The Samson and Delilah ink sketch – cutting Rubens to the quick

Today, in a sale of old master drawings (and on an estimate of £1.5m -£2.5m), Christie’s is offering large claims for the artistic and historical significance of a small (roughly 16cms square and shown here at Fig. 1) pen and brown ink drawing:
“This is the only known preparatory drawing for Rubens’s Samson and Delilah in the National Gallery, London (inv. NG 6461), and it was followed by a modello oil sketch now in the Cincinnati Art Museum (inv. 1972.459). Commissioned by Nicolaas Rockox (1560-1640), who was Rubens’s most important early patron, this powerful composition dates from shortly after the artist’s return to Antwerp from Italy, where he had been from 1600 until 1608, and provides a valuable insight into his developing style and preparatory processes.”
This account is conventional but, nonetheless, contentious. No hint is given that the relationships between these three linked works are highly problematic or that all three have suffered cuts or thinning. The authorship of this group has been contested for over two decades. On February 19 2004 the Daily Telegraph published a letter from ArtWatch on the painting’s problems (“Is the National Gallery’s Samson and Delilah another copy?) We have published two special issues of the Artwatch UK Journal mounting challenges (Figs. 2 and 3) and have written a number of articles on the subject for the Art Review. The principal challenges to the attribution came from two artist/scholars, initially, Euphrosyne Doxiadis, whose findings (made with fellow artist Steven Harvey and Siân Hopkinson) were compiled in a report (see this website) that was submitted to the National Gallery in 1992 and later covered in the Times and the Independent. In 1997 researches by Kasia Pisarek, prompted two articles by the Sunday Times’ art critic, Waldemar Januszczak (“A Rubens or a costly copy?” and “National’s £40m Rubens could be fake”). In the latter article, the then director of the National Gallery, Neil MacGregor, conceded that the evidence “is respectable, and the scholar raises some serious questions that I cannot easily answer”. Those questions have never been answered. In October 1997 the National Gallery issued a press release in which it was said that:
“Debates of this sort require patient consideration of different sorts of evidence. The best format is for this evidence to be presented at some length for public discussion – and the National Gallery will be arranging such a lecture and debate over the next few months.”
A debate that has yet to take place
Within a few days the commitment was dropped when the press release was re-issued and the debate never took place. To this day there remains an enormous accumulation of problems with the National Gallery’s “Rubens” Samson and Delilah and, therefore, with its two closely associated works – the ink drawing and the oil sketch. All three works, which are dated to 1609-10, have unusual and anomalous features – and all appeared only in the 20th century. The modello arrived last without name or history in 1966 and was upgraded by Christie’s to Rubens even though it is painted on a soft wood and not the oak which Rubens invariably used.
Ludwig Burchard’s cunning plan?
Behind the successful 20th century elevation of this trio, is the fact that both the drawing and the large finished painting in the National Gallery were attributed to Rubens barely two years apart by the same man, Ludwig Burchard. Burchard was a great authority on Rubens who, notoriously, was unable to publish his life-long Great Work on the Artist for fear of having to de-attribute very many paintings for which he had supplied unwarranted certificates of authenticity. In the ArtWatch UK Journal No. 21(Spring 2006) Kasia Pisarek, whose PhD Dissertation was on Rubens and Connoisseurship, identified over sixty Burchard Rubens attributions that had subsequently been demoted in the Corpus Rubenianum itself.
Dr Pisarek felt that the year of launch for the picture now in the National Gallery might be signicant. As she put it:
“That year 1929 was not free of strange coincidences. By a bizarre stroke of luck, the painting re-emerged 48 years after its disposal by the Prince of Liechtenstein in Paris in 1881 (not 1880, as is commonly said), the exact same year as the deaths of the Prince Johannes II, the previous owner of the painting, and of his picture adviser Wilhelm von Bode, the then General Director of the Berlin Museums. The former died in February 1929, the latter a month later, in March. Moreover, we know that the Prince himself had weeded out a considerable number of pictures, Samson and Delilah included. He also financed many research projects, and the collection was accessible to scholars. The art historian Wilhelm von Bode published (in 1896) the first comprehensive and illustrated book on the Liechtenstein collection, so he could have been aware of the Samson and Delilah’s disposal. Why didn’t he identify the picture as the long lost Rubens if he was also a Rubens expert and had even co-signed certificates of authenticity with Ludwig Burchard?
In 1927 the drawing was bought from a private collector by a scholar of drawings and prints, I.Q. van Regteren Altena, for 26 guilders as a Van Dyck (whose initials it still bears). It was promptly upgraded to Rubens by Burchard, who then cited it as such in his 1930 certificate of authenticity for the Honthorst on offer by a Berlin dealer that is now in the National Gallery as an entirely autograph Rubens.
A precursor or a successor – or both?
