeneration of blanched natural resin varnishes with solvent vapour’ Hochschule für Bildende Künste – Dresden/ The ENCoRE Symposium: Recent development in conservation-restoration research 19-21 June 2001.

Eastlake, C. L., (1847) Methods and Materials of Painting, Dover Publications, New York (2001)

Favre-Félix, M. ‘On the recipe for a varnish used by El Greco’, Conservar Património 26 (2017) pp. 37-49 – ARP – Associação Profissional de Conservadores-Restauradores de Portugal

Holmes, M. (1999), ‘Amber Varnish and the Technique of the Gentileschi’, in Artemisia Gentileschi and the Authority of Art: critical reading and catalogue raisonné, R. Ward Bissel, Pennsylvania State University Press, pp. 169-182.

Kirby, J. (1999) ‘The Painter’s Trade in the Seventeenth Century: Theory and Practice’- National Gallery Technical Bulletin, Vol 20.

Leonard M., Khandekar N., Carr D.W. (2001) ‘Amber Varnish and Orazio Gentileschi’s Lot and his Daughters ’, The Burlington Magazine Vol. CXLIII, pp. 4-10

Merrifield, M. P. (1849) Medieval and Renaissance Treatises on the Arts of Painting, Dover Publications, New York (1999).

Pes, J. (2018) ‘The National Gallery’s New Artemisia Gentileschi Should Be a Triumph – But Clouds Are Forming Over Its Ownership During WWII’, December 12, 2018,

Pfister, P. (2011) ‘Régénération : l’emploi des vapeurs d’alcool et les dangers des alcools liquides’ / Kunsthaus – Zürich / Nuances 42/43, pp. 24-29.

Ruhemann, H. (1968) The Cleaning of Paintings. Problems & Potentialities. Praeger Publishers.

Two “As Good as New”, Colour Contrasting Velazquezs ~ Thanks to the Fundación Iberdrola

19th December 2011

Following the recent National Gallery restoration that launched a £1.5 billion Leonardo blockbuster, comes news of a brace of corporately funded restorations (- two pictures and a classical sculpture) at the Prado. The energy corporation Iberdrola is acting through its cultural off-shoot, “Fundación Iberdrola”, to “Raise awareness about the contribution of the IBERDROLA Group to society” by allocations of pre-tax profits dedicated “decisively to the promotion and dissemination of art and culture in the territories where IBERDROLA operates”. As part of its green image-building campaign of Good Works (which includes illuminating buildings at night and training young researchers in the energy field) the foundation has made itself “Protector Member” of the Prado’s “Restoration Programme” which encourages restoration research and trains restorers within the museum to the tune of 300,000 euros per annum.

Our prime fear with Art/Business relationships in the treatment of unique, historic and irreplaceable works of art has always been that the commercial tail might wag the custodial dog; might press for excitement and drama over minimalist judicious and restrained intervention; for more frequent rather than less frequent interventions – in effect, might expect big bangs for big bucks. A second concern is that corporate restorations receive over-hyped and propagandistic “Good News” promotion as instruments of miraculous “recoveries” and “discoveries”. This last practice compounds the problem of chronic unaccountability in art restoration. Restorers who work in-house at museums are, for obvious reasons, given full political protection for their actions however controversial or demonstrably harmful they might be. Museums rarely concede that even long-past restorations were harmful and almost never admit to recent – let alone current – injuries.

When reporting the Prado’s latest restorations, the online, echoes Iberdrola’s own website account of the treatment of the monumental, paired paintings “Philip III on Horseback” and “Margaret of Austria on Horseback”. As the proselytising global energy giant puts it: “These paintings have been rehabilitated by the art gallery’s team of technicians with the backing of Fundación IBERDROLA as a supporting member of its restoration activity. The labour of specialists has allowed for recovering the original values of both portraits, which were significantly affected by the accumulation of dirt and the alteration of the varnish that had buffered the contrasts of colour”. Both accounts fail to appraise the Prado’s own before and after cleaning photographs which show changes that seem arbitrary and artistically injurious (see Figs. 1, 4 & 5). Such preparedness to accept on trust that the latest photographically recorded states are the best, most “advanced”, most reliably truthful – and even “original” – conditions of historic works of art reflects a wider and dangerous absence of properly critical appraisals of restorations. It would seem axiomatic that if works of art are to be altered (and then re-altered by the next generation) the processes concerned should be absolutely transparent and freely discussed. Artdaily trills that the paintings have been “fully restored to their original appearance” by the removal of a “veil” of dirt and “altered” varnish; and, that the restorers “allowed for the recuperation of [the pictures’] original values”. This is naive and illogical: if removing the “veil” had revealed the original paintwork, what would need to be recuperated?

