(PDF) Leonardo da Vinci's Madonna of the Yarnwinder. A Historical and Scientific Detective Story | Thereza Wells and Martin Kemp - Academia.edu
Skip to main content

Leonardo da Vinci's Madonna of the Yarnwinder. A Historical and Scientific Detective Story

This paper
A short summary of this paper
33 Full PDFs related to this paper

READ PAPER
Academia.edu

Leonardo da Vinci's Madonna of the Yarnwinder. A Historical and Scientific Detective Story

Leonardo da Vinci's Madonna of the Yarnwinder. A Historical and Scientific Detective Story

Although every effort has been made to establish copyright and to contact copyright holders prior to printing, this has not always been possible.The publishers would be very pleased to rectify any omissions or errors brought to their notice at the earliest opportunity.There are many more people who have also given us a helping hand along the way. That the book has finally appeared serves as a collective vote of thanks to all of them. input throughout from Thereza Wells, who is effectively its co-author. Since the 'detective story' is told in the first person by Martin Kemp, we might best say that it is written 'with' Thereza Wells. In the later parts, which contain a detailed account of the history of the pictures, including provenance and photographic documentation, the division of labour is almost precisely reversed, with Thereza Wells undertaking the lion's share of the responsibility for the research and writing.However, we assume joint and undivided responsibility for the book as a whole, as equal co-authors. PROLOGUEThe Madonna of the Yarnwinder is among the most clearly documented of Leonardo's paintings. We know that it was commissioned in Milan by The lists were drawn up in 1525 so that Salaì's assets could be divided equitably between his sisters following his death the previous year. 2 If this is the Madonna of the Yarnwinder -and there are no other obvious candidates -Leonardo must have produced two paintings of this subject, one delivered to Robertet and one remaining with the artist. As we will see, this is consistent with the existence of two prime versions. For twenty years or so the little composition was as influential as anything that Leonardo ever painted, affecting the whole course of the portrayal of the Madonna and Child in High Renaissance art and inspiring many imitations. Then, the historical trail becomes problematic. It is difficult to trace either of the Madonnas after they left the hands of Robertet and Salaì. Of the very many copies and variants that survive, it has long been recognised that two are of major account. One has since 1756 been in the possession of the Dukes of Buccleuch. The other was once in the distinguished nineteenth-century collection of the Marquesses of Lansdowne. Pictures are often named after their original patrons or particularly notable owners. In neither case is the first option currently available. The best designations -so that we do not habitually have to refer to this 'version' or that -are the 'Buccleuch Madonna' and the 'Lansdowne Madonna' , and these are the names adopted in this book. All these enterprises and events bring with them their own interests and 'angles' , and all leave some imprint on the collective perception that determines how we look at the paintings. It seems more honest to disclose the complexity of the intersections of interest and knowledge in this particular case, in order to show how the various hypotheses have arisen and how the conclusions have assumed their present character. Thus we have included more about the making of the 1992 exhibition at the National Gallery of Scotland than would be customary, including things that would not normally be considered 'relevant' . The whole 'industry' of the exhibition is as much a part of the history of the image in its changing public identities as any aspect of its patronage and provenance. We give an account of how the knowledge is coloured by the human contexts in which it was generated. Writing up research in this way is something of an experiment, but the spirit of experiment is in keeping with the nature of the enterprise and is consistent with Leonardo's own tastes, as will become clear in what follows.We are aware that this approach involves some shifts of literary gear, as we move rapidly through passages of narrative only to be faced with some sustained passages of historical analysis, pushing our way up hills and over bumpy terrain. But a long journey across a flat plain at the same speed can be very tedious.The more personalised account, much of it cast in the present tense, should, at the very least, convey some sense of the sheer delight and excitement of undertaking state-of-the-art detective work on a fascinating painting, in collaboration with some remarkable experts.The histories of the pictures delivered to Robertet Our account of the prolonged detective work, spanning almost twenty years, is written as a personal narrative. As we have indicated in 'Background' the main story in the first part is cast in the first person by Martin Kemp, though the endeavour was fully shared by Thereza Wells (née Crowe) almost from the beginning. Sober and objective historians are not meant to write up their research in this way; however, we have decided to break the unspoken rule for two reasons.The first concerns the accessibility of complex material. The Lansdowne Madonna was subjected to a battery of eleven scientific tests, which involve some inevitably difficult technical details unfamiliar to non-specialists. The Buccleuch Madonna has been less extensively examined, but has been the subject of X-ray and infrared tests. By telling the story of the paintings' involvement with the ever more refined and elaborate processes of examination, the nature of the tests, their ways and means, their aspirations and the results gradually become clear; a barrage of technical terms, data, diagrams and images that need specialised decoding would only be daunting.The issues involved in scientific examination, and how they can be intermeshed with the questions tackled in more conventional art history, emerge more naturally through the story of their interaction.The second reason is that the knowledge produced by any piece of research bears the imprint of the manner of its production. This is normally concealed behind the façade of sober historical or scientific language, which suggests that the knowledge just 'arrives' as an independent body of fact. Studies in the history of science have turned their attention to the modes of the production of knowledge in laboratories, and looked at how these modes are present tacitly in the published results. 3 THE STORY BEGINS: THE SCOTTISH PICTURE 19 18 FORENSICS as a Trustee of the National Galleries of Scotland I was well placed to seed the idea. I did so because I was notinclined to accept what was then said about the Duke's picture and its most notable rival, a painting in private hands, generally billed as being in a 'New York Private Collection' . The standard view, inasmuch as any set of opinions on such a contentious attribution can be said to conform to a standard, was that among the many copies and variants of this highly popular composition the Buccleuch and New York pictures were the finest copies of a lost original. But there had been occasional claims that both were products of Leonardo's own hand, or alternatively of him working with his studio, that is to say produced by assistants or pupils working under his supervision. And a few had advocated Leonardo's sole or part authorship of one or both of the paintings. The fluctuating history of the attributions will later warrant full review in its own right.The three of us look first at the picture by eye. We can readily see the discoloured varnish, some signs of damage and some scattered retouching. It is evident that the paint surface is in good condition for a picture of its age. Under strong lighting, the image comes vibrantly to life. The back of the panel, which is quite thin, reveals inscriptions and labels, which tell something of its later history and the exhibitions of which it has been part. It is also pock-marked with shallow holes, which, it is generally said, were made to facilitate penetration of the wood by chemicals designed to resist infestation, above all by woodworm. Indeed, there is some evidence of localised woodworm damage around the edge of the panel, but the main areas are unaffected. However, it seems more likely that the holes were drilled to discourage warping after the panel had been thinned to its present thickness.Our intention is to go much further than the unaided eye. The Madonna is to be subjected to scientific examination. But before going any further with our story, it will be helpful to give some background on the development of the scientific techniques which have shed light on old masters during the twentieth century.The science that can be applied to paintings made great strides in the second half of the twentieth century.The relatively new and rapidly improving technique of infrared reflectography (IR) was yielding particularly striking results with Renaissance paintings. The longer-established method for attempting to catch a technological glimpse of what was happening beneath the surface of old master paintings was X-radiography.We had X-rayed in the Western Infirmary the Chardin paintings in the Hunterian Museum of the University of Glasgow, using the same equipment that revealed the entrails of human patients. X-rays pass through much of the surface of paintings, as well as through the panels and canvases on which they are painted, though it is often possible to see ghostly traces of wood grain of panels and, in the case of canvases, of the thicker priming in the interstices of the weave. Where a substance is opaque to X-rays it shows up clearly on the plates -as light in the negative (or dark if reversed as a positive image). Nails used in the construction of panels or to secure canvases leave sharp silhouettes.White lead or other pigments composed of elements with heavy atomic weights register clearly on the X-ray plate because they are X-ray opaque; lead was 2. THE STORY BEGINS: THE SCOTTISH PICTURE At 10 a.m. on Wednesday 16 May 1990, John Montagu Douglas Scott KT, the 9 th Duke of Buccleuch and Queensberry, arrives at the new premises of the Conservation Department of the National Galleries of Scotland at Belford Road in Edinburgh. This is where the Scottish National Gallery of Modern Art has recently opened to the public in the elegantly refurbished George Watson's School, a classical building with grand portico worthy of the ' Athens of the North' . In the back of the Duke's Land Rover is a beautifully made wooden box, lined with velvet. The precious cargo is carried gingerly through the security doors of the special entrance to the Department, which is housed in the basement. In the conservation studio, John Dick, the Chief Conservation Officer, supervises the removal of a picture from this protective case. It is small -painted on a panel that is a little under 19 inches high and 14½ inches wide (48.3 × 36.9 cm) -and set within a fine gilded frame. The picture is one of the two best versions of a composition undoubtedly conceived by Leonardo da Vinci and known as the Madonna of the Yarnwinder.It was so named because of the cruciform object held by the infant Christ, which is recognisable as a shaft on which yarn was traditionally wound.It has also been called the 'Madonna of the Distaff ' . Oddly, the Italians generally call the composition the Madonna dei Fusi, that is to say the 'Madonna of the Spindles' , although no spindles are actually visible. It should be the Madonna dell'Aspo. As it happens, however, spindles will come to feature prominently in the scientific story as it unfolds.Working in Scotland, first in Glasgow and then in St Andrews, I had known John Dick for some years, first as the assistant to the redoubtable Harry Woolford, a conservator of great tact and sensitivity. John was a worthy heir to Harry's judicious skills but did not aspire to emulate his mentor's colourful personality. I had suggested to the Duke and to John that the painting would be worth examining by the latest scientific techniques. Having served earlier THE STORY BEGINS: THE SCOTTISH PICTURE 21 20 FORENSICS and still is used as a barrier to protect us from nuclear radiation. In the case of one of the small pictures by Chardin we could see where the artist had originally sketched the cellar boy's left arm in a different position, using white lead among the other pigments. 1 The boy originally appears to have been holding up a glass of wine for scrutiny. This was later covered over, but the painter's secret pentimento or change of mind had been uncovered by X-ray analysis. Pentimenti are much sought-after by art historians, since they are often taken as a sign that the painting in question is the 'original' , not a slavish copy, and they were to feature prominently in our subsequent quest.Another of the Hunterian Chardins, the Lady Taking Tea, was being conserved by Harry Woolford in Edinburgh, and when the relining canvas (added by an earlier restorer to reinforce the original canvas) was removed, an original inscription by the master was discovered, dating the painting to 1735. 2 The date on the front of the painting was only semi-legible and had been read in various ways.The New York picture (i.e. the 'Lansdowne Madonna') had been already examined by X-radiography and with ultraviolet light in the early 1930s by Wilhelm Suida, Robert Bigenberger and the artist R. Maurer. 3 Suida correctly stress-ed that 'no historian in matters of art can afford, especially when important works of art are in question, to ignore the modern techniques at our disposal' . He reported that several pentimenti were visible in the X-ray images: Christ's left shoulder was overlapped by his head; his right hand had been adjusted; his left cheek was less 'heavy'; his body was fuller; the space between his left thigh and abdomen was less compressed; and the Virgin's robe was cut lower; but there was no trace of the basket of spindles described by the first eye-witness. What Suida observed were areas in which the white lead used in the flesh tones recorded changes of mind relatively late in the painting process.Looking at the reproduction of Suida's X-ray in his published essay, we can confirm most of the changes he noted, though the position of Christ's right hand is difficult to read ( fig. 1). The right profile of the Virgin's head and her hairline also appear to be distinctively different in the X-ray. The lack of definition in the lower parts of Christ's legs suggests some changes of mind by the artist. Although making stylistic judgements in X-rays is a hazardous business, some of the passages appear quite heavy-handed, and may reflect assistants'interventions before Leonardo resumed control of the finished picture. We can also observe the outline of the stretcher and the profiles of the nails that are attaching the lining canvas to it, following the removal of the paint layers from the original wooden panel earlier in the century. We will return to the conservation history of the 'Lansdowne Madonna' in due course.The examination with ultraviolet light, which is good at picking up re-touchings, indicated that the Lansdowne Madonna had not suffered from too much over-painting, but that the Virgin's extended palm, the Child's right cheek, the lower part of his body and left arm showed clear signs of later intervention. Suida's analysis of the ultraviolet examination is discussed later (pp. . He aligned the features he observed in Vinci, 1939. 'the first rough sketch' as revealed by X-radiography with a drawing in the Uffizi, which he attributed unconvincingly to Raphael. We were later to realise that the believed that the Buccleuch picture was entirely the work of a pupil. As will be seen, the later X-ray exami-nation that we conducted on the Lansdowne Madonna did not disclose what Suida described, but it is likely that subsequent conservation procedures (most notably its re-mounting on dense composite board) now prevent X-radiography from revealing much of what it did previously.The Buccleuch Madonna had also been X-rayed in Milan in 1938 for a major Leonardo exhibition (fig. 3) and again for the exhibition in Edinburgh. 4 The X-ray images prominently feature the succession of shallow holes that had been drilled in the rear of the panel. There are clear signs of a canvas imprint in the Buccleuch picture, which indicates that Leonardo adopted the common expedient of using a fine canvas or canvas strips to stabilise the ground over the panel.There were no obvious pentimenti visible in the poses of the figures, though the area around the Virgin's head bears witness to some manoeuvring, as John Dick has noted:the somewhat flattened back of the head has been caused by clumsy outlining. Also an intruding tongue of paint at the back of the neck gives a bobbed appearance to the hair, distinguished by a now hidden, crackle pattern. Otherwise X-ray opaque paint is notable throughout with idiosyncratic crackle patterns. 5 The absence of any clear signs of the landscape in the X-ray images is consistent with the idea, to be explored later, that the background was added relatively late in the making of the picture. For the flesh and other lighttoned features, relatively sparse quantities of white lead were mixed thinly with other pigments in the binding priming, are picked up by a camera sensitive to their wavelength. At the time when the Buccleuch Madonna was in the Edinburgh studio, the results were recorded on a screen in the same way that cathode rays created pictures on televisions at that time. The image was composed of a series of lines that were sufficiently fine to convey the effect of a reasonably continuous picture when viewed from a suitable distance. As the camera moves over the surface of the painting, a sequence of images flickers into view, each recording only a portion of the painting (figs. [4][5][6][7][8]. To record the reflectograms the screen is photographed for each exposed patch. In 1990 the only way to assemble a complete picture corresponding to the whole painting was to assemble a photographic mosaic, which, like any mosaic, consists of a collage of abutting parts with the seams or joints readily visible.Looking at the resulting images, whether on the screen or in printed form, involves a good deal of skill in selective viewing. With both X-rays and infrared, the features of interest may be relatively indistinct and fugitive, and need to be teased out from the general visual 'noise' . The same applies to an X-ray of a patient to whom the doctor carefully demonstrates how the image is to be seen or 'read' . For Bellini in Venice was able to achieve comparable levels of visual magic in oil paint at this time.By contrast with X-rays, infrared reflectography picks up different signs from the underlayers of a painting.Invisible light emissions of a wavelength of 0.7 microns to about 1 millimetre, lying immediately beyond the boundary of red in the visible spectrum, are able to penetrate some of the pigment and binder mixtures.Where they are able to penetrate, they rebound from the white gesso ground that was almost invariably used in the Renaissance. Gesso is a plaster made from gypsum (calcium sulphate dihydrate) bound with size, and was applied layer by meticulous layer first to panels and later to canvases to provide smooth and brilliant surfaces on which to lay colours. Leonardo, as will become apparent, appears to have covered this already white layer of gesso with a coating of white lead to enhance its brilliance and perhaps to reduce its absorbency. Infrared radiation is absorbed by carbon-rich pigments, most notably those used for underdrawings in such media as charcoal and ink rich in carbon black. As soon as infrared analysis began to be used on early paintings, a remarkable and surprising series of complete or partially visible underdrawings emerged, unseen since the very earliest stages of a painting's development and perhaps only previously seen by the artist himself.The IR technique involves a device that emits infrared radiation, much as a torch emits visible light. As far as we know, neither X-rays nor infrared have any detrimental effects on paintings. The reflected rays, having passed through the pigment layer and bounced off the The results are exciting and encouraging, since they suggest that the Duke's picture was not simply a routine copy, transcribed feature by feature or even traced from a prototype. The presence of pentimenti is not definitive in proving that the work in question is the 'original' , or even painted by the master himself.A copyist may well manoeuvre for position as he or she lays down the composition, either to manipulate the contours into the right position or even to make deliberate changes. However, the adjustment to the Virgin's thumb, and its relationship to other versions, argues against the pentimenti being mere copyist's adjustments. Nothing conclusive at this point, but what we found certainly provides grounds for continuing the research.The provisional airing of the findings takes place a few months later, on 25 Mantua in February 1500, and stayed there for a total of less than three months, but it was a stay that has important consequences for our story, as the next chapter will show.sharply foreshortened hand, a particularly difficult thing to render convincingly, has been subject to careful adjustment ( fig. 7). Perhaps most surprisingly, and certainly most intriguingly, her thumb has apparently been more tucked in towards her palm. This is recognisable as a feature that occurred in other variants, which as a group did not seem to be notably close to either the Buccleuch or Lansdowne versions ( fig. 108).It is also apparent in the Uffizi drawing ( We can indeed confidently infer that Isabella admired the portrait since she asked him to portray her. The immediate result was a drawing on a large enough scale for direct transfer to a painting. 3 The Marchioness is shown in profile, a consciously more formal pose than that of Cecilia, and pointing to a book that is resting on a ledge, details that are more visible in an assistant's version in Oxford (fig. 9) than in the Louvre original, which has been trimmed at its base. 4 When Leonardo departed for Venice, he left one drawing with her (probably the one in the Louvre), and took another version with him (possibly the one in Oxford), presumably to be used as the basis for a painted portrait in due course. The drawing Leonardo carried with him was seen and admired in Venice. 5 Once Leonardo was settled again in Florence, where he had been trained and where he forged his early career, Isabella began to apply pressure. She asked Fra Pietro to report on what Leonardo was doing, because it was already well known that the artist scattered his talents across a wide field of practical and cerebral pursuits, and that finished paintings were hard to come by. This was the beginning of a campaign to pester Leonardo over a number of years. One reason why she was keen to press the painter was that her husband had given away Leonardo's drawing of her, apparently to her dismay, and that she needed a replacement, either another drawing or, ideally, a painted version. 6 She also harboured a desire to possess a painting by the renowned if difficult master on another subject. She had shackled some of her commissioned artists with pedantically precise instructions, presenting Perugino in Florence with a very detailed programme for aBattle between Chastity and Lust, and even sending the harassed painter a length of string to show how tall the largest figures should be. 7 However, she had come to understand that some of the new breed of artists were more reluctant to be told precisely what to do. Bellini, in particular, was quoted as reacting badly to Isabella's instructions because 'his way of working, as he says, is to wander at will in his pictures, so that they can give satisfaction to himself as well as to the beholder' . 8 In Leonardo's case, she knew that she would do well to get a painting of anything, and three years later she had been forced to settle for the vague Leonardo's portrait of her to compare with portraits by Bellini. 1 Cecilia, a lady of considerable cultural accomplishments who was by now married to a Lombard nobleman, wrote elegantly to comply with Isabella's request but warned her that the portrait did not look like her. 2 She explained that the fault was not the painter's but that it had been made some years ago when she was 'unformed' (imperfecta) and that she had since changed. It seems likely that Cecilia's portrait, shoing her in a gracious, turning pose, with a sidewardglance and slight smile, cradling an ermine symbolic of purity and moderation, was sent to Mantua for the Marchioness's informed scrutiny. We may well imagine that she was greatly taken by the beauty and innovative complexity of Leonardo's characterisation of his courtly sitter. Given Isabella's interest in Leonardo's art, it was not surprising that he should be welcome in Isabella's city after Ludovico's fall. 30 FORENSICS suggestion that he might like to provide an image of unlike the more prosaic rendering of such features by even his most able followers. And there was one feature I had not observed previously. A startling series of vivid, blood-red filaments of yarn snakes across the strata of rocks on the right, superbly executed with both minuteness and vitality ( fig. 18). However, the rocks themselves were painted in a routine way -very much as standard rocks of a generally Leonardesque type -betraying none of the brilliant geological observations evident in the Buccleuch strata.Even without exercising elaborate judgements of connoisseurship, it becomes apparent that the relationship between the two pictures is complex. Some features in one painting speak of higher skill and thought than are discernible in the equivalent passages in the other, while in some respects the other painting exhibits higher accomplishments. It was clearly not a simple matter of original and copy, nor, probably, a matter of two uneven copies. But this latter point still needs to be investigated. Among the drawings connected with the composition, only one is indisputably autograph, that of the torso of a turning woman in Windsor, executed delicately in red chalk and silver-point on pink prepared paper ( fig. 20). 8 The subtlety of motion and modelling, firm but soft -in keeping with his insistence that the anatomy of women and young people should not be characterised too emphatically -spoke of its autograph status, and it has never been doubted. What has occasionally been questioned is whether it is a direct study for the Virgin, since it does not precisely match the pose in the painting. My view is that it is a study from life, of the kind he recommended painters to make once the composition had been resolved, and that it was made specifically with the painting in mind. CREATING THE CONTEXTNaturally enough, it needed to be subject to adjustment when it was fitted into the underdrawing for the painting on the panel.Leonardo being Leonardo, we will not be surprised to learn that the observation and subtle recording of gradations of light and shade on such a form is not enough in itself. The rule that determines the levels of light and degrees of shade needs to be understood. The rule he formulated is that the intensity of light was proportional to its angle of impact; that is to say if it strikes its 'blow' or 'percussion' at right angles to the surface it will exercise its greatest effect, whilst a glancing blow an intensity, moistness and translucent sheen that allows it fully to sustain its pivotal role in the narrative (fig. 15 Passing across her forehead, it catches the light and appears paler than her skin, while it assumes the guise of a dark filter against the brilliance of the pale mountains. These are just the kind of relative effects of light and dark that he spent so much time discussing in his notebooks. The lumpy clumps of trees correspond precisely to his discussion of how detail is progressively lost as we see trees at increasing distances. The closest trees above the Virgin's knee, which are already quite far away, exhibit selective highlights on clumps of leaves, while the more distant ones are just modelled in generalised shapes ( fig. 17). Even in the areas that were quite abraded, such as her foreshortened hand or the modelling of her chest above the delicate fringe of her dress, it is possible to discern a magical subtlety both in the main study, drawn with much subtlety, and also in the more summary explorations of the angles of the Virgin's head, neck, shoulders and torso in the left margin of the sheet. 10 itself to the body below. 16 The complex passage of the blue swathe of drapery across the Virgin's shoulder, around her waist and over her bent knees, responding to her twisting motion, would have necessitated a comparable study or studies, as would her hair and headdress ( fig. 28). 17 Leonardo also avidly studied rock formations, and he could have drawn on his bank of images for the rocks on right, below the foot of the yarnwinder. 18 Similarly, he might well have to hand drawings of the precipitous peaks that he favoured when he wanted to endow a background with a 'timeless' feel (figs. 29 and 30). 19 Although they look somewhat improbablemountains of the mind rather than of reality -they are concocted from observations of geological features he is known to have studied, including the mountains north of Milan and elsewhere in Italy, and some strange vertical formations in the area of Arezzo in Tuscany, that we can glimpse from the train as it passes between the Tuscan cities. 20 ing already supplied to one of his courtiers, but the Renaissance attitude to such transactions was very different from our own, and it was not uncommon for a contract to stipulate that a painting should follow the 'manner and form' of an admired prototype. Odd though this document was, in form, content and manner of publication, we are inclined to think it genuine, and it is duly discussed in the Edinburgh catalogue.However, doubts were beginning to be expressed by the art historian who was undertaking fundamental research into Robertet, Reluctant though any historian is to surrender any attractive and apparently informative document, she questioned its authenticity. We have since come to share her doubts and would not now draw confidently upon it to illuminate Robertet's collection. We agree that it is as much a work of fiction as fact. It seems to be a fictionalised construction, assembled from elusive cores of fact to which Chesneau had access. In any event, it tells us nothing about the Madonna of the Yarnwinder.In order to set the context for Robertet's commissioning of the Madonna for the Edinburgh exhibition, we successfully request a series of items directly con- What could these conjunctions in the underdrawings mean? There seem to be two obvious possibilities.Either the New York and Edinburgh versions were both based on the same lost prototype -a finished painting, a shared cartoon, or developed drawing; or they underwent parallel developments in the studio. At this stage, it is the manifest quality of the quick underdrawings (seemingly not copied or transferred in a pedestrian manner from any prototype) and of the finished paintings that favours the latter possibility.This startling new technical evidence emerges as the catalogue is nearing completion, and not all its implications can be fully digested and investigated. Further, we have not taken on board the clues about a figure group in front of the arch. What is apparent, however, is that the 'New York' version seems to have under-gone more radical changes than we had managed to observe in the Buccleuch picture -though drawing firm conclusions from any comparison is unsafe, given the different capabilities of the equipment involved in the examinations. In any case, neither picture shows all the customary signs of being a routine copy or studio replica. In the section of the catalogue discussing the results of the IR examinations, I argue that the two paintings had essentially been produced alongside each other in the workshop, one to fulfil Robertet's commission and the other to be sold to a suitable client when the opportunity arose. This means that there was no lost 'original' , and that the whole concept that there must be just one autograph version of a picture, with all others relegated to the status of copies by followers, is deeply flawed. It is a concept that meets the needs of the trade, dealers, auction houses, collectors and curators, who want to think that they have access to the exclusive 'original' . It does not correspond to the nature of the production of small-scale devotional pictures in the studios of those Renaissance masters whose works were in great demand.If this interpretation is right, the questions remain as to how the pictures were actually produced -what roles were played by Leonardo and his assistants -and which one was delivered to Robertet in 1507. Answering the first question involves connoisseurship and the second requires research into the provenance of the paintings. EYE VS SCIENCEConnoisseurship is not an exact science or even a craft that inspires consensus, as can be seen from the dis- During the court proceedings, which at one point decamped to Paris to undertake a direct comparison between the two pictures, Duveen's defence lawyers paraded a series of art experts to bolster his assertion. 4 These included Berenson, who was being paid handsomely by Duveen for attributions that enhanced the value of paintings in his hands. Beginning his testi-mony with confidence, asserting that he knew because he knew, Berenson found his case gradually unravelling under cross-examination from Hahn's insistent attorney, who was little impressed by the intangibility and lack of intellectual rigour with which Berenson expounded the basis of his expertise. Berenson was particularly vulnerable to the charge that he ignored all forms of technical examination (then in their infancy)and was not concerned about media, supports and all material aspects of a painting's construction, about which he boasted that he had 'not the faintest idea' . 5 For the great attributer, 'eye' was all.Connoisseurship fared badly in Berenson's hands under legal scrutiny. Duveen eventually settled out of court, but the Hahn family were still not able to find a buyer for their painting. It has recently been sold at Sotheby's for the not inconsiderable sum of $1.5 million, under the correct designation of 'copy' . 6 It was probably made by one of the major French court artists, and is of very high quality.Few if any historians would now dismiss the value of technical evidence in support of an attribution, but too many still do not give it enough credence.The battery of tests that has become available since Berenson's fumbling court testimony has increased vastly. X-ray images have improved; infrared reflectography has made a big impact; pigment analysis has made great strides, using both techniques that rely on small paint samples, such as gas chromatography and microscopic examination, and non-invasive methods such as spectral analysis. The non-invasive methods will feature prominently in later sections of this book.This does not mean that science answers all the attributional questions. It can be definite in precluding the possibility that a particular work is by the master in question. If, to take an obvious example, the predominant yellow pigment in a prospective Leonardo is identified as chrome yellow, which was one of the many new colours synthesised in the early nineteenth century, rather than the lead tin yellow pigment used in the Renaissance, the painting cannot be by him.Often, the scientific analysis only produces a nil obstat, that is to say that there is nothing that stands in the way of the painting being by Leonardo or an artist of his time. Even the most positive kind of evidence, such as the notable pentimenti that we were seeing, does not produce absolutely conclusive arguments.The underdrawing in the 'New York' picture could have been done by a pupil or follower, and, even if by Leonardo, none of the final painting might have been undertaken by him. Only if we can identify technical characteristics shared by definite Leonardo paintings and not shared by works of even close associates can we begin to work towards secure conclusions on attribution using scientific evidence. This will only become possible with the assembly of much more data than we currently have available, and this was the aim of the Universal Leonardo project, about which we will shortly be hearing more. Even then, the final judgement that this painting does or does not 'look' like a Leonardo will never quite go away as one of the criteria in constructing his oeuvre, that is to say the body of work generally accepted as being by him. to completion with Leonardo's participation in their design and his supervision (at least) of its final execution in such a way as to ensure that the execution was 'assai valente' [skilful enough] to be fully acceptable as works by Leonardo. 