It is claimed that Rubens’ characteristic stylistic development through stages of work is evident in the three works’ sequence, when the essential motif remains remarkably constant throughout. In fact, the modello (see Figs. 5 and 7) is so like the finished work that one supporter of the attribution, the former senior curator of the National Gallery, David Jaffe, has suggested that this oil sketch might be a ricordo – a record of the finished painting[!] However, if the presently accepted 1, 2 and 3 sequence of drawing, oil sketch, finished painting were to become 1, 3 and 2, it would make nonsense of the National Gallery’s technical reports which stated that the finished picture’s uncharacteristic thin, swift and little-revised paint work – paint work which today remains preternaturally fresh and unblemished (see Figs. 10 and 11) – was a product of the fact that Rubens had made such an unusually complete and resolved oil sketch that he had been able to paint the larger panel (which, the gallery claims, itself resembles a large sketch) out of his head and at a stroke and without any need for his customary revisions. Then again, the ricordo suggestion constitutes, perhaps, a kind of insurance policy, a way of covering against the possible outcomes of an eventual debate and presentation of evidence? If so, the sequence 1, 2, 3 and 2 again, would make a kind of institutional sense? This might indeed constitute a veritable “belt and braces” insurance: given that the gallery has admitted that its large finished panel is so very swift and sure-footed in its execution (or uncharacteristically sloppy and out-of-character to its critics), that it is itself but an over-blown sketch, the formulation 1, 2/4, 3/2 and 2 might serve perfectly to cover all eventualities.
The evidence of our eyes
The Samson and Delilah ink sketch, as a drawing, lacks the customary force, focus and eloquence of design seen in Rubens’ initial compositional ideas (- see Figs. 8b, 9a and 16). This supposed preliminary study has a curiously finished, pictorial air. Iconographically it has a pronounced “portmanteau” quality, showing, for example, Delilah’s draped right leg as seen in the secure Rubens oil sketch of 1609-10, The Taking of Samson in Chicago, while her draped left leg is as seen in the insecure National Gallery picture. Most disturbingly (to this draughtsman, at least) is that fact that when looking at the drawing in the flesh it is impossible to read an order or purpose to which its many and various components might have been made or to locate the essential, determining compositional and figural point at which Rubens always and brilliantly drove (see Figs. 8b and 16).
A ruled ink border surrounds and compositionally confines the ink and wash drawing (Fig. 1). When seen in reproduction, this border gives an impression that Rubens designed a format from the outset precisely in order to achieve an effect that is the single most problematic feature of the finished painting – the fact that the toes on Samson’s right foot were cropped at the edge of the painting. The border, like the drawing, is drawn in brown ink but clearly, as Christie’s describes, it can be seen by eye to comprise later framing lines. However, while this usage is seen to be common in the collection where the drawing has lived since 1927 – and while the border lines themselves can be seen to pass over a number of tiny losses on the edges of the sheet – the particular placement of the border is disquieting because the sheet on which the drawing was made has been trimmed at either the outside edges of the border or even within the border lines themselves. Why and when was this done? While some of the ink lines of the drawing can be seen by eye to run into the ruled borders, we cannot calculate where they might have terminated because of the severity of the sheet’s cropping. For whatever reason, this is now an artificially constrained and possibly edited image.
Flouting historical evidence
While the toes on Samson’s right foot are cropped at the edge of the National Gallery painting (Fig. 12), both of the contemporary copies that were made of the original Rubens painting show the foot, as painted by Rubens, to have been both whole and set well within the right-hand edge of the painting (see Figs. 4, 5 and 6). It is hard to see on what grounds this testimony might be disregarded: the first copy, an engraving (see Fig. 14), was made in c 1613 and very possibly under Rubens’ instruction. The second was a painting in oil commissioned by Rockox to show off his collection of paintings in the grand salon of his home (see Figs. 6 and 13). Is it conceivable that he – and Rubens, who was still alive – would have permitted a man famous for the accuracy of his records, to make a gratuitous, out-of-character “improvement” to the Rubens painting that occupied pride of place above the mantelpiece? Because of the inked box and the trimmed sheet it is not possible to determine whether the drawing’s author might originally have drawn the foot whole.
The panel support of the modello, as reproduced in the catalogue (see Fig. 7), is seen to have been cropped on its vertical edges since being sold to the Cincinnati Art Museum by the removal of two strips of wood, thereby conferring a clear crop onto Samson’s foot and bringing it into accord with the foot seen in both the National Gallery picture and the ink drawing. At one point the Cincinnati Museum claimed that the oil sketch’s panel was made of oak. When the picture was loaned to the National Gallery we asked if the panel was oak or softwood. It was not possible to say, we were told, because the back of the frame was enclosed and the gallery was not permitted to remove it. The museum today ducks the issue by saying that its painting is “on panel”.
The National Gallery’s picture was doctored at some undisclosed point by planing rather than cutting. The gallery restored the picture after purchasing it and reported that the panel had been planed down to a thickness of 2-3mm and set into a sheet of block-board. We knew for technical reasons that that was most unlikely: block-board is held together by its outer veneer layers and cutting one of them away would have had catastrophic structural consequences. When pressed, the gallery acknowledged that the planed-down panel had in fact been glued onto, and not set into, a larger sheet of block-board, with its edges being concealed by a bevelled putty. The restorer, David Bomford (now of the Museum of Fine Arts, Houston), said in his report, that the planing had taken place at some point in the early twentieth, or possibly during the late 19th century. That, too struck us as improbable: could there be no record of the back of a panel bought for a world record price (£2.5m) for a Rubens? Had the gallery not made a record of condition when the picture was loaned to it before the sale at Christie’s? We asked Neil MacGregor, if the gallery had any record of the back – and he said not. We asked if we might see picture’s conservation dossiers and there found Burchard’s 1930 certificate of authenticity, which described the panel as being intact and in excellent health.