The over-selling of restorations can distort scholarship itself. Where Artdaily describes the two pictures as having been painted by Velazquez with the help of assistants, Iberdrola speaks with possessive proprietary pride of “emblematic works of Diego Velazquez” now rightfully displayed in the same gallery as “Las Meninas” and “the other three renowned equestrian portraits of the artist”. This inflation traduces the labours of scholars. In her 1948 book “Velazquez”, Elizabeth du Gué Trapier (a member of the Hispanic Society of America) said of the Philip III that it is:

by an unknown artist, or according to Beruete by Bartolomé González, retouched by Velázquez…Beruete wrote of Philip III’s portrait: ‘The greater part of the horse, the retouches of the armour, the horseman’s right arm, leg and foot; the stirrup, bit, the ornaments which hang from the horse’s croup, and the retouching of some parts of the sea-scape in the distance, are undoubtedly by his hand; one feels in them the lightness of his touch and his habitual precision and vigour. On the other hand, the forehead and the nostrils of the horse, as well as a great part of the background, were doubtless executed by the pupil Mazo.’ Beruete thought that Veláquez left the head of the King in its original state; others are of the opinion that he repainted it.”

More recently, in his posthumously republished Catalogue Raisonné (Taschen/Wildenstein, 1996), José Lopez-Rey describes a bewildering array of attributions and summarises that most scholars are agreed “on the strength of visual evidence that this equestrian portrait […had been] executed by a lesser hand and later reworked by Velázquez or under his direction”. Lopez-Rey adds that “Whoever the original painter of Philip III on Horseback, the painting has been visibly repainted, mainly the head, chest, forelegs and tail of the horse, possibly by Velázquez or an assistant in about 1634-35” and that by 1772 both Philip III and Queen Margaret had been extended from vertical to roughly square formats by additional vertical strips of canvas on both sides. Re Iberdrola’s hyperbole, Lopez-Rey draws distinctions between the entirely autograph Velazquezs “The Surrender of Breda” and the equestrian portraits of Philip IV and Prince Balthasar Carlos; the equestrian portrait of Queen Isabel, where Velazquez’s hand is “recognisable”, and the two recently restored works under consideration here, where that authorial hand is present only to a “lesser degree” in pictures “which were mostly executed by other painters”.

Given this consensus of uncertain authorial contributions, Iberdrola’s attempt to spin authentic Velazquez silk seems brazen when the Prado’s own before and after restoration photographs of the Philip III show so many artistically disturbing changes. During their latest restoration/recuperation, the sections of canvas that had been added to the sides of both paintings during the 18th century were cut off. Artdaily says that these additions had been made to make the paintings compositionally compatible with other works when installed in a new room in the Royal Palace in Madrid. The Fundación Iberdrola justifies stripping these historical extensions on the grounds that their removal has created a greater compositional contrast between the two pictures and the three great autograph Velazquez equestrian portraits of Philp IV, Isabel of Bourbon and Prince Baltasar Carlos. Artdaily describes the cleaning and the cropping together as having achieved the greatest possible recuperation of the “original perceptual conditions“. What is not acknowledged by either party is that this removal of historically resonant material has served to eliminate possibly discomfiting visual testimony to the original condition of the paintings (- at least insofar as it had survived into the 18th century) and made the task of gauging the effects of successive restorations almost impossible.