7 Reading the Edinburgh catalogue again now, it seems that the arguments were heading generally in the right direction, though in retrospect I think Leonardo's direct participation in the actual painting of both was more thorough-going than I was prepared to argue at that stage. I was also inclined to think that the Buccleuch Madonna was the earlier to be completed, given the 'late' quality of the Lansdowne landscape, and that it would have been the picture delivered to Robertet in 1507. Determining which picture might actually have been delivered clearly has a crucial role to play alongside the visual judgements. What we now know about the paintings' histories is the subject of Part III of this book. COMMERCE AND CLAMOURA great deal of work and effort goes into the organisation of any exhibition that involves old master paintings and drawings. Various organisations and personalities play key roles in determining what is and what is not in the show. And the way the artworks are actually exhibited affects how we perceive and understand them.Exhibitions are expensive and they are big business, though this one is relatively modest in scale. A budget is set, a little below £ 70,000 (a sum that now seems tiny), with a sponsorship target of £ 50,000. Internal staff costs are not included in this total. It has been decided not to charge specifically for admission, with the intention of raising the necessary income from sponsorship and sales. Much of the elaborate preparation and administration involved in mount-ing an exhibition is not apparent to exhibition visitors, who simply see some items that have been assembled, hung and located in cases. Assembling the items is itself a long affair. Diplomatic visits often need to be organised long before the loan request is submitted. I need to produce 'justifications' for each of the prospective loans, which the Gallery can incorporate into loan request letters. Unexpected issues arise. One anxious lender fears that it might be a 'selling exhibition'; that is to say, the Duke will use it to promote the sale of his picture. We assure the enquirer that nothing is further from the truth. As it happens, the owner of one of the lesser versions of the Madonna offers to sell it to the Duke, an offer that he unsurprisingly declines.The publicity machine is set in early motion, and stories are developed to attract the attention of the media. In this case, pre-exhibition publicity tends to well up spontaneously. A number of early stories centre upon the 'news' that the Duke's picture is to be examined. As early as 22 COURIERS AND CASHOwners of agreed loans might ask for conservation costs to be met, and charge an exhibition fee. Expensive specialist packers and transporters are employed, and costly insurance arranged. Curatorial time is spent inspecting the items before dispatch and doing a report on every detail of their condition, in case a dispute later arises about whether damage has occurred and who is responsible. Many works are only lent on condition that they are 'couriered' , that is to say accompanied by someone from the gallery or owner who will be with their work from 'nail to nail' , watching over it from its dispatch to its safe location in the exhibition. The processes will be repeated at the end of the show.Curators in the collections might bid to be couriers to the most attractive venues and vice versa. Paying for a courier's air travel, with a neighbouring seat reserved for the artwork, and meeting the costs of good accommodation for them until the show opens is not cheap, and the accumulative cost of many couriered works can be formidable. New display cases might need to be provided to meet the lender's requirements (perhaps with special temperature and humidity control). Light levels are specified, typically 50 lux for works on paper, set in relation to the eye's acuity at what is essentially a compromise level, above which extra light produces relatively small gains in legibility. To embark on this process, we have to be sure that bringing the star pictures together serves a real purpose. The scholar needs to have a persuasive turn of phrase.In the gallery itself, a designer or designers need to be employed (in-house or from outside firms) for all aspects of the exhibition design, from wall finishes to graphics. The same designers might or might not be responsible for publications, brochures, posters, banners and so on. Branding is important. A catchy title is We look forward to raising our glasses in celebration at the opening of the exhibition. ' Although we do not promise Martini that a 'new' Leonardo masterpiece will be revealed, there has to be a clear sense, arising from our research, that we are not showing a series of studio remnants. SETTING UPThe galleries earmarked for the show need to have their existing contents removed, walls refurbished and repainted, and the whole lighting system adjusted or even renewed. Special handling teams will be involved in unpacking and hanging or placing items, under the fretful eyes of the couriers. Detailed condition reports will again be taken and agreed, so that any new scratch or previously un-noted blemish is duly recorded. Guest and in-house curators will discuss and sometimes dispute the best location for each item. A carefully hung painting will suddenly be seen to be in the wrong place, once other items are positioned. Where something is hung, and the company it keeps, can radically affect what it looks like. Dealers fully realise this when they display paintings to potential clients in rooms set up specifically for the purpose. The handling teams need patience as the curators change their minds.There is nervousness that someone more senior, perhaps the Director, might arrive and say that we have everything in the wrong place. As it happens, Timothy Clifford is away so that the Keeper, Michael Clarke, with my intermittent advice, proceeds on the basis of his own good judgement.At the same time, special temporary shops need to be built at the entrance of the exhibition rooms, with stands for books and less easily displayed materials.The shops will of course need special staffing during the show. Security is reviewed, with constant surveillance during the night. The scene in the days leading to the actual opening is invariably full of frenetic activity, crises, stress and short tempers. Most items do not arrive early, and late arrival triggers anxiety. On 13 May, less than 36 hours before everything is due to be complete, the debris of installation is everywhere. As it is being removed, and everything made impeccable, press photographers and some early-bird reporters arrive. The photographers want to set up their photographs carefully, and to take many shots from which they will choose the best. Exhibitions tend to bring out the arty instincts that press photographers regularly have to suppress. They take a long time to arrange the set-up, with their lights on tall stands. It is easy for a scholar to dismiss the commercial and promotional aspects of the whole exhibition process as, at best, a tiresome necessity in the face of the inadequate funding of our public galleries. But, in essence, Robertet and Do such details properly belong to the historical story? What I am doing here, rather than conventionally presenting only the results of the show, serves to acknowledge how the past and present histories of an artwork belong to a continuous fabric that tells of its commissioning, reception, collecting, sale, financial value, critical reaction and public viewing over the years. We were not intervening in the material fabric of the pictures, but we were effectively altering how they look, and everyone, from scholar to sponsor, plays a role in that collective process. It is a process in which many people have different kinds and degrees of interest.The exhibition opens to the public on 15 May 1992, precisely one month after Leonardo's birthday. The story that most of the papers would like to run is the discovery of an unknown Leonardo, or at least the recognition of a little-considered picture as a masterpiece by the painter of the world's most famous work of art. In this context it is necessary that the Buccleuch and Lansdowne paintings should be locked in mortal combat, in which the attribution of one rises triumphantly over the vanquished body of the other. The feature writers want to cite the black-and-white opinion of the expert to tell the public which is and which is not by Leonardo. We have explained carefully that neither picture is the previously 'lost original' , because there never was such a thing in the terms meant by that phrase. But academic subtleties do not make good press stories, and the framework for the pieces that appear on the news and features pages tends to play to stock expectations, like the stereotype of the Rembrandt found in the attic of a council house.Drumlanrig is hardly a humble attic, but the principle remains the same. I have dealt enough with the press over the years to understand how reporters have to do their job, and a story with a misplaced emphasis is better than not being noticed at all. At least I always ask the reporters to read back any direct quotes they might be using. The national and international coverage is far greater than a modestly sized show in Edinburgh might expect.The Scottish press is naturally keen to think that the exhibition provides the apotheosis of the Buccleuch contender as Scotland's genuine Leonardo. The his own earlier thoughts -that the faces look more in proportion when viewed from a glancing angle rather than when viewed flat on. 6 Although Leonardo knew about the trick images, called anamorphoses, that were designed perspectivally to be looked at from shallow angles, there is no evidence that such optical games were ever played with devotional images in the Renaissance, and I am not inclined to pursue this particular suggestion.An important review is commissioned by Apollo from Cecil Gould, senior Leonardo scholar and former curator at the National Gallery in London. Gould acknowledges that 'scientific material which became available at the eleventh hour demolished once and for all the theory of "either/or", and showed that the situation was far more complicated' . He cleverly identifies the still puzzling traces of figures at the base of the arches in both reflectograms as an 'embryonic Nativity' , but repeats our failure at this stage to make the tie-in with the baby-walker in the Edinburgh picture. Gould also proposes the ingenious but overcomplicated theory that 'what Fra Pietro saw was neither Leonardo's painting nor the definitive cartoon but rather a preliminary cartoon' . The same magazine later carried a characteristically thoughtful piece by Michael Bury of the University of Edinburgh, which contributed to an understanding of the paintings' content, while trying to sustain the old argument that all the versions stemmed from a lost prototype.In his earlier monograph Gould had recognised the qualities of the Buccleuch picture, arguing that it 'has long seemed to me to be demonstrably of Leonardo's execution in the main. The reason why it is not uni-versally acknowledged as such may be that the image of an autograph Leonardo painting of his maturity has been so greatly elevated that hardly any existing picture could live up to it. ' 7 He now suggests that the Lansdowne Madonna was the one delivered to Blois, and that Charles d'Amboise asked for the Buccleuch version to be finished for himself. Working on the principle that he regards himself 'more as a lifetime protector rather than an owner of one version' , the Duke outlines the advantages of retaining some important works in the hands of private owners whose houses are open to the public, not least because the visitors to historic houses are not identical to those who attend public galleries. He also points out that seeing pictures in the room setting of a great house sets them within 'living history' rather than placing them in 'sanitised isolation' .A visit is arranged for Prince Charles. It is part of his shoehorned schedule, orchestrated by clock-watching courtiers. He seems genuinely interested, having some talents as a watercolourist and a known appreciation for historical art. He has been briefed, of course, and knows that no less than nine of the exhibits are owned by his mother. He discusses the items on display in an engaged manner, which seems to go beyond a royal's normal professionalism in looking interested in whatever they encounter. He makes particularly perceptive comments about the use of silver-point in the drawing of the woman's torso ( fig. 20). When hurried away by his programme managers, he shows some reluctance to leave but a schedule is a schedule, and that is that.The exhibition remains open until 12 July, by which time it has attracted 66,155 recorded visitors. The supporting programmes run throughout the exhibition, including talks by myself to business and other interest groups that the Gallery considers to deserve special attention for one reason or another. FORENSICS CONCLUSIONS AND CONSEQUENCESFor the curators, the exhibition serves a special purpose, beyond the inherent satisfaction of seeing a public project realised and well received. The ultimate justification for subjecting objects to travel and exposure together is that each will gain (or even lose) in varying degrees from the others. The contextual material in the 'Robertet room' is good to see, and sets the stage effectively.But it is the 'Leonardo room' that involves more active and sustained looking. Looking at the two main protagonists side by side is deeply informative but poses problems. As we have noted, the one is bright and clean, but with a CONCLUSIONS AND CONSEQUENCES 71 surface that has suffered from various abrasions; the other is mellow in an old-masterly way, and its underlying paint surface has suffered relatively little. As it happens, the illustrations of the two paintings on facing pages in the catalogue underplay the differences, since someone at the printers has decided to bring the tonal range of the Lansdowne picture closer to that in the Buccleuch Madonna. On the other hand, the strong, even and equal lighting in the gallery enhances the differences. Somehow the Buccleuch painting continues to look more like the Leonardos we know, veiled by dark varnish and muted by the patina of time. Yet the Lansdowne picture asserts its special qualities, and I know that its colouristic vitality is matched by what is emerging from the cleaning of the Last Supper.A series of close comparisons is made, always trying to filter out the superficial differences resulting from their condition. The process of production is more in keeping with the commissioning of a superbly made chair from a major designer-craftsman than it is designed to facilitate the art historian's conventional task to tease out different 'hands' in a painting. We do not ask if a certain glued joint in the chair was made by the head of the workshop or one of his assistants -providing the joint holds and looks good. Our collector-like concerns with precise nuances of who laid on which brushstroke are out of kilter with the context of the works' production, and the procedures set up by Leonardo are designed to frustrate rather than serve such concerns. To pass muster, which I am convinced both the pictures do, Leonardo must be seen as active and in control of every facet of their conception, planning and execution. If this has to be translated into such a phrase as 'largely autograph' , so be it, but I would prefer that it is not necessary to resort to such anachronistic formulas.None of the issues is definitively resolved when the exhibition closes, but some of the simpler ideas have earliest IR images. And, more significantly, the feature that we had rather overlooked -the tangle of brushdrawn contours at its base -suddenly makes sense.There, beyond doubt, are the kneeling Virgin, twisting Christ and standing woman of the baby-walker group in the Edinburgh variant and other paintings of this type ( fig. 2). Joseph is harder to tease out, but the bent leg with which he holds the baby-walker steady can be seen. In the upper right corner of the group, there appears to be the head of a horse, ass or ox. It also appears that some kind of platform or broad step had been placed at the base of the arch, just behind the figures. A comparable platform with a low picket fence is apparent in the Edinburgh picture. Looking back at the IR reflectogram of the Lansdowne Madonna -now that we have had more time to digest its implications than when we were rushing to write the catalogue -the presence of the baby-walker group becomes obvious. It is less clear than in the new Buccleuch image, butthere is no doubt about the key features. There is, however, no indication of the animal or the platform.Curiouser and curiouser. The main body of the essay then reviews the emerging evidence of the parallel development of the paintings, reinforcing the hypothesis that had been necessarily provisional at the time of the Edinburgh exhibition. It also allows some of the points in the more substantial reviews to be taken on board; most of the reactions, as we have seen, were naturally concerned with authorship, but Michael Bury, a lecturer at the University of Edinburgh, made an elegant contribution to the paintings' iconography, to which we now turn. ICONOGRAPHIES 77 ICONOGRAPHIESThe baby-walker group was clearly integral to Leonardo's early thinking about the composition. The meaning and symbolic content of the pictures had undergone a dynamic and innovatory development no less extensive than the formal aspects of the composition.Fra Pietro da Novellara's account of the painting in his 1501 letter to Isabella, quoted and discussed in Chapter 3, is both valuable as a contemporary reaction and notably informative about how to read the new kind of meaning -or rather the new way to convey established content. His percipience is all the more striking because he would not have encountered anything quite like this previously. He well knew that incorporating portentous symbols of the passion into Madonna and Child compositions was not unusual in itself.Goldfinches were, for instance, introduced for this purpose. The red 'stigmatum' on the bird's head makes reference to the wounds that Christ carried after the Crucifixion. Additionally, the tiny bird was also reputed to live on thorn bushes, an allusion to the crown of thorns placed on Christ's head, mocking him as 'King of the Jews' . The lamb in the cartoon seen by Fra Pietro is also a standard enough reference to Christ's sacrifice on behalf of mankind.The key symbol in Leonardo's painting for Robertet is the yarnwinder, which has a cruciform construction and makes obvious allusion to Christ's death.The spindles and the possible basket are secondary accompaniments to the yarnwinder.Compared to the comparatively common goldfinch, the yarnwinder is an unu- ICONOGRAPHIES 79In the Madonna of the Yarnwinder, the ambiguity of the Virgin's expression and the indeterminate motion of her right hand serve to leave open even key motives at the centre of the narrative. Is she trying to take away the yarnwinder, as Fra Pietro reasonably suggests, or is she reacting uneasily to an incompletely formed premonition, to the shadow of the cross falling across her consciousness? Christ's eagerness and intense gaze probably signal that he is fully and precociously aware of what he is doing, but such a reading is irredeemably subjective, and his reaction could be seen as childish curiosity, in which case we are aware of something that he is missing. This latter possibility seems unlikely.When we turn to the secondary group that Leonardo was unquestionably planning for the middle-ground of the composition, we are faced with a specific motif that has no known precedent. The practice of adding ancillary narrative scenes in a kind of 'genre' mode The New York drawing, which relates to some smallscale paintings from the Leonardo orbit, seems to have played a role in the generation of the second version of show devoted to one of the great iconic figures. Rather than trying to gather in one place all the works that reluctant owners were willing to lend, we devised a scheme for simultaneous exhibitions across Europe linked by travel and IT. The venue was not to be Milan or Florence or Paris or London. It was to be Europe.The research core of the project was to undertake a full set of scientific examinations of all the Leonardo paintings and those paintings deemed closest to him.The principle was that the same or essentially identical equipment should be involved, and that an inner group of experts would compare the results. This approach was devised specifically to combat the uneven results that had bedevilled our earlier research.If necessary we would move equipment rather than paintings. If a piece of equipment is damaged or lost in transit, it can be mended or replaced. If a masterpiece is lost or damaged, that is that. As it happens (and the story is full of happy coincidences), Artakt was asked to conduct further research into the Lansdowne Madonna, as the owner recognised that after a gap of ten years since the last research there was likely to be scope for new inves-tigations using the latest scientific techniques, some of which had been refined and some newly invented. We plan to see the exhibition and arrange for the painting to be inspected by a leading private conservator, Zahara Veliz, who works in London and is married to David Bomford, senior conservator at the National Gallery in London. We endeavour to fly to Florence but are diverted to Genoa, finally arriving by coach and taxi at 3 am. In Arezzo the same day, we look together at the picture, out of its frame, using low magnification. Marina Wallace, a trained painter, notices that Leonardo has made little hooked and wavy brushstrokes in the fringes of the landscape where it abuts Christ's curls, effecting a kind of optical liaison. We leave Zahara taking notes for a condition report which, among other things, will be used to judge whether any further conservation work should be undertaken, not least to tidy up the discoloured retouchings. Zahara is visibly moved by the quality of the painting, as happens characteristically for someone who knows it only through reproductions. Her resulting report is consistent with what we have already observed, but gives much greater precision to the reading of what has happened to the paint surface over the years. As someone well versed in the deterioration of old paintings, she is pleased rather than shocked by its condition. She confirms that the pigment layers are very stable, and we are not inclined to recommend any restoration at this stage. POLITICS AND PROGRESSBringing all the international galleries and museums on board proves arduous to a degree beyond even our worst expectations. National temperaments and working methods swing into stereotypical operation.In one country, calm reasonableness and a collaborative spirit are overtaken by concern for petty details undermined by parochial and personal inflexibilities.In another, sustained discussions seem to get bogged down at the same level, in an endless diplomatic dance, with no one prepared to take the decision to move on to the next stage. Another country expresses welcoming enthusiasm, formulating imaginative schemes, but nothing concrete happens in spite of continual prodding. One major gallery in another country begins by being recalcitrant and suspicious but comes entirely on board, albeit with no funds to participate. Only one country moves expeditiously and efficiently, but very much on its own terms, without involving the full collaborative process with the Laboratory as a whole.Some countries covertly voice suspicions that the Universal Leonardo is a money-making venture, even though the Bureau is a non-profit-making entity, heavily subsidised by the work that I and others put in for no or inadequate remuneration. If I cared to stop and think, I would realise that I am badly out of pocket on the venture. But I prefer not to think about this, and decide to ignore slurs reported by third parties. At the same time we are conducting fundraising, with all the elaborate rituals involved in the cultivation of what are called 'prospects' , corporate and private. Some of those whom we visit, establishing apparently good relations, making tailored presentations, and agreeing parameters for possible assistance and for what we will deliver on our part, make encouraging noises and even vague promises, followed by discourteous silences. We are encouraged to talk in millions by the Italian branch of a major IT company, only to find that not only the noughts disappear rapidly when any decision becomes potentially imminent. One sponsor, whose equipment is intimately involved in the technical examination, brandishes largesse with apparently firm promises, only to retreat to an unacceptable scheme to subsidise a press conference in Milan.I wonder if these people behave with comparable duplicity in their main business dealings. Undeterred THE UNIVERSAL LEONARDO AND THE LABORATORY 91we press on, confident that the merits of the project and the quality of those involved will win through. Key private supporters fortify our resolve and prop up the project's wobbly finances.In contrast to such frustrations, the research in Florence produces a feast of exciting results, reported enthusiastically back to London and Oxford by Thereza Crowe, as revelation follows revelation. Some of the tests, including a new method of infrared reflectography, are The relief pictures are not dissimilar in type to those customarily produced by raking light, but are more precise, distortion-free and quantifiable.We are confident that many of the secrets of the picture will be disclosed. However, it is in the nature of the pioneering ambitions of the Leo Lab (as it comes to be called) that some of the techniques will yield more than others. Some new methods are being tried, and some devices specially adapted and calibrated. It is also evident that some of the results will mean relatively little in isolation beyond their immediate existence as records of data. Only when comparative material has been obtained from an adequate number of paintings will some of the findings coalesce into meaningful sets from which art-historical conclusions can be drawn.The methods only assume utility and value in direct proportion to the quality of the questions being asked of them through informed collaboration between the scientists, the specialists in conservation and the historians of Leonardo's works. And what we see depends on the expertise, experience and persistence of the observers.We cannot here undertake a full report of the results obtained by the diverse techniques. Rather I will outline the main consequences as they appear to me from my standpoint as a historian of Leonardo. This is not, of course, the only standpoint from which the scientists' work is of importance. What follows draws on the scientists' dedicated work, concentrating on those analyses that delivered the most immediately significant findings. THE INFRARED TESTSIt is inevitable that pride of place goes to the results The most striking result has been the extensive use of lapis lazuli or ultramarine (finely ground from the semi-precious stone) for the blues throughout the picture. That it is present in the Virgin's garments is no great surprise, since it was regarded as fitting that the Queen of Heaven should be honoured with the very finest and most costly pigment. However, its use in the landscape is more unexpected. Good-quality lapis could be more costly than gold leaf and was sometimes used as a final layer over azurite. It would be normal to detect a cheaper and more readily available blue pigment, normally azurite, in the secondary areas. The only definite signs of non-lapis blues, in this case cobalt, were detected in areas of re-touching.This indicates the status Leonardo accorded the picture, and his willingness to run up additional costs.It was clearly intended to be a precious, intimate picture on which he devoted much meticulous effort, and which was to be cherished in private by a connoisseur.Its artistic and technical qualities are fully consistent It has been claimed that two of Leonardo's portraits, the Cecilia Gallerani and the so-called La Belle Ferronnière, are painted on panels from the same block of wood, and it will be important to correlate all the data we can obtain on the wood used by Leonardo in Milan and elsewhere. 4 At present, the information on the wood grain can be added to a bank of information that may in due course yield significant results.What the IR data and other analyses show is that much more information remains available in the surviving surfaces and layers than has previously been thought in the cases of paintings subject to harsh transfer from panel to canvas with the removal of most or all of the gesso ground. And above all they have confirmed that we are dealing with a remarkable product generated using remarkable procedures. DAYLIGHT ROBBERY AND MORE FORENSIC WORKThe next step for the Laboratory, logically, is to re-examine the Buccleuch Madonna using the same equipment, and since the painting is not in the care of a ponderous official body but is in the hands of an enthusiastic private owner, our hopes are high. The Duke agrees in principle that his painting can join in the programme, and expresses a wish that the panel should not move from Drumlanrig, where he had just re-installed it with special security measures. The equipment will travel to Scotland, if we can find the funds to take it. Then, on 27 THE THEFTThe press story inevitably runs in two phases. First comes the narrative of the dramatic snatch. The police issue information. They are seeking a white Volkswagen Golf GTI, five-door saloon car, registration number h596 vrp, which contained four men, one of them wearing a large white hat. The car was last seen locally on the Thornhill to Durisdeer Road at about 11:15 hours. Two of the men are described: 1. in his early 40s, 5 feet 10 inches tall, slim build and clean-shaven, he was wearing brown shoes, cream trousers with black belt, a cream T-shirt, brown Nubuck leather jacket, a light-coloured brown baseball cap and round-framed glasses; 2. in his late 40s, 5 feet 10-11 inches tall, slim build and clean-shaven. He was wearing black trousers, black shoes, cream long-sleeved shirt, sleeveless taupe-coloured safari-type jacket with lots of pockets and a light cream-coloured wide-brimmed hat. CCTV pictures and e-fit images are issued, with appeals for information. One shows the man in the flat-brimmed white hat and his partner in crime scuttling into the white Golf before speeding to the exit of the estate.Next, experts are wheeled out to assess the painting, with him before and found him to be a serious and responsible journalist. 3 I explain that there is no lost 'original' but that both paintings are of high status. He reads back my quotes for his piece, and I confirm that they are OK. There are things in the story that I would question, but I did not write it and I recognise that he is bound to report views with which I might not agree.In the various press reports, the quoted value of the painting fluctuates wildly according to whether it is considered 'autograph' or a studio 'copy' . The upper estimates are in the £ 50 million range. The Duke is, as I expected, deeply affected by the loss, and feels that he has let Britain down. The greatest immediate worry is that it has been stolen by lower-league thieves trying to move into the big time, who will find that the work is unsaleable and will decide that it is best to destroy it.The myth of the 'Mr Big' who is hoarding a great secret collection by commissioning thefts remains a fantasy. We pray that the Buccleuch Madonna will be restored to its owner. And we wait. The Duke follows up a suggestion that the painting is in Milan. This leads nowhere. A reconstruction is shown in September on the BBC programme Crimewatch, in the hope of jogging witnesses' memories. My best guess is that after a gap of some years, the new possessors of the painting will, via intermediaries, contact the loss adjusters who have been acting for the insurers and that they will attempt to strike a lucrative deal for its return. The thieves may, in the meantime, have been using their 'asset' as security in major criminal deals, as in the drugs trade. At least this will mean that the painting is being looked after. As damaged goods it would be worth less to them. He knew that there were moves afoot, but did not live to know definitely that his beloved Leonardo would be restored to his family. The panel, apart from possible scuffing in one corner, appears to have suffered no damage either from its wrenching from its display case or during its subsequent life in captivity. The fragile paper labels on the reverse are still intact. It is obviously the real thing, but the police are interested in the hard forensic evidence that I had promised. We take the painting into a As we move the camera over the surface, we not only see features of the underdrawing that were earlier apparent (but unpublished) but also other aspects that the older equipment had not picked up. We complete the inspection for the immediate purposes of the identification, but do not at this stage record everything we will need to bring the examination up to the level we had conducted on its sister painting. I leave Edinburgh, now minus its fog, in high spirits. We can see that the cantilevered roof above the arch on the left is less defined than in the Lansdowne version, and no supporting struts are visible. There might be some hints of other arches to the right, but these are too elusive to be confirmed. The figures in the baby-walker group below the arch can be seen more clearly to conform to those in the sister picture, though the pose of Joseph seems to be less resolved. The head of an animal, probably the ass that transported the mother and child into the wilderness, is visible just to the right of the Virgin's head in the baby-walker section of the underdrawing. The ass appears on the right There is plentiful evidence of rapidly skilful underdrawing in the Madonna's head, with a conspicuous alternative contour down her left cheek. We know that Leonardo often fussed with this contour in faces, seeking ways to make the contour turn into depth. The adjustments are more conspicuous than in the other underdrawing. The Child's head is drawn in the same confident manner, with lovely rhythmic strokes in his curls, which can also be seen to have been reinforced by a later hand, especially to the right of his forehead and cheek. RESTORED TO THE DUKEThere is a vertical slit at the top of the Madonna's right breast. It seems to be quite deep-seated. The binocular microscope later reveals a weave pattern in a rectangular area across the slit. It is likely that a fine linen patch was placed over an area of damage or irregularity in the original panel when it was being prepared for painting. Another smaller patch is visible beside it.Perhaps the most surprising new findings concern the yarnwinder and the Child's left arm (see detail p. 212). Wallace was the first to notice them. These look to be original. They have not been seen elsewhere in Leonardo's paintings, but they may not have been observed.Are they an attempt to convey the reflection of the colour of the sky in lit surfaces? It seems likely that this is the case. We know that he was keenly interested in the reflections of colour. For instance, he noted the blue shadows cast by intervening objects on a white wall, which he explained as the mirroring of the colour of the sky in areas occluded from the direct rays of the sun. 4 As we prepare the new technical material for this book, we become aware that the subtle traces we have observed on the screen of the computer in the gallery do not reproduce well. We can only hope that readers will trust what we are recording.All in all, the quality of key areas of the picture is clearly evident, both in the underdrawing and final pigment layers. The relationship between the design processes in the two pictures emerges with even greater complexity. Clearly, one of the pictures is not perpetually in advance of the other, with its companion following in the manner of a copy. We either have two originals or neither is the original. The problem is with our use of the term 'original' . It simply does not work well with such a production method. FORENSICS IN THE DOCKThe last act in the drama to date involves the trial of the five men charged by the police. They were accused of conspiring to extort Callum Jones in the same firm. The next move was to make contact with the loss adjuster, Mark Dalrymple, who was acting for the Duke's insurers.Dalrymple, who seems to have been the only one who can be said to have behaved entirely properly, immediately informed the police, on the understanding that strict secrecy must be maintained. The police then set up their own sting, with officers acting as intermediaries and representatives of the Duke. It was agreed by the intermediary that he was to pay £ 2 m into the law firm's client account. He guaranteed that 'he and no other person acting on his behalf has given any notification or information relating to the terms of the said agreement (nor will do so until after the completion date) to the law enforcement agencies' . In order to massage the transaction, Ronald The figures underwent some substantial changes in the underdrawing, more in keeping with sketching on paper than working on a panel. Even more radically, the props and backgrounds were changed in such a way that motifs appeared and disappeared in an extraordinarily experimental manner. The yarnwinder was adjusted in orientation, and threads were hung from its arms only to be eliminated. A series of spindles were once scattered across the rocky ledge at the foot of the yarnwinder. It is possible that the basket of spindles at the foot of Christ, mentioned by Novellara, was entirely painted out, leaving an area that appears unusually illegible in the examination and exhibits an unusual concentration of brown, iron-based pigment. However, it seems increasingly likely that the basket, if there ever was one, was not in this position but was at the foot of the yarnwinder, as in one 'family' of copies. 10) landscapes of rounded peaks were laid in very roughly in both paintings; 11) the decision was taken to eliminate the subsidiary group with the baby-walker, architecture and animals in both paintings, replacing it with a landscape containing a bridge; 12) the painting proper began, probably concentrating on the figures at first, as was normal; 13) the general direction of the successive changes was to eliminate distracting detail, such as the pile of fusi and maybe a basket 14) the landscapes moved in the direction of their final appearance. The Lansdowne panel was endowed with the most advanced kind of Leonardo microcosmic topography, while an assistant painted the Buccleuch landscape more or less along the lines with the underdrawing, but with the bridge excluded; 15) one of the pictures was completely finished and dispatched to Blois, while the other was to be finished for the King or 'whomsoever he wishes'; 16) the version not sent to Blois was subsequently finished and remained in Leonardo's (and later Salaì's) hands. Since the landscape in the Buccleuch picture was eventually completed in a routine manner in the studio, it may be that the Lansdowne version was that sent to Robertet. In attempting to answer this final question as to which panel was actually delivered to Robertet, we need to turn to the study of the paintings' provenances.NOTES TO PART I 121 14. Barone, 2010, pp. 19-22. 15. The Mystery, 1992, pp. 76-7, no. 20, citing an observation made by Johannes Nathan. 16. Kemp and Barone, 2010, pp. 19-22. 17. Jean Paul Richter, The Literary Works of Leonardo da Vinci: compiled and edited from the original manuscripts (1883), 2 vols., 3 rd ed., London, 1970, no. 50; Martin Kemp and Margaret Walker, Leonardo on Painting: An anthology of writings by Leonardo da Vinci with a selection of documents relating to his career as an artist, ed. Martin Kemp, New Haven, 1989, p. 196, no. 505. 18. Richter, 1970, no. 1349Walker, 1989, p. 259, no. 619. 19. Villata, 1999, pp. 286-8, no Kemp, New Haven, 1989, p. 22, no. 571. 16. Clark andPedretti, 1968, no. 12521;and The Mystery, 1992, pp. 66-7, no. 15. 17. Ibid., no. 12533;and ibid., pp. 58-9, no. 11. 18. Ibid., 1968, no. 12394;and ibid., 1992, pp. 70-1, no Studies, xxiv, no. 2, 1994, pp. 259-74, figs. 29-32. 24. Villata, 1999, p. 213, no. 247. 25. Ibid., p. 208, no. 240. 26. Richter, 1970, no. 1349and Kemp and Walker, 1989, p. 259 London, 1975, p. 107. 8. John Rowlands, 'Two Ladies' , Antique, vii, no. 3, 1992; and the Duke of Buccleuch,Antique,vii,no. 4,[n.d.], [n.p.]. There are many ways a painting can reveal itself to us. As the previous chapters have shown, scientific examination can be one of the most rewarding in giving us an idea of how a work was made, what materials were used and what has happened to it over the years. In the case of the Lansdowne and Buccleuch Madonnas, it can be a way to observe how paintings made by the same artist relate to one another as well as how an artist approaches the same subject at different times. CONCLUSIONS AND CONSEQUENCESOne of the most fundamental concerns in investigating a painting is to establish not only its current physical condition, but the history of this physical condition, that is to say its conservation history. This would include work carried out on the painting in order to stabilise or restore it. There is very little documentation describing early conservation and restoration work undertaken on the two Madonnas. Published articles and catalogues have not, for the most part, dealt with these issues. What means can we employ to discover this fundamental history? And how can it aid us in determining issues of attribution? CONSERVATION: THE LANSDOWNE MADONNAOne of the most rewarding avenues of research on the Lansdowne Madonna has been the study of photographs taken of it over a period of more than one hundred years. One might think that this would reveal no more than developments in the field of photography. Certainly, a cursory glance over the dozen or so photographs shows differences between black-and-white and colour photographs, and between images taken for exhibitions or for the owner's home. However, on closer inspection, we find an extraordinary series of changes in not only the condition of the painting but the composition itself (see figs. [53][54][55][56][57][58][59] THE ROLE OF THE CONNOISSEURThe subject of attribution was a hot topic then as now, and comparisons between the two versions of the Virgin of the Rocks were also discussed in detail in the catalogue. In fact, many of the paintings in the exhibition were seen as having contentious attributions, as the catalogue entries show. The exhibition took place at a time when respect for the judgement of the 'connoisseur' was at its height; the connoisseur combined historical knowledge with a trained eye and an indescribable 'gift' for discerning the authorship of works of art.The note at the beginning of the catalogue explains: Of the 76 paintings exhibited, ten were attributed to Leonardo, including our two Madonnas, and two paintings given to the 'school of Leonardo' .Whereas today there are now only twenty or so paintings attributed to Leonardo, it is evident that in the nineteenth century there were thought to be many more 19 He made a number of observations that are discussed in the first part of this book.Suida went on to discuss the ultraviolet examination CONSERVATION AND ATTRIBUTIONHe also suggested that Leonardo's hand could be seen in the Christ's head and in the landscape. 2) the fore-and middle fingers of the Virgin's left hand previously extended below Christ's right arm, cradling the left side of his abdomen; 3) in 1893 a thin veil of looping drapery passed over Christ's left thigh and lower abdomen, concealing his genitals; 4) in the painting as it is now, the drapery on which Christ sits and which tumbles over the rocky ledge is broader and far more defined than before; 5) there are now numerous passages of more pronounced definition, especially in the draperies and the hair; 6) there are now clear signs of damage along vertical cracks from the shin of Christ's left leg downwards as well as a pronounced vertical mark to the left of these cracks; 7) the Virgin's eyes, especially her right eye, now have more defined pupils with indications of a reflection in the right eye. The same is also apparent in Christ's eyes; 8) the curve of the outside of Christ's right arm was softer than it appears now. His ear is less visible than before, with more curls of hair covering it; 9) the Virgin's hair is now slightly fuller, with an extra strand of curls falling along the inside of her right side and the outside of her left side. INTRODUCTIONThe provenance of a work of art tells us where the painting has been and who has owned it. It can reveal whether or not it has travelled and through whose hands it has passed as the result of gift, swap or sale. It can also serve as a significant factor in establishing an attribution for a painting. If, for example, a painting can be traced to where the artist or patron is known to have lived or worked, an attribution to that artist can be made stronger. In the case of the Unlike the Buccleuch Madonna, which has remained in the same family for more than 250 years, the Lansdowne Madonna has passed through at least nine owners during a 200 year period. As noted, the history of the painting can be firmly traced to the early nineteenth century. Three auction sales from this period give us clues as to who the owners were as well as where the painting had been before. This is the best place to begin establishing the provenance of the painting.The first auction took place in 1833 in London at the auction house of Henry J. and George Henry Robins.The auction was held from 12 to 18 The starting point for finding the answers to these questions is to establish information on presumably the earliest known owner of the painting, Lord Darnley, from whom Lansdowne is cited as having bought the painting in 1809. Who was he, did he actually know Lansdowne, and was he likely to have bought a precious painting such as the Madonna of the Yarnwinder? Next, we will look at the Lansdownes to see who they were and how they fit into the story. A DARNLEY CONNECTION?The English now seem to be governed entirely by the He also appears to have visited Naples. 