At Christie’s we asked, and were kindly permitted, to examine the back of the drawing which is said to bear other drawings. A little (unintelligible) drawing is present but most of the surface bears the remains of a second sheet of paper to which the ink sketch had once been pasted. Effectively, the drawing’s verso is invisible – just as is the back of the National Gallery’s picture, any evidence on which has ceased to exist.
As for the contention – made against the evidence of the contemporary copies – that Rubens deliberately cropped Samson’s toes at every stage of the work, we know that he was very attentive to his toes. When drawing one of Michelangelo’s ignudi in the Sistine Chapel, he ran out of room on the paper for the toes on one of the feet and then drew them separately elsewhere on the sheet. On his return from Italy, and virtually simultaneously with working on the Samson and Delilah, Rubens made the magnificent Michelangelesque study of a nude man kneeling shown at Fig. 17. On that sheet, the right foot was truncated by the edge of the paper and, again, Rubens redrew the whole lower leg so as to include the foot and toes.
What kind of artist was Rubens?
The National Gallery has admitted that its painting is not typical of Rubens’s oeuvre, which fact it attempts to explain by claiming that immediately after his return to Antwerp from a long stay in Italy, Rubens was working “experimentally”. Unfortunately, it so happens that at the date of the Samson and Delilah’s execution, Rubens was also working on the very large altarpiece The Raising of the Cross (see Fig. 10). No one has ever suggested that that great work occupied a position in some experimental mode. To the bizarre and unsupported suggestion that Rubens, on his return from Italy, simultaneously worked experimentally and not-experimentally within the same brief period, Christie’s lend support with a contention that:
“The exact date of Samson and Delilah is unclear, partly because Rubens experimented with two very different approaches to the same subject in these post-Italian years.”
The truth is that attempts to keep this Burchard-initiated show on the road require that everything today be considered part of a moveable feast. It is neither a satisfactory situation nor a tenable position. Attribution is a difficult and taxing activity at the best of times and there is no shame in admitting error – and least of all with Rubens. As we put it in the 2006 Spring Journal:
“The upgrading of copies or studio works to autograph status frequently flouts the most elementary visual and methodological safeguards. Identification of the autograph hand of a master requires a ‘good eye’, sound method, and a recognition that comparisons are of the essence, that like should be compared with like. Procedural fastidiousness and visual acuity are nowhere more essential than with Rubens, who not only ran a large studio of highly talented assistant/followers but who famously placed a very high premium on studio works that had been modified or finished off by his own hand. When wishing to claim unreserved autograph status for a ‘Rubens’, it would seem imperative that some plausible connection between the aspirant and an unquestionably secure work be established. With the National Gallery’s Samson and Delilah, exemption is claimed on grounds that this work was special product of a peculiar moment in the artist’s career. Unfortunately for the attribution – and the picture’s supporters – this special ‘moment’ coincides precisely with a work of bedrock security – The Raising of the Cross of 1609-1610. An artist’s designs and motifs are easily replicated – and with Rubens, were often intended to be so ‘in house’. Pronounced similarities of subject matter or motif, therefore, are no guarantors of authenticity. What is most distinctive to a master and impossible to replicate – even by close associates within his own studio – is what is termed his touch, his individual, characteristic manner and speed of execution. Artistic mastery lies in some particular combination of technical fluency and commanding thought. The quality of an artist’s thoughts and his authorial ‘fingerprints’ are certainly made manifest in and through material – it cannot be otherwise – but only in material as handled, not in terms of its intrinsic, chemically analysable composition. A flat-footed analysis of the material components of pictures can no more corroborate authorship than they can validate a restoration. There are no material tests for authenticity…”
Update:
16.00, 10-07-14. The editor of Jackdaw, David Lee, writes to point out that, R W P de Vries, the person who sold the Samson and Delilah ink sketch produces this note, when Googled:
“Reinier Willem Petrus de Vries Jr. (Amsterdam , March 3, 1874 – Hilversum , 27 May 1953 ) was a Dutch artist. He was a painter , illustrator , book cover designer , and made ??etchings and woodcuts . He was a student at the State Normal School in Amsterdam, obtained his MO drawing. From 1913 to 1935 he was a teacher at a secondary school in Hilversum.”
The Jackdaw’s distinguished editor reflects: “An artist and secondary school teacher who flogs drawings. Not exactly what you’d expect…” No, indeed, but precisely the kind of thing about which we have learned not to expect to be given information.
Michael Daley
Comments may be left at: artwatch.uk@gmail.com
Discovered Predictions: Secrecy and Unaccountability at the Metropolitan Museum of Art, New York

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: artwatch.uk@gmail.com