It can be assumed that when those strips were added, their values matched and seamlessly extended the then extant values of the two pictures (as with the repair to Michelangelo’s Sistine Chapel frescoes discussed in our April 1st 2011 post). What, therefore, might account for the mismatched values that had emerged and persisted (see right) until the recent twin restoration? The strips would likely have been painted by a single artist, in a single campaign and without need for making the numerous revisions and re-revisions by the assorted, variously attributed hands that are evident in the paintings themselves. It is unlikley that any paint layers in the extensions would have been made over already varnished paintwork and therefore be at risk of floating off during cleaning. Conspicuously, the most obviously exculpatory explanation for the mismatch of tones – that the additions had matched the values of the picture when under an old darkened varnish – has not been offered. The restorers reportedly attribute the mismatches solely to the technical fact that the bordering strips had been painted over darker ground colours than those in the paintings, and that this had somehow caused “the pigments in the two areas [to have] behaved differently over the course of the centuries”. This does not hold water: while darker grounds would certainly come to influence the values of the extensions as the overlying paint became translucent with age (as can most clearly be seen in the “see-through” of the first state of Philip III’s horse), it could not alone have done so to such striking (and varying) extents as have occurred in the two paintings. Whereas, the fact that darker grounds were used on the extensions would itself suggest an initial perceived need to match the then darker values of the paintings.

There is another reason for disregarding the current restorers’ explanation. If the dark ground theory were correct, the mismatch phenomenon would have arisen very slowly over time and not have – as the photographic record shows – lurched into being on successive restorations (see Figs. 2, 6 & 7). Artdaily has not shown before and after photographs of the Queen Margaret, which painting Lopez-Rey describes in the 1996 edition of his catalogue raisonné as having been restored in 1968. His post-1968 illustration is shown here in Fig. 2. However, in the 1978 edition of the book an apparently earlier (and presumably pre-1968 restoration) state is reproduced (see Figs. 6 & 7 for details of that state). At that date, the left-hand extension read only fractionally darker than the painting, and although the right-hand extension was appreciably darker it was less obtrusively so than was the case after the 1968 restoration. As described opposite, and as can clearly be seen in Fig. 7, the impact of the dark ground was neglible at the brightest part of the sky, on the horizon, where it might have been expected to be most evident. The original dramatically escalating darkness in the sky above those points should, therefore, properly be taken as part of the original tonal schema – and not as either accumulated filth or ground paint see-through. The similarity of the states in both paintings, as seen in Fig. 2, might suggest that the Philip too had been restored around the late 1960s. The differences between the pre and the post 1968 restoration states of the Margaret speak of massive changes of value being made during a single “treatment”. The horse, for example, was reduced from a deep rich chestnut to a tan colouring. That the earlier chestnut appearance had not been a by-product of some filthy misleading “veil” is demonstrable: the whites on the horse’s muzzle and upper left leg read as white not yellow or grey or brown. Whatever might account for the radical changes, it was not consistent with the removal of some overall disfiguring layer.

Characterising the surgical elimination of material that bears awkward testimony of an earlier, now irrecoverably lost state, as a recuperation of a painting’s “ original conditions” is naive and seriously misleading. With every restoration – however funded – the most urgent critical question must always be: “did it do any harm?” To answer it, we must begin by using our eyes and, perhaps, by heeding the advice of artists, one of whom reportedly askedWhat’s that dirt called that the restorers clean off? Oh, that’s right – burnt umber.”

Michael Daley

Comments may be left at:

Printer-friendly PDF version of this article

Above, Fig. 1: “Philip III on Horseback”, before (left) and after (right) its recent restoration as shown in, with photographs by courtesy of the Prado. Consider this: if restorations did no more than simply remove dirt and discoloured varnish, then with each restoration a work would be returned to its condition when last cleaned. This does not happen – every restoration produces a new, different work – but were it to occur, there could be no news-grabbing claims about miraculously recovered original conditions. Were it to occur it would also be possible, by placing photographs of today’s restored works against photographs of the works when last restored, to reassure art lovers that nothing untoward happened during restoration. Such photographic comparisons are not volunteered because they cannot be; because, as frank curators admit, in every restoration there are losses as well as gains; there is always a trade-off between improved legibility and losses or corruptions of artistic content.
Above, Fig. 2: “Philip III on Horseback” (left) and “Margaret of Austria” (right) as shown in the 1996 Taschen/Wildenstein edition of José Lopez-Rey’s “Velazquez ~ Painter of Painters”. With regard to the reliability of these photographs, unless otherwise established, we take them to have been judged by the authors and publishers to have derived from reliable sources and to constitute fair depictions. Concerning the testimony of photographs generally, they do not need to be absolutely accurate and truthful (although that is, of course, a desirable goal – see Gareth Hawker). In two respects their testimonial value is essentially relative. That is, it is the patterns of values within a photograph that are crucial – the linked ratios of tone/hue values. (If a tone which goes into a restoration darker than its neighbours emerges lighter then there is, ipso facto, cause for investigation and explanation.) Similarly, much of the testimony of individual photographs lies in their appearance vis-a-vis earlier and later photographs. Both of the two photographs above are markedly different – and in the same way – from their earlier counterparts in the 1978 Italian edition of Lopez-Rey’s book. Both of the paintings then were much darker (and handsomer – see Figs. 6 & 7), and in both cases the discrepancies between the original paintings and their 18th century extensions were then less pronounced and artistically disruptive. In the photographs above we can see that although the extensions are now markedly darker than the centre sections, they are not uniformly so. The darkest passages occur in all of the corners of the Philip and in three of the corners in the Queen Margaret picture. At that date artists worked hierarchically, with figures – especially monarchs – pre-eminent, and backgrounds (landscapes, seascapes, skies) subservient. The lighter tones of heads and their fineries were commonly thrown into relief and pictorial promininence by conveniently dark clouds or foliage (see Fig. 6). Artists did not transcribe their values from camera obscuros, they manipulated them to suit their own purposes. Had the darkness of the extensions been a consequence of underpaint show-through, its disfiguring presence would have been most apparent where the lightest values fall, which is in the bottom of the sky at the horizon. In fact, any see-through distortion is least apparent at those points (see below). It seems safe to conclude, therefore, that paint, previously present elsewhere, has been lost. By an accident of chemistry, it is the darker pigments, such as burnt umber, that are most vulnerable to cleaning solvents.
Above, Fig. 3: “Philip III on Horseback”, attributed to Velazquez and others, as shown in the 1996 Lopez-Rey. To appreciate the scale of changes that restorers’ make, the photograph above, which predates 1996, has been cropped to the format shown in the before and after photographs shown above and below (Figs. 1, 4 & 5) which record the effects of the painting’s recent treatment and surgery. The shifts of value are enormous. The surgery was radical – the 18th century additions were removed. The claims made by the restorers’ are matchingly enormous. These paintings are said to be “fully restored to their original appearance”. It is claimed that “While it was previously difficult to appreciate the merit of these works within the group (other than for their purely iconographic merits) , these are now absolutely evident.” It is specifically claimed that a new dynamism has emerged in the Philip III and that this has been transmitted “through the foreshortening of the horse and the luminous sky behind the figure“. The foreshortening of the horse was a done deal before this restoration began. As can be seen in Fig. 5, however, the painting has acquired an overall, florid colouring. The sky may be a cooler more uniform blue but it has lost its previous theatre and differentiations. The lights in the painting came from two sources. There was a brilliant, off stage-right “spotlight” that lit the front of the horse and the monarch, whose own brilliantly illuminated and modelled leg bisected the horse’s primary lights and shades. Additionally, there was a raking general light from the horizon that reflected in the water, throwing the horse’s darker tones into relief. These effects, as they survived in Figs. 3 & 4, have not been enhanced by restoration or recuperation, they have been weakened. By liberating blueness at the expense of tonality, the sky, instead of receding and throwing the horse and rider into relief, now simply obtrudes.
Above, Fig. 4: “Philip III on Horseback”, before the recent restoration, as shown in One the hardest things to establish in restorations is the extent to which restorers wield their own brushes. To spot their tiniest but often crucial interventions (on the most sensitive and expressive parts of faces, for example) requires the provision of detailed photographs recording after-cleaning and after-cleaning-and-retouching states. The two artdaily photographs are sufficient, however, to show that massive retouching (“recuperation”?) occurred in one zone. Before this last intervention, the originally positioned baton carrying arm of Philip showed through the painting of the sky. Now (as seen below) it does not. What other additions were made? What proportion of this now “fully restored to… original appearance” painting is a product of the 21st century?
Above, Fig. 5: “Philip III on Horseback”, after the recent restoration, as shown in One noteworthy injury in this painting can be seen at the horse’s left knee. Previously, as seen in Figs. 3 & 4, the contour of the leg was articulated by the contrast between the light upper surface of the leg and the adjacent dark of the distant town across the water. With the eradication of that strategically placed, pictorially functional dark zone, an ungainly thick dark bar has emerged that attaches to neither the horse nor the town but instead floats on the picture surface. In all probability its original function at an early stage of painting had been to serve as a shorthand guide to the artist of the tonal requirement that would later be needed to throw the advancing horse into relief.