6 Darnley's Rome visit is recorded in the extensive jour- where I could scarcely allow myself to see the famous pictures of Rubens, etc. However, I did not only see them, but like a very bad boy made some purchases, for which I hope you will not be very angry, as I think you will like them when you see them, and they are certainly so cheap, that I can always get more than my money for them again. 14 For £ 500 Darnley was able to buy several paintings by Rubens, a landscape by Claude Lorrain, a deer painting by Snyders, and a Dutch landscape by van de Velde. 15 By far one of the most celebrated collections which appeared at this time came from the duc d'Orléans, Philippe Egalité. In the 1790s, in an attempt to raise money, Orléans approached London with a view to selling works to a syndicate which involved the Prince of Wales as well as the auctioneer Christie. When that fell through, he sold part of the collection, consisting of the French and Italian works, to a Belgian banker by the name of Walkeurs for 750,000 livres, who then sold it to a Laborde de Méréville for 900,000 francs. The collection was eventually moved to London when the political situation became too dangerous. 16 According to Buchanan, these paintings were consigned to a house of eminence in the city of London, and it is believed that they were in the hands of that house when a treaty was entered into by the There is no reason to think that Robert was any different. Might he have obtained the version in the Salaì list, or that owned by Robertet, who died in 1527? The picture has never been in a cleaner's hand, andMr. Hamilton alone has cleaned and repaired it.When you consider the antiquity of the painting, I think its preservation remarkable, and no material part of the picture had suffered. It is in good harmony, and he will not part with it under 800 guineas. It was sold for 1,000 sequins in its uncleaned state, and as the chance has been in favour of the purchaser, I do not think that he overpays himself; but certainly it is a picture which might be put at the head of any collection. And in case you had been in cash I name it to you. King's Mews 58The letter is a fine example of salesmanship, but also reveals that it was cleaned by Hamilton himself, who, have been enabled to procure on advantageous terms … I am not apt to throw away money on the fine arts, but the uncommon distress of Paris afforded me an opportunity which to a certain extent I really thought it worth while to improve. 66 He appears to have had his father's love of sculpture (from Rome he writes, 'everything I see here confirms the preferences which I have ever given to sculpture over painting'), 67 but painting is also well regarded, as other letters describe. In a letter from Florence dated 16 July 1794 he writes, 'My short stay in Bologna gave me a more exhalted [sic] idea of the art of painting than any which I entertained before. Though I have not been ten hours at Florence I have already paid a first visit to its celebrated gallery [the Uffizi] in which Lord Holland was so good to conduct me. ' 68 Wycombe was also sensitive to the contemporary culture of Rome as he quite harshly describes it in a letter of 8 November 1794:Inhabiting Rome is like conversing with the dead, it exhibits a people, as well as a city in ruins, it has no Society, no Spectacle, no commerce, no industry, and consequently no animation. The objects of antiquity however, and those which the fine arts afford make ample amends to the stranger whose views do not extend to residence. 69 Athough none of Wycombe's letters mentions Gavin Hamilton, he does note meeting with William Hamilton, as well as the dealer Thomas Jenkins. Wycombe occasionally includes descriptions of particular works that might interest his father and also comments on the market. One letter, written from Turin on 23 December 1795, is a good example: I can assure you that you had better have nothing to do with [Thomas] Jenkins who has … reduced his traffic to Cameo Intaglios and who is now as assiduous in proffering rings of 4 or 5 guineas and even of less value, as a new beginner in a Bond Street Toy Shop … As to Pictures I know of none in the Market worth your notice unless it be at Bologna where the prices asked are immense; several thousand pounds being asked for one Corregio by a man who offers two for sale. The same man has an Angelica and Medoro, a pretty picture which he says a Dominichino, and a specimen of Andrea del Sarto. It is said that he bought his fine things in some obscure quarter for a mere trifle. 70 As well as providing evidence of the inflated prices demanded for famed old masters, the letter continues by describing an incident, probably not too uncommon, when a friend was sold forgeries: Wycombe may have portrayed himself as an impostor but his extensive travels abroad had made him more knowledgeable than most.MrIt was in Ireland that Wycombe first met Maria Arabella Giffard, who was then married to Sir Duke Giffard.The earliest reference to the Giffards is from 1 March 1798 when Wycombe mentions having returned from Dublin via Castle Jordan, which was owned by Sir Duke Giffard, and writes of his desire to purchase some land near there. By mid-April, Wycombe is referring to them as his good friends and asking Holland's advice on a case the Giffards had brought to him concerning a Captain Maddock (Maria's maiden name was Maddock, so the Captain was perhaps a relative). The letter goes on to describe the Giffards:Sir Duke Giffard is almost without politics too, but standing well with the lower classes as an enemy to persecution and indulgent to the poor. In short a mild man of ancient family just come into a large fortune by the death of an old father, and governed by a buxom spouse of about 28. 77 It has been noted that Maria had daughters, but Wycombe also mentions a son, who was ill. In the same letter, Wycombe's interest in Maria is more apparent when he describes her as 'a woman of exquisite sensibility and so desirable …' (Wycombe to Holland, 27 The original portrait of Napoleon which Wycombe refers to could be the same painting that was listed in two auction sales in 1834 and 1835, following the death of Maria Arabella. Another reference to a painting that also appears in both sales, this time a Salvator Rosa, occurs in a description Wycombe gives of two victims of a shipwreck he had seen in Ireland. The second corpse, more exposed to the fury of the breakers, was rendered strangely horrible perhaps from injury, and could be compared to some such ghastly images as the genius of The Marquis of Lansdowne's Gothic Castle at Southampton, which we cannot help considering as a proof of the wealth rather than the taste of the owner, is carrying on with spirit; and yet we question whether part of the work will last until the whole is finished, another settlement having lately been discovered in the advanced South wall, and that weighty mass of masonry is prevented, by props, from burying in ruins its beggarly neighbourhood. 93 The rooms are crowded with fine things of every description brought from Lansdowne House [London].One room is filled with curious portraits all of which had been sent for. She calls them Lord Lansdowne's Mariette was even less forgiving in his annotation, writing that it did not look to be a Leonardo and the baby appeared stillborn. 10 In the review that follows, we will encounter a range of possible lines of descent into the Tallard collection.None of them can be regarded as proven at this stage, not least because of lost or laconic inventories. We will endeavour to outline the most likely routes, as well as The answer is certainly yes. Indeed, there was a famous Madonna and Child painting by Leonardo da Vinci in France that was first described in 1651. Its circumstances are crucial to the history of the Madonna of the Yarnwinder and will now be discussed in detail. It will be seen that not only was the Madonna and Child mentioned in 1651 directly connected to FlorimondRobertet, but also it is quite possible that it is connected to the Tallard painting as well. As the eldest child, it would make sense that François might inherit the painting following the death of his mother, Isabeau, in 1626. However, her will, drawn up in May 1625, provides a possibly different scenario. 31 Although the will does not mention specific paintings, it does reveal that she left her second son, Henri, the enormous sum of 40,000 livres. To François, Isabeau bequeathed only a diamond once belonging to her mother. 32 Charles's will is discussed below, but a thirty-fivepage post mortem inventory of his Château de Jouy dated December 1666 will be mentioned first. 49 The quick sale that followed the marquis's death indicates that the family were rather more in need of cash than family heirlooms. However, they did not sell Gaujacq immediately and this will be discussed later.If the Madonna of the Yarnwinder cannot be found in existing family archives, it is possible that it was sold by one of the de Sourdis members. Charles de Sourdis is perhaps a likely candidate. He was known to buy and sell his own works of art. The collector Thomas London, 1961. 55. Noel Joseph Desenfans (1746-1831, whose collection was the foundation of the Dulwich Picture Gallery. 56. Bowood House archives. 57. C.J. Greville appears to have been the nephew of the diarist C.F. Greville, whose correspondence to Rutland is also included in the Rutland Papers (see following note). There is also a C.T. Greville who writes to Rutland, also London, 1922, p. 174. 107. Battersea, 1922. I am grateful to Alexander Reford for this account of the studio. 15-22 November, Lot 19. See Lugt, 1983, no. 3053). It is attributed to Andrea del Sarto, but is described very much like the Madonna in the Tallard sale ('La Sainte Vierge tenant sur ses genoux l'Enfant Jésus contemplant une croix qu'il tient dans ses mains'), has the same dimensions ('18 sur 13 de larg'), and is described as coming from the Tallard sale. There is no other painting in the Tallard sale that matches this description, but as the Madonna is documented in the Montagu inventory of c. 1770, this painting in the 1779 sale must be a different one. Perhaps it was another variant of the Madonna of the Yarnwinder and was confused with the version in the Tallard sale. 1453-1715, 1966, p. 429. 31. Last will of Isabeau Babou, Archives Nationales, Paris. M.C. Etude. xlv 154. 32. Ibid. However, she then asks him to give the diamond to a nun so that prayers may be said for her. She apologises for giving François so little but has to look after her heir, Henri, who is in debt. Schnapper, 1994, p. 174. 40. 'Le Marquis de Sourdis a fait descendre partie des tableaux de la grande gallerie alors … a entedu aud … qu'ils ayent estes bailles par feu monseigneur le cardinal de Sourdis en faveur de ses successeurs archeveques que le chapitre anota … à la confirmation d'iceliy alors archevesque et d'empecher les tableaux accroches tant à la chapelle archevesque … Monseigneur marquis de Sourdis fasse remettre les tableaux qu'il a fait enlever de la grande gallerie …' Tuesday 16 January 1646, Archives de Bordeaux, Chapitre de San André, Actes Capitulaires, 1642-1657, p. 361. As with all paintings created by Leonardo the Madonna of the Yarnwinder has been copied many times. The works of other famous artists, particularly their smaller-scale devotional pictures, were quite heavily replicated and imitated, but Leonardo was probably copied to the greatest extent. Leonardo's own studio procedures seem to have encouraged such replication. The followers of Leonardo have emulated the composition, sometimes in great detail, and at other times combined it with motifs drawn from a separate source. A number of the works are very similar in style and some compositions appear to have employed a cartoon, perhaps one taken from Leonardo's studio, or used other methods of precise replication, such as tracing. Given the lack of proof that Leonardo used a cartoon for the whole composition, tracing is to be favoured as an explanation. THE BUCCLEUCH MADONNA AND Will of CardinalA number of Leonardo's close followers took up the subject of the Madonna.The resulting body of paintings includes artists from Italy, Northern Europe and Spain. They are now in private and public collections throughout Europe as well as in North and South America. Locations for some are now not known.Nearly forty versions of the Madonna of the Yarnwinder have been found.Two publications have previously reproduced a number of these versions and one of them is used as reference in the figure captions. 1 They can be roughly arranged into five groups or 'families' , which exhibit striking or subtle variations in composition and iconography, which clearly originate in different geographical centres. There is some cross-over and several pictures could belong in more than one group, especially those attributable to Spanish artists. The baby-walker group appears closest to one of the earlier stages in the development of the original paintings, before the figure group was painted over. The interesting and is possibly explained by the use of a drawing or cartoon which was turned over so that the image visible on the back of the paper was reversed.The reversal could also have resulted from a tracing process. Such a reversal also appears in some of the Spanish paintings. GROUP WITH ROCKY OUTCROPThe There may well be yet further copies and variants to be found. It is also likely that other versions have been lost over the course of around 500 years. All this points to the extraordinary influence of Leonardo's inventions across Europe. Artists and collectors clearly responded eagerly to his memorable compositions. It is always said that Leonardo completed few pictures.This relative scarcity cannot be said of the Leonardo 'brand' , which comprises an enormous number of 'Leonardos' (as described in inventories) and pictures that exploited his unique manner of portraying conventional subjects. In terms of workshop production, two things are exceptional and surprising about the Madonna of the Yarnwinder. The first is that two pictures were begun at the same time and underwent a kind of syncopated development in his workshop. One picture seemed to be in the lead at one point, and then the other, as motifs were tested and sometimes eliminated. By con-trast, when Botticelli or Bellini began their Madonnas, they made no really radical changes once they started painting. The second surprise is that the intermediate states of the paintings, which may have been visible for some time given Leonardo's notorious slowness in bringing works to completion, assumed authority for his pupils and assistants, in such a way that they based their own paintings on the unrevised and unfinished panels. This means that inventions that Leonardo himself was to discard leaked out into other 'families' of copies.The technical examinations to which the pictures have been subjected are probably as extensive as ever applied to Renaissance Madonnas. It may be that sustained examination of other artists' paintings will yield comparable surprises, but we doubt it.What our study also clearly shows is that scientific examination is not an end in itself -fascinating though it may be -but it has an impact upon our understanding of the production, patronage and reception of pictures, aspects of art history that are generally Inevitably, the story is not definitively concluded.As the book was going to press, we learnt that the Lansdowne Madonna was again resident in the premises of the Opificio delle Pietre Dure in Florence.Should we wait to publish? The problem is that any new findings that emerge with enhanced techniques 1525On Salaì's death a list is made of his possessions (known in two copies) so that his assets can be divided between his sisters. Among the paintings is a 'Madonna with a Child in her arms' , valued at 20 scudi, compared with the top valuation of 200 scudi for the Leda. What appears to be the Louvre St John (larger than the Madonna) is valued at 80 scudi. 1527Death of Florimond Robertet. c.1650Florimond Robertet's great-grandson, the Marquis Charles de Rostaing (d. 1660), commissions Henri Chesneau to fabricate a mémoire inventory of the family's art collection by the widow of Florimond Robertet, Michelle Gaillard, as a way to increase the standing of his family. The Madonna of the Yarnwinder is not mentioned. The auction entry describes the painting:The Virgin seated, carrying on her lap the Infant Christ who is holding a cross. The background is a landscape; the height is eighteen inches [pouces] by thirteen inches [pouces] wide. This painting is from the best period of the Master and is perfectly preserved. Nothing is so rare, and at the same time so esteemed as the paintings of Leonardo da Vinci. This excellent artist who had the highest understanding of Painting, did not make many Pictures, but still went very far in his career. Constantly looking into his imagination or in nature for things to give his Works all of the perfection possible, he employed far greater time to his theory of art than in the execution of his paintings. He was therefore able to give his figures the character of life unknown to almost all painters who preceded him.(La Sainte Vierge assise, aïant sur ses genoux l'Enfant Jésus qui tient une croix. Le fond du tableau est un païsage; sa hauteur est de 18 pouces de haut sur 13 pouces de large. Ce tableau est du meilleur tems de ce Maître & d'une conservation parfaite. Rien n'est si rare, ni en même tems si estimable, que les Tableaux de Léonard de Vinci. Cet excellent Artiste qui avoit la plus haute idée de la Peinture, n'a pas fait beaucoup de Tableaux, quoiqu'il ait poussé fort loin sa carrière. Sans cesse occupé à chercher dans son imagination, ou dans la nature, des objets capables de donner à ses Ouvrages toute la perfection dont il les croïoit susceptibles, il emploïoit à la théorie de son Art beaucoup plus de tems qu'il n'en mettoit à l'exécution de ses Tableaux. Aussi étoit-il parvenu à donner à ses An annotated catalogue of the sale notes 'Remi pour l'Angleterre' alongside this painting as well as six others, all of which entered the Montagu collection. CHRONOLOGY OF THE LANSDOWNE AND BUCCLEUCH MADONNASof the Rocks in the National Gallery and St Jerome in the Vatican' , The Burlington Magazine, cxlvii, no. 1228, pp. 450-63. Thomas, Ann, The Painter's Practice in Renaissance Tuscany, Cambridge, 1995 Trapier, Elizabeth du Gué, Luis de Morales and Leonardesque Influences in Spain, New York, 1953. Trustees of the National Gallery, The National Gallery Report, April 1992-March 1993, London, 1993 Thomas, Art in England: 1800-1820, New York and Cambridge, 1928 Whitley, William Thomas, Artists and their Friends in England, 1700-1799, London and Boston, 1928 Wildenstein, Daniel, and Jean Adhémar, 'Les tableaux italiens dans les ventes parisiennes au xviiième siècle' , Gazette des Beaux-Arts, t. ii, 1982. Williams, Julia Lloyd, Gavin Hamilton 1723-98, Edinburgh, 1994 Wingfield-Stratford, Esmé, The Lords of Cobham Hall, London, 1959 We know that Leonardo da Vinci finished a small painting of the Madonna and Child with a crossshaped yarnwinder. In a quite new way Leonardo showed mother and baby dynamically reacting to each other and to the ominous presence of the cross.Has the painting survived? There are many copies and versions, but can one of them be identified as the original? There are two prime candidates. If they are subjected to an unprecedented battery of scientific tests, will one emerge victorious? Or will neither emerge with credit?In fact, as we learn in the extraordinary detective story told in this book, the forensic work reveals that Leonardo himself was involved in creating not one but two paintings of the subject. We can in effect witness the two paintings progressing side-byside in his workshop. Equally intriguing is the search for their later histories through French and British collections. And the story is not over … Art Theory/Historythe book how John and Richard, successive Dukes of Buccleuch, have encouraged and even initiated the research and display of their treasured painting. They and the owner of the other painting have been selflessly willing to subject their precious masterpieces to detailed forensic analysis in the laboratories, knowing that one unfavourable finding, such as discovering that the painting was executed in pigments not known in the Renaissance, could wreck their beliefs and expectations. A series of specialists in the scientific examination of paintings have played crucial roles in a notably enthusiastic manner. John Dick at the National Galleries of Scotland first showed that the Buccleuch picture contained secrets hidden beneath its surface. The comprehensive examination of the Lansdowne Madonna in the laboratories of the Opificio delle Pietre Dure (OPD) in Florence was undertaken by Cecilia Frosinini and Roberto Bellucci under the directorship of Cristina Acidini. They collaborated with a comprehensive group of scientists, including Alfredo Aldrovandi and Ottavio Ciappi, also at the OPD; Mauro Bacci, Andrea Casini, Lorenzo Stefani, Franco Lotti and Marcello Picollo at the Istituto di Fisica Applicata 'Nello Carrara' the Ente per le Nuove Tecnologie l'Energia e l'Ambiente (ENEA), La Casaccia, Rome. Together they undertook what was at that time the most exhaustive set of examinations of any painting. The results, as we will see, are remarkable. Finally, when the stolen Madonna was recovered, Lesley Stevenson at the National Galleries in Edinburgh ensured that we obtained data that could be fruitfully set beside the findings of the OPD in Florence. Those involved in the scientific examinations could reasonably be listed as co-authors of key sections in this book. Since Part I is cast as a narrative, it will become clear how many people have contributed constructively to the research, and their later appearance in our story can serve as a form of sincere thanks, without listing them formally at this stage. Many people have helped during the course of the historical research. We would like to thank Thierry Morel, Caroline Brooke and Martine Stroo for their enthusiasm and diligence. Joseph Baillio provided early and recent research. Francis Ames Lewis's sup-BACKGROUND ix viii ACKNOWLEDGEMENTS port of the book enabled the research to continue to publication stage. At estate and family archives we have been aided by Alexander and Nikola Reford, Kate Fielden at Bowood House, Sandra Howat at Bowhill, and Andrew Fisher at Drumlanrig. For nearly ten years the support of Nanne Dekking has been indispensable. This book could not have been published without the generous support of the British Academy, Pieter and Olga Dreesmann and the Research Department at Central Saint Martins College of Arts and Design, University of the Arts, London. Our collaboration in writing this book is part of a larger group of people who have all made significant contributions on practical, academic and personal levels. They include Gaia Maiolo, Faye Gordon and Jane Scarth. Julie Hill's elegant design work speaks for itself and we are grateful for her diligence and care. Eileen Cadman has brought great skill to the final editing of the book, which has been unusually demanding. We are grateful for her patience and good humour. We are especially grateful to Marina Wallace who has provided continued support on all levels. Martin Kemp's Personal Assistant, Judith Flogdell, has played a vital role in orchestrating his various tasks so that the book could reach its final form. Michael Clarke of the National Galleries of Scotland has provided encouragement at key stages, as will become apparent during the course of our narrative, and Timothy Clifford has played a characteristically positive role. Rachel Billinge and Luke Syson of the National Gallery in London, together with Ashok Roy, Larry Keith and other staff in the gallery, have been very helpful at key points in the story. BACKGROUNDThis book has been long in the writing. This is in part due to the extended nature of the programmes of research into the two paintings of the Madonna of the Yarnwinder. The detailed technical examinations of both pictures have involved a series of skilled collaborators over a number of years. Research into the provenance of works of art -trying to track the ownership and collections through which an item has passed -is notoriously time-consuming, not least given the often sketchy descriptions in early inventories and sale catalogues. Sometimes it feels like finding the proverbial needle in a haystack. Circumstances also intervened. In 2003, the painting owned by the Duke of Buccleuch was stolen from Drumlanrig Castle in Scotland. This precluded our bringing the technical analyses of the Buccleuch Madonna to a point comparable to those already conducted on the Lansdowne version. Above all, we had not managed to obtain a full set of highresolution images using infrared reflectography that we could set beside those that had been taken of the other picture, which had been meticulously examined in the Opificio delle Pietre Dure in Florence. Only with the recovery of the Buccleuch Madonna in 2007 could we rectify this discrepancy, courtesy of the conservation staff in the National Galleries of Scotland. The appearance of this book at this particular time has been prompted by the run-up to the exhibition Leonardo da Vinci: Painter at the Court of Milan at the National Gallery in London, in which the Buccleuch Madonna is to feature (though it is not, as we will see, a Milanese picture but was undertaken for a member of the French court and executed in Florence). The book falls into four parts. The first contains the narrative account of some twenty years of research conducted by Martin Kemp, who was joined by Thereza Wells (née Crowe) at an early point in the quest. The first fruit of our collaboration was the exhibition Leonardo da Vinci. The Mystery of the Madonna of the Yarnwinder at the National Gallery in Edinburgh in 1993. The actual writing of the first part was undertaken predominantly by Martin Kemp, but with vital FlorimondRobertet, Secretary of State to the invading French king, shortly before Leonardo's departure from the city in December 1499. The artist was actively working on the picture in Florence in 1501, and it can reasonably be identified with the 'small picture by his hand that has recently arrived here [in Blois]' , recorded in a letter of 1507 from the Florentine Ambassador to the French court. 1 A 'Madonna with a Child in her arms' subsequently appears in two inventories of the possessions of Leonardo's long-term assistant Salaì. and owned by Salaì are investigated in the second part of this book. The research involves moving chronologically forwards from the early sixteenth century and backwards from what is known of their ownership in more recent hands. As we will see the older and newer stories do not join up seamlessly. There is nothing unusual or suspicious in such gaps in the provenances of old pictures. This book deals comprehensively with the techniques used in modern art history to match the historical record with the surviving objects in order to bridge the gaps. Above all it records the application and testing of both old and new techniques of scientific analysis to understand the nature of the surviving pictures and whether they exhibit special characteristics that can point to their origins. In the case of the present research, the lengthy process involved has exercised a huge influence on what we have discovered and what we have believed to be true. The Madonnas have been enlisted in various enterprises, including major exhibitions, commissioned research, academic conferences, international arts diplomacy, the programme of the Opificio delle Pietre Dure (the state conservation institute in Florence), and the activities of research scientists in various laboratories. Most recently, the story of the Buccleuch Madonna has even involved a dramatic theft with all its attendant publicity, and the equally sensational recovery of the picture by the Scottish police. Fig. 1 X1-ray of the Madonna of the Yarnwinder (The Lansdowne Madonna) reproduced in Wilhelm Suida, 'Leonardos Madonna mit dem Kreuzstab (oder dem Garnwinder)', in Miscellanea di studi lombardi in onore di Ettore Verga, Milan, 1931, pp. 333-9. Fig. 3 X3-ray image of the Madonna of the Yarnwinder (The Buccleuch Madonna), plate 2. Image taken in Milan in 1938 and published in Wilhelm Suida, 'La Scuola di Leonardo', in Leonardo da Fig. 22After Leonardo da Vinci (Fernando Yáñez de la Almedina?), Madonna of the Yarnwinder, oil on panel, 62 x 48.8 cm, Edinburgh, National Galleries of Scotland, (Leonardo e il Leonardismo 53). 24 FORENSICS THE STORY BEGINS: THE SCOTTISH PICTURE 25 the untrained viewer, distinguishing what is significant in the unfamiliar landscape of ghostly traces is difficult if not impossible. Even doctors and surgeons have trouble in reading X-radiographs and scans produced by radiologists and radiographers with whom they are unused to working. In the case of the Buccleuch Madonna, we not only need the expertise of John Dick in reading radiograms in general but also his specific knowledge of what the gallery's equipment delivered. The infrared set-up in the Edinburgh studio at this time is less advanced than that in Scotland's sister National Gallery in London, or in a few of the world's other leading galleries. John's knowledge needs to be coupled with the historian's awareness of what we might be seeking in terms of what we know about the picture. To a large extent, we see only what we are actively looking for. With the picture placed securely on John Dick's easel and clamped in place, the IR apparatus is wheeled into position, so that the lamp that emits the rays and the camera that receives them can scan the painted surface. We cluster around the screen, watching the tantalisingly indefinite images arise in jerky sequence as John Dick pans the camera over the surface of the panel. We become increasingly naturalised in the strange landscape of streaky greys, and begin to pick out key features of the painting as geographical markers. Naturally the heads of the Virgin and Child are of considerable interest, both as reference points and as areas into which the artist was likely to have put a lot of work. Clearly visible on the screen are firm, simple, confident lines of underdrawing, as the artist mapped out the main contours of eyes, nose, cheek and other crucial boundaries of the human forms. They show up less well in the photographs. There are some slight pentimenti, particularly along the left side of the Virgin's cheek. It seems that cheeks -as the roundness of the face retreats away from the spectator's gaze -often gave Leonardo trouble. As we move to more subsidiary areas of the picture, more striking evidence of changes of mind becomes evident. The Child's right arm appears to have been relocated (fig. 6). And the Virgin's medium of oil. Leonardo was thus able to overlay a series of translucent glazes of thin colour on the brilliant white or ivory priming of the panel. The optical results of the selective overlaying of the translucent, semi-opaque and opaque glazes endow the description of forms and surfaces with an elusive beauty. Perhaps only Giovanni Fig. 44Infrared reflectogram detail (head of Virgin), Leonardo da Vinci, Madonna of the Yarnwinder (The Buccleuch Madonna), plate 2. 1990. Fig. 5 Infrared reflectogram detail (head of Child), Leonardo da Vinci, Madonna of the Yarnwinder (The Buccleuch Madonna), plate 2. 1990. 26 FORENSICS THE STORY BEGINS: THE SCOTTISH PICTURE 27 have done because we do not know what we are looking for. fig. 85) and Edinburgh version (fig. 2). This is the first indication that there was some kind of unexpected relationship between the picture on the easel and various families of variants; and, as we know only in retrospect, we are seeing the first sign that there had been some kind of 'leakage' of the composition into other families of copies in the early stages of both prime versions. As we move into the middle-ground and background of the picture, more surprises are in store. To the left of the Virgin's head is a recognisable arch (fig. 8), and to the left, signs of a landscape with higher hills than those in the final painting. There are also indications of other features in the lower region of the arch, but we fail to read them clearly or understand their significance. In part, we do not see what we might Fig. 6 Fig. 767Infrared reflectogram detail (right arm of Child), Leonardo da Vinci, Madonna of the Yarnwinder (The Buccleuch Madonna), plate 2. 1990. Infrared reflectogram detail (hand of Virgin), Leonardo da Vinci, Madonna of the Yarnwinder (The Buccleuch Madonna), plate 2. 1990. Fig. 88Infrared reflectogram detail (arch), Leonardo da Vinci, Madonna of the Yarnwinder (The Buccleuch Madonna), plate 2. 1990. Fig. 99After Leonardo da Vinci, Portrait drawing of Isabella d'Este, black chalk on two joined pieces of off-white paper, 62.9 x 48.4 cm, Oxford, Ashmolean Museum, University of Oxford. 3. ISABELLA'S AGENT Isabella, of the house of d'Este in Ferrara, had married Francesco Gonzaga in 1490 at the age of sixteen and was using her base in his court to acquire what would now be called a great collection of art. The most expensive items were ancient marbles and gems, in which there was already a flourishing market, particularly of newly excavated items. There were also signs of a rising industry of fakes. Isabella and other courtly patrons were customers for beautiful bronze statuettes by Pier Jacopo Alari Bonacolsi based on Roman carvings; not for nothing had Pier Jacopo earned the nickname Antico. Isabella was also an avid acquirer of works by the leading contemporary masters -not just those by artists readily available locally, the most notable of whom was Andrea Mantegna, the Gonzagas' established court painter. Isabella sought works by leading artists from far afield, such as Giovanni Bellini in Venice and Pietro Perugino in Florence. A series of highlevel contacts in other cities effectively served as a network of 'agents' for her acquisitions and commissions. She already knew of Leonardo's talent through her contacts in the Sforza court. In 1498 she had written to one of Ludovico's former mistresses, Cecilia Gallerani, specifically asking to borrow Fig. 1616Detail of Virgin's headdress and hair, Madonna of the Yarnwinder (The Lansdowne Madonna), plate 1. Fig. 1515Detail of Child, Madonna of the Yarnwinder (The Lansdowne Madonna), plate 1. Fig. 1717Detail of close and distant trees, Madonna of the Yarnwinder (The Lansdowne Madonna), plate 1. Fig. 1818Detail of filaments of yarn, Madonna of the Yarnwinder (The Lansdowne Madonna), plate 1. Fig. 1919Portrait medal of Florimond Robertet, 1512, bronze, Paris, Bibliothèque Nationale de France. Fig. 2121Leonardo da Vinci (?), Preparatory drawing of the Madonna, c.1501, red chalk and metalpoint on pink prepared paper, 25.7 x 20.3 cm, Venice, Gallerie dell'Accademia. in the Virgin of the Rocks, and had become something of a Leonardo trademark. He was obviously pleased with it, not only because it was a virtuoso piece of draughtsmanship but also because it functions effectively in the pictorial field to define an area of spatial tension, potential action and imminent contact. In the Madonna of the Yarnwinder it serves to express the Virgin's surprise and uncertainty -a critical emotional moment -as she witnesses her son's impulsive embrace of the implicit crucifixion. It was apparent however that this drawing was not by Leonardo himself but by one of his more able followers. As such, it is just one of many indications of how 'authorised' motifs by Leonardo became the stock-in-trade for Milanese artists, available for insertion in compositions as required. In fact, the Venice variant is used in a series of paintings of the Virgin with the holy Children at play (fig. 24), the composition of which seems to have been a Leonardo invention, though none of the surviving paintings is good enough to be regarded as a product of his immediate studio. 13 This suite of drawings, only two of which are directly connected with the genesis of the composition, represents slim pickings compared to the range of studies Leonardo is likely to have undertaken. Perhaps the most notable losses are the kinds of brainstorm sketches that survive for an earlier project for a Madonna and Child, probably undertaken around 1481. The British Museum holds a series of sketches for a composition of interaction between mother, child and cat, including vivacious studies of Jesus and the struggling animal (fig. 25), and a remarkable sheet on which he has scribbled his developing ideas for the complete figural arrangement (figs. 26 and 27). 14 Having densely overlaid alternatives on the page, Fig. 2626Leonardo da Vinci, Study of the Virgin and Child with a cat, c.1481-3, pen and brown ink, and brown wash, over stylus under-drawing, 13 × 9.4 cm, Popham no. 9a (recto), London, British Museum. Fig. 2727Leonardo da Vinci, Study of the Virgin and Child with a cat, c.1481-3, pen and brown ink, over stylus underdrawing, 13 × 9.4 cm, Popham no. 9b (verso), London, British Museum. Fig. 2525Leonardo da Vinci, Two studies of the Virgin and Child with a cat, and three studies of the Child, c.1475-83, pen and brown ink over black chalk and leadpoint,28 × 19.7 cm, Popham no. 11 (recto),London, British Museum. Fig. 2424Follower of Leonardo da Vinci, Madonna with the Holy Children at play, oil on panel, 72.2 × 50.5 cm, Oxford, Ashmolean Museum, University of Oxford. tion of the limbs of your figures and attend first to the movements appropriate to the mental attitudes of the creatures in the narrative rather than to the beauty and quality of their limbs. You should understand that if such a rough composition turns out to be right for your intention, it will all the more satisfy in subsequently being adorned with the perfection suitable to its parts. I have in the past seen clouds and wall stains that have inspired me to beautiful inventions of many things. These stains, while wholly in themselves deprived of perfection in any part, did not lack perfection in regard to their movements or other actions. 15 The next step for the figures would have been studies from life, including details of limbs and hands, and of the draperies. We know that Leonardo, like other Florentine painters in the late fifteenth century, made studies from linen draped over lay figures and soaked in plaster. He also drew some wonderfully compelling studies of compressed layers of cloth that spoke eloquently of the forces that folded and bunched the hillocks and valleys of the fabric as it accommodated Fig. 31 Jacques31Androuet du Cerceau, The château of Florimond de Robertet at Bury near Blois, c.1573-5, ink and blue, yellow and grey wash on vellum, London, British Museum. EDINBURGH, VINCI, NEW YORK: LOOKING, PLANNING AND RESEARCHING 51 50 FORENSICS to the extent that it was becoming difficult to tease out what was what, he turned it over and resumed the brainstorming process, with the composition in reverse. No finished painting seems to have resulted from all this effort. We can well imagine that the complexity of motion and emotional interaction in the Madonna of the Yarnwinder would have required sketches of this type. Leonardo himself provided a clear justification for this design process, which he had essentially invented:Do not draw the limbs on your figures with hard contours or it will happen to you as to many different painters who wish every little stroke of charcoal to be definitive. … Have you never reflected on the poets who in composing their verses are unrelenting in their pursuits of fine literature and think nothing of erasing some of the verses to improve upon them? Therefore, painter, decide broadly upon the posi- Fig. 2929Leonardo da Vinci, A mountain landscape, c.1510-15, black chalk, 8.8 × 14.5 cm, The Royal Collection © 2011 Her Majesty Queen Elizabeth II (RL 12407). Fig. 3030Leonardo da Vinci, Rocky peaks rising from a plain, c.1505, red chalk, 8.7 × 15.1 cm, The Royal Collection © 2011 Her Majesty Queen Elizabeth II (RL 12406). Fig. 28 Leonardo da Vinci, The head of St Anne, c.1510-15, black chalk, wetted in places, 18.8 × 13.0 cm, The Royal Collection © 2011 Her Majesty Queen Elizabeth II (RL 12533). Fig. 3232Exterior view of the townhouse of Florimond Robertet, Blois, France. Fig. 3333Courtyard of the townhouse of Florimond Robertet, Blois, France. LATE SURPRISES FROM NEW YORK 55 54 FORENSICS 5. LATE SURPRISES FROM NEW YORK For a time we hear nothing from New York about the Lansdowne Madonna. Then we learn that the 'anonymous' owner had arranged for the small panel to be examined using infrared (IR) reflectography in the conservation studios of the Metropolitan Museum, where the equipment is as good as any. Disappointingly, the results reported to the National Gallery Keeper Michael Clarke by Gary Tinterow of the Met on Michael's visit to New York are not thought to be encouraging. He writes to me on 31 January 1992: 'There is apparently little underdrawing visible and what is there is uncharacteristic of Leonardo. There is not a great deal of later restoration/repaint. There is, however, an underdrawing for a small hut in the upper left. ' The significance of the 'small hut' is not yet clear. Perhaps we are dealing with a studio replica after all, or even a good later copy. Somehow this seems unsatisfying, and not the place to end the quest. Maryan Ainsworth of the Met picks up the technical research on a private basis and works to assemble the IR images into a mosaic of the whole picture. But the date for the opening of the show is getting ever nearer, and the catalogue is being drafted. Then the results of Maryan's examination arrive, and they are remarkable -as surprising as anything I have seen when paintings have been examined using IR technology. It transpires that the examination at the Metropolitan Museum had not been as discouraging as first reported. Strange! I receive photographs of the infrared mosaic from the second examination (fig. 34). Large portions of the underdrawing are visible, some lines apparently drawn in charcoal or black chalk and others with the brush. Pentimenti abound. Some seem to correspond to those we had observed in the Buccleuch painting, though they are far clearer than those we could see using the National Gallery's less-developed equipment. The Virgin's turned-in thumb is evident, as is some evidence of the revised position of the Child's right arm. This is puzzling, and we begin to consider the possibility that both paintings had undergone parallel development of a kind not previously documented in the analysis of versions of compositions by any artists. In our research for the exhibition, we pay particular attention to what seemed to be a remarkable document, an inventory-cum-description apparently drawn up in 1532 by Robertet's wife, Michelle Gaillard. It did not survive in any sixteenth-century original, but was printed in Henri Chesneau's Bury Rostaing around 1650. 