Above, Fig. 6: Artdaily does not show photographs of the restored Margaret but an earlier restoration seems apparent in a comparison of the plates in the 1978 and the 1996 editions of José Lopez-Rey’s book. We see in the details above left and below left the painting before its 1968 restoration. The differences between them and their later counterparts are remarkable. No educated eye could equate the differences between these states with a straightforward removal of an overall film of dirt, or discoloured, “veiling” varnishes. The changes are variously too abrupt, too arbitrary and too injurious to the force and strength of the picture. No overall covering could return the washed-out post-1968 restoration state to its former vigour. No discoloured varnish could now “recuperate” the former chestnut of the horse from the succeeding tan colouring, on the one hand, and/or simultaneously restore the wooded valleys of the bottom right to their former vigour and panache. A single, seemingly intractable passage of paint gives the lie to successive treatments of this picture. We see in the tiny recalcitrant dark remnants of trees that still cling to the edge of the side of the second hill down from the horizon, the last corroboration of the values in the original picture that had been matched on the extension canvas. We see them here, as in the aftermath of the 1968 restoration, but will not, presumably, see them again, given the recent “recuperating” removal of this mute information-rich canvas extension.

Above, Fig. 7: the right-hand extension of the “Margaret of Austria on Horseback”, before (left) and after (right) the 1968 restoration.
Click on the images above for larger versions. NOTE: zooming requires the Adobe Flash Plug-in.


Discovered Predictions: Secrecy and Unaccountability at the Metropolitan Museum of Art, New York

25th February 2011

Impeccable condition in a painting is more of a goad than a deterrent to restorers. When the youthful Thomas Hoving was appointed director of the Metropolitan Museum in 1967, he formed a respectful – even deferential – alliance with the (then) head of picture conservation, Hubert von Sonnenburg. Two decades earlier in London, the National Gallery’s director, Philip Hendy, forged a similarly dependent relationship with the German émigré restorer Helmut Ruhemann. Ironically, von Sonnenburg had presented as the heir-apparent to Johannes Hell, another German émigré to Britain who’s mild and gradual cleanings were widely preferred to Ruhemann’s controversially swift “total cleanings”.

Hoving and von Sonnenburg together stalked one of Velazquez’s finest portraits, his Juan de Pareja, which the Met acquired in 1970 for a world record $4.5m. Although, on their own testimony, that picture was in superb condition and had never even been lined, on acquisition it was whisked to Wildenstein and Company, “for secrecy”, as Hoving later admitted. There, Von Sonnenburg secretly “proceeded to discover”, as Hoving put it, “everything he had predicted he’d find”.

It was not unprecedented for a museum director to have a major acquisition secretly restored. Sir Charles Eastlake, scorched by National Gallery cleaning controversies in 19th century Britain, had his acquisitions cleaned in Italy before bringing them to the gallery. Secrecy in conservation can seem systemic: in 1960, when the National Gallery constructed “modern” purpose-built conservation studios, part of one was partitioned by a wall, behind which the chief restorer could work on projects of “particular difficulty or confidentiality”, as a then National Gallery restorer, David Bomford, put it in 1978.

Eastlake made no photographic record of the pre-restoration condition of his acquisitions – even though he happily used photographs for attributing paintings, and must, as president of the Royal Photographic Society, have appreciated photography’s unprecedented testimonial capacities. Fortunately, photographic records of the Sonnenburg/Hoving Velazquez restoration were kept and published by the Metropolitan Museum (in an undated booklet – see right). While these photographs may not be of the highest, digital age, standards, they are nevertheless “of a piece” and permit comparisons between recorded states to be drawn.