27 It takes the form of a rather personalised list of the sumptuous possessions and furnishings in the château. These include an otherwise unknown painting of the Madonna of the Sorrows by Michelangelo, about which Michelle waxes lyrical, as well as the bronze David. It is possible that Robertet had obtained another Michelangelo, and Michelle says, plausibly, that it came via Cardinal Brissonnet, but no trace survives in copies or other records. The Leonardo Madonna is nowhere to be found in the account, which would almost certainly mean that it was not in the possession of Robertet's widow at the time of her writing. nected with him -a copy of a lost portrait of 1510, the medal of 1512 (fig. 19), a tapestry with Robertet emblematic devices, three drawings by Du Cerceau of the Château de Bury, and a necklace of the Royal Order of St Michael to which Robertet belonged. 29 It was Francis I's interest in the order that led to his commissioning of the Raphael St Michael, which, as we have seen, was to become one of the earliest subjects of Picault's transfer process. Additionally, the loan of some manuscript and medallic material is sought, to introduce visitors to the three monarchs served by their faithful secretary. The show is shaping up nicely. Fig. 3434Infrared reflectogram mosaic of the Madonna of the Yarnwinder (The Lansdowne Madonna), plate 1.1992. Even more remarkably, other areas of the picture have undergone substantial changes. The upper cross-member of the yarnwinder appears to have been planned at a different angle, and a series of straight threads hangs from it. On the rocks below is a series of curved shapes which can be seen in the underdrawing, but it is not clear at this stage what they are. To the left of the Virgin some architectural structures can be discerned, drawn rapidly with a brush. There is an arch in a wall, which we had seen in the Buccleuch underdrawing, and a few inches higher a structure like the overhanging roof of a porch, cantilevered out on two diagonal braces. What we miss, as we had in the IR image of the Buccleuch Madonna, is the significance of the tangle of brush-drawn contours towards the base of the arch. FROM NEW YORK 59 cussion in the second part of this book. It would be nice to be able to do without it. It traditionally relies on what is called having an 'eye' , that is to say being able to make perceptive judgements based on innate and acquired visual discriminations and upon a sustained experience of looking at art works in the original. At one time, it seemed to be a foundational skill in art history. It was what certified which works were produced by which artist, and when. With the rise of the social histories of art, connoisseurship has moved towards the peripheries of academic research in universities, though it is still a mainstay for historians working in galleries and for dealers and auction houses. Indeed, this association of connoisseurship with 'the trade' has served to bring it into disrepute in some facets of academia, as being tainted and playing to a set of values that 'fetishises' the art work rather than analysing it in its social and economic settings. But often, even the social art historian needs to rely on judgements of who actually did what. In the field of Leonardo studies, the building up of a reasonably secure corpus of autograph drawings and their arrangement in chronological order was essential after the chaos of centuries of accumulation and misattribution. Kenneth Clark's great 1935 catalogue of the drawings in the Royal Collection at Windsor accomplished this considerable task on the basis of identifying some core drawings with known, dateable projects, and analysing others in relation to this core according to criteria of style. 1 Clark's exposition of Leonardo drawings, often based on visual intuitions, has stood up remarkably well in the face of later research. His elegant and perceptive mono-graph, published in 1939, also helped establish an inner group of autograph Leonardos in a convincing chronological order. 2 Ultimately, anyone tackling Leonardo, or any other 'old master' for that matter, has to rely to a greater or lesser extent on judgements of style and quality. We have already seen how we cannot understand the drawings relating to the Madonna without making such judgements or relying on someone else's. Bernard Berenson, the famous and now notorious authority on Italian Renaissance masters, referred to connoisseurship as the 'sixth sense' . He did so in the surprising context of a New York court of law (figs. 62-64). In 1921, Madame Andrée Hahn, who owned what she claimed was the original of Leonardo's La Belle Ferronnière in the Louvre, sued Lord Duveen, international dealer and art world grandee, for saying in the press that the Hahn painting was a mere copy. Duveen did not mince his words: 'the painting in Kansas City is a measly copy of which a hundred have been made of this and other subjects by Leonardo da Vinci and offered to the world as genuine. Leonardo never made replicas of works and the real original La Belle Ferronnière is in the Louvre. ' 3 He had never seen the disputed work. As already indicated, the conventional categories of 'original' and various kinds of lesser 'copies' of versions COMMERCE AND CLAMOUR 61 60 FORENSICS used in constructing oeuvres generally does little to respect what we know about the production of various types of picture in busy Renaissance workshops. With this in mind, I applied connoisseurship to the two star exhibits, and to the orbiting set of more remote variants that we were borrowing for the Edinburgh show. Some excerpts from the catalogue, Leonardo da Vinci. The Mystery of the Madonna of the Yarnwinder, give a flavour of my argument. It would be possible to draw up an account of 'plus and minus' factors for each painting -with, for example, the rather spare landscape in the Buccleuch painting as a 'minus' factor and the beautiful handling of Christ's features as a 'plus' factor. … [but] The results of such a balance sheet are unlikely to be decisive if we insist on asking in a crude way which is the 'original' and which is the 'copy' . … I should like to keep open the possibility that both might be 'Leonardos' in the broadest sense of pictures that went out from his workshop as the result of procedures he devised for the making of paintings. … I suspect that both Madonnas of the Yarnwinder were brought required, and 'Leonardo da Vinci. The Mystery of the Madonna of the Yarnwinder' is agreed. The catalogue needs to be designed and produced, and reproduction fees met. Jerry Cinamon of Cinamon and Kitzinger in London works skilfully on the design of the catalogue and the exhibition graphics. Merchandise is sought to sell in the shop -no shortage of that in Leonardo's case, though good-quality items are less plentiful. Education provision is planned, with packs geared to school syllabuses. Attractive public lecture series are to be organised -at a cost where star speakers are involved, especially those who only travel business class. The curators are involved in the planning of all kinds of printed and other outputs -labels, wall panels, brochures, audio tours, press releases -all under the contract for modest fees that would not provide pocket money for other outside consultants employed by galleries on activities that are less central to the gallery's main purpose. Why should this be? By the time this exhibition is under discussion, no gallery has the budget to finance fully its exhibition programmes. Most charge for entrance (which we are not doing) and make money from the shop and the catering. A sponsor is needed, and the curator might need to make a sexy presentation. After discussion with various potential backers, the Gallery cunningly arranges for Martini to be sponsors, to the tune of £ 25,000, under the new sponsors scheme run by the Association of Business Sponsorship in the Arts, which awards additional funding to match the sponsor's contribution. £ 50,000 is in the bag. Martini, prominent Italian drinks company, is hardly a new sponsor in its own right. But they are launching a new product, Martini Brut, a fizzy wine that I am told is not meant to be an affordable kind of Champagne, and somehow they manage to corral their sponsorship into the fold of the new drink. The Association of Business Sponsorship in the Arts seems happy with this device, and so are the Gallery and Martini. Suitable proclamations are made. Martini, as an Italian company, is delighted and honoured to be sponsoring this intriguing and magnificent Leonardo da Vinci exhibition at the National Gallery of Scotland in May. This coincides with the launch, in Scotland, of our new quality spumante, Martini Brut. We are confident that the exhibition will be a great success. The support of Martini made the front page of Sponsorship News together with a picture of the 'Scottish Leonardo' . Sponsoring Leonardo fits in with Martini's image, and the launch of the new drink, but, as invariably with sponsorship, one or more crucial people in the company must be personally enthusiastic about the show and its subject -in this instance, most notably David Rutherford, Director of External Affairs, Martini & Rossi UK. Timothy Clifford, Director of the National Galleries, responded in kind. 'We thank Martini Brut for their generous support, 64 FORENSICS COMMERCE AND CLAMOUR 65 and wish them well with the launch of Martini Brut. They want an image of myself as curator, standing between the battling babies. It is taking too long. One of the gallery's staff pulls the plugs on their lights. The Scotsman photographer, Ian Rutherford, suddenly sees his picture. I am silhouetted in the darkened gallery by the spots still on the paintings. The photographer from The Times, Denzil McNeelance, piggy-backs on his idea, and a series of pictures is taken. Reporters conduct rushed interviews and jot down points in their notebooks. We hope they get things right, but do not have high hopes. On that evening of Thursday 14 May the sponsors' and lenders' dinner is to take place in the Gallery itself, fol-lowing the main private view. The dinner is not to be in the galleries where the exhibition is mounted. The exhibition itself is located in two rear octagonal rooms on the ground floor in the harmoniously classical building. The room on the left is devoted to Robertet and the French context, while the Leonardo paintings and drawings and related material are on the right. It looks good, after all the anxiety and chaos. The dinner is held in the first-floor octagonal room immediately over that housing the Leonardos. This is hardly a second-rank choice, since we are in the company of Gauguin, Van Gogh and other great masters working in late nineteenth-century France. The gallery looks splendid, not least courtesy of £ 721.45-worth of flowers. I am tempted to feel uneasy at the use of galleries in this way, but I remember that many of the paintings we now revere and sanitise in galleries were once part of the living fabric of houses and other buildings with civic, private and ecclesiastical functions. The Duke's picture normally belongs in one of his grand Scottishhouses, that is to say, in a domestic setting, though he has always been keen to grant public access to his treasures. RECEPTIONAngus Grossart, Chairman of the Trustees, arrives on 66 FORENSICS COMMERCE AND CLAMOUR 67 a plane from New York on the day that the exhibition opens. A copy of The Times is available on the flight. On the front page, no less, is Denzil McNeelance's picture taken after the pulling of the light plugs. It is a striking image, as it has to be if it is to meet the criteria of newspaper picture editors. But it is also testimony to the reach of Leonardo, beyond the specialist pages devoted to the arts. He is always news, and the silly season never closes for him; any theory about the Mona Lisa is newsworthy, however daft. The Scotsman had used Ian Rutherford's silhouette image on its front page the day before, in advance of the public opening, accompanying a story by Gillian Harris under the headline 'Madonnas mystery solved at the double' , and reporting my view that 'both Madonnas were in fact the work of the master' . On the arts page, Allen Wright outlines the findings at greater length alongside a large colour reproduction of the Buccleuch Madonna. Scotsman critic, Edward Gage, in a very enthusiastic notice of the show as a whole, following up the paper's pre-opening notices, comes down clearly in favour of the Duke's picture: 'moved by a gut reaction plus a singularly convincing presence of sfumato -Leonardo's most subtle handling of tone -I would place my bet on the Buccleuch version' . 1 Martin Gayford in the Telegraph covers the arguments well, and finally comes to a similar conclusion. 2 It is clear that the brightness of the Lansdowne Madonna is telling against it beside the old-masterly mellowness of its rival. The professional critics make a reasonable job of handling the evidence that there is not one 'original' . James Hall in The Independent tells his readers that 'Kemp spins his own ingenious and rather convincing yarn' . 3 However, the critics find it hard to resist the temptation to give one of the versions precedence over the other. An interesting alternative to the Scottish angle comes from the Globe and Mail in Canada, where the earlier presence of the Lansdowne Madonna in Reford's Montreal collection provides a rationale for running a story on 26 May about a show that few Montrealers are likely to see. Oddly, Artists and Illustrators Magazine manages to reproduce the image of the Lansdowne picture in reverse. John McEwen, a Scotsman based in London, reviews the exhibition for the Sunday Telegraph, missing the point spectacularly. 4 He reports that 'both pictures have recently been subjected to the latest scientific tests in a competitive spirit which demands a winner. What an anticlimax therefore when the referee, Professor Martin Kemp, calls a draw by declaring bits of both pictures to be by Leonardo. ' After running the sporting analogy further than it ought to go, he adds that 'some obviously duff copies make up the numbers, and a further room is devoted to the life and times of the obscure French nobleman who may have been the original owner. It makes this show the most laughable muckle ever made of a mickle. ' It is to be hoped that Telegraph readers have access to a Scots dictionary. I later notice that his name had earlier been crossed off the list of possible invitees to the dinner -maybe an unwise move, in retrospect. Inevitably questions are asked about the 'Private Collection' from which the Lansdowne Madonna was borrowed. Geraldine Norman, an indefatigable art market sleuth, then writing for the colour supplement to the Independent on Sunday, identified the owner as Wildenstein and Co. In her review of the show on 17 May she reports that 'my inquiries in New York suggest it is owned by Wildenstein's, the art dealers whose secret stock of masterpieces is legendary' . As Joseph Baillio of Wildenstein's resignedly wrote to Michael Clarke, 'there was obviously no way to prevent Geraldine Norman from asking questions on this side of the Atlantic' . Her lively and well-informed review, headlined 'Hunting Leonardos' , focuses on the whole question of attributions and financial value. The main BBC TV arts review programme The Late Show arranges for the filming of the paintings. I speak on camera in a London studio kitted out with a framed reproduction and a clutter of Leonardesque paraphernalia. I am asked by the producer not to 'ham it up' as much as I am apparently doing, a request that is surprising given the setting. I do my best to explain the complexities of the issues in the face of the sound-bite culture that even the producers of The Late Show have to satisfy. I decide that I prefer radio. The reviewers writing in the specialist arts periodicals generally do a better and more extensive job, as they should, in addressing the arguments in the catalogue. 5 The experts weigh in. Carlo Pedretti in his own journal, Achademia Leonardi Vinci, and in Il Giornale dell'Arte draws upon his knowledge to provide wideranging and informative assessments of the evidence and to suggest new avenues for enquiry. He is provisionally happy with the hypotheses put forward in the catalogue about the status of the two pictures. Pedretti draws attention to a contemporaneous article by Paul Joannides on optical distortion -picking up one of COMMERCE AND CLAMOUR 69 68 FORENSICS One of the livelier episodes involved a review of the 'Two Ladies' in Antique by John Rowlands of the British Museum, who had been vital in the process of approving the loan of drawings from the Museum. Weighing the evidence carefully, Rowlands tends to favour the Buccleuch picture as involving Leonardo's direct participation, but the magazine contrives to caption the Lansdowne picture as being owned by the Duke and vice versa. What the magazine later acknowledges as 'the hideous mistake of transposing the captions' triggers an article from the Duke of Buccleuch, declaring himself happy at the outcome of the investigations. 8 The two prime versions are together for the first time since their exhibition in London at the Burlington Fine Arts Club in a show of Pictures by Masters of the Milanese and Allied Schools of Lombardy (May-July 1898). Now they are displayed in the National Gallery of Scotland with material relating to theirtechnical examination and with a cluster of variant pictures. The latter are interesting in their own right, not least in relation to the strange development of the composition that is emerging from the infrared analyses, and they are not to be dismissed as devoid of merit as paintings. They also serve less flatteringly to offset the hugely involving and affecting quality of the Buccleuch and Lansdowne paintings. Even the version owned by the Duke of Wellington (fig. 93), which looked good when we saw it in the private rooms of Apsley House in London, is palpably diminished in the presence of the two main pictures. The drawings work well with the pictures to give a sense of the emergence of the dynamic composition from Leonardo's innovative processes of design.The autograph drawing for the woman's torso does seem to be closely related to the pictures now that it can be seen close to them. One of the exhibited drawings now looks more likely to be a forgery than anything else, but in the absence of proof the suspicion is best left unvoiced. The landscape drawings and the background of the Lansdowne Madonna seem entirely compatible, as does the drawing of rock strata from Windsor with the rocks in the Buccleuch Madonna. The key areas in both pictures, namely the two figures and their accoutrements, seem not only to be of high quality but to be essentially similar in handling if not identical in detail. Any pluses and minuses are marginal and not to be confidently adduced given the differences in condition. As I acquire greater facility in making visual allowance for the differences that had been foisted on them by later intervention, I increasingly recognise in both pictures the strange element of tense vitality that characterises Leonardo's handling of trademark details. It involves the endowing of static features, like hair and drapery, with an uncanny sense of mobility that came from his understanding of the forces that had shaped them. The cascading rivulets of curling hair, in particular, speak the same graphic and physical language as his studies of water in turbulent motion. As he wrote:Note the motion of the surface of water, which conforms to that of the hair, which has two motions, oneof which responds to the weight of the strands of the hair and the other to the direction of the curls; this water makes turning eddies, which in part respond to the impetus of the principal current, while the other responds to the incidental motion of deflection. 1 The twisted veils of the Virgin's headdress, looped through each other just above the parting of her hair, play variations on the helical theme. The structure of the loop is more readily and beautifully apparent in the Buccleuch version, in which the defining highlights have survived more wholly intact. The result, on reflection and repeated viewing, supports my argument in the catalogue that both are 'Leonardos' , that is to say paintings of a quality in invention and execution suitable for presentation to major patrons. Indeed, I become less cautious about Leonardo's leading role than before. In making this judgement, I would like to set aside the normal jargon applied to matters of attribution: autograph,partly-autograph, original, studio, workshop, school, pupil, assistant, follower, imitator, copy, version, vari- ant, after, based on -and so on. The process of the works' production is designed to produce remarkable paintings in which Leonardo's brilliant reshaping of the formal, emotional and spiritual dimensions of the Florentine Madonna and Child is presented in a fully effective manner. If Leonardo involved more than one person in their making, the procedures would not be to the detriment of the functions that the image per-72 FORENSICS formed, either for Leonardo or its intended recipient. been comprehensively dispelled, and terrific areas of further research have been activated. As the Keeper Michael Clarke wrote in his letter of 8 September to the Duke of Buccleuch, thanking him for his generous loan, 'the fascinating debates continue on the little picture which Leonardo painted for Florimond Robertet' . In a hand-written postscript, he adds, 'I look forward to seeing your picture at Trafalgar Square. ' After the exhibition closes on 12 July -accompanied by the inevitable sense of emptiness that always follows the ending of a show integral to a period in one's life -the Duke of Buccleuch and the National Gallery in London (Trafalgar Square) agree to put his Madonna on show for a few months. The Duke had written to Neil MacGregor, Director of the Gallery, on 12 August suggesting that it might be 'of interest for you to display it until then [April 1993], perhaps in the grotto near your Cartoon with which it seemed to me to have an affinity' . I have no direct role in this decision, which is being facilitated by Michael Clarke, but express enthusiasm for conducting further examination of the picture. It is good that the Edinburgh exhibition has initiated additional research and further exposure of the picture and that the large number of visitors to the National Gallery in Trafalgar Square should have an opportunity to see it. A SURPRISING POSTSCRIPT While at the National Gallery in London, the Buccleuch Madonna is subject to further analysis using infrared reflectography. The examination is conducted by Rachel Billinge, Leverhulme Research Fellow, under the direction of Martyn Wyld of the conservation laboratory and the curator Nicholas Penny, who had reviewed the show in The Burlington Magazine. The Gallery staff in London know that the equipment and the computer processing they can apply to the image should clarify and extend what had been achieved in Edinburgh. I am in Australia when the examination is conducted, but slides of those portions of the IR images that yield the most interesting results are forwarded to me. They are even more surprising than those we saw in the IR mosaic from New York (figs. 35-40). The pentimenti we had already observed are clarified, as might have been expected. The fine, certain underdrawing in the heads of the Virgin and Child has become more clearly apparent, particularly in places where the painted contours differ discernibly from those originally sketched in. There might be hints of spolveri (dots made during transfer from a pricked cartoon) in the contours around Christ's eye, but they are too elusive to be confirmed. The pentimento of the Virgin's thumb is clarified, and there seem to be two earlier stages, with the thumb tucked in to greater or lesser degrees. The yarnwinder does not in itself apparently exhibit the kinds of major pentimento found in the Lansdowne version. But, on the rocks at the right of the foot of its shaft are the same elongated oval contours as we had observed in the Lansdowne IR reflectogram. Very surprising. Even more remarkable is what emerges in the middleground on the left. First, at the level of the Virgin's right hand are very clear signs of the multiple-arched bridge with which we are familiar in both the underdrawing and final state of the Lansdowne Madonna. It was once intended to be present in the Buccleuch picture, but was entirely painted over. Above the bridge is the arched portal, already observed in the CONCLUSIONS AND CONSEQUENCES 73 Fig. 35 Infrared reflectogram mosaic, detail: head of the Virgin, Madonna of the Yarnwinder (The Buccleuch Madonna), plate 2. 1992. Courtesy of the National Gallery, London and Buccleuch Living Heritage Trust. Fig. 3636Infrared reflectogram mosaic, detail: head of the Child, Madonna of the Yarnwinder (The Buccleuch Madonna), plate 2. 1992. Courtesy of the National Gallery, London and Buccleuch Living Heritage Trust. Fig. 37 Infrared reflectogram mosaic, detail: Virgin's thumb, Madonna of the Yarnwinder (The Buccleuch Madonna), plate 2. 1992. Courtesy of the National Gallery, London, and Buccleuch Living Heritage Trust. Fig. 3838Infrared reflectogram mosaic, detail: yarnwinder and rocks, Madonna of the Yarnwinder (The Buccleuch Madonna), plate 2. 1992. Courtesy of the National Gallery, London, and Buccleuch Living Heritage Trust.. All the normal expectations and rules are being broken. Both pictures exhibit complex underdrawings that appear to have been very similar but not identical at a very early stage in their development. The subsidiary narrative group in the left background, and other shared preparatory moves, indicate that they have undergone parallel pro-cesses of compositional simplification. The bridge, which could not have co-existed with the baby-walker group and the arch, seems to have entered as a second thought in both pictures. But it was to survive into the finished painting only in the Lansdowne picture. These remarkable findings lend welcome support to the hypothesis that the two paintings were commenced more or less simultaneously. The underdrawings were laid down in essentially the same way, with a few experimental variations (as in the animal and platform). At some relatively early point, it was decided to eliminate the baby-walker group, probably because it was overcomplicating the effect of the main group in such a small painting. Initially, in both, it was replaced with the type of bridge that features in the Mona Lisa. Then the Lansdowne picture proceeded (probably not rapidly) to its finished state, with bridge and mountainous landscape, while the Buccleuch version, perhaps later, gained its simpler landscape of low hills and water. Earlier suspicions that the composition in the unfinished Madonnas had leaked out -either via pupils' knowledge of the paintings at an early stage of their CONCLUSIONS AND CONSEQUENCES 75 Fig. 39 Infrared reflectogram mosaic, detail: bridge in the left middle ground, Madonna of the Yarnwinder (The Buccleuch Madonna), plate 2. 1992. Courtesy of the National Gallery, London, and Buccleuch Living Heritage Trust. Fig. 4040Infrared reflectogram mosaic, detail: arch and figure group on the left, Madonna of the Yarnwinder (The Buccleuch Madonna), plate 2. 1992. Courtesy of the National Gallery, London, and Buccleuch Living Heritage Trust.. 76 FORENSICS development or via drawings (or perhaps via a cartoon) -are also confirmed. The shared presence of the baby-walker group in the two underdrawings and in the 'family' of pictures represented by the Edinburgh variant gives this 'family' an altogether more important status. In the Edinburgh catalogue I had suggested that the Edinburgh Madonna originated with Fernando Yáñez, a Spanish painter who appears to have been in Leonardo's workshop in 1505. 2 Whether this precise attribution is right or not, the scenario makes sense. Someone in Leonardo's studio knew the paintings in their first state, and had perhaps left the studio before their modification, subsequently utilising his knowledge as the basis for one or more paintings. If the connection with Yáñez is right, the paintings remained in their first state until at least 1505. This would fit with the idea that at least one of them was only completed (perhaps rapidly) in the period immediately before its delivery to Blois in 1507. An opportunity to publish the new observations is provided by an invitation to contribute to a volume of the Journal of Medieval and Renaissance Studies being dedicated to John Spencer, the distinguished historian of Renaissance art at Duke University, North Carolina. Apart from my not knowing him personally, he had contributed an unflattering review of the Leonardo exhibition I had curated with Jane Roberts at Windsor in 1989, but it seems ungenerous to say no, particularly since the editor is Louise Rice, a respected friend. Rather than just reporting these new results as a kind of appendix to the Edinburgh exhibition, reinforcing the arguments I had made on that occasion, I decide, as a kind of homage to John Dick and his fellow conservators in London, to effect a conceptual marriage between technical examination conducted in galleries and one aspect of the social history of art favoured in universities. I had also, in the meantime, explored and photographed Robertet's palace at Blois, and trekked fruitlessly around the fields where his Château de Bury must once have been. In the essay, 'From scientific examination to the Renaissance market: the case of Leonardo da Vinci's Madonna of the Yarnwinder' , I argue that Technical examination of pictures can have very direct implications for our understanding of patronage … I am claiming that unless we use all the available skills to know what kind of objects we are studying -in terms of their past and present makeup, their date and authorship -we will be on insecure ground in examining the whole range of questions about the production and patronage of artefacts. 3 sual symbol. The only possible precedent known to me occurs in the Virgin and Child in gilded marble carved by Alberto Arnoldi in 1364 for the chapel of the Compagnia della Misericordia beside what is now called the Loggia del Bigallo in Florence, in which the Virgin and Child seem mutually to hold a rather decorative yarnwinder above an orb. 1 There is a possible link suggested by Leonardo's commission for the Compagnia della Misericordia in Milan, which resulted in one of the two versions of the Virgin of the Rocks. Was the 78 FORENSICS yarnwinder a symbol favoured by the Compagnie? Or was it known in France, given the French patronage of the picture? Whichever, the identity of the object in the 1364 Madonna is not entirely obvious. By contrast, Leonardo's yarnwinder, however unusual, is as clear in its form as it is in its meaning. Should anyone be tempted to miss the point, the blood-red filaments of yarn that trickle across the rock strata below the yarnwinder in the Lansdowne Madonna (fig. 18) leave us in no doubt about what was meant. In one of the 'families' of copies, represented by the Edinburgh picture (fig. 2), essentially the same iconographical reference is made by the twisting vine motif that wends its way up the shaft of the yarnwinder. The vine bears bunches of grapes as an allusion to the red wine of the Eucharist as embodying the blood of Christ. What is wholly exceptional in Leonardo's composition is what we have already called 'dynamic symbolism' . Rather than existing inertly as a symbolic reference inserted somewhat arbitrarily into the composition, the yarnwinder is incorporated into the physical and psychological action of the figure group. Fra Pietro understood this precisely. He recognised that something was happening, that he was witnessing a deeply meaningful action, which implied both before and after. Thus, looking at the cartoon that he saw on his first visit, His Mother, half rising from the lap of St Anne takes hold of the Child to separate him from the lamb (a sacrificial animal) signifying the Passion. St Anne, rising slightly from her seated position, appears to want to restrain her daughter from separating the Child from the lamb. She is perhaps intended to represent the Church, which would not have the Passion of Christ impeded. 2 His uncertainty as to whether St Anne represents the Church indicates that this kind of 'action Madonna' set up something of an open field, into which the spectator could project a variety of possible readings. The cartoon has not survived, but the related painting in the Louvre (fig. 41) shows the fluency with which Leonardo orchestrated the integrated motions and emotions of the protagonists. Fig. 4141Leonardo da Vinci, Madonna, Child, St Anne and a Lamb, c.1510, oil on poplar, 168 cm × 13 cm, Paris, Musée du Louvre. in devotional pictures is not exceptional, and may have been popularised in Florence by artists in the Netherlands. A particularly apposite example is the Portinari Altarpiece painted for Florence by Hugo van der Goes, in which the Annunciation to the Shepherds occurs in the landscape of the main panel (behind the adoring shepherds), while Joseph can be seen helping a weary Virgin along the road in the background left-hand panel of saints and donors. The Holy Family is fleeing to Egypt in the face of Herod's massacre of the innocents. The Portinari Altarpiece was a famous object in Florence, and much admired by artists of Verrocchio's generation. Ghirlandaio had borrowed its rustic-looking shepherds for his Adoration in the Ospedale degli Innocenti. Although pausing to make a baby-walker is not a known episode from the Family's journey into exile, it could well have been intended and recognised as an ingenious recasting of the 'Rest on the Flight into Egypt' . The meeting of Christ and St John, accompanied by the angel Uriel, as depicted in the Virgin of the Rocks, was an event that also occurred during the period in exile -according to later embroideries of the biblical account -and Leonardo was well aware of the story-telling potential of that phase in Christ's life. Three of the figures in the planned group are readily recognisable, the Virgin and Child with Joseph. The third woman, standing behind the Virgin, may either be one of the other Maries, St Anne (the Virgin's mother), or, most probably, one of the midwives often inserted into Holy Families on the basis of apocryphal later gospels and stories. The object Joseph is making is not, as far as I know, present in any earlier or contemporary paintings or any other records. Given the popular tendency to credit Leonardo with the invention of any device that he represented, we might feel tempted to see the baby-walker, with its open frame and small wheels, as another, if domestic, example of him as a man 'ahead of his time' . There is something nice about the idea of Ludovico's or King Louis's children toddling around precociously in Leonardo's walking frame. I suspect, however, that it is simply one of those everyday devices that had escaped the historical record before Leonardo's time -or at least the historical record known to me. Either way it is a diverting idea for a Madonna and Child composition, and unquestionably appropriate for Joseph, who was a carpenter by trade. There is a later Florentine drawing 80 FORENSICS that depicts a baby-walker constructed by Joseph. The Child, held by his mother while she is seated on the lap of St Anne, eyes the device with obvious excitement. Perhaps there are more examples to be found. In his review of the Edinburgh show, Michael Bury provided the textual key that brings together the activities of the Virgin and her husband in the underdrawings. 3 In the popular Meditations on the Life of Christ by an unknown Franciscan author called the pseudo-Bonaventura, details of Christ's early life in exile are imaginatively inserted into the large gaps left by the biblical narrative. The author, as quoted by Bury, supposes that Mary 'provided the necessities for herself and her son with the spindle and the needle; the Lady of the world sewed and spun for money. … The old St Joseph practised the art of the carpenter. ' The Virgin, seated directly on natural rock ledges rather than a grand throne, is a recognisable variation of the 'Madonna of Humility' , in keeping with the emphasis upon her undertaking humble handiwork in the Meditations. The architecture in the eliminated underdrawing, which does not appear to represent a complete building, is familiar in kind from Adorations, Nativities and other subjects drawn from Christ's very early life, alluding to the rude setting of Christ's birth, the humble circumstances of his infancy, and sometimes to the origin of a new world from the wreckage of the old. A more elaborate version appears in Leonardo's unfinished Adoration of the Magi in which a hugely elaborate structure of arches and staircases represents the ruins of antiquity, while a rustic shed (albeit of grand proportions) rises to the right. Something more like the modest structure in the underpainting of the yarnwinder compositions occurs in the drawing of the Madonna with the Holy Children in the Metropolitan Museum in New York (fig. 42), which experiments with compositions involving one of both of the holy babies. 4 Such structures are very much stock elements in the scene-setting for a Virgin and Child or a Holy Family. Fig. 4242Leonardo da Vinci, Study of the Madonna with the Holy Children, c.1494, metalpoint reworked with pen and dark brown ink on pink prepared paper, 19.3 × 16.2 cm, New York, The Metropolitan Museum of Art. ICONOGRAPHIES 81the Virgin of the Rocks in London. When an IR scanner borrowed from the Istituto Nazionale di Ottica in Florence delivered its infrared reflectogram of the National Gallery painting, we were surprised to see clear signs that Leonardo had considered radically revising the composition (figs.43 and 44). 5 The drawing on the gesso priming of the panel reveals that the Virgin's pose -with her right arm extended and left hand on her chest -combines motifs from the New York study and the St Philip in the Last Supper. In the event, Leonardo sensibly decided not to re-invent the whole composition but settled on retaining the composition of the Louvre version of the Virgin of the Rocks, with minor adjustments. The obliterated underdrawing in the London painting provides the closest parallel to what happened during the preliminary stages of the Buccleuch and Lansdowne Madonnas. Given the ingenuity of the planned baby-walker group -an ingenuity recognised in the versions that retain it -why did he paint it over? Although its planned inclusion testifies to Leonardo's complex ambitions for this small panel, and Leonardo was never short of complex ambitions, it would have created a diverting but inappropriate focus of interest within the limited compass of a little picture already loaded with things demanding sustained attention and interpretation. Some of Leonardo's schemes for allegorical decorations in praise of his patrons are horribly overloaded with symbolic motifs and arcane allusions. In this instance he seems to have thought better of being too clever and overtaxing what he required of the spectator: thus, out go the baby-walkers, and in come the landscapes. Even as it is, the deeper meaning needs someone as per-ceptive and well-informed as Fra Pietro to make full sense of it. LANDSCAPES The final landscapes in the two pictures may also be endowed with meaning, beyond their presence as relatively novel outdoor settings for the Madonna and Child. The mountainous background in the Lansdowne Madonna is a recurrent feature in Leonardo's works, making its first appearance through the arched windows in the early Madonna and Child with a Carnation in Munich (fig. 14). Its most famous manifestation is of course in the Mona Lisa. In that painting it almost certainly alludes to Leonardo's beloved analogy between the 'lesser world' of the human body and the 'body of the earth' -that is to say, the ancient theory of the essential unity of the microcosm and the macrocosm. The idea appears in various guises in his notebooks and drawings. The most complete and mature expression occurs in the Codex Leicester, and can be dated to c. 1506-8: Nothing arises in a place where there is neither sensitive, vegetative nor rational life. Feathers arise upon birds and change every year; hair arises upon animals and changes every year, except some parts, like the hair of the beard of lions and cats and similar; grasses arise upon the meadows and leaves upon the trees and are renewed in great part every year; so that we might say that the earth has a vegetative soul and that its flesh is the soil, its bones are the arrangements of the connexions of the rocks of which the mountains are composed, its cartilage is the tufa, its blood is the veins of waters; the lake of the blood, which is throughout the heart, is the oceanic sea; its breathing 82 FORENSICS Fig. 43 Infrared reflectogram, Leonardo da Vinci, The Virgin of the Rocks (The Virgin with the Infant Saint John adoring the Infant Christ accompanied by an Angel), c.1491-1508, 2005. Courtesy National Gallery, London. ICONOGRAPHIES 83 Fig. 44 Outline diagram of the first underdrawing of The Virgin of the Rocks, fig 37. Courtesy National Gallery, London. 84 FORENSICS and the increase and decrease of the blood through the pulses in the earth is thus: it is the flow and ebb of the sea; and the heat of the soul of the world is the fire which is infused throughout the earth; the seat of the vegetative soul are the fires, which breathe, in various parts of earth through baths and mines of sulphur, and in Vulcanus and Mount Aetna in Sicily and in many other places. 6 The parallels between the bodies of humans and the earth are less obvious in the Madonnas than in the picture of the living woman seated magnificently against the 'bones, flesh and blood' of an ancient tract of land. However, any landscape depicted by Leonardo was underpinned by how he looked at the 'body of the earth' . The meandering 'vein of water' (again Leonardo's term) passing under the bridge and the fertile flesh of the soil-clad slopes and plateaux are all expressions of the analogy. In the Lansdowne Madonna the analogy is probably not central to the painting's meaning, but endows the landscape with a philosophical gravity appropriate to the depiction of an incarnate (embodied) divinity and his virgin mother. Christ is the Word made flesh. Moreover, God's first acts of creation had been to separate the light from the darkness, the heavenly bodies from the world, and the waters from the earth. Leonardo came to reject the biblical account of the Deluge as solely responsible for the shaping of the earth and for the presence of fossils in high mountains, but he did not doubt that an ineffable God had designed the whole system and was the author of the microcosm-macrocosm analogy. Indeed, such grand designs were incorporated into his creation precisely so that we could understand his transcendent purpose. The incarnate or 'en-fleshed' nature of the divine spirit of Christ is also evident in the portrayal of his wholly naked body. It was the real presence of his body and his blood that was celebrated in the bread and wine of the Eucharist. Originally, his overtly displayed genitals would have left us in no doubt that he had been placed by God on earth as a man, fully clothed in human flesh. The displaying of Christ naked or semi-naked, with his genitals on display, was a common feature in Renaissance Madonnas, and it does not require recourse to theological texts to understand that his essential humanity is being made evident. 7 In the Lansdowne Madonna, as is evident from the technical examination and the photographic record of the painting over the years (as we will see later), drastic things have occurred in the regions of Christ's penis. Veiling draperies have come -presumably added in prudish ignorance of his genitals' theological import -and gone. 8 His penis has been re-made -and later un-made. Paint loss through vertical cracking is also evident. The result is an incoherent dark blur, containing only ghosts of anatomical forms. The best we can say is that the damaged area does not attract too much attention. His genitals in the Buccleuch Madonna are intact and fully apparent, though in a region of soft shadow that leaves them relatively reticent. The landscape in the Buccleuch Madonna is notably less characteristic of Leonardo than is the Lansdowne mountainscape, less characterful and less engaging. It is pleasant enough, with its water and rolling hills, and something similar seems to be present in the underdrawing to the right of the child in the Lansdowne Madonna. The aerial and colouristic perspective is ICONOGRAPHIES 85 nicely observed, particularly in the generalised island, though the effects are somewhat vitiated by darkened varnish. Even if it is relatively routine in appearance, however, the Buccleuch landscape is not without its point. It may well refer to the well-known description of Mary as 'Star of the Sea' -the stella maris of the 'Song of Solomon' in the Old Testament. This biblical tract, sometimes called the 'Song of Songs' , a paean of love directed (not unerotically) at an unknown beloved, was a rich quarry for Marian symbols. The rose without thorns and the lily that does not labour are among the most well-known. The angel in Leonardo's early Annunciation proffers the inevitable white lily. One traditional way of making the allusion to the 'Star of the Sea' was to include a prominent star on the Virgin's garment, as happens in the Botticelli Madonna purchased a few years ago for the National Gallery in Edinburgh, in which a gilded star is placed on the shoulder of the Virgin's rich blue cape. Botticelli has also included a fine trellis of roses. Leonardo was little given to such static allusions as a star on a garment, and would have preferred to incorporate the reference into the naturalistic fabric of the image. If this is right, the decision to paint out the baby-walker group, the architecture and the arched bridge would have been Leonardo's. If it was he who actually painted the new landscape, he did it in a tired and unengaged mode. It is easier to see it as added by someone in the workshop when he decided to finish the picture, probably either around 1507 if it is the picture sent to Robertet or later if it is the version that remained in his studio. Regardless of the niceties of authorship, we can say with complete confidence that in 1501 Leonardo was rewriting how the Madonna and Child were characterised with respect to both meaning and form. Indeed the formal and iconographical revolutions are absolutely integrated in his conceptual process. This rewriting is possible as the result of his revolutionary design methods -his graphic brainstorming -in which form and content evolve in an indissoluble unity. It is exceptional to see so much brainstorming continuing on the surface of the primed panel itself. publication in 1994 of the article on the science and social history of the Madonnas in the Journal of Medieval and Renaissance Studies, it seemed that was that. Although, inevitably, there were further questions to be answered, not least in relation to Leonardo's studio practice and to that of Florentine artists in general, we had taken the evidence that emerged before and immediately after the Edinburgh exhibition about as far as it would go. What was needed was technical evidence of comparable quality from paintings by Leonardo and artists in his immediate circle. But I had no means of triggering such an investigation. BILL GATES'S CODEX Then, six years later in 2000, an opportunity arose that provided an ideal and unexpected way of taking the investigation forward. Four years earlier I had been the consultant and part-author of a CD-ROM for Bill Gates, who in 1994 had purchased the Codex Leicester of Leonardo for $ 30,802,500. The Codex was largely dedicated to the movement of water and the 'body of the earth' , and had been owned since the eighteenth century by the Earls of Leicester. In 1980 it had been sold at auction for $ 3.5 million to the petroleum magnate and egotist Armand Hammer, later a convicted fellon, who used his company's money to buy it. It was renamed the Codex Hammer and removed from its binding so that it could be exhibited as a series of separate sheets. After Hammer's death the 'Codex' remained in the galleries built to house his uneven collection of art, located behind Occidental Petroleum's office block on the Wilshire Boulevard between Los Angeles and Santa Monica. Characteristically for someone who consistently disappointed those organisations to which he enjoyed promising his largesse, the museum was not adequately endowed, but was confusingly tied in with the finances of Occidental Petroleum, resulting in a complex legal case. In due course the University of California at Los Angeles was persuaded to adopt Hammer's collection if it came with suitable THE UNIVERSAL LEONARDO AND THE LABORATORY 87 financial terms -courtesy of the sale of the Codex, again at auction. This time it was acquired by a purchaser who has treated it with admiration and respect. Under the aegis of Corbis, the company he had set up to handle the photographic reference archives he was acquiring, Bill Gates asked Curtis Wong of Microsoft to assemble a high-level team of producers, programmers, authors, researchers, scholars and scientists. Together we pioneered ways of displaying, analysing and demonstrating what this demanding Codex was about. Consisting of 72 pages devoted very largely to Leonardo's scientific interests, particularly technical questions of water in the 'body of the earth' and the geological history of the world, it consists largely of dense pages of text with marginal illustrations.The Codex Leicester (as it is now called again) is difficult for the non-specialist. It is not really a portfolio of exhibitable 'drawings' , but rather a set of draft pages from a disputatious treatise, with characteristic excursions into other areas. We devised a variety of ways, including animations of Leonardo's drawings and video clips of his experiments recreated in the laboratory, to bring the Codex to life and to produce accessible insights into its content. It remains a pioneering and exemplary work of its kind. We are now working on an innovatory edition of the Codex for the internet.Bill Gates had been active in lending the Codex to galleries and museums around the world, each of which accompanied the exhibited folios with some other exhibits that they felt appropriate, and the results were uneven. Fred Schroeder, who handles the exhibition programme for Gates, asked on his behalf if it would be possible to suggest a different and more innovatory way to display the Codex sheets. At this time I was beginning to work with Marina Wallace of Central Saint Martins in London on a series of projects for exhibitions, and we had decided to set up a company, Wallace Kemp Artakt, to take forward the plans without the encumbrance of institutional inertia. We met with Fred Schroeder and devised a scheme in which Bill Gates's question would be the catalyst for rethinking the whole concept of the 'blockbuster' The genuine comparability of the results would overcome the kind of problems we had encountered with the piecemeal examinations of the two Madonnas. The major state conservation institute in Florence, the Opificio delle Pietre Dure (named because the workshop it occupies 88 FORENSICS was previously devoted to the 'hard stone' mosaics that became a Florentine speciality), agreed to provide the core staff of what was to be called the Leonardo Laboratory. Cristina Acidini, Director of the Opificio, agreed to become the head of the Laboratory, with Cecilia Frosinini (an art historian in the Opificio) and Roberto Bellucci (renowned specialist in the conservation and examination of Renaissance panel paintings) playing leading roles. Bill Gates was keen on the idea, and provided essential funding for the research and development of the project, at the same time making it clear that he would not bank-roll the whole enterprise. Once this first phase was complete, and when some of the tricky initial diplomacy had been conducted with the main participants, the Council of Europe provided settingup funding, which was used to establish the Leonardo Bureau as the central coordinating body. This was located in the newly established Innovation Centre in Central Saint Martins, which also provided the home for the company Wallace Kemp Artakt in reconfigured form as Artakt at Central Saint Martins. Since I remained a full-time employee of Oxford University, Marina Wallace within the College became its sole Director, and I became involved on a contractual basis, as was the case with my other work outside the University, particularly for the curation of exhibitions. in the story. Having worked in the Museum of Fine Arts at Houston, she had begun to investigate the possibility of employment in Britain at the very time when the new research was being commissioned. She naturally became the first employee to work on the Leonardo research and a founding member of the Universal Leonardo team.She subsequently assumed special responsibility for the 'Lab' and she was our representative in Florence while the tests were being made, liaising closely with the Opificio and the scientists.CLAIMED BY AREZZOThe next significant episode in the story of the two Madonnas involves, somewhat surprisingly, the Tuscan city of Arezzo, about 35 miles from Florence.Famed for its fresco cycle of the True Cross by Piero della Francesca in the church of S. Francesco, and noted as the birthplace of Giorgio Vasari, sixteenthcentury biographer of artists and himself a leading entrepreneur of painting, sculpture and architecture, Arezzo did not seem to have any special claim to Leonardo. However, a local geologist, Carlo Starnazzi, was arguing that the archetypal Leonardo landscape of jagged mountains was based on topography near Arezzo and more specifically that the bridge in both the Mona Lisa and the Lansdowne Madonna is recognisable as the Ponte di Buriano. In general terms, these observations are fair enough, providing we realise that Leonardo was 're-making' nature in his painted THE UNIVERSAL LEONARDO AND THE LABORATORY 89 Arezzo, the consortium of leading businesses in the city and its environs. The Commercianti are opening a newly refurbished exhibition space in the Palazzo dei Priori, the old seat of government, high on the steep main road that rises to the Cathedral, where one of Piero della Francesca's frescoes, Mary Magdalene, is located. Piero had also provided his great series of murals on the story of the True Cross for the choir of S. Francesco in the city. The Aretines know that having Leonardo on show will bring considerable prestige and many visitors to the centre -Piero's frescoes clearly do not do the publicity job on a big enough scale. Whether the Commercianti asked for the loan of the Mona Lisa is not known to me, but they successfully obtain the Lansdowne Madonna -so that the landscape might be seen to come home, as it were. The painting is hung in its own small room, selectively spot-lit, and looks superb. The other exhibition rooms contain items, texts and reproductions of related material. The handsome catalogue is written by Carlo Starnazzi with contributions by Carlo Pedretti. Starnazzi the geologist expounds his ideas about the landscape, with nice juxtapositions of photographs of topography near Arezzo and Leonardo's drawings and paintings. Pedretti undertakes an extensive and fair-minded review of the history of the painting and its attributions to date. Following this initial examination by eye, the owner very generously consents to the Lansdowne Madonna being the first subject for the Leonardo Laboratory's research programme, which the Opificio had defined as involving eleven separate procedures. Some real courage is required on the owner's part because, even in the face of what we already knew, one definitively 90 FORENSICS negative finding could still transform the painting from a Leonardo into an also-ran. Since the owner does not have the facilities for the examinations, the panel is taken to the high-tech premises of the Opificio, and it remains in Florence for a period of seven months from November 2002 to June 2003. To have a painting available over an extended period for such a battery of examinations is wholly unprecedented. This generous dispensation allows the tests to be refined and each technique to be calibrated, if necessary by working out how an existing method could be revised or a new technique could be set up to give optimum results. Effectively, the Leonardo Laboratory researches both the picture and the techniques, establishing the standards and the methods for the programme as a whole. This book is one of the ways in which the standards are being recorded and made available, both for the Universal Leonardo project and beyond. conducted in-house by the Opificio. The IR scanner has been developed for them by the Istituto Nazionale di Ottica Applicata (INOA) at the University in Florence and is unrivalled in the world. It can uniquely produce a full-sized, continuous image that can be judged and juxtaposed or overlaid on a one-to-one basis with the painting. The resolution is outstanding. For other tests, the precious panel has to travel short distances. The nuclear accelerator that generates the proton induced X-ray emission (PIXE) occupies two large rooms at INOA, and certainly does not qualify as portable equipment under any rubric. Working with the scientists is a joy after the factionalism and pettiness of the art world. They are, one imagines, competitive and territorial beasts in their own jungles, but here they let their enthusiasm motivate action.Major galleries in other countries tend to claim that they do not need the INOA scanner, since their own equipment is as good -which is simply not true. It is interesting that the National Gallery in London eventually imports the Florentine equipment to examine its Virgin of the Rocks, with spectacular results, as noted in Chapter8. The results of the examination in the OPD (discussed elsewhere in this book) are so striking that we organise two meetings to review them. The first is held in London on29 September 2003, and involves presentations from all those responsible for the tests. The audience includes those directly involved in the project, and representatives of those institutions that hold important material, with the notable exception of the Louvre, which has so far failed to send anyone to any of the setting-up and coordination meetings. We are learning a great deal about the rhetoric and realities of internationalism. The exchange of views at the first meeting is stimulating and enthusing. A second symposium is organised in Paris in December under the auspices of the Centre de Recherche et de Restauration des Musées de France headed by Jean-Pierre Mohen. Although the Centre occupies newly refurbished premises in the Louvre, they are not the conservation department of the Musée du Louvre, but rather a national service to which the Louvre has access. At this meeting, we find we are talking to two sets of people with quite different interests and financial imperatives. Again, high-level exchanges serve to increase confidence in the quality and utility of the Universal Leonardo project, if not to overcome all the barriers erected by political agendas to which we are not privy. The Opificio delle Pietre Dure, under Cristina Acidini's skilled direction, has an established record of collaboration with science research centres in Italy to advance the examination of works of art. The techniques that now achieve such striking results in conservation lab-92 FORENSICS oratories of museums and galleries were not originally invented specifically with art in mind. The research and development money for the equipment requires mainstream scientific and medical applications; whereas the art world is not big enough or rich enough to warrant the investment. However, once a technique has been developed and applied in its intended area, it may well be useable, in existing or adapted form, for other purposes. Thus, X-rays, discovered by Wilhelm Röntgen in 1895 during his research into invisible emissions in Crook's tubes, and progressively applied for noninvasive medical analysis of living bodies, were found to provide remarkable disclosures with some old master paintings. This applies to all the techniques we are using, though with less delay between the inventions, their intended application and their subsequent transfer to museum laboratories. The project involves a series of leading research centres at the University of Florence, the National Institute of Applied Optics (INOA), the Laboratory of Applied Spectroscopy in the Institute of Applied Physics, the Nuclear Accelerator Laboratory in the Department of Physics, and Siemens Medical Systems in the Careggi Hospital, together with the Institute for New Technology, Energy and the Environment in Rome and the Department of Chemistry in the University of Perugia. If only institutions in the art world demonstrated comparable habits of ready collaboration, our research would be faster and more far-reaching. The set of tests devised by the Opificio delle Pietre Dure and their partners as the standard for the Universal Leonardo project fall into three main types, classified according to the subject of enquiry, that is to say the ends, rather than the scientific means. IN THE LAB 1 Beginning with those methods that are most familiar and which yield the most spectacular visual results, the first type aims to reveal what lies below the surface of the varnish and paint films. X-radiography and infrared reflectography are the standard techniques. Both are available in Florence using the most advanced equipment and achieving the most precise results. The wholly new method involves the application of an extra-high-resolution CT (Computed Tomography) scanner under development by Siemens Medical Solutions in conjunction with the Florentine Hospital at Careggi (site of one of the villas of the Renaissance Medici). The method involves the use of X-rays with highly sophisticated computer analysis and imaging to build up a slice-by-slice layer of the object under investigation (originally intended to be the human body) to produce a complete spatial rendering of the object and, if desired, this may be viewed in animated sequences. The limiting factor is the thinness of the slices, and the new machine achieves a remarkable degree of resolution that yields results with subjects as thin as the layers of pigment and priming on a panel or canvas. The second type of test informs us about the substances from which the painting is made, both originally and as the result of later interventions; if the latter can be distinguished from the former, so much the better. The intermittent presence of cobalt blue, for example, signals areas of retouching, since cobalt was not used THE UNIVERSAL LEONARDO AND THE LABORATORY 93 until the nineteenth century. One of the established techniques involves taking small paint samples, usually from damaged areas of the picture, and subjecting them to analysis. Fragments and sections are viewed in a microscope, disclosing the structure of pigment particles and binding media. Different pigments exhibit recognisable particulate forms and crystalline structures. They can also be subject to analysis by chroma-tography, which tells us about the elements present in the sample. Mapped on a graph, the profiles of the elemental composition of the pigments point to the likely identity of each pigment. A blue made from azurite (a crushed mineral), for instance, will exhibit a quite different profile from one made from powdered lapis lazuli (ultramarine), a much more expensive pigment. The Opificio recommends that we do not promote the idea of using paint samples. The method is destructive, even if on a tiny scale, and is limited to those portions of a picture from which samples can readily and unobtrusively be taken. These portions might well not correspond to the areas in which we are most interested. The Opificio proposes alternative methods, which can deliver much, if not all, that pigment sampling can achieve. The relative disadvantage of the non-destructive methods, which is being progressively overcome as they are developed, is that they are less direct in giving access to deeper layers of the pigments and priming. The great advantage is that analysis can be conducted on any area of the painting and attention focused intensively on regions of particular interest. The methods recommended by the Opificio with its collaborators comprise a daunting list. They are: ), and proton induced gamma emission. Each of this formidable-sounding battery of tests is all directed at detecting the elemental composition of the strata in the painting, either involving the surface layer or penetrating into the deeper strata. Some of the equipment is spectacular. It is dramatic to see the small, 500-year-old panel poised under the searching electronic eye of the PIXE scanner, at the output end of two rooms, worth of nuclear accelerator in the Department of Nuclear Physics. We drive up to the Department, which is still housed in its historic premises high on a Florentine hill, although it is shortly to move. The genius loci who greets us, Professore Pier Andrea Mandò, is an established authority on PIXE analysis, which until now has only been used on manuscripts, inks and ceramics; he is immensely enthusiastic to discover what it might deliver when applied to paintings. We sense that Leonardo would have approved of being associated with the cutting edge of science. All the scientists we encounter, in each institution, share a similar enthusiasm to see what their techniques might deliver in the service of our uomo universale. The remaining techniques are primarily directed at recording and understanding the surface appearance of the picture. They comprise multispectral ultraviolet fluorescence, laser micro-profilometry and laser-line 94 FORENSICS Fig. 45 Profilometric image, Madonna of the Yarnwinder (The Lansdowne Madonna), plate 1. 2002/3. Courtesy of the Opificio delle Pietre Dure, Florence. THE UNIVERSAL LEONARDO AND THE LABORATORY 95 Fig. 46 Infrared reflectogram, Madonna of the Yarnwinder (The Lansdowne Madonna), plate 1. 2002/3. Courtesy of the Opificio delle Pietre Dure, Florence. scanning. It is vital to have the best possible record of the visual appearance of the picture flat on, reproducible to exact dimensions without any of the even slight distortions that arise with the best photographic equipment. Significant features of the existing surface can be detected. Along with its collaborators, the Opificio has pioneered the use of scans to obtain full-scale images with no distortion, which permit the precise overlaying of the results of different kinds of examination. Of the techniques used in their most developed form, ultraviolet is invaluable in picking up areas of visible paint and varnish that date from different eras, while the laser profiling (fig. 45) gives us a detailed relief map on which overpaints show up as gross scabs on the skin of the original surface. provided by the remarkable scanner developed by university physicists in Florence and housed in the Opificio. The IR reflectograms are not only vivid and relatively explicit from the general public's point of view, but they also provide the specialist with the most telling evidence, through the underdrawing, about the artist's thinking as the picture evolved on the priming of the panel. We can see, on exactly the same scale as the panel itself, an incredibly clear picture of what lies directly on the surface of the gesso priming -not everything, to be sure, but enough to tell us some very remarkable things (fig. 46). Experienced though they are, Cristina Acidini and her staff at the Opificio, most notably Cecilia Frosinini and Roberto Bellucci, have never seen anything quite like it. There are inevitably areas where the underdrawing is not readily apparent, especially below thick layers of THE UNIVERSAL LEONARDO AND THE LABORATORY 97 relatively opaque paint. Parts of the flesh tones present something of a barrier, and the thickly painted dress of the Virgin tends to mask what might lie beneath. Like any technically generated scan operating with emissions that lie outside the visible spectrum and requiring electronic devices for the translation of data into visible form, IR reflectography demands informed looking. It will thus be worth describing and assessing systematically what can be seen in terms of the historical evidence we have already encountered. Some of the detected features were visible in the mosaic of the New York IR reflectogram, but much of what we could then see dimly is now clarified, and important new features have emerged. It is apparent that two types of underdrawing are present. Most immediately apparent are broad strokes of wash, laid in with a brush. More detailed drawing, particularly in the contours of foreground figures and objects, is accomplished with fine lines. I suspect that they are produced by the tip of a small brush, but they might conceivably have been drawn in charcoal or black chalk. In tracking the underdrawings, let us begin towards the top left and work our way steadily up and down and across the panel. Along the upper border are irregularly shaped darker patches, indicating loss of paint and priming, with jagged vertical fissures following the grain of the original panel. They are obvious but not too severe. The porch roof is a thick protrusion, cantilevered on two diagonal struts. The vertical edge of the wall of the building is apparent below its right boundary, the arch, outlined with fine lines, is shaded in broadly. Inside it seems to be a smaller arched aperture or recessed section, somewhat like the niche of a shrine. The general outlines of figures in front of the arch, brushed in with haste, conform to those in the known baby-walker groups, as in the Edinburgh variant, but Joseph's activity is not clear in itself. It is just possible that the group has not yet achieved its definitive identity. Perhaps we can see the child lying on the ground and being adored by three figures. But what I am looking at is probably the base and wheels of the contraption. With elusive passages such as this, we tend to see what we want, and I prefer to think that the actual baby-walker was sketched in a way that does not show up well in the reflectogram. The impression is very much work in progress, with possibilities being explored in an improvised manner. Fig. 4747Detail of infrared reflectogram (outlines of fingers near bridge), Madonna of the Yarnwinder (The Lansdowne Madonna), plate 1. 2002/3. Courtesy of the Opificio delle Pietre Dure, Florence. 98 FORENSICS Below the figures is the bridge of the final painting, probably laid in at a relatively early stage, once Leonardo had decided not to take the figure group any further. More surprisingly, to the right of the bridge, are what appear to be the dotted outlines of fingers (fig. 47). They are significantly higher and to the left of the painted hand. Does this indicate the use of a pricked cartoon? Spolvero marks, deposited on the priming of a panel when charcoal dust is pounced through the pin-pricks in a cartoon, tend to be very ephemeral, being easily brushed off, dispersed or obscured during painting. Have the spolveri simply disappeared from elsewhere across the surface, with this possible exception? Or was a part-cartoon used for this particularly tricky motif of the foreshortened hand? The outlined hand, as far as it can be separated out, looks notably like that of the Virgin's right hand in the composition of the Madonna with the Holy Children at play, known in a number of versions at greater and lesser degrees of remoteness from Leonardo (fig. 24) -but in reverse. It is this hand, as we have seen, that is the subject of a drawing by one of Leonardo's followers in Venice (fig. 23). 2 This resemblance confirms my suspicion that Leonardo had developed a procedure of 'cutting and pasting' authorised motifs when making pictures of this kind, even using techniques of transfer such as cartoons to reverse them. The hand as painted also shows the pentimento of the more tucked-in thumb, which is in the wrong place to belong to the outline of the upper hand. The area in the region of Christ's right foot is not easy to read. The image is complicated by thick paint, deterioration and abrasion. There does seem to be the shadow of a darkish form just below the level of the Virgin's knee. Its top edge, curving upwards from left to right at an angle of about 40°, seems to be marked by spolveri. Is this the open lid or rim of the elusive basket described by Fra Pietro but not visible in any of the other versions? The presence of brown pigments detected in this area reinforces the possibility, as we shall see. As for Christ's right foot, which would have been in or on the basket, we can see fairly notable adjustments. The lower boundary of his leg beneath his knee was originally much higher, and his leg much straighter. Although the underdrawing of the leg disappears under the thickly painted drapery, his foot would have fallen within the contours of the dark patch of the possible basket. The lower contour of his left leg has also been noticeably adjusted, but less radically than his right. The vertical paint losses across his abdomen are clearly evident, as is the painful damage to his genitals as the result of successive interventions. Oddly, one of the variants (fig. 108), which otherwise appears to be relatively remote from Leonardo and does not seem to belong to a clearly defined 'family' , exhibits the straighter, higher right leg of Christ. The Virgin's left thumb is also in the higher position. Additionally, the position of the Virgin's right hand has clear affinities with the Lansdowne underdrawing. Once again, a discarded motif seems to have leaked out into one of the variant pictures. Now that we have clear evidence of the higher, straighter leg in the Lansdowne underdrawing, a previously unnoticed feature in the 1939 and later X-rays of the Buccleuch Madonna comes into play. In the region THE UNIVERSAL LEONARDO AND THE LABORATORY 99 of Christ's right knee is a short, curving horizontal ridge. It looks as if this is a trace of the lower contour of his knee in the same original position as it had occupied in the earliest stages of the Lansdowne Madonna, but the evidence is suggestive rather than conclusive. Moving upwards to look closely at the Virgin's head, we can confirm that the hair and headdress are drawn with extraordinary rhythmic delicacy using the tip of a fine brush. The fine band with frilly or serrated edge that arches over the upper part of the twisted veil is a particularly elegant touch, and serves to emphasise the roundness of the ensemble of diaphanous cloths. More remarkably, a series of rapidly drawn swirls and streaks shows that the headdress was once intended to be more ample with, perhaps, some portions hanging free. The effect is very close to that in a drawing at Windsor (fig. 28), usually said to be for St Anne in the Louvre painting of the Virgin, Child, St Anne and a Lamb. In the light of what I have argued, it may be that such a specific identification of a drawing as destined for one particular head in one particular painting is not strictly warranted. On the other hand, the style of the Windsor study seems to be later than 1501. In any event, it is clear that Leonardo constantly worked on themes and variations throughout his pictures for Madonnas on various scales. The initial idea for the headdress in the Lansdowne picture, discarded in this case, may have been revived for the later picture. In terms of quality, Christ's head is the most exciting passage in the IR image. It has an intensity and nervous life that shouts Leonardo's presence. Although the curls of the hair may have been reinforced in places, there are many signs of Leonardo's characteristically vivacious vortices, here rendered with exquisite refinement. There is obvious manoeuvring to position his left cheek, as the artist searches for precisely the right bulge to convey the pudginess that he so liked. The damage running from the corner of his mouth and across the left margin of his chin is apparent, as are at least two lumpy repaints. The region of the Child's right arm and the Virgin's hand reveals a good deal of nice underdrawing. The outlines of his biceps, forearm, wrist and hand are sketched in finely and with confidence. However, the Virgin's hand, as she cradles his lower torso and upper abdomen, is a real puzzle. Her thumb, nail and all, is clearly visible at a higher position, across his bosom, but it is difficult to attach it securely to what else is visible of her hand. Then, as her fingers pass under his arm, they fail to emerge on the lower side, as they surely should. The area in question is quite damaged. Maybe they have simply become lost, like Christ's genitals. But looking at other versions raises doubts. Sometimes the fingers are present, as in the Buccleuch version, while in others, including the Edinburgh 'family' of variants, they are absent. The explanation may lie within this 'family' , all of which exhibit the higher thumb across Christ's breast, and a correspondingly higher hand position, which would mean that the Virgin's fingers could legitimately be wholly concealed behind the Child's arm. It looks as if the artist decided to lower the hand, to correspond to the hand that he was painting in the Buccleuch Madonna, but did not carry through all its consequences. At it happens, we only become aware of the absence of the fingers once 100 FORENSICS our attention is drawn to it. But it is strange, whatever the explanation. No bigger surprises are awaiting than in the underdrawing of the actual yarnwinder. Some of the changes had been in part apparent in the New York IR mosaic. Now we can see them in their full extent. The left limb of the upper crossbar has been revised at least once, while an alternative, apparently one-limbed bar protrudes from the right at an acute angle. This protruding bar seems to be pressed by the tip of Christ's extended index finger, which was originally pointing upwards in a stiff vertical, a motif familiar from the Last Supper and the St John the Baptist. There is no visible underdrawing for the extended finger as painted, when it adopts a somewhat more relaxed position. The drawing of his other fingers and knuckles varies distinctly from that of the hand as finally painted. Vertical lines hang from the armature of the yarnwinder, presumably unwound threads. Then, as we look down the shaft, we encounter another single-limbed crossbar, just below his elbow, which itself seems to have been repositioned. Perhaps there is a trace of a diagonal structure passing upwards from the limb to the shaft, or maybe it is just other strands of yarn. By contrast, the lowest crossbar, set at right angles to the upper, is drawn in much the same position as it occupies in the painting. Finally, at the foot of the yarnwinder is a series of fine lines describing oval-ish contours with pointed ends. They are what scientists call fusiform in shape. I realise, as I should have done earlier, that they are indeed fusi, the spindles on which yarn has been wound in the standard way, fat in the middle and tapering towards the ends of their shafts. Here, then, is the Italians' Madonna dei Fusi. But the spindles seem to have progressed no further than simple outlines, and they are not in the right position for the basket described by Fra Pietro as being at Christ's feet. However, they are in the same general position as in other variants. The cluster of fusi raises the possibility that Fra Pietro meant to say that it was the foot of the yarnwinder not that of Christ that was placed in the basket of spindles.Finally, looking into the middle-ground and background beside and behind Christ's arm and the yarnwinder, we can see a substantial hill-cum-mountain rising towards the right. It is more dramatic than the hills in the Buccleuch picture, but it is far from the clustered peaks of the painted landscape in the final painting. It bears some resemblance to the hill behind Christ's head in the baby-walker 'family' of paintings, but it is not possible to see any trace of the arched dam beside the hill. The dam is, however, a surprising motif and fits well with Leonardo's involvement with hydraulic engineering. There is no sign of the painted peaks in the underdrawing anywhere in the panel.Then, crammed uncomfortably between his forearm and the edge of the panel, is a reclining animal, a horse or an ass. The latter would make sense in the context of a 'Flight into Egypt' , supporting the idea that the baby-walker group was so intended. The animal's exaggeratedly large ears also reinforce its identification as the ass that carried its precious burden into exile.I can think of no Florentine Renaissance picture that has undergone such remarkable changes during the course of its execution, though Leonardo's unfinished THE UNIVERSAL LEONARDO AND THE LABORATORY 101 Adoration and St Jerome exhibit so many fluid and undefined passages that they may well have progressed in a comparably improvised way had they reached completion. We also know that the underdrawing in the London version of the Virgin of the Rocks posited an alternative composition. TECHNIQUE, PRIMING AND SUPPORT The conclusions that can safely be drawn from the analyses of the material composition of the painting are necessarily provisional, since full interpretation will rely in due course upon comparisons with other paintings yet to be examined. But it is already clear that the painting technique used by the artist is extraordinarily refined and probably idiosyncratic. It has been supposed that to achieve his effects of elusive sfumato Leonardo laid on his paints in successive layers of very fine glazes of oil containing relatively small and dispersed quantities of particulate pigments. Results from the analysis of the Madonna in the Alte Pinakothek in Munich, conducted at the Doerner Institute, suggest that this is indeed the case, as does the recent analysis of the Mona Lisa. 3 Although Jacques Franck has claimed that Leonardo used 'micro-pointillism' (painting with many thousands of tiny dots, like a precocious Seurat endowed with microscopic eyesight), it is clear that effects like those in the Mona Lisa must have been built up painstakingly and agonisingly slowly with the application of successive layers of fine glaze. The analyses have confirmed the thin and very refined stratification of the paint layers in the Munich Madonna. The result is a magic elusiveness that physically evokes the soft pli-ability of living flesh and the veiled blue atmospheres of distant mountains, while psychologically invoking the complex mobility of inner thought, both that of the subject and the spectator. The essential ambiguity of the Virgin's reaction, as recounted by her downcast eyes and almost open mouth, relies upon an indefiniteness of contour that no earlier artist had dared to leave. No one achieved a sfumato that looked exactly like Leonardo's. Raphael and Andrea del Sarto learnt crucial lessons from it, but their sfumato somehow has a different look, both with respect to each other and in relation to Leonardo. His immediate followers aped the effect by adding blurred and masking shadows in a heavy-handed way. Hosts of Lombard Madonnas smile their penumbral smiles and cradle their pneumatic babies in a dispiriting and self-parodying array. Leonardo always maintained some subtle quotient of openness and reflected glimmer in the shaded portions of lighter-toned surfaces, such as flesh. He understood the optical phenomenon of 'percussion' , as light and colour rebound from one surface to an adjacent one. He was essentially painting the passage of light across and between forms, whereas his followers made shadow adhere to the surface of flesh and cloth like an inky stain. It is difficult to imagine any other artists' works being executed with the persistent delicacy of Leonardo's paint layers -other, perhaps, than the Netherlandish painters, who used successive glazes for rather different effects. The pigment analyses, arrived at using the varied techniques outlined below, are consistent both with each other and with the supposed date of the picture. Fig. 4848Computed tomography scan, Madonna of the Yarnwinder (The Lansdowne Madonna), plate 1. 