Much as von Sonnenburg thrilled over an impeccably preserved, never-lined canvas, he could not resist tampering with it. Two of its edges had been folded over on the stretcher. This fact was presented to Hoving as a “discovery”, even though it had been reported by the Velazquez specialist José Lopez-Rey seven years earlier. The folded canvas strips were opened, flattened and reinforced with new canvas to extend the picture’s format and diminish its subject, shifting him leftwards and downwards (see right). The justification for this compositional “recovery” was that original paint had been applied to the folded strips, but the pictorial testimony of that paint, when first revealed, was not photographically disclosed – see account on the right.

Von Sonnenburg, it seemed, could not resist the urge to “liberate” the painting’s supposed “pure flesh tones” and thereby leave the dark-skinned servant’s face lighter and pinker. By stripping off “varnish” von Sonnenburg also caused previously unified components to detach themselves from each other:

the rounded shape of Pareja’s forehead, for example, is defined only by a large spot of impasto-crisp in the center, bordered by dragged spurs – applied directly on the thin underpainting. When seen close up, the highlight seems to be floating over the paint in an almost measurable distance…

This was a classic restoration apologia. Even the emergence of a formerly hidden streak of flesh-coloured paint on the background was presented as an act of liberation and recovery:

Attention should be drawn to the single dragged brushstroke of light skin colour in the center of the background at the right…Unquestionably, this randomly applied paint is original, and shows how Velazquez chose to try out his loaded brush on the background…Such spontaneity, combined with the greatest subtlety of color and technique make the Juan de Pareja one of Velazquez’s most painterly works.

Convinced that Velazquez had happily left his own brush-wipings visible on one of his two finest portraits (the second being his Pope Innocent X), and that he had used glazes less than Titian, von Sonnenburg was not dismayed when his cleaned painting betrayed markedly less colouring and reduced to a “predominantly gray color scheme”. His rationale for losses of colour and of spatial and plastic coherence; for the flattening of a formerly prodigiously well-modelled and sympathetically lit head; and for the spatial inverting of a background that formerly receded, was audaciously lame: in 1938 an English restorer, Horace Buttery, had described the doublet as “dark gray”. Despite recognising that the painting had – miraculously – shown “no signs of ever having been abused by solvent action during the past”, von Sonnenburg nonetheless contended that it must have been cleaned and varnished “at times”. On that basis, he speculated that it could therefore safely be assumed to have been so restored by Buttery, and, therefore, to have enabled him, on that occasion, correctly to have read the doublet’s true colour. This hypothetical daisy-chain was presented as a proof, despite the fact that before and after Mr Buttery, the garment had always been described as a “green doublet” – not least by Velazquez’s biographer, Antonio Palomino who in 1724 precisely reported “a muted green for Juan’s doublet”.

After their stripping and repainting of pictures, restorers invariably apply fresh varnishes… which in turn discolour and thereby serve as a pretext for another “restoration”. With successive varnish removals, solvents deplete, embrittle and optically alter paint films. When penetrated by solvents, paint films heat, swell and soften so that even the friction of cotton wool abrades them – as the restorer Caroline Keck admitted. Soluble plastic components of the paint itself are carried off by evaporating solvents. Restorers sometimes claim that because old paintings have so frequently been abused in the past, there is nothing left to extract today – but with the Juan de Pareja, no such claim could be made. At the same time, they sometimes admit that cleaning pictures with thick paint is easier than cleaning ones with thin paint. (If cleaning methods really were as safe as is claimed, it would not matter whether the paint being treated was thick or thin.) When stripped to a restorer’s conception of “clean”, the remaining paint is left parched, absorbent, matt and in need of “nourishment” by varnishes.

When new varnishes (i.e. resins dissolved in solvents) are applied, they penetrate and amalgamate with the parched paint thereby making the next cleaning the more hazardous, and so on ad infinitum. If we are lucky, von Sonnenburg will have used a natural resin varnish. If not, if he subscribed to the Met’s then hi-tech enthusiasms, he will have used a synthetic resin in the confident but erroneous expectation that it would not discolour and that it would remain easily soluble.

In 1966 a restorer at Moscow’s Pushkin Museum of Fine Arts deplored the use of untested synthetic materials, judging them “all right for space ships” but not for old master paintings. By 1995 a conservation scientist, Tom Learner, reported that synthetic resins which had “appeared to offer” advantages over natural ones “are now known to be inherently unstable”. In 1998 the Met’s (present) Chairman of European Painting, Keith Christiansen, admitted that synthetic varnishes used at the Metropolitan Museum had turned not yellow but grey and had “cross-linked with the pigments below, meaning that removal is, if not impossible, extremely difficult”.