2002/3. Courtesy of Siemens Medical Solutions, Florence. 102 FORENSICS Fig. 49 Computed tomography scan, Madonna of the Yarnwinder (The Lansdowne Madonna), plate 1. 2002/3. Courtesy of Siemens Medical Solutions, Florence.with a prestigious and sophisticated international client, whether Robertet, the King or other prominent member of Louis's court.The detection of a white lead layer (or layers) beneath the main pigment surface suggests that the ground and the underdrawing were covered thinly with white, leaving the drawn features sufficiently visible to guide the painter. This white lead priming corresponds to the unexpected white lead layer over the final plas-ter in the Last Supper and to the white lead 'wash'(as yet incompletely published) detectable in the lower strata of the Adoration of the Magi. Maurizio Seracini of Editech in Florence, who is an independent researcher contracted by the Soprintendente to examine the Adoration in the wake of controversy over its possible restoration, has found that a thin, relatively translucent white lead layer was brushed over the first underdrawing. It appears likely that Leonardo wanted to enhance the assertively light-toned backing for his pigment layers, on which he could achieve the luminous effects of his glazes, which, as we have noted, rely for much of their radiance on light reflected off the ground. There is some indication that lead tin yellow was present in the white priming layer, endowing it with an ivory hue. Spectroscopy also indicates the presence of clay materials (kaolin) in the ground, which suggests that Leonardo was going to great lengths to achieve precisely the kind of support for his pigment layers that would serve his ambitious optical ends.The effect, not so remote from that of stained glass, was perhaps intended to emulate the brilliance of theglazed terracottas for which the della Robbia family had become famed in Florence and beyond. The hightoned brilliance of colour in the restored Last Supper supports this less-than-conventional interpretation of Leonardo's ambitions as a vibrant colourist. The analysis of the distribution of lead in the paler areas of the flesh and sky, and in the paler brown and green areas, taken together with the x-ray images, indicates that in the lighter areas his glazes contained fine mixtures of the tinting pigment (e.g. cinnabar for flesh tones or resinate pigments for the greens) with small THE UNIVERSAL LEONARDO AND THE LABORATORY 103 Fig. 50 Computed tomography scan, Madonna of the Yarnwinder (The Lansdowne Madonna), plate 1. 2002/3. Courtesy of Siemens Medical Solutions, Florence. quantities of lead white. This practice of relying on the ground and tinted glazes for luminosity is generally akin to that of Antonello da Messina and Giovanni Bellini in Venice, and reflects Netherlandish methods. But it does not precisely match any of his precedents. Another unexpected feature is the presence of the kind of bistre pigments used in ink in the browns. It may well be that the pigment used in the underdrawing was also used in mixing the paints when a particular kind of brown was required. The presence of bistre is most conspicuous in the thickly painted area beneath the Child's outstretched foot, although I am not now inclined to think that is where Fra Pietro's basket would have been. Until more paintings by Leonardo and his contemp-oraries have been examined in comparable detail, it is difficult to be certain about the degree of idiosyncrasy in his technique, but other technical and documentary evidence paints a picture of continual and sometimes unsuccessful experiments to extend traditional methods. Both the Last Supper and Battle of Anghiari departed radically from accepted practice. For the former he used a method more akin to tempera on a panel than the normal fresco for murals, while for the latter he tried to exploit a technique closer to that of oil painting. The results in both cases were not propitious for their survival. The flesh in the Munich Madonna is puckered and wrinkled, as are passages in the landscape, as he has laid down successive veils of thin pigment mixed with fatty binder, which have contracted at different rates as the painting dried. Experimental oddness, as he stretched the bounds of established technique, may well be one of the hallmarks of autograph Leonardos. The scanning techniques have also yielded significant findings about the physical properties of the picture. The profilometric analysis shows that the canvas imprint visible on the surface of the painting comes from two separate canvases: perhaps those used as the new support behind the paint layer and to face-up the surface during the stripping of the original wood support. However, the Paris 're-transfer' notes, as transcribed later, record the presence of strips of canvas used by Leonardo under the gesso priming of the panel. This was not unusual, and was recommended by Cennino Cennini in his Il Libro dell'arte when the painter was faced with knots and potentially unstable grain. As we noted, the X-rays of the Buccleuch Madonna reveal a comparable use of fine canvas -in this case in a painting that has never been removed from its panel. It may be that one of the imprints in the Lansdowne picture results from Leonardo's own use of canvas to supplement the wooden support. It has been assumed for pictures transferred from panel to canvas that little data would remain about the original support. The characteristic cracking and paint loss of a wooden panel would be discernible, but little else. The Siemens CT scans, never before applied to a painting, have provided astonishing results. They have revealed both outer and inner profiles of the strata of the painting, passing successively through the pigment layers, what remains of the priming, imprints of canvas, and the dense composition board, backed by the criss-cross struts of wooden cradling designed to prevent warping. Vertical sections in the CT scans disclose the fine strata of the painting and clearly reveal the threads of canvas used to support the picture, probably from the relining canvas but possibly from the canvas reinforcement used by Leonardo himself during the original priming. Looking through the successive horizontal sections, there is something particularly moving when we witness the scarred skin of the priming layer with its ghost of mother and child, like a faded fragment of ancient parchment (fig. 48). Around the margins can be seen the nails that pin the fillets of wood around the edges of the later board. We encounter what look like relief maps of weatherbeaten planks, in which we can just discern the highlighted portions of the mother and child (fig. 49). The slightly thicker paint of the brighter areas of flesh stands relatively proud of the surface, while thinner mixtures have tended to sink into the layer immediately below, being partly absorbed by the gesso. Move a little deeper with the scanner and we encounter a series of vertical striations in what remains of the shaved priming (fig. 50). Somehow, through the vicissitudes of successive transfers, the imprint of the grain of the wooden panel has survived. It represents an extraordinary study of information long since assumed to be lost. The sense of fragility and the miracle of survival over 500 years speak more impressively through the CT scans than any images of a painting I have seen before. August 2003, I am sitting under an umbrella on the terrace of the Villa Vignamaggio above Greve in Chianti, a villa once owned by the Gherardini family and haunted by the shade of a famous daughter known as Mona Lisa, when Thereza Crowe calls to report the theft of the Duke's treasured picture. The news is as yet hazy. It seems that some men driving a VW Golf GTI had abruptly removed it shortly before the rooms were to close to the public that day. They had overpowered the female custodian and threatened her with a knife. I receive the call when I am in the process of writing a new book on Leonardo for Oxford University Press, which involves, of course, a discussion of the Madonna of the Yarnwinder. 1 A coincidence of the worst kind. not least to deal with its disputed attribution. A series of arbitrary pronouncements is made by scholars, some of whom have not seen the painting in the original, and invariably without reference to the technical evidence. They should know better. The pronouncements resemble those made about the Raphael Madonna of the Pinks that the National Gallery is trying to purchase rather than letting it go to the Getty Museum in California. I happen to think that the Getty Museum would be a good place for it, but, in any event, the quality of the quoted comments on whether it is really by Raphael are for the most part dire. The same applies to the reactions, even by accomplished scholars, about the recently discovered portrait of a young lady on vellum. 2 I decide to do just one interview, with Godfrey Barker of the Sunday Times, not because I like the paper (and I have been misquoted and badly treated on three occasions by the daily Times) but because I have dealt The police contact me to ask if there are features that will infallibly identify the painting as the real thing if they recover it or what purports to be it. Not too dif-ficult, I explain. Two officers from the Dumfries and Galloway Constabulary visit me in my house and I take them through the evidence of the scientific examinations. They are fascinated, not least because they can recognise similarities with the kind of scientific forensic work they use in their own investigations of crimes. Unhappily, the question about means of identification is framed in anticipation of the Duke's panel being found eventually, and not because they have recovered it. There is still no sign of the painting, and the trail seems to have gone cold. For the moment, and we hope it is only for a short time, we have to rely on comparing the results of the Florence analysis of the Lansdowne Madonna with the existing IR examination of the Duke's stolen painting and the old X-ray plate. In the meantime, on Friday 5 October, the recovery of the Madonna is announced to the press. the Guardian announces: 'Police in south-west Scotland said last night that they had recovered the painting, in an operation also involving detectives from the Scottish drugs enforcement agency, the Scottish organised crime agency and Strathclyde police. Chief Inspector Mickey Dalgleish, who led the investigation, said the force was "extremely pleased"at its recovery. "For fouryears police staff have worked tirelessly on the theft and with help from the public we have been able to track down the painting. "' Four men have been arrested, three from Lancashire and one from Glasgow. The Lancashire trio are later named as Robert Graham, John Doyle and Marshall Ronald, a solicitor from Skelmersdale. Another man from Glasgow, Michael Brown, later appears in court. More pictures of cars are issued, including a blue Rover, together with a CCTV image of a male person whom the police want to interview. The man is described as being about 5 feet 10 inches to 6 feet tall, slim build with dark hair and a beard.It is subsequently learnt that David Boyce, a corporate partner in the respected Anglo-Scots law firm HBJ Gateley Wareing, has been charged, and that the painting wasrecovered from the lawyer's West Regent Street offices. Another lawyer who worked for the firm, Callum Jones, has also been implicated and is said by the Scotsman to have been acting as a go-between for two parties by scrutinising a contract that would have allowed an English firm to 'secure legal repatriation' of the painting from an unidentified party. The picture is now accorded a value of £ 37 million, though other papers cite widely divergent values. The trial and its strange result will feature later in our story.However, joy at the recovery of the Madonna and the arrests is tempered. Johnnie, the Duke of Buccleuch, had died a month earlier on 4 September 2007, aged 83. to Scotland as agreed, taking off from Birmingham for Edinburgh. Dense fog intervenes, and we are diverted to Prestwick on the west coast of Scotlandprecisely the wrong side of the country. A cold wait for a bus ensues, followed by an uncomfortable journey to Edinburgh airport, where I am met by a police car to ferry me to the conservation studios of the National Gallery. There, entering precisely the same room in which we had first inspected the picture seventeen years earlier, my heart leaps to see the small panel lying on a covered table with the delighted conservation staff, Jacqueline Ridge and Lesley Stevenson, and two members of the team of detectives. Even from a distance, its radiant intensity declares that the police had indeed got their woman (and baby), so to speak. Fig. 5151Infrared reflectogram, Madonna of the Yarnwinder (The Buccleuch Madonna), plate 2. 2007/8. Courtesy of the National Galleries of Scotland and Buccleuch Living Heritage Trust. smaller room down a corridor, where a new and more potent IR apparatus is now housed, linked directly into a computer that delivers the images to its screen (fig.51, p. 12 and p. 212). Lesley and Jacqueline could not be more helpful and skilled. Iwrite a report for the police, summarising the results of the analysis. I am also asked to provide a formal statement, covering everything I knew and could remember from before the theft onwards. Two officers from Dumfries visit me in Oxford, and painstakingly record in longhand my answers to a series of questions in chronological order. My answers are finally read back to me with ponderous care, and I sign the statement as true. The idea that I might be a suspect briefly crosses my mind, but it is apparent that the police are determined that there should be no loose ends and no gaps if they are to achieve the prosecutions they desire. The next step in the research is initiated by Richard, the incoming Duke of Buccleuch, who regards the picture and the technical investigations with no less enthusiasm than his father. He sets up a more sustained examination of the panel in the conservation studios of the National Gallery. Thereza Wells is to join us, as is Marina Wallace, and we meet in an Edinburgh hotel on11 March to walk to the Gallery.The Duke's daughter, Louisa, a portraitist practising in Edinburgh, is also present to witness the action and lend her lively support. Lesley Stevenson guides us through the examinations. We use a newly acquired 'toy' , a very high-quality binocular microscope, to look at the surface under a variety of magnifications, which both clarifies some issues and raises fresh questions.We also renew the IR examination, this time moving systematically across the surface, taking notes and comparing the emergent underdrawing with that of the Lansdowne Madonna, which I have as a highresolution file on my computer. A series of exciting new observations emerges, as well as data that confirms what we had seen less clearly with the more primitive equipment.The following account generally tracks our actual process of inspection, as we work in vertical bands from left to right. In making comparisons with the IR image of the Lansdowne picture we still need to be cautious.The equipment in Edinburgh and that in Florence are different. The latter gives an unrivalled degree of tonal resolution. I need to sharpen the Edinburgh images using Photoshop. However, the new Edinburgh equipment does allow better comparisons to be made than was previously possible. edge of the picture in the Lansdowne Madonna. There appears to be a platform sketched in perspective at the level of the kneeling Virgin's thigh. This is too high to accommodate the group, and may be a first thought for the positioning and support of the figures in the fragmentary architecture or represent a stage in which only architecture is to be included. The 'Mona Lisa' type arched bridge that survived into the final painting of the other version is more clearly apparent but did not seemingly progress beyond the outline underdrawing. It is likely to have been considered only after Leonardo had decided not to persist with the baby-walker group. The bridge and the group could not have co-existed in any coherent spatial relationship. The loop visible below the Virgin's left knee in the Lansdowne underdrawing is not visible. There is a good deal of damage below the knee in the region of the Child's right foot, especially in the rocks, with a pronounced vertical section of restoration. The alternative position of the Virgin's thumb is confirmed, but there is no sign of a radically different position for her fingers. There are no pentimenti in the Child's right or left legs, but changes in the contours of his right arm and wrist are clearly apparent. Underdrawing is apparent in the Child's belly and the Virgin's fingers are visible supporting his body. At the bottom edge of the panel, vertically below the Child's extended foot, is a feature than can be read as a set of toes. Thereza is keen on this reading. I see two problems: firstly the foot is too high to be reconciled with the position of her knee; and secondly the Lansdowne Madonna shows that in both the underdrawing and finished painting Christ's foot is braced (with his large toe pressed upwards slightly) against a rocky ledge that continues to the bottom of the picture. However, Thereza is unmoved. This divergence of reading confirms the elusive nature of the scanned image and the difficulties in arriving at a definitive reading. A series of complex marks is apparent in the sky to the left of the Virgin's head, but they do not all coalesce into a recognisable image. A darkish, curved mark seems to embrace the right of the baby-walker group but is difficult to read as a specific feature. The Virgin's headdress is laid down in the underdrawing with brushstrokes of notable refinement, and carries over into the sky to left, aligning it (as is the case with the Lansdowne underdrawing) with the St Anne study at Windsor (fig. 28). The top of the vertical shaft of the yarnwinder was at one point conspicuously to the left of the painted version, with threads hanging diagonally and vertically from the left arm of the cross-member. A clearly drawn contour for the Child's arm, well inside the painted one, confirms that he was once holding the cross closer to his head. There is some indication in the Lansdowne Madonna that the shaft may have been marginally to the left, and we can clearly see threads matching those in the Buccleuch underdrawing. The drawing of Jesus's right hand is more resolved in the latter, with his index finger looped convincingly over the vertical cross-member. As in the Lansdowne version, outlines of spindles (fusi) are visible around the base of the yarnwinder, though the resolution of the image is less good than in the other reflectogram. There is no obvious sign of the basket mentioned by Fra Pietro, but the elaborate drawing and painting of the rock strata in this region dominate the IR imaging.The binocular microscope discloses a series of delights.The heads of the Virgin and Child, her headdress, her forehead veil, the neckline of her dress, the highlights of his curls and his moist eyes are handled with magical delicacy. The edge of the translucent veil above her eyebrows is painted with a tiny brush with dark pigment as a calligraphic series of broken ripples, brilliantly conveying its delicately scalloped margin. We marvel at the ability of Leonardo to see and operate at such tiny scales, an ability confirmed by his often tiny drawings in his manuscripts.One feature we observe late in the examination.The microscope reveals some tiny hooked blue brushstrokes in the curls of Christ's hair, and a few more in the vegetation to the left of the Virgin. Marina £ 4.25 million for the return of the Madonna of the Yarnwinder. The trial took place at the High Court in Edinburgh over eight weeks in the spring of 2010. Those who appeared in the dock were a motley crew. The account that emerged in the trial is riven with inconsistencies, and I doubt whether the full truth is yet known, but the broad outline of events is reasonably clear. The two Liverpool men in the dock were operators of a kind that Salaì would have recognised. Robert Graham, a publican, and his friend John Doyle had set themselves up as private investigators and established an online business, Stolen Stuff Reunited, which aspired to get crooks to return stolen material that proved to have little value (other than to the original owners). They clearly had their underworld contacts, and it seems to have been Doyle who indirectly or directly established a line of communication with the thieves. Graham then involved the Lancashire solicitor Marshall Ronald, who might best be described as a maverick. Realising that they needed Scottish expertise, Ronald contacted David Boyce of the lawyers HBJ Gateley Wareing in Glasgow, with whom he had earlier done business. Boyce in turn involved in the painting's return. There was the possibility of a substantial reward and an expectation of some other kinds of payments. The English trio were clearly hoping to profit handsomely at the end of the day. There is of course also the excitement of being involved with such a big story and the glamour of association with Leonardo. Even a sober Glasgow solicitor can be forgiven a frisson of excitement at the thought of a stolen painting by Leonardo being handed over in his office. We might ask, however, whether members of the legal profession should be forgiven for not informing the police. The five charged gave what seem to me to be thin excuses for not doing so.After complex negotiations a meeting was set up in the solicitors' offices in West Regent Street in Glasgow, with the promise that the painting itself would be present.The undercover police officers, whose true identity was not known to the other participants, attended the meeting. Jones told them that the lady was close by, as Ronald had assured him. Jones introduced the officers to Ronald, Graham and Doyle in the boardroom of the law firm. The painting was then extracted from a black case carried by Doyle. It was obviously the real thing.Having blocked all routes of escape, the police moved briskly in, surprising the men assembled there with their precious charge. The Dumfries and Galloway Constabulary were rightly proud of their operation.As one of the officers said to me, 'I normally deal with muggings and slashings. 'In March I travel from Italy, where I had been staying for two months at the Villa I Tatti owned by Harvard.I had been asked to testify about the painting itself and the advice I had given the police. I was rather surprised to be called to testify, since whether the picture was actually a Leonardo or not did not seem to me to affect the case of extortion. The court chamber is unexpectedly cavernous, with serried ranks of what I take to be lawyers -considerable numbers of themtogether with other observers, including the press. I can see why the law is so expensive. The accused sit in a glum row. I give an account of the visual evidence we have assembled about the painting's authorship, under examination by the prosecutor. I try to give a lively and accessible account of the painting's qualities and of the scientific examination, explaining how we can be sure that the work that has been recovered is the same as that which had been stolen.I sense that those in the court are pleased to have a diversion into a realm outside the main legal grind. The painting itself is on show in the National Gallery at the Mound in Edinburgh, where it remains for the time being. I go to pay my respects. The panel is beautifully framed, mounted and lit. It looks stunning, and more than holds its own with Raphael's Bridgewater Madonna in the same room. Raphael's composition, in which the infant Christ surges across his mother's lap, is the younger artist's own homage to Leonardo's seminal invention. At some point the main players in the court decamp to the gallery to become directly acquainted with the star actor in their drama. 116 FORENSICS On 22 April I am sent the news in Italy that the jury had returned its series of verdicts on the previous day. I am astonished to learn that not proven majority verdicts had been given to Ronald, Graham and Doyle, while the Glasgow solicitors, Jones and Boyce, were found unanimously not guilty. 'Not proven' in Scots law allows a jury to say that the case against the accused has not been demonstrated to their total satisfaction but stops short of declaring the accused incontrovertibly not guilty. I can only imagine that something had gone badly wrong in the prosecution. Perhaps the extortion charge was not the right one. Doyle argued that if they had been found guilty, no stolen art would ever come back again. It seems to me that the actions of the English trio were such as to encourage thieves to think that stealing paintings could be a profitable enterprise for all involved, including the intermediaries. In any event, the result confirms my view, from involvement in other cases of art in court, that the law at this level involves elaborate and horribly costly rituals acted out by lawyers according to arcane internal rules in which commonsense and natural justice are all too readily obscured. On 6 July it is reported that the police have interviewed Ronald over allegations he embezzled £ 800,000 from a client to fund the recapture of the painting. I feel for the police, who are still doing their best. Worst of all, at the end of the day, the actual thieves have not been brought to justice. SUMMING UP 117 12. SUMMING UP There is no Florentine Renaissance Madonna (and a considerable number have now been examined) that shows such radical design decisions being taken during the course of the execution of the painting, with the possible exception of the second version of the Virgin of the Rocks. The underdrawing not only shows those detailed adjustments that might be expected from a major master searching adventurously for the best arrangement of his figures -moving a hand or finger here, a leg there, adjusting the contour of a head or nose, re-arranging drapery and so on -but also exhibits radical changes of mind about the overall appearance of the composition. Leonardo seems to have used the underdrawing on the gesso priming (perhaps with an upper wash of white lead tinted with lead tin yellow) as a laboratory for the invention of the radically new kind of Madonna and Child composition that he was developing on his return to Florence in 1500. The Madonna of the Yarnwinder was to revolutionise the way the subject was handled by Michelangelo, Raphael, Fra Bartolommeo, Albertinelli and all those lesser Florentine artists who were open to new ideas. The relatively constricted space between the left edge of the picture and the Virgin's right shoulder was at one stage occupied by the 'baby-walker group' , comprising Joseph, who is busied in the making of a walking frame with its four small wheels, the Virgin and the infant Christ (who twists to watch Joseph's activities with eager curiosity), and a second woman. Behind the group in the Buccleuch version are what may be traces of animals, perhaps the ox and ass familiar in nativity scenes. A ghostly animal also seems to have been sketched into the mid-ground on the right side of the landscape near Christ in the Lansdowne picture.Above the group on the left of both underdrawings is a semi-circular arch in a wall of masonry. A cantilevered roof like that of a porch juts out above the top of the arch in the Lansdowne version, supported by diagonal struts. The arch and roof are comparable to those in the Metropolitan Museum drawing for the Madonna with the Holy Children at play in the 1490s. The reading of the baby-walker group in the underdrawing is informed by those variants by followers, most notably that in the National Gallery of Scotland, which include the figures in apparently identical form. Since none of the later variants with the group includes the architecture, it might just be possible that the figures and arch were alternatives, rather than intended at any stage to appear together. In which case, the arch with the roof, a motif already used by Leonardo, would have come first, and was supplanted by the idea for the little narrative scene. In any event, both the architecture and the figures were entirely eliminated by overpainting, though the group 'escaped' into the versions by followers (presumably those young painters who were in or had access to the studio in the early years of the century). All this was to be totally covered by a more familiar landscape of precipitous mountains melting into plains with meandering watercourses, arched bridge and cliffs, familiar from the Mona Lisa. The bridge was once present in the Buccleuch underpainting, but was eliminated in the final version. It seems likely that the profusion of secondary motifs and centres of attention in the underdrawing, which obviously seemed attractive at design stage, proved too much in the compass of such a small panel as it was progressively painted in all its detail. It was in keeping with this progressive elimination of 'clutter' that the spindles (fusi), distributed over the rocks near the base of the yarnwinder, and the possible basket were also eliminated. The general direction of Leonardo's creative impulse as the painting evolved was towards consolidation and focus -concentrating on the main figure group, with its intense emotional interplay, and its atmospheric setting in a landscape that evoked the wonder of the 'body of the world' as analogous to the miracle of the body of Christ. It cannot be stressed enough that such a fluid process of design is exceptional in underdrawings or underpaintings for Madonna paintings by any Renaissance master. If Leonardo used a whole or part-cartoon, he certainly did not feel constrained to follow it precisely. It seems entirely unlikely that a studio assistant would have been entrusted with such radical decision-making in a work for an important patron, or would have been capable of playing a determining role in the forging of the new type of Madonna. SUMMING UP 119 The relationship of the design processes in the Lansdowne Madonna and Buccleuch Madonna is complex. The technical evidence clearly indicates that the evolution of the pictures occurred in tandem. If we suppose a phase in which both paintings looked very similar as underdrawings (architecture, figure group and all), it does not appear that one painting was consistently in the lead, to be followed by the other. The more extensive pentimenti in the face of the Virgin in the Buccleuch picture indicate that it served as the locus for the defining of her features and expression. On the other hand, there is no evidence at present that the Buccleuch picture exhibits the large pentimenti when the Child's extended right leg was moved. The yarnwinders each seem to undergo their own development, once the basic position was established in the Buccleuch version, before converging on the final simplified solution. On the other hand, the rocks below the yarnwinder in the Buccleuch painting seem to have provided the prototype for the less geologically defined features in the other picture, but the brilliant filaments of red yarn falling across the rocks are only present in the latter. There is no sign that the Buccleuch picture has been abraded in such a way as to lose the threads. It is somewhat surprising that the brilliant landscape in the Lansdowne version was not taken over into the Buccleuch painting, and was replaced by a somewhat barren seascape. However, it appears from the underdrawing above Christ's head that the Lansdowne picture was once to be given a landscape of rolling hills or rounded mountains closer to that now in the Buccleuch version. The shared bridge in the under-drawings may have arrived in the latter from the Lansdowne picture, in which it became a definitive feature, only to be eliminated in the final version of the Buccleuch landscape. The divergence of the final landscapes was probably intended to differentiate between the versions. It was not uncommon in Renaissance practice for landscapes and subsidiary features to be changed in variants, while the figure group remained the same. It looks very much as if there was a coming and going between the panels as they evolved, with one advancing beyond the other, and then the other not only catching up but proceeding further, and so on successively, until both paintings were finished. These observations can only be convincing if they can be reconciled with a feasible model for the two paintings' progress. Although such a model must necessarily remain hypothetical, it should exhibit an air of reality to work at all. In the model I am proposing here, the precise order of some of the steps could be adjusted without detriment to overall direction of the scenario: 1) the basic concept of the mother and her child in the painting was developed in drawings, with the group designed in the scribbled manner of the British Museum drawings for the Madonna, Child and Cat and the National Gallery cartoon; 2) the drawings indicated an arched structure in the background as in the Metropolitan study for the Virgin with the Holy Children, probably, with the figure group and animals; 3) details of the main components were studied from life, as in the drawing of the woman's torso at Windsor; 4) the scheme was resolved to the point where Leonardo was confident to begin laying in the composition on the prepared panels. Whether this involved a cartoon or part-cartoons, a modello or simply a developed drawing is unclear, though a few surviving lines laid down with spolveri indicate that one or more of the component parts might have been transferred directly; 5) the underdrawings were established, using a variety of thin lines to define contours of figures and objects, with broader brushed lines in the middle and background. At this stage the two pictures stood beside each other as sketchy twins, not quite identical but very similar; 6) the baby-walker group and attendant animals were added either in front of the arch or replacing it; 7) adjustments were made to the Virgin's features in the Buccleuch picture, either before or after the Child's pose was adjusted in the Lansdowne version, and various key elements in the figures in both pictures, such as their hands, were manoeuvred into position; 8) other adjustments in the main figure group and the props were made in parallel in the pictures, with the possible elimination of the 'Novellara basket'; 9) the form and position of the yarnwinder, attendant threads and spindles underwent experimental development in both paintings, with the Buccleuch version in the lead initially; and recorded about the conservation and restoration of the Lansdowne painting, provide a picture of an object that has undergone dramatic changes. Not only has it survived two complete transfers of the pigment layers, but entire elements of the composition have also been removed and then returned. One could say that the painting we see in front of us today is, in some ways, a different painting from the one seen by a nineteenth-or early twentiethcentury viewer. This leads to fascinating questions as to how and why these changes occurred. It also affects the questions of attribution, which are discussed in detail, mainly in relation to the Lansdowne Madonna, later in this chapter. In the case of the Buccleuch Madonna, relatively little has happened to it in terms of conservation. It has not undergone any transfers, but, as with any painting of this age, there has been some damage and restoration.THE ISSUE OF ATTRIBUTIONIn dealing with the history of attribution of the two Madonnas in the nineteenth and twentieth centuries, we have to take into account their relationship with each other because, as early as 1898, comparisons were drawn between the two paintings. The Buccleuch painting had been bought at auction in Paris in 1756 as a Leonardo and, within the family which has continuously held it, the attribution had remained more or less constant as a Leonardo, or occasionally as aSodoma. The first recorded expression of doubt about the Buccleuch Madonna appears to have been during an exhibition of 'Old Masters' at the Royal Academy in 1872, described in The Times in January of that year. This was the third in a series of old master shows where several hundred paintings were gathered from collections of the most prominent patrons. It is the first record we have of the painting being exhibited, and the Duke of Buccleuch was one of the largest lenders to the show. The extensive review in The Times, which came in three instalments, notes 'two singularly interesting examples of the Milanese school' , one of which was the Buccleuch Madonna. It goes on to describe the painting: one is a small Madonna and Child (117) of the Duke of Buccleuch's, in which the richly toned, softly rounded limbs, the full, dreamy, almost voluptuous, features of the mother, the exquisite drawing and expression of the child's upturned face, his look fixed on the Cross, and the mother's action of drawing the child to herself, as if in involuntary shrinking from that fate which he foreshadows but she cannot, all tend to justify the ascription to Leo-nardo himself, while yet there are defects in the drawing of the Virgin's face, right hand and arms that seem to forbid such an authorship. 1 The anonymous reviewer leaves it at that, perhaps allowing the exhibition's visitors to make the final judgement. The first known public showing of the Lansdowne Madonna took place in London in 1880 during the Royal Academy's winter exhibition of old masters. The previous year the painting had been bought at Christie's auction house by Cyril Flower, later Lord Battersea. THE LANSDOWNE AND BUCCLEUCH MADONNAS: CONSERVATION AND CONNOISSEURSHIP 129 The painting had been attributed to Leonardo in the auction and was so again at this exhibition, although perhaps with some reservation, as the caption read 'Lionardo da Vinci (ascribed to)' . The earliest image of the Lansdowne Madonna found to date is a photograph from 1893 at the Witt Library in the Courtauld Institute, London (fig. 53). The photograph is simply mounted on board and accompanied by a caption from an exhibition that was held at the New Gallery, London, in 1893-4 entitled Exhibition of Early Italian Art from 1300 to 1550. The caption shows that the painting was still in the collection of Lord Battersea and this time Leonardo's name is not accompanied by 'ascribed to' . The image is one of the most important found because it shows the painting on its original panel. In 1898 the Lansdowne and Buccleuch Madonnas were exhibited together for the first time, at the Burlington Fine Arts Club exhibition of 'Milanese and Allied Schools of Lombardy' . The large catalogue by Herbert Cook contained fine photographic reproductions of the paintings on show, including the two Madonnas presented alongside each other (figs. 52 and 54). It is apparent from the reproduction that the Lansdowne Madonna is still on panel and does not appear to have undergone any changes in the preceding five years. The catalogue also included photographs that were exhibited instead of the original paintings, such as both versions of the Virgin of the Rocks (figs. 60 and 61). Fig. 52 1898 reproduction of Leonardo da Vinci, Madonna of the Yarnwinder (The Buccleuch Madonna). Reproduced in Burlington Fine Arts Club, Illustrated Catalogue of Pictures by Masters of the Milanese and Allied Schools of Lombardy, reproduction of Leonardo da Vinci, Madonna of the Yarnwinder (The Lansdowne Madonna). Witt Library, Courtauld Institute of Art, London. Fig. 54 1898 reproduction of Leonardo da Vinci, Madonna of the Yarnwinder (The Lansdowne Madonna). Reproduced in Burlington Fine Arts Club, Illustrated Catalogue of Pictures by Masters of the Milanese and Allied Schools of Lombardy, London, 1899. Fig. 55 c.1931 reproduction of Leonardo da Vinci, Madonna of the Yarnwinder (The Lansdowne Madonna). Reproduced in Bodmer, Leonardo, 1931. Image from the Harvard University Center for Italian Renaissance Studies, Villa I Tatti, courtesy of the President and Fellows of Harvard College. Fig. 56561931 reproduction of Leonardo da Vinci, Madonna of the Yarnwinder (The Lansdowne Madonna). Reproduced in Giorgio Nicodemi, ed., Miscellanea di studi lombardi in onore di Ettore Verga, Milan, 1931. Fig. 57 c57.1939 reproduction of Leonardo da Vinci, Madonna of the Yarnwinder (The Lansdowne Madonna), Reford Family Collection. Fig. 5858Post-1939 reproduction of Leonardo da Vinci, Madonna of the Yarnwinder (The Lansdowne Madonna). Witt Library, Courtauld Institute of Art, London. Fig. 59 2003 reproduction of Leonardo da Vinci, Madonna of the Yarnwinder (The Lansdowne Madonna), plate 1. 