Dr Christiansen has yet to reply to the question ArtWatch and ARIPA put to him on February 6th, concerning the Met’s intentions towards its new, miraculously well-preserved Perino del Vaga painting.

Michael Daley

Comments may be left at:

Printer-friendly PDF version of this article

Above, Fig. 1: Velazquez’s oil on canvas portrait Juan de Pareja, before treatment at Wldenstein and Company and when still “covered by a yellowish brown varnish and with the top and right edges folded over” (- as described in an undated, untitled Metropolitan Museum booklet that contained essays by the museum’s President, Douglas Dillon; Director, Thomas Hoving; Vice-Director, Curator-in-Chief, Theodore Rousseau; Curator-in-Charge, European Paintings, Everett Fahy; and Conservator, Hubert von Sonnenburg).
Above, Fig. 2: the Juan de Pareja during treatment when the top and right margins had been unfolded; paint losses had been filled with putty; and new red-brown paint covered some of the fillings and was to serve as a priming for the final “touching up“. At this point, the “discolored varnish” had been removed from the right-hand side of the picture. (This photograph and the one above were by courtesy of Wildenstein and Company.)
Above, Fig. 3: the Juan de Parejaafter cleaning and restoration“. (Photograph was by Taylor & Dull, Inc.)
Above, Fig. 1, detail showing warm orange-red layer over the background and tonal modelling on the lace collar.
Above, Fig. 3, detail showing loss of red on background and the new flattened, whiter-than-white collar.
Above, Fig. 1, detail. Note the relatively small area of lights on the face. Note, too, the large proportion of warm reds and in particular their deployment at all the points at which the forms of the face turn away from the viewer at the profile. At this stage, none of the light passages in the face abutted the black of the hair, the warm red mid-tones formed transitions between the brightest lights and the darkest darks.
Above, Fig. 3, detail. Note, in comparison with the untreated image above it, the profound transformations of pictorial values and language system that occurred as a consequence of this so-called “restoration”. In the post-cleaning and post-restoration state (for both activities took place extensively), the new lighter, cooler background asserts its presence more, in an entirely historically innapropriate modernist, “abstract” fashion. The former, highly selective and focussed placement of the the brightest lights on the collar (which articulated the forms) have been lost in the great expansion of whiteness. The general lightening of the background around the head introduces a halo-effect not previously present. The new light passages to the (viewer’s) left of the hair now nearly meet the lighter passages to the right of the head with most unfortunate and unoriginal consequences. Previously, the head emerged towards the viewer out of a warm dark enclosing space. Now, given the great lightening of the background around the head, the black of the hair can be read as a void in in a light coloured wall. One of the commonest signs of restoration injury to a face is present here: the contrasts between the blacks and the whites of the eyes are intensified regardless of the general system shading that had applied to the head. In the unrestored state, the most brilliant lights present were in the reflected lights of the dark irises, not in the whites of the eyes themselves. Another common loss that is seen here occurs in the tonal modelling around the eyes, which is used to establish the forms of what are essentially a pair of balls set in two sockets. The notorious carelessnes of restorers with anatomical features, is matched by an obvious indifference to shapes. The loss of the coherence of the former treatment of the hair is an almost universal restoration short-coming. Note the extent to which violence has been done to the former linked areas of hair in the subject’s side-burn and beard. Now the lighter mid-tones of the flesh at the cheek race through, breaching the hair like water out of a dam. In order to be complicit with such injuries one would have to subscribe to a fairy tale – one would have to believe that all the previously superior articulations of form, physiognomy, space, atmosphere and pyschological insight, were the unintended, undesigned, fortuitous benefits of some physical degeneration of a layer of varnish. In fact, one would have to subscribe to two fairy tales. One would have to believe that if the present varnish were to be left in place for long enough, it too would improve the drawing and modelling of the present state of the painting; that it would impart red-ness here, and green-ness there to stunning pictorial effect as it gradually turned into a yellowish brown covering.
Click on the images above for larger versions. NOTE: zooming requires the Adobe Flash Plug-in.