55 56 No laws (so called) of artistic genealogy, no list of plagiarisms or tricks of style can pierce the secret of a masterpiece by Luini, Gaudenzio or Leonardo. True criticism must be gifted and trained to recognise the touch of a master like the handwriting of a friend, and his palette like the timbre of a voice. In discriminating between true and false, between good and bad, between beauty and ugliness we may have our preferences, but we must be catholic not partisan, disinterested not egoistic, interpreters not dogmatists. 2 Fig. 60 189860reproduction of Leonardo da Vinci, The Virgin of the Rocks, c.1483-1486, Paris, Musée du Louvre. Reproduced in Burlington Fine Arts Club, Illustrated Catalogue of Pictures by Masters of the Milanese and Allied Schools of Lombardy, London, 1899. 132 CONSERVATION AND ATTRIBUTION Fig. 61 1898 reproduction of Leonardo da Vinci, The Virgin of the Rocks (The Virgin with the Infant Saint John adoring the Infant Christ accompanied by an Angel), c.1491-1508, The National Gallery, London. Reproduced in Burlington Fine Arts Club, Illustrated Catalogue of Pictures by Masters of the Milanese and Allied Schools of Lombardy, London, 1899. THE LANSDOWNE AND BUCCLEUCH MADONNAS: CONSERVATION AND CONNOISSEURSHIP 133 Fig. 6363Joseph Duveen, Courtesy of Gimpel Fils Gallery, London, © Gimpel Fils. THE LANSDOWNE AND BUCCLEUCH MADONNAS: CONSERVATION AND CONNOISSEURSHIP 135 sculpture and who began courting not only collectors, but also scholars. He invested heavily in buying up collections, a gamble which eventually paid off when he sold them. Eventually Joseph (fig. 63) became head of the Duveen empire in which his siblings were also involved. (fig. 67) which he believed showed the painting to have been undertaken by two hands and to have later additions. Darker areas in the ultraviolet image indicate later retouchings, of which there are a number as described by Suida: [The ultraviolet] shows the epidermis of the picture and enables one to distinguish clearly between the old condition and additions to it. Sky and landscapes are well preserved; there are minor retouchings scattered here and there over the sky; in the landscape they are both few and unimportant. The rocks on the right are in good condition. Shadows in the lower part of the landscape are partially faded. Very striking in the ultraviolet photograph, but also easily apprehended in the picture itself, is the difference in tone between the head of the Madonna and the head and upper part of the child's body. The head of the Madonna is perceptibly darker and the reason for this is presumably different paint materials. … In some places the folds of the blue drapery have been repainted. The palm of the right hand of the Madonna has been greatly retouched. The child's face is well preserved. On the right cheek there is an old injury and on the curls some small shadows have been brought out more fully but the structure is original. The contour of the left arm of the child and the fingers have been repainted and the same thing can be noticed in the cross. The right arm of the child goes into the shadows. The lower part of the child's body has on the whole been considerably changed, though it is possible that much retouching here was superfluous. 20 Suida's pioneering observations are generally consistent with those we made in Chapter 1, though the darker passages in the Virgin's head seem to signal its restoration not an original difference in 'paint materials' . His final comment on the lower part of the Child's body could in part refer to Christ's genitals which have changed from being distinct in the 1931 Bodmer/I Tatti reproduction to having a rather different appearance (figs. 55 and 56). Suida concluded that the Lansdowne painting was finished by a student of Leonardo and it was the same person who was the author of the Buccleuch and Louvre versions (fig. 91). when we were then closely involved in researching and examining the painting, as is described in the first part of this book. The resulting historical, technical and scientific research has shed considerable light on the painting, its place in Leonardo's oeuvre as well as giving further insight into Leonardo's exceptional working practices and the materials he used. It has also shown how precarious stylistic judgements can be when technical evidence is unavailable or not fully taken into account. 1. 'Old Masters at the Academy' , The Times, 20 January 1872. 2. Herbert Cook, Illustrated Catalogue of Pictures by Masters of the Milanese and Allied Schools of Lombardy, viii, Burlington Fine Arts Club, London, 1899. Madonna of the Yarnwinder, placing the painting in Italy, where the painting was made, or France, where Florimond Robertet, the patron who commissioned the painting, lived, could strengthen its attribution to Leonardo. With the Lansdowne and Buccleuch Madonnas, it can be a way to develop a theory as to which was destined for the original patron and which was painted for someone else.Finding the provenance of a painting depends on finding the evidence to tell its story. The Buccleuch Madonna, discussed in the next chapter, appears to have been in the same family since the middle of the eighteenth century, when it was apparently bought in France. The Lansdowne Madonna, the subject of this chapter, has changed hands a number of times, and can only definitely be traced as far back as the early nineteenth century in London. Beyond this, we have to rely on educated supposition based on scattered sources. The two paintings have very different provenances. At the same time, they are equally fascinating in the way their stories not only reveal their history but also the history of the people who owned them. Because of the rich history surrounding the periods and people discussed, some detail of this is given as a way to provide context and greater depth to the story. In turn, each owner represents a certain period in history. This can tell us a great deal about how the painting was perceived, if it was fashionable to own it, what patrons of 14. THE LANSDOWNE MADONNA: A PICTURESQUE PROVENANCE Detail of the Madonna of the Yarnwinder (The Lansdowne Madonna), plate 1.the art looked for, and how art fits into the society of the time. July and comprised sixty-three works from the estate of 'the late Marchioness Dowager of Lansdowne' . Lot 47 lists a Leonardo da Vinci, 'Madonna and Child, an uncommonly fine production, formerly the property of Lord Darnley' . It was withdrawn from the sale. The next auction that mentions the painting was held the following year on 3 May, this time at the auction house of Christie, Manson and Christie. The title page highlighted both the Lansdowne name and our Madonna: ' A Catalogue of a very valuable Collection … a part of which were the property of the Dowager Marchioness of Lansdowne, Deceased; and sold by order of executors. … The collection contains a Virgin and Child by Leonardo da Vinci. ' Lot 112 describes the painting: 'Leonardo Da Vinci … The Virgin, with the Infant Christ, contemplating the cross, in an extensive landscape, -painted with great sweetness of character. ' The painting was the last lot but, as with the 1833 sale, it was withdrawn. A final sale on 25 July 1879 lists the painting once again, and again at Christie's. The Madonna was included in a group of five paintings from 'The Property of a LADY, Deceased: #334 L. DA VINCI. THE MADONNA, with the infant Saviour holding a cross, with landscape background. Purchased by the Marquis of Lansdowne in 1809, bequeathed by him to his wife, from whom it passed to the present owner.' An annotated catalogue at Christie's archives, London, tells us several more things. The painting was brought in on 21 June 1879 by Mssrs Arber, Rutter & Waghorn of Piccadilly. The picture was part of the estate of Miss Giffard (deceased) and was put into the sale on 25 July 1879. It was sold to a Mr Cooper, for Cyril Flower, later Lord Battersea (1843-1907), for a grand total of £ 147. These sale catalogues give us vital clues as to where to begin researching the Lansdowne Madonna's provenance. First, we know that the widow of one of the marquesses of Lansdowne once owned the painting. We also know that it was described as having been in the collection of Lord Darnley. The 1879 sale explains that the painting was bought in 1809 by the Marquess of Lansdowne and went from his widow's collection into Miss Giffard's. Finally, we can establish that the painting went from the 1879 sale into the collection of Lord Battersea. The first question to ask is: who was the Marquess of Lansdowne in 1809? Actually, there were two Marquesses of Lansdowne that year. The 2 nd Marquess John Henry Petty (1765-1809), died in November of that year and was succeeded by his half-brother, THE LANSDOWNE MADONNA: A PICTURESQUE PROVENANCE 151 Henry, the 3 rd Marquess (1780-1863). The published provenance for the Madonna describes the painting as being in the collection of the 3 rd Marquess. However, this can be discounted as Henry's wife, Louisa Emma Fox-Strangways, died in 1851, long after the 1833 and 1834 sales. How did the 2 nd Marquess obtain the Madonna of the Yarnwinder? Did he indeed purchase it in 1809, or did he inherit it? If the painting subsequently belonged to the widow of the 2 nd Marquess, Maria Arabella, how did it end up with Miss Giffard, and who was Miss Giffard? The 1833 auction describes the painting as coming from the Darnley collection. If this is true, who was Darnley, and what role does he play in the story? Vanity of possessing Capital works of Art of a few favourite Masters, and the pictures of these Masters at their best time … Pleasing compositions are what are most greedily sought after … Young St. Johns -Virgins and Child … are all the rage … The leading Masters are Titian and Rubens … The other fashionable Masters at present likewise are the Carracci, Leonardo da Vinci, Domenichino, Guido, Murillo, and very fine pictures of Albano. -William Buchanan to James Irvine, 6 June 1803 1 In 1809 Lord Darnley was John Bligh, 4 th Earl of Darnley (1767-1831) (fig. 68). He was born in Ireland and contemporaries drew parallels between his family and the equally prominent Lansdownes, as can be seen in a newspaper clipping c. 1800: This family, like that of the Marquis of Lansdowne is chiefly indebted to Ireland for its fortune. The ances-Fig. 68 Thomas Gainsborough, John, 4 th Earl of Darnley, 1785, oil on canvas, 76 × 63.5 cm, Washington D.C., Widener Collection. Image courtesy of the Board of Trustees, National Gallery of Art Washington. nal of Elizabeth Gibbes who travelled with her family to the Continent from August 1789 until at least October 1790, and it is from this source that we can glean infor-are likely to have also been noted by Darnley, as they were considered the places to visit and important artists to see. They include the fashionable artists described in Buchanan's letter to Irvine and within the next decade are found represented in Darnley's ever-growing collection.In her entry for25 March Elizabeth notes having 'dinner at the home of Lord Darnley' . A couple of weeks later, on6 April, 'Lord Darnley came to take his leave' and on10 April Elizabeth mentions that she 'saw Lord Darnley's chimney piece' . To date, no journals or letters from Darnley during this period have been found. Elizabeth Gibbes's descriptions of Italy, its collections and the people there give a good idea of the world Darnley inhabited. Many of the places she mentions were standard stops on the Grand Tour itinerary and give a good picture of who and what the young lord, already collecting for his estate in Kent, Cobham Hall, would have seen for himself. Elizabeth had arrived in Italy by September 1798, and mentions having seen in Verona 'the picture of St. George by Veronese at St. Georgio [the spectacular Martyrdom of St George, c. 1564, San Giorgio in Braida]' . In Venice she was busy visiting San Marco, the Ducal Palace and the church of San Marco, among other places, and she 'Rowed in a Gondola … A gondola is a very disagreeable thing. ' After Ferrara and Bologna -where she saw works by Correggio and THE LANSDOWNE MADONNA: A PICTURESQUE PROVENANCE 153 Guercino -she visited the 'Palazzo Caprasa' , 'a most magnificent Palace' , and saw there 'a beautiful holy family by Raffaello' . Elizabeth travelled to Imola, Rimini ('the picture of St. Jerome and the Angel by Guercino, very much spoiled by an attempt that was made to clean it' [The Vision of St Jerome, Confraternita di S. Girolamo]), Pesaro, Loretto, Spoleto, and was in Rome by February. Although 'disappointed by the piazza diSpagna' and frustrated by the complications of finding accommodation, Elizabeth began to settle down in Rome, receiving visits from fellow countrymen. One such person was a 'Miss Jenkins' , probably the niece of the dealer Jenkins, who was then based in Rome and who would dine with the family during their stay. She also mentions visiting the studio of the Scottish history painter, archaeologist, and dealer Gavin Hamilton (1723-1798), who will feature later in our story. Of the extensive list of churches and palaces she saw in Rome, of particular interest is the Palazzo Colonna, where she mentions 'a large picture by Salvator Rosa of Regales [Regulus] at the moment the Carthaginians are putting him into the Barrel filled with nails' . (This painting is now in the Virginia Museum of Fine Arts.) The Colonnas had been major patrons of Salvator Rosa, who was well represented in their magnificent gallery. This particular work was celebrated at the time and would become part of Darnley's collection. A standard excursion from Rome was Naples, and Elizabeth Gibbes and her family visited the home of one of its more famous ex-patriots, Leo. da Vinci. The hands well managed, Virgin's left hand on the arm of the Magdelaine: her right hand up in the attitude of a person who talks earnestly … Magdelaine profile body, three quarter face …' (17 April 1790). A few days later another work attributed to the same artist is spotted in the Villa Aldobrandini: 'portrait of Joan of Naples by Leo. da Vinci in black -beautiful right hand which holds little urn' (23 April 1790). By the end of the month the family had quit Rome and made their way to Florence, by way of San Lorenzo and Siena. The journal mentions visits to the Uffizi, Pitti Palace and Boboli gardens among other sites, and describes seeing friends. Darnley was back in England by 1791, preparing for his marriage to Elizabeth Brownlow, daughter of the Right Honourable William Brownlow of Lurgan in County Armagh. 7 The wedding in Lurgan appears to have been a quiet affair without even Darnley's mother present. In a letter dated 29 August he describes the event: 'We were married at Lurgan on Friday Morning last of the following day's letter, he elaborates: 'I was married in an old rusty coat and she in an old green riding habit. ' 9 The Irish aspect is worth pointing out in terms of possible connections with the Lansdownes. The above letters were written from Clifton Lodge, in County Meath, one of Darnley's Irish homes; the Lansdownes also had properties in Meath. Other Irish properties in the Darnley family during the 4 th Earl's lifetime could be found in Athboy (also County Meath), Trim, Clonteagh, Dublin and Longford. 10 Back in England, Darnley concentrated on Cobham Hall, where he continued years of rebuilding and redesigning the estate begun by his father, employing the English architect James Wyatt (1746-1813) for the house (as his father had done) and Humphry Repton (1752-1818) for the landscaping. 11 Wyatt completed the music room, which had originally been a banqueting hall, and it became known as the Gilt Room, because of its richly decorated interior. Wyatt was also involved in creating, in the upper storey of the Elizabethan north wing, a long picture gallery with a vestibule at either end. Windows along the north side were bricked up to provide wall space for hanging pictures along the 50 by 134-foot passage. Money does not appear to have been a problem for Darnley, and Cobham Hall became noted for its splendour, as an 1806 description of the house, which also reveals the taste of its owner, attests: The vestibule, which opens from the lawn, is partly fitted up in the Turkish and partly in the Italian manner … the Music Room … has been fitted up and furnished by his present Lordship, at an expense of more than 20,000l … the Gallery … is intended to be filled with pictures … The apartments in the south wing are decorated with many fine paintings … Among the most eminent of those now to be seen, is a large picture of the Death of Cyrus by Rubens, for which Lord Darnley has refused 2000 guineas; a most spirited sketch of the Lion Hunting, by the same artist; the call of Samuel, by Sir Joshua Reynolds, an exquisite painting of the Pascal Lamb, represented with the crown of thorns, and lily. St. John and his Mother, by Titian; a Nativity, with a great variety of figures, finely grouped and coloured; Judas Betraying Christ; the Finding of Moses; and a large piece of Fishermen in a Storm by Salvator Rosa, the colouring in the sky in which is extremely fine. 12 This is the earliest description of some of the paintings to be found at Cobham Hall and it is unfortunate that many of them had been moved to London at the time of the visit. Perhaps the Madonna of the Yarnwinder, which is not mentioned in the description, was in that London group. If so, she would have been very close to her next owner, as Darnley's London home was, like Lansdowne's, in Berkeley Square. 13 The 1790s was a decade during which a great deal of art from the Continent appeared in London as a result of the upheavals of war and revolution, and there were spectacular opportunities to buy top-quality works, though the prices for the most highly rated artists became very inflated. Darnley was heavily involved in making purchases as well as selling through agents, auctions and on his own. He was not afraid of going to the Continent during these troubled times as we see in a letter to his wife from 1793: THE LANSDOWNE MADONNA: A PICTURESQUE PROVENANCE 155 I reached Antwerp last night, where we slept and 'late Mr. Bryan [the dealer], as authorized by and on the part of the late Duke of Bridgewater, the present Earl of Carlisle, and the Earl Gower, now the Marquis of Stafford, for the ... price of43,000l sterling. 17 Which was the 'house of eminence' described byBuchanan, and was Darnley involved? Darnley's name is not mentioned here but a 1792 newspaper clipping does state that 'Lord Darnley is the purchaser of all the small pictures in the collection of the Duke of Orleans at the price of 20,000l. '18 Perhaps this explains his visit to Antwerp described earlier, or perhaps it relates to the earlier failed syndicate, but it is clear that Darnley was involved and, as we will see, he did own works from that collection. By 1798, a great sale which lasted six months was held in the rooms of Mr Bryan in Pall Mall, with Darnley as a noted buyer of some dozen paintings.19 As for the Dutch, Flemish and German part of theOrléans collection, that too, came to London, when Orléans sold it for 350,000 francs to Thomas Moore Slade, one of London's major art dealers, Lord Kinnaird, Mr Morland and Mr Hammersley, 20 who then sold them in London the following year. Thomas Slade was involved in selling to Darnley not only works from the Orléans collection, but also paintings from the Venetian Vitturi family collection, which he had acquired during a three-year stay in that city from 1774 until 1777. 21 The Vitturi family had been involved in purchasing works from Italian families and had produced a collection which became a target for other dealers such as Thomas Jenkins -the dealer in antiquities and art whose famous marble dog was recently purchased by the British Museum -as well as the West India merchant Robert Udney (whose works went into the Darnley collection). Slade eventually managed to secure the Vitturi collection for himself and brought it to his home in Rochester, not far from Cobham Hall. He held the collection for fifteen years 156 PROVENANCE before breaking it up, but selling it first to his friends, as he describes: Most of the principal pictures I let my good friends the Earl of Darnley and Sir Philip Stephens have … Most of those to Lord Darnley are still in his lordship's grand gallery at Cobham Hall, particularly theTitians, and the fine Pythagoras of Salvator Rosa … 22 These paintings give an idea of Darnley's taste, which mirrored the fashions of the time. In Buchanan's letters to his partner in art dealing, James Irvine (who was travelling around Italy picking up works to send to London), he describes the trends of the day: Vanity principally prompts the English to buy -and that Vanity leads the purchasers to please the prevailing taste and fashion of their friends, or be governed by the whim and voice of Artists. Of the popular Masters at present, Titian and Rubens take the lead. The former of these in his fine works large and small will bring any money. In the same list I shall likewise rank Leonardo da Vinci, the Carracci -Vandyck's Compositions, Historical -Guido's fine pictures, Claude's very capital pictures -Domenichino -Murillo's capital pictures -Albano's very fine pictures -Rembrandt, and beyond this very short list I hardly think it safe to go unless for a highly celebrated and well known Correggio, Raffaelle, or M. Angelo -which indeed could they be procured would bring any money. 23 Buchanan's letter could have been describing Darnley's own collection, with all of the artists mentioned eventually being represented. The 'Leonardo' painting mentioned by Buchanan, comes up again several times in his letters and was included in Gavin Hamilton's compilation of the best Italian works (Schola Italica Picturae, 1776). Although Darnley did not manage to secure this one for himself, he did buy a well aware of Darnley's spending power and kept note of it. In writing to Irvine in January 1804, he mentions that Darnley is in London as a possible target for the sale of some paintings, and in April, when discussing a potential buyer for a van Dyke Venus, Darnley's name is brought up: 'have Lowther or Darnley (the latter of whom you know is fond of Rubens and Vandyck and is now returned to London) seen her?' 25 Darnley had business with other dealers and agents, including the artist William Etty, who is noted as buying for him in France; 26 John Breen, who appears as a buyer for Darnley in the auctions; 27 Seguire, who bought on his behalf the Rubens painting of two boys blowing bubbles (£ 69 6s) which had been in Joshua Reynolds's collection and which appears in the 1833 Cobham inventory; 28 and Michael Bryan, who bought for him in 1802 two Giulio Romano paintings sold by Bridgewater and from the Orléans collection. 29 Darnley's active buying is also apparent in other sales: the 1801 sale from the Purling collection which created a scandal for the high prices fetched (£ 10,870 9s 6d for 107 lots), 30 an 1806 sale from the estate of Richard Hulse, which was noted for its important works, 31 THE LANSDOWNE MADONNA: A PICTURESQUE PROVENANCE 157 and the major 1819 sale from the Panné collection, where he purchased a Schiavone (Andrea Meldolla) Scourging of Christ for £ 69 6s. 32 Darnley was also involved in selling at this time, which brings into play the possibility of his selling the Madonna of the Yarnwinder. The likelihood of his actually knowing Lansdowne because they were neighbours in London is confirmed by a note from Lord Byron who mentions them both attending a dinner at Lord Cowper's London home, albeit in a humorous light: 'small but select, and composed of the most amusing people. The dessert was hardly on thetable when I counted five asleep; of that there were … Lord Lansdowne and Lord Darnley…. ' 33 If Darnley sold the painting to Lansdowne, it could have been a private transaction, which is certainly not impossible, as Darnley is known to have done this. In a letter to Sir William Hamilton, for example, Charles Grenville writes, 'January 5 th 1795 … I have been to Vandergutch. … There is a S. Rosa which Ld Darnley bought at Rome to which gbp300 is affixed. ' 34 The Madonna does not appear in sales where Darnley was known to sell his paintings. 35 Darnley's collection can tell us about his taste and appetite for collecting, but not very much is known about the man beyond that. He was involved with the Royal Academy and attended dinners there in 1796 as described by the painter and diarist Joseph Farington. 36 Later, in 1816, Darnley supported the Royal Academy and its creation of a school of painting by lending works from his collection as models for copying. 37 But Farington does not appear to have had a very good impression of him when he was described by the sculptor Richard Westmacott who spoke of Lord Darnley as being very proud, with the high notions of the Old Nobility. At Cobham Hall, they dine at a round table Lord Darnley sometimes sitting in one part & sometimes in another, & Lady Darnley always on His right Hand. When His Lordship is seated the Chaplain and Tutor places Himself opposite to His Lordship & the Children next to the Chaplain. The manner of Lord Darnley is uniform, cold and reserved. He has £ 20,000 a year. 38Farington also notes several occasions when Darnley was buying and selling paintings: in an 1802 sale: 'Lord Darnley bid 255 guineas for it [a Rubens Loves of the Centaurs] but Sir Wm Hamilton bought it for 260gs. ' 39 Darnley's art collection also consisted of works by contemporary painters such as Hoppner, and again Farington helps us. In 1794 he writes, 'Hoppner called. He has been in the Country painting, -at Mr. Bowles, Lord Darnleys, &c. &c. ' 40 At Darnley's, Hoppner would have seen portraits by Joshua Reynolds of the 3 rd Earl Darnley and the 4 th Earl's younger brother, and Gainsborough painted a portrait of the 4 th Earl (fig. 68). Hoppner also executed a portrait of the 4 th Earl as well as his son at the age of six 41 and the 1833 inventory (see below) lists about fifteen other portraits by the artist. Cobham Hall was the centre of Darnley's collection and an 1833 inventory, made following Darnley's death in 1831, describes and values over 160 paintings, but does not cite the Madonna of the Yarnwinder. 42 The inventory, drawn up by the artist Douglas Guest, gives a 158 PROVENANCE total value of £ 20,456 to the collection. Descriptions of the works Darnley bought and accounts of his contemporaries provide ample evidence of the Earl's passion for collecting what became 'one of the most considerable [collections] in England' . 43 His taste indicates that a painting such as the Madonna of the Yarnwinder would be at home in his collection. He has been found to have several links with Lansdowne, including an Irish background and estates in the same counties; both had an address in Berkeley Square; they were contemporaries in developing their collection; each is recorded to have works from the other's collections; and they shared similar tastes -buying from the same sources and dining together. There seems to have been ample opportunity for Darnley to have bought the Madonna of the Yarnwinder, but did he, and how did it pass to the Lansdownes? How could the 1833 sale description, citing the Madonna of the Yarnwinder as coming from the Darnley collection, fit in with the story? The possibilities are that the picture was obtained privately from Darnley by one of the Lansdownes, or that the picture was bought by a Lansdowne from an as yet untraced sale of items from Darnley's collection. We also have to allow the possibility that the reference to its ownership by Darnley is simply wrong. If Darnley was the previous owner, there is a route to his ownership which did not involve him as an active collector. LORD DARNLEY AND A FRANCO-ITALIAN CONNECTION Before exploring the Lansdownes and their role in the story of the Madonna of the Yarnwinder, travelled extensively in Italy. As early as 1491 he was an ambassador in Milan to Ludovico Sforza, Leonardo's patron. From 1494 he was part of the army of Charles VIII of France who, with the backing of Sforza, was making claims on Naples. D'Aubigny is known to have been in Naples, Florence (where at one stage he was ambassador), Rome, and Genoa. He spent long periods in Italy and was Governor of Calabria (in 1495). As Lieutenant General of the French army he won the Battle of Seminara against Ferdinand II of Naples in June 1495 and was appointed Grand Constable of Naples. He returned to Italy in 1499 with the army of Louis XII and was one of the three leaders -with Louis of Luxembourg (Count of Ligny) and Gian Giacomo Trivulzio -who invaded Milan, where Leonardo was living at the time, in September of that year. Trivulzio was to become a patron of Leonardo along with Charles II d'Amboise who acted as Lieutenant General to the duchy of Milan from 1499 to 1511. Ligny was known to Leonardo and may have collaborated with him on a secret undertaking to Naples. 45 It is therefore likely that d'Aubigny would have also been in contact with Leonardo, either in Milan or later during his movements around Italy. D'Aubigny wrote two treatises on war 46 and as Leonardo was deeply involved in military designs, machines and strategies they would have had at least this interest in common (fig. 69). THE LANSDOWNE MADONNA: A PICTURESQUE PROVENANCE 159 D'Aubigny was in Italy again in 1501, when he led the invasion of Naples. He was then in Rome where he crossed paths with Cesare Borgia (who employed Leonardo the following year). He was in further battles against the Spanish, culminating in the later (1503) Battle of Seminara which ended with a French defeat. D'Aubigny was a member of the Stuart family and this prestigious Scottish connection explains why he was also once Captain of the Scots Guard. His position gave him opportunities to represent France in dealings with Scotland. In 1508 he was sent Lumbardy' , 47 and compared him to the great classical military leaders. Not long after his arrival d'Aubigny developed a fatal illness and was buried in the church of the Blackfriars, Edinburgh. A further potential connection between the d'Aubigny family and Leonardo lay with Bernard's cousin, sonin-law, and heir, Robert Stuart, 4 th Seigneur d'Aubigny (c. 1470-1543), who also campaigned in Italy (1509, 1511-13, 1515) and was made Maréchal of France. As first Captain of the Guard he was in close attendance on the King and Queen and was given the honour of Knight of the Order of St Michael. This was a position also given to Florimond Robertet, the original commissioner of the Madonna of the Yarnwinder. One can probably assume that Leonardo was at least known to the d'Aubigny family, either in Italy or in France when Leonardo moved there in 1516. Could one of them have been the recipient of a variant of the Madonna of the Yarnwinder? French ministers, such as Florimond Robertet, and military figures, such as Pierre de Rohan,who will be discussed later, were well known to be collecting and commissioning art during the campaigns. Fig. 69 A69nobleman [presumably Bernard Stuart] and his servant receive written orders from a royal clerk in the presence of the French king and his courtiers, Fol 24v from Bernard Stuart, Traité sur l'art de la guerre, Illuminated manuscript on vellum, second quarter sixteenth century, 20.6 × 14.8 cm, Northern France (probably Paris). Courtesy of Jörn Günther Rare Books. Darnley, is descended. 48 Although this route for the Madonna cannot as yet be documented, it is entirely possible that the Madonna of the Yarnwinder simply came through the Darnley family before entering the Lansdowne collection.LEONARDO AND THE LANSDOWNESReturning to the Lansdowne family, it will be remem-bered that the sales of 1833, 1834 and 1879 all place the Madonna of the Yarnwinder firmly in the Lansdowne circle. In the first two, the painting is listed as from the estate of the Marchioness Dowager of Lansdowne and, in the third, the description of the work cites the Marquess of Lansdowne as a previous owner. Who were the Lansdownes during these key dates and how were they related to each other? What sort of people were they, and were they likely to have had an interest in collecting old master paintings? The starting point for research is the members of the Lansdowne family most closely connected to the clues we have. Investigation of this prestigious family provides a much clearer picture that supplies pieces of the provenance puzzle. The Marchioness Dowager of Lansdowne in 1833 was Maria Arabella, widow of John Henry, 2 nd Marquess of Lansdowne (1765-1809). By the time of the 1833 auction, the Lansdownes were already established as one of the leading families of the day. It is worth explaining something of the family history in order Fig. 70 Giuseppe70Cades, Gavin Hamilton Leading a Party of Grand Tourists to the Archaeological Site at Gabii, 1793, pen and ink and wash over pencil on paper, 44.90 × 58.30 cm, Edinburgh, National Gallery of Scotland. THE LANSDOWNE MADONNA: A PICTURESQUE PROVENANCE 161to understand more fully the setting in which this painting was found and to discover how the Madonna might have come into their hands.The father of the 2 nd Marquess of Lansdowne was William Henry Petty (1737-1805), the 1 st Marquess of Lansdowne and 2 nd Earl of Shelburne, who was a well-known but not always popular politician. As a stockholder in the East India Company, a large landowner in Ireland, and particularly as an official dealing with the American colonies, he was involved in the major issues affecting the British Empire in the mideighteenth century. In England, he was President of theBoard of Trade in 1763 and became Secretary of State twice -from August 1766 to October 1768 (resigning because of opposition to his support for conciliation with the American colonies), and from March to July 1782. He even enjoyed a short-lived stint as Prime Minister from July 1782 until February 1783, during which he completed the negotiation of the Treaty of Versailles, which ended the American Revolution. Shelburne was also a well-known patron of the fine arts and literature with a famed collection of antique sculpture. According to William Buchanan, the painter Henry Walton (1746-1813) was known to be an artistic adviser to Shelburne. Buchanan noted further, 'he [Walton] has everything to say with Lansdowne [Shelburne] -indeed it was he who formerly purchased all Lansdowne's pictures' . 49 Shelburne's homes, at Bowood in Wiltshire and Lansdowne House in London, became centres of cultivated and liberal society. 50 Following the death of his first wife (searching for old master pictures to sell to his wealthy clients. Together, Hamilton and Shelburne devised a scheme to create a hall of antique sculpture for Lansdowne House that would draw fromHamilton's extensive excavations in Rome. They drew up a memorandum which described the acquisition of '16 fine antique statues' , twelve antique busts, and twelve antique bas-reliefs for the Lansdowne House gallery, together with '11 large historical pictures' and '4 landscapes with figures relative to the Trojan war' , to be painted by Gavin Hamilton. 51 The archives at Bowood House include many published letters from Hamilton regarding the sculpture, 52 but there are also letters, unpublished, describing his search for pictures for Shelburne. They give an insight into Shelburne's taste for old master paintings as well as the style of Hamilton's picture dealing. They also remind us that these objects were very much planned to be integrated into the Lansdowne homes, rather than seen as pure investment. Furthermore, the letters give a lively sense of the atmosphere and the tastes of those involved in collecting old master paintings. Hamilton makes detailed suggestions to Shelburne, in letters that portray him as dealer, interior designer, antiquarian, restorer and artist. In one such example, a letter written from Rome on 8 August 1776, Hamilton wrote 'would not the St. Michael of Guido do over the chimney' , when discussing the ante-room at Lansdowne House in London. And for the drawing room he suggested, 'the space over the chimney merits as your Lordship observes a fine picture of Guido or Guercino and I wish I may be so fortunate as to meet with something that will suit' . The space above the chimney in the drawing room appears not to have found a picture by 1783, as a letter in the Bowood archives from later that year reveals: I am about purchasing a small collection near Venice where there are several fine things, in particular one that I imagine will suit the chimney piece of your drawing [room], it is an upright picture representing a Diana and Endymion, a beautiful picture and in fine preservation, and will not cost more than £ 120. I have just finished one of the same s ubject for which Mr. Beckford gives me £ 150. So that this picture will be a bargain and [I] should advise your Lordship not to let it escape you, as pictures of the old masters are now grown very scarce … Gavin Hamilton appears to have had access to a Leonardo manuscript during his searches. In a letter to William Hamilton, he wrote, Rome, 24 th April, 1779 … The principal intention of this letter is to beg that you would bestow some attention on a curious manuscript of Leonardo da Vinci, which is in the possession of the Abbate Corazza, now at Naples, who will wait upon you with it when you will think proper to acquaint him of your intentions. He lives in the Casa Orsini, being preceptor to a young man of that family. Perhaps a work of that sort wou'd be agreable to the King, as he is already in possession of a similar manuscript traiting on anatomy, the Abbate Corazza will give you all the particulars relating to this work … 53 THE VIRGIN OF THE ROCKS Although we cannot confirm that Gavin Hamilton sold the Madonna of the Yarnwinder to Shelburne, he did sell him another, most significant painting by that artist, which supports the argument that he may also have purchased the Madonna. In 1785, Hamilton had made his way to Milan where he came across and bought Leonardo's Virgin of the Rocks (National Gallery, London) at the Hospital of S. Caterina alla Ruota. The complex history of this painting has been deeply researched by Martin Davies and later scholars. 54 However, an unpublished letter at Bowood House is worth adding to the mix. The undated letter was written by Hamilton in London to Lansdowne and reads: I had yesternight a visit from Mr. Desenfans 55 the picture dealer who had heard of my Leonardo da Vinci, he has seen it and is desirous of being the purchaser, paying me part in money and part in pictures given in exchange. I have put off determining anything until tomorrow at two o clock when I must go to his home to settle this affair. I therefore esteem it my duty to acquaint your Lordship of what passes and offer it to your consideration, the price is £ 800 and as it is a most capitol performance of the master and never to be got again. I could wish it in your Lordship's possession, as a beginning of a collection which might be one day rendered the best in England and at moderate expense. If your Lordship chosed on this occasion to discount the sum I laid out in pictures for you at Bologna I shall be willing to take them on my own account, when your Lordship has considered my proposal. I should be glad of an answer before two o clock tomorrow when I must go to Mr. De-THE LANSDOWNE MADONNA: A PICTURESQUE PROVENANCE 163 senfans, all my ambition is that your Lordship should have the preference of this as well as every other fine thing that may offer to me … No 12 Poland St. Wednesday Morning 56 As it is undated, there is no firm tie with the 1785-6 period when Hamilton bought and sold the Virgin of the Rocks. However, there are two reasons to think it refers to the Virgin of the Rocks rather than the Madonna of the Yarnwinder. First, the price tag of £ 800 is very high and it would be very unlikely for a small painting such as the Madonna to sell for this amount. Secondly, another letter, this time from C. J. Greville 57 to the Duke of Rutland describing the Virgin of the Rocks, notes a price of £ 800 for the work.The letter, dated 24 December 1785, was written soon after the painting had arrived in London: I know so much the ardour and liberality with which you pursue all your objects that I can believe the purse is not more weighty than it used to be, and that your extravagance takes a more royal range than virtu requires. The reason, however, I write at this time is to inform you that I saw this day a picture which Mr. Hamilton, of Rome, purchased at Milan on his return to England.It is a public picture, painted by Leonardo da Vinci for the Church of St. Francis at Milan, and removed from thence when that convent was suppressed to the Foundling Hospital, which was allowed to sell it to make up a deficiency in the fund of that charity.It was painted on a board about 6 feet 8 inches by 3 feet 6 inches, contains a Virgin and Child, and a St. John and an angel. The figures are all beautiful, and it is the finest picture I have seen of the master out of Italy, and contains remarkably the expression and sentiment for which that master is celebrated. being an artist, was likely to have been capable of such an undertaking. Both the Hamilton and Greville letters show that the work was being touted at a speedy pace. Shelburne became the buyer by February 1786 and it appears to have stayed in the family until after his death, when it was presumably sold in the 20 March 1806 Lansdowne sale, where it was inaccurately described.59 The Virgin of the Rocks was documented in Lansdowne House in 1802/3 by the German travel writer Christian August GottliebGoede (1774Goede ( -1812). 60 He wrote about his travels through England, Wales, Ireland and Scotland, detailing visits to the estates and homes of the aristocracy. In London he visited Lansdowne House which he considered 'one of the greatest sightsof London' . 61 He enthusiastically noted the impres-sive architecture with an equally impressive collection of sculpture and paintings. Included in the listing of highlights is the Virgin of the Rocks, which Goede described as 'One of the largest and most admirable masterpieces of this artist. ' 62 He goes on to outline the work's composition and concludes with a description of the emotional intensity conveyed in 'the intimacy of a mother's love and pious joy over her Divine son' . 63 Could Gavin Hamilton have also found and sold to Shelburne the Madonna of the Yarnwinder? We have yet to find evidence of this. What is clear, however, is that Shelburne was closely connected with one of the key providers of Italian old master paintings to England, and purchased at least one Leonardo painting from him. 64 THE 2 ND MARQUESS OF LANSDOWNE AND THE GIFFARDS Upon the death of the 1 st Marquess in May 1805, his son, John Henry Petty, Earl of Wycombe, became the 2 nd Marquess of Lansdowne. Because of estate debts, many of the collections of the 1 st Marquess were sold at auction. The huge collection of manuscripts and books alone took thirty days to sell. They were followed by the collections of maps, prints, political and historical papers, and coins and medals. A number of paintings from Lansdowne House were also sold, but the antique sculpture remained in the family when the 2 nd Marquess bought the collection from his father's executors. Perhaps because Wycombe held the title of Marquess for such a short time, from 1805 until his death in 1809, he has become one of the lesser-known Lansdownes. However, he is key to our study as it was his widow's estate which attempted to auction the Madonna of the Yarnwinder. The archival records, which include numerous letters, show him to have been an eloquent and humorous individual. Although never able to please his father, he appears to have found his own niche with a somewhat eccentric lifestyle. He was given a liberal education at Bowood, taught by the philosopher and self-taught chemist Joseph Priestly (1733-1804) who was then living at Bowood (it was at Bowood that Priestly discovered oxygen). In 1781, the utilitarian philosopher and social reformer Jeremy Bentham (1748-1832), an habitual guest at Bowood, made the following comments: Of Lord Fitzmaurice [i.e. Wycombe] I know nothing, but from his bust and his letters: the first bespeaks him a handsome youth, the latter an ingenious one. He is not sixteen, and already he writes better than his father. He is under the care of a Mr. Jervis [Rev. Thomas Jervis, 1748-1833], a dissenting minister, who has had the care of him since he was six years old. 65 Wycombe followed his father's political footsteps only so far, acting as MP for Wycombe from 1786 until 1802. He was a keen sailor and would later spend much of his time on the coast, near Southampton. He travelled widely on the continent, and from 1793 to 1796 he spent much time in Italy, where he was known to have visited Florence, Siena, Venice, Pisa, Rome, Naples, Sicily, Genoa, Savona, Milan, Lucca, Bologna and Turin. He returned to Italy again in 1800. While in Florence, he commissioned a portrait from Francois Xavier Fabre (1766-1837), which now hangs at Bowood House. In 1794, while in Naples, he met William Hamilton THE LANSDOWNE MADONNA: A PICTURESQUE PROVENANCE 165 who was known for his collections of antiquities and paintings. Was Wycombe a keen collector of art like his father? Could he have found and purchased the Madonna of the Yarnwinder while in Italy? There is no suggestion that Wycombe became as well-known a collector as his father had been, or as his younger half-brother would become (Henry Petty, later Petty-Fitzmaurice, 3 rd Marquess of Lansdowne). But there are references in his letters to purchases made, and it can be established that Wycombe created an art collection of his own. Most of the Wycombe letters at Bowood House are addressed to his father and they deal chiefly with the politics of the day. However, letters written during his travels also include descriptions of a more artistic nature. Indeed, he took it upon himself to buy paintings for his father. Writing from revolutionary Paris in 1792, Wycombe addressed his father: I have taken the liberty of drawing upon you for the sum of one hundred pounds in return for which I send you pictures which, through the misery of the times and the unaccountable panic which prevails, I Manfredini was taken in some time since at Florence by purchasing as originals of Andrea del Sarto, two little pictures successively with the sanction of the Director of the Gallery, which pictures proved to be the work of a poor devil who said he meant no offence to Mr. Manfredini, but had reason to complain of the Director of the Gallery and had taken this way of being revenged upon him. 71 Wycombe's interest in creating his own collection is confirmed in a letter he wrote to his younger brother (Henry Petty-Fitzmaurice, the future 3 rd Marquess): 'I consider myself as having compleated by my purchases last spring, a small collection of historical and other pictures by old masters, which I have been some years in forming.…' 72 These lines confirm that Wycombe did indeed collect works by old masters and also make the suggestion of his being the owner of the Madonna of the Yarnwinder quite plausible. There is also a plausible explanation for the Madonna of the Yarnwinder entering the 1833 estate sale of Wycombe's widow, Maria Arabella. Wycombe's will at Bowood shows that he left all his possessions to Maria Arabella and in the case of her death, to her daughters from her previous marriage to Duke Giffard. The Giffard name emerges as a key clue in the provenance as it was the name of the seller in the 1879 auction of the Madonna. It now appears that one of Maria Arabella and Duke Giffard's children could have owned the painting. A copy of Maria Arabella'they maintained their close friendship after their return to England. The Holland House Papers include dozens of letters from Wycombe to Holland during the period 1797-1809. Further letters from Wycombe to Lady Holland cover the period from 1805 until his death in 1809. These entertaining letters have helped to construct Wycombe's movements, thoughts and relationships with family and friends. They have also given some insight into his collecting and have shown how and where he met Maria Arabella Giffard, his future wife. Returning from his Italian travels in 1796, Wycombe was met by a disapproving father who had heard about his son's affair with Mrs Wyndham, the wife of the English representative in Florence, and was anxious that his son find a career for himself. Probably because of his strained relations with his father, Wycombe did not spend much time in England. By October 1797 he had gone to Ireland, where the family had a number of estates. He appears to have decided to turn over a new leaf, with exaggerated promises to Holland of a reformed life: 'it would ill become you to talk of my unbridled passions. … Hence forward my conduct will THE LANSDOWNE MADONNA: A PICTURESQUE PROVENANCE 167 be such as to defy wonder. I shall become the pattern and example of constancy' (Wycombe to Holland, 30 October 1797, Dublin). 74 However, he kept up his relationship with Mrs Wyndham, who had followed him from Italy and would continue to be a troubled part of his life for several years, which ended in a suit brought against him by her husband. Dublin was not a centre of the arts and Wycombe does not appear to have attempted to make it one: People I think should take up the fine arts in Italy, as foreigners drink tea in England. … It would be as absurd to visit the County Mayo, or the County Tipperary, with a lamb after the Parma Corregio, as to arrive in Rome with a tub of whiskey (Wycombe to Holland, 27 September 1797, Dublin). 75 He made joking claims to Holland of taking advantage of the local lack of knowledge in the arts, by posing as a connoisseur: 'I encounter the mixed companies of Dublin where there is scarcely such a thing as a picture, by pretending to be a Connoisseur in the fine arts, whereby I have imposed on many worthy people' (Wycombe to Holland, 20 December 1797, Dublin). 76 SalvatorRosa might have introduced into an Incantation (Wycombe to Holland, 14 March 1800, Dublin), 89 By the end of 1800 Wycombe was required to go to Florence in order to find a witness for his defence in Mr Wyndham's case against him. Although he makes no mention of buying art while in Florence, he did make some purchases in Paris, on his way home: 'I have brought from Paris with me a number of trinkets and other objects d'agrement a part of which you may have if you chuse …' (Wycombe to Holland, 1 May 1801). And another reference to a French work is made a few days later: 'You should have walked upstairs to see my picture of Marat which is an undoubted original and a great curiosity' (Wycombe to Holland, 8 May 1801, London). 90 The following year Wycombe can be found in London, contemplating the purchase of another painting: 'There is at Mr. Bryans a picture of Susanna and the Elders well worth your looking at. I rather think I should buy it myself if my Bullocks were a year older' (Wycombe to Holland, 27 January 1802, London). 91 The 1834 and 1853 auction sales list a Susanna and the Elders by Jordaens, but we cannot be certain if this is the same work. Fig. 71 James71Gillray, 'A cockney & his wife going to Wycombe', 1805, hand-coloured etching and aquatint, published by Hannah Humphrey, 10 June 1805, © National Portrait Gallery, London. Fig. 7272Unknown artist, View of the Marquis of Lansdowne's Castle, etching, Southampton City Council Arts and Heritage. 170 PROVENANCE That spring, Lady Giffard's husband died. The couple had amassed a certain amount of money due to some successful litigation they were carrying out, but it is not known how wealthy a widow she was. Her release from marriage did not allow for Wycombe to wed her immediately, even if he had wanted to. His father appears to have been against the union, as was society in general, and it would not be until the Marquess's death in 1805 that Wycombe and Maria would marry. They wasted no time, however, and were married only three weeks after his father's death in May. At the time Harriet, Lady Bessborough, mentioned the occasion in reviewing recent weddings: The only extraordinary one is the present Ld Lansdowne and Ly Giffard -a Vulgar Irish woman near fifty and larger than Mrs. Fitzherbert … I suppose it is a point d'honneur, for she has liv'd with him publickly as his Mistress for some years past (19 May 1805). 92 This notably unflattering description is made more vivid in the satirist James Gillray's cartoon depicting their marriage (fig. 71). The death of the 1 st Marquess had brought with it several years of wrangling with the executors of his estate. He had left his affairs in disarray, with large debts that had to be paid off by selling the collections of sculpture, paintings and books. Wycombe managed to buy back the sculpture collection, but it is not known which paintings he purchased. He chose not to be closely involved in the handling of the estate process, feeling it more prudent to let the executors carry out the formalities. During this time, he was based mainly in Southampton, where he was busy constructing a mock Gothic castle from a ruin he had purchased in 1804 (fig.72). It was later badly damaged by fire and has not survived. His letters mention the problems he was having with the construction work and it was even noted in the papers: Fig. 73 David73Wilkie, Maria, Wife of John Henry Petty-Fitzmaurice, 2 nd Marquess of Lansdowne, 1808, oil on panel, 32 × 35 cm, Dublin, © National Gallery of Ireland. THE LANSDOWNE MADONNA: A PICTURESQUE PROVENANCE 171 He no doubt planned to base part of appear to have led a busy life, moving between Southampton, London and sometimes Bowood. A quite eccentric picture of the Marchioness appears in a description given by Jane Austen's nephew (Jane Austen's stay in Southampton overlapped with the Lansdownes' and she was lodged in Castle Square). He wrote: The Marchioness had a light phaeton drawn by six and sometimes eight, little ponies, each pair decreasing in size and becoming lighter in colour … It was a delight to me to look down from the window and see this fairy equipage put together; for the premises of this Castle were so contracted that the whole process went on in the little space that remained of the open square. 95 A portrait of Maria Arabella with her dwarf depicts a lady not unlike the descriptions we have found (fig. 73). It was painted by the Scottish artist David Wilkie (1785-1841) who was commissioned to carry out the work in May 1808. 96 Wilkie's diaries describe a pleasant stay of about two weeks at Southampton where he was often asked to dine with the Lansdownes and met a number of their friends, including the artist Chalon who was also carrying out work for the Marquess and his family. 97 Wilkie also painted for them The Sick Lady, now lost, and both paintings were delivered to Lansdowne House in London by February of the following year. A description of Wycombe during this period provides a picture equally eccentric to that of his wife: He was a very tall and thin man, riding on a long lean horse, and had following him a very little page, called his dwarf, mounted on a diminutive pony. The knight, the dwarf, and the castle, seemed made for each other. He must, in the main, have been a good sort of man, as the people about here, although they laugh at the castle and castle-builder, all speak well of him, and are hardly willing to admit that he was mad; but then, as I have observed before, the qualifications required for acknowledged insanity, are by no means easily attained in England, where a greater latitude is granted for whims, fancies, and eccentricities, than in other countries. 98 It is clear that Wycombe was a keen collector of paintings and an active patron. There is ample evidence to show that his interest in the arts was more than just a passing one. Although we know from the 1879 auction catalogue that he once owned the Madonna of the Yarnwinder, the surviving records do not reveal how it came into his hands. Did he purchase it in 1809, the year of his death, as stated in the 1879 catalogue, perhaps from Darnley? Perhaps he acquired the painting during his many travels in Italy (or elsewhere on the Continent). He could have inherited the painting from his father but did not allow it to be included in the sales of his father's collections (it is not mentioned in these catalogues). The last two possibilities would, of course, bring the picture close to Italy, its original home. Did Lansdowne buy the picture from Darnley? These questions will have to remain unanswered until further information comes to light. Wycombe did not outlive his father by many years. The financial problems with the estate, a full life as husband and stepfather to six daughters, Jane (who married in 1792, so presumably not his responsibility), Ann, Eliza, Harriet, Louisa and Maria ('my house is little better than a magazine of young virgins' , Wycombe to Holland, 17 June 1805, Southampton), 99 must have taken their toll and by the summer of 1809, he was suffering from a severe liver infection. The last letter before his death, written on 10 October from Lansdowne House in London, was written for him by Harriet. He was still very ill, being fed by his wife, but wanted Holland to look at two enamel works that had come from his father's collection, one attributed to Cellini. He goes on: 'My antiquities are well preserved. In modern works I am less fortunate, having had no less than two effigies of his present majesty mutilated at Southampton' (Wycombe to Holland, 10 October 1809, London). 100 This refers to a fire which nearly destroyed their home. Wycombe's will mentions only one painting, a portrait of Marat which he left to Lord Holland. Other than the usual monetary bequests, much of the estate was left to Maria Arabella, including 'my Plate, Pictures, Statues, carriage …' , 101 so presumably the Madonna of the Yarnwinder as well. Wycombe was only forty-four and as he left no male heir, the Lansdowne title and estate devolved to his younger half-brother, Henry Petty. Following the death of her husband, Maria Arabella appears to have continued with an active social life in both Southampton and London, as various items of news in the social columns attest. In 1811 Lady Bessborough, writing after a visit to her at the Castle, noted Fig. 74 Frederick74Sandys, Portrait of Lord Battersea (Cyril Flower), 1877, chalk on paper, 109.8 cm × 74.9 cm, Norwich, Norfolk Museums & Archaeology Service (Norwich Castle Museum & Art Gallery). been sold off in that year. By at least 1814 the widow and her daughters had a London home on Albemarle Street where they gave balls and remained prominent in London society. 103 In 1829 Wycombe Lodge (demolished in the nineteenth century) in Kensington was constructed and Maria Arabella was its first occupant. 104 It was here that she died on 24 April 1833. Although we know from her will that Maria Arabella left everything to her surviving daughters, Ann Bagot, Eliza, Harriet Mellish (a widow), Louisa and Maria, it appears that the estate was possibly in financial trouble as only three months later the auction which first mentions the Madonna of the Yarnwinder took place. The painting was withdrawn and then reappeared as the star lot in the 1834 sale, when it was once again withdrawn. A further sale of fifteen paintings from Maria Arabella's estate took place on 2 May 1835 at Christie's which included paintings from the previous sale, but not the Madonna. 105 It should be noted that a Mellish, perhaps Maria Arabella's daughter, Harriet Mellish, bought five paintings at the 1835 sale. They include a Head of St Peter by Guido, The Return of the Prodigal Son by G. Bassano, A Sibyl by Benvenuti, Portrait of a Young Venetian Noble by Giorgione and a picture of a Dentist Extracting a Woman's Tooth by G. Dow. Both the Benvenuti and the Dow appeared in the 1833 sale. It could be that the Madonna of the Yarnwinder was considered too precious to the daugh-ters to be sold at auction. At any rate, its appearance in the July 1879 auction from the estate of a 'Miss Giffard' now makes sense. The 'Miss Giffard, deceased' is most likely to have been Louisa Giffard, the last surviving daughter of Maria Arabella, who died on 6 May 1879, two months before the auction. The Bassano and Dow paintings bought by Harriet Mellish are also listed in this sale, which confirms the connection between families and paintings. The nineteenth-century provenance up to 1879 can therefore be summarised as follows. Some time during or before 1809 Wycombe bought the Madonna of the Yarnwinder. It was bought either from Lord Darnley or came with a Darnley provenance. At the death of Wycombe in 1809 the painting passed to his widow, Maria Arabella. When Maria Arabella died in 1833 the painting remained in the hands of her daughters until the death of her last surviving daughter, Louisa Giffard, in 1879 when it was sold at auction. LORD BATTERSEA AND THE MADONNA'S PUBLIC DEBUT At the 1879 sale the Madonna of the Yarnwinder finally left the Lansdowne circle. As noted earlier, the painting was sold for £ 147.00 to a Mr Cooper for Cyril Flower, later 1 st Baron Battersea (1843-1907) (fig. 74). Fig. 75 Nathan75Wildenstein, 1900. Courtesy of Wildenstein and Company. THE LANSDOWNE MADONNA: A PICTURESQUE PROVENANCE 175 never blinded his eyes for the men of greatness of his own days. He was one of the very earliest to recognise the genius of Burne-Jones and to become a close friend of the artist. ' 110 Other artists with whom he was in contact included Millais, Sandys, Tissot and Whistler. The Madonna of the Yarnwinder stayed in Battersea's collection until his death in November 1907. In February the following year his widow Constance sold the painting.IN THE HANDS OF DEALERS As mentioned earlier, the purchasers of the painting were the art dealers Nathan Wildenstein (fig. 75) and René Gimpel. Nathan Wildenstein grew up in the Alsace region south of Strasbourg. He came from a family who for generations had bought and sold horses. After the Franco-Prussian war he moved to France so that he could remain a French citizen. It was there that he met his wife, Laure Lévy (1856-1937). He began selling paintings by chance when an acquaintance asked him to sell one for him. Nathan visited the Louvre to research the painting and there he found his passion for art. With his wife in charge of the bookkeeping, Nathan opened his first shop buying and selling French paintings in Paris in the late 1870s. By the turn of the twentieth century he had become one of the most successful dealers of old master paintings in Europe. By 1903 Nathan had opened a gallery in New York followed by another in London in 1925. He saw an opportunity in the newly rich industrialists on both sides of the Atlantic, to one of whom he would eventually sell the Lansdowne Madonna. He established his own collection of paintings, many of which came from some of the most prominent collections in Europe. His son, George Wildenstein (1892-1963), inherited his father's eye and a passion for the study of art history, creating a library and photographic collection focusing on nineteenth-century art. George's own son, Daniel Wildenstein (1917-2001), continued the family tradition. Educated at the Sorbonne, Daniel published a number of catalogues on various nineteenth-century artists. Today Wildenstein is a global firm with galleries in New York and Tokyo as well as a research institute in Paris. Nathan Wildenstein's partner in the purchase of the Lansdowne Madonna, René Gimpel, was himself an established dealer. Gimpel's wife was Florence Duveen, youngest sister of Joseph Duveen, one of the most successful dealers of the time and who features in the history of the Buccleuch Madonna (see Chapter 13). Gimpel's sons, Charles and Peter, continued with the family tradition and founded Gimpel Fils in 1946. It still exists today in London with a focus on contemporary art. Fig. 7676Elsie and Robert Reford, Reford Family Collection. Gimpel and Wildenstein took the Madonna of the Yarnwinder with them to their Paris gallery. In 1911, they took the decision to transfer the painting from its original panel on to canvas. This procedure, discussed in Chapter 13, was a perilous one and is now undertaken only in extreme situations. The reason would have been to provide a more stable support for the painting (perhaps the wood panel was suffering from woodworm or was cracking). Another dramatic event in the Leonardo world also took place in Paris the same year. In the early morning of 21 August 1911, Vicenzo Peruggia, a thirty-year-old Italian painterdecorator working at the Louvre, took the Mona Lisa from its frame, hid it in his coat and disappeared. The news of the theft caused an overnight sensation. The director of the Louvre was immediately dismissed and a search, which captured the imagination of millions, began. Months of searching proved futile and by 1913 the Louvre had removed the painting from its catalogue in the belief that it would not be found. However, Fig. 77 The studio, showing the Madonna of the Yarnwinder, Montreal residence of Elsie and Robert Wilson Reford, c.1950, Reford Family Collection. THE LANSDOWNE MADONNA: A PICTURESQUE PROVENANCE 177 that same year the painting was finally recovered when Peruggia, hoping to secure ransom money, took it from Paris to Florence to strike a deal with the Uffizi and a Florentine art dealer and was caught. The joyous global reaction to the return of the painting helped secure the eternal fame of the Mona Lisa and Leonardo. THE REFORD MADONNA The Madonna of the Yarnwinder remained in Paris with Wildenstein's until the early 1920s, when it was transferred to their New York gallery. The move is likely to have been commercially motivated. The New York gallery catered to newly wealthy industrialists who required works of art to raise their cultural standing. By this time, the painting had lost its Leonardo attribution, however, as described in Part II, and was accepted as a Sodoma (1477-1549), a contemporary follower of Leonardo. It was as a Sodoma that the painting was eventually sold in 1928 to Robert Wilson Reford (1867-1951) and his wife Elsie Stephen Meighen (1872-1967) (fig . 76). Reford was a Canadian industrialist and shipping tycoon and bought the painting for $ 10,000. Reford had travelled to Europe in 1885, 1887 and 1888, where perhaps he acquired his passion for art. A 1912 Canadian Men and Women of the Time listed his hobbies as fishing, hunting, golf and art. Reford was a conservative but enthusiastic collector. His European paintings included works by Bronzino(portrait of Ferdinand Medici, now at the National Gallery, London), Willem Key, other minor artists. He also established a substantial collection of Canadiana which included prints, drawings, maps, medals and historic documents. Reford's relationship with Wildenstein was longstanding. He maintained contact with both their New York and Paris galleries and appears to have had a friendly relationship with both Nathan Wildenstein and his son George. Joseph Stransky, another dealer at Wildenstein's, was expected to deliver the Madonna of the Yarnwinder to Reford in Montreal in May 1930. The arrival of the Madonna appears to have been delayed, but Stransky remained in Montreal to work with Reford in rehanging the collection to accommodate his new acquisitions. The home in which the Madonna found itself was on Drummond Street, the Montreal base of the Reford family. It was hung in what was called the studio, which was used as a library/reading room by Robert and Elsie (fig. 77). 111 The room had been expressly built to house the art collection a decade or more after the building of the main house in 1904. There were cases and a large, walk-in closet to store the Canadiana collection. Both Robert and Elsie spent a great deal of time in the studio, and Robert especially would devote many hours to inspecting pieces from his collection at a large table. The room was occasionally used to welcome visitors, and was much less formal than the sitting room. While Elsie spent the summers outside Montreal, Robert remained in town to work. During these times, he spent most of his free time in the studio, poring over his collection and books. Following Reford's death in 1951, the Reford estate, including the Madonna of the Yarnwinder, went to his widow, Elsie. Several years after Elsie's death in 1967, the family decided to sell the painting. In December of 1972 the picture was sold at auction and purchased by Wildenstein again, by whom it was sold to the present owner. The Buccleuch Madonna has enjoyed the luxury of staying within the same family for more than 200 years. During this time it remained with the Dukes of Buccleuch and Queensberry (Montagu Douglas Scott) and had a permanent home at Drumlanrig Castle in Scotland. The Buccleuch family's collection has accumulated over many centuries and was obtained through personal purchases by family members, inheritance and marriage. The collection includes works by Rembrandt, Holbein, Gainsborough, El Greco, Murillo and Van de Velde among others, which are housed at Drumlanrig, Bowhill in the Scottish Borders, and Boughton House in Northamptonshire. A significant part of the Buccleuch collection was acquired in the eighteenth century through the marriage of the 3 rd Duke of Buccleuch to Elizabeth Montagu in 1767. In the decades before, Elizabeth's father, George Montagu (previously Brudenell), 4 th Earl of Cardigan and later 3 rd Duke of Montagu and 6 th Baron Montagu (1712-1790), her mother Mary, and her brother John Montagu, Lord Brudenell (1735-1770), had been active collectors. When Elizabeth inherited the collection, it passed into the Buccleuch family. Among the Montagu paintings which came to the Buccleuchs was the Madonna of the Yarnwinder. It had been purchased in 1756 in Paris from the estate of an eminent French collector, Marie-Joseph duc d'Hostun et de Tallard (1684-1755). Unlike the Lansdowne Madonna, this provenance takes the Buccleuch painting to France, the country of Florimond Robertet, the powerful French minister who first commissioned the Madonna of the Yarnwinder. Information from the 1756 sale allows us to gather leads on possible previous owners of the Madonna as well as possible links to Robertet and his descendants. the duc de Tallard. 1 Following the death of the duke the year before, the estate had compiled some 900 lots to be sold over the next two months (22 March-13 May). Pierre Rémy, who created the catalogue with Jean-Baptiste Glomy, also acted as a buyer. The paintings were valued by 'Mr. Collins' and the drawings and prints by 'Mr. Joullian' . The catalogue, long in descriptive details, was an example of an emerging trend in sales catalogues which sought to provide extensive information on the works for sale and their owners. 2 On the opening day of the sale, Rémy no doubt attended, apparently as a buyer for the Montagu family. Lot no. 1 one was the Madonna of the Yarnwinder now in the Buccleuch collection. Around forty-three copies of the sale catalogue survive, many of which include annotations with prices, buyers, and comments about specific works. These tell us that although the painting was valued at 1,600 livres it was bought for half that amount by 'Mr. Rémy, painter and dealer in paintings and other curiosities'. 3 Rémy went on to purchase at least six other paintings including the Portrait of Saskia as Flora by Rembrandt (no. 156, now at the National Gallery, London) and Rubens's Watering Place (no. 141, also at the National Gallery, London). 4 All seven of the purchases by Rémy appear in a c. 1770 inventory of the Montagu collection, so it is reasonable to assume that Rémy bought them for the family during the 1756 sale. 5 Attributed to Leonardo da Vinci, the Madonna of the Yarnwinder was given an extended description which began with the following: The Virgin seated, carrying on her lap the Infant Christ who is holding a cross. The background is a landscape; the height is eighteen inches [pouces] by thirteen inches [pouces] wide. This painting is from the best period of the Master and is perfectly preserved. Nothing is so rare, and at the same time so esteemed as the paintings of Leo-nardo da Vinci. This excellent artist who had the highest understanding of Painting, did not make many Pictures, but still went very far in his career. 6 Fig. 78 Back of the Madonna of the Yarnwinder (The Buccleuch Madonna), © Marina Wallace. The description reflects not only the great popularity of Leonardo, but also the rarity of his works. Having achieved half the estimate, the purchase price of 800 livres was not unreasonable for a small painting. One of the surviving copies of the catalogue includes annotations by the renowned collector and dealer Pierre-Jean Mariette (1694-1774), who himself amassed a collection of prints and engravings while cataloguing and researching artists. 7 Mariette's comments throughout the catalogue are rather critical and his notation for this painting was no different when he wrote, 'The flesh is the colour of gingerbread, which is not favourable to the painting. ' 8 Other works in the sale were perhaps not so unappealing. Mariette himself bought no fewer than eighteen paintings and works on paper at this sale.The Madonna of the Yarnwinder was not the only Leonardo painting for sale at the auction, but it appears to have been the only one now thought to be by Leonardo. Lot no. 2 two was given a Leonardo attribution and described a scene of Jesus in the manger with the Virgin and St Joseph. The background contained two shepherds and angels atop clouds. 9 Despite its much larger size ('3 pieds 11 pouces de haut sur 3 pieds 3 pouces de large') and Leonardo attribution, it was valued at 900 livres and sold for only 601 livres. The painting was sold to the possibly gullible Maximilien Radix de Sainte-Foix, a politician and financier. OWNERSHIP BEFORE THE DUC DE TALLARD How did the Madonna of the Yarnwinder come to be in the collection of the duc de Tallard? Was it a known painting and if so, where did it come from? There are normally three ways in which a painting can become part of a family's collection. It can be inherited, it can be purchased or it can be gifted. In the case of the Buccleuch Madonna of the Yarnwinder, we will look at all three of these possibilities. Suitably for a painting by Leonardo da Vinci, this investigation will involve the highest levels of the French aristocracy, including the original commissioner of the painting, Florimond Robertet, as well as the Church and the Crown. indicating others that should remain open until firmer proof becomes available. Marie-Joseph duc d'Hostun et de Tallard was the only son of Camille d'Hostun, duc de Tallard (1652-1728), a diplomat and military commander who became Maréchal of France. In 1677 Camille d'Hostun married Marie-Catherine de Grolée de Viriville La Tivolière. They had one daughter and two sons. Camille d'Hostun was an active and often highly successful commander of important battles of the period including during the War of the Spanish Succession (1701-1714). He was a less successful commander of the forces against the Duke of Marlborough at the Battle of Blenheim in 1704, where he was roundly defeated and taken as 182 PROVENANCE THE BUCCLEUCH MADONNA AND A FRENCH CONNECTION 183 prisoner to Nottingham until 1711. After his return to France, Camille d'Hostun was made a duke and in 1715 he became a Peer of France. In 1724 he was elected President of the Academy of Sciences. He died in Paris in 1728 at the age of seventy-six.In 1713 Camille's son Marie-Joseph married the fourteen-year-old Marie-Isabelle Gabrielle Angélique de Rohan (1699-1754) at the Palace of Versailles in the presence of Louis XIV. They had one child, Louis Charles, Comte d'Hostun, who predeceased both his parents in 1739. Marie-Isabelle was the granddaughter of the famous Madame de Ventadour (1654-1744), governess of the royal children (Gouvernante des enfants royaux), including Louis XV. In 1725 Marie-Isabelle became a lady-in-waiting to Louis XV's queen consort, Marie Leszczynska (1703-1768). When her grandmother retired in 1735, Marie-Isabelle took on the role of royal governess, a position she held until 1754. The Tallard family was therefore firmly positioned within the inner circle of the French Crown. The wealth of the Tallard family is clearly reflected in the contents of the 1756 sale. Nine hundred and ten works of art from the family collection were sold in this sale over a period of nearly two months from 22 March to 13 May. Not only were there two paintings credited to Leonardo, the collection also included hundreds of Italian works including ones assigned to Michelangelo, 11 Andrea del Sarto, Raphael, Domenico Fetti, and Correggio to name but a few. A number of the works can be identified in collections today. 12 One copy of the auction catalogue includes annotations which tell us that the collection was formed by Tallard with the help of 'M. Auberti lawyer from Rome' , but gives no further details. 13 It does reveal, however, that Tallard was directly in touch with collecting in Italy and not only buying second-hand in France. A collection of this scale reflects not only the taste of the owners, but also their level of wealth. To find a painting by Leonardo in a collection of this kind is perhaps a natural thing. Marie-Isabelle's position as royal governess allowed the couple to have a residence at Versailles and they were therefore in close contact with the royal family. Marie-Isabelle acted as governess to the ten children of Louis XV and Marie. The Tallard relationship with the Court at Versailles will be discussed later in the context of possible sources for the Madonna of the Yarnwinder's presence in the Tallard collection. But first, let us look at possible connections within the family itself, in this case with Marie-Isabelle's family. Marie-Isabelle was the daughter of Hercules Meriadec de Rohan, duc de Rohan and Prince of Soubise, and Anne-Geneviève de Levis-Ventadour. As a member of the Rohan family, Marie-Isabelle was descended from the famous Pierre de Rohan, Maréchal of France (Maréchal de Gié, 1451-1513) to Kings Charles VIII and Louis XII. In this capacity he took part in the French invasions of Italy during Leonardo's lifetime. Indeed, Marie-Isabelle was his direct descendant. 14 As mentioned in Chapter 3, Pierre de Rohan had been keen to obtain from the Signoria in Florence a commission from Michelangelo to produce a copy of Donatello's bronze David, a request he first made in June 1501. 15 This was just three months after Fra Pietro de Novellara described Florimond Robertet's commissioning of the Madonna of the Yarnwinder from Leonardo and reveals the great appetite the French had for Italian art. After a sustained campaign on the Maréchal's part, during which the Signoria was also involved in securing Michelangelo to carry out his famous marble David, the commission for the bronze David was initiated. However, Pierre de Rohan was not to get his statue. His political downfall, before the bronze was eventually completed by Benedetto da Rovezzano in 1508, meant that he could no longer claim the work. Instead, the sculpture was successfully obtained by none other than Florimond Robertet. The incident illustrates quite clearly Pierre de Rohan's keen interest in obtaining works by the foremost artists in Italy. He would certainly have known of Leonardo, and may have known him personally. They were both present in Milan during the French invasion of the city in 1499. It is not inconceivable that Pierre de Rohan, having heard of or seen Robertet's Madonna of the Yarnwinder, sought to acquire one himself. As can be seen from the example of the David, obtaining a copy of a work was in no way considered second best by Pierre de Rohan, especially if the work were by a great artist. It is therefore possible to consider that Marie-Isabelle was the owner of a Leonardo Madonna of the Yarnwinder through inheritance. Another possibility to consider is that the Madonna of the Yarnwinder came into the Tallard family through a purchase or a gift. Were there any known Madonna and Child paintings by Leonardo da Vinci in France during this time or previously and if so, can they be connected to the one found in the Tallard collection? biography of Leonardo da Vinci in the first printed edition of Leonardo's Treatise on Painting. 16 Du Fresne's Life of Leonardo was drawn mainly from Vasari and Mazzenta, early biographers of the artist, as well as his own reading of the treatise. It was the first theoretical biography of Leonardo and referred to his anatomical and water manuscripts while giving importance to him as an artist. In the biography, du Fresne described some of the Leonardo paintings in French collections. The list recounted that in Paris, in the cabinet of the marquis of Sourdis, there was a Madonna 'di riputatione' by the artist. 17 Several years later, André Félibien (1619-1695), biographer of French and Italian artists and historian of French royal collections, embarked on his seminaland influential ten-volume history of painting from antiquity to the seven-teenth century. In it, he wrote a biography of Leonardo, which also described works by the artist in French collections including the de Fig. 79 Peter79Paul Rubens, The Coronation of Marie de Medici (1573-1642) at St Denis, 13 May 1610, 1621-25, oil on canvas, 394 x 727 cm, Paris, Musée du Louvre, The Bridgeman Art Library. by Alessandro Turchi (1578-1649), 27 in which his coat of arms is displayed, and commissioned a sculpted group of the Annunciation by Pietro Bernini, 28 the father of Gian Lorenzo. This was to be placed in the church of San Bruno (1611-1620) in the Carthusian monastery which he had founded in Bordeaux. 29 Other paintings brought back from Rome for the churches of Bordeaux include a school of Caravaggio painting, St Jerome, a Christ on the Road to Calvary by Frans Franken II (1580-1642), a Resurrection by Alessandro Veronese, and a St Francis. 30 François also modified the archbishop's palace in order to house his own collection by creating a gallery. His time in Italy had given François an intellectual and aesthetic training which he utilised in France while his connection to the throne remained. In 1607 he was given the duty of baptising the duc d'Orléans, second son of Henri iv, and he took part in the coronation of Marie de Medici (1573-1642) in 1610. The event was depicted by Peter Paul Rubens in his Coronation of Marie de Medici (fig. 79). Here, François can be seen in his cardinal's vestments alongside the bishop of Paris, Henri de Gondi (1590-1659). In 1615 he officiated at the weddings of Elisabeth of France with the Infant Felipe (later Philip IV of Spain) and of Louis XIII with the Infanta Anne of Austria. Was the Madonna of the Yarnwinder ever in François's possession? It would certainly have been at home in his collection and his interest in Italian art was great. It lists a large number of tapestries, furniture and objets d'art as well as brief descriptions of a selection of paintings. Nineteen portraits are mentioned in a single sentence. Three large paintings showing the Delphic Sibyl, the King of England ('le roi dangleterre') and a landscape are also listed in another room. The portrait of the King of England could be connected to the Raphael St George and the Dragon Charles was known to have purchased after 1651 and which was formerly in the collection of King Charles I. No mention of a painting by Leonardo da Vinci can be found in the Jouy inventory. An inventory of Charles's Paris home was also drawn up in December 1666. However, this sadly was destroyed during the Paris Commune in 1871 and it is not known if another copy exists. The nine offspring of Charles and Jeanne lived varying lives. Three of the daughters joined a convent and a fourth, Isabelle (d. 1644), married in 1637 Martin second 'family' , in which all the works depict a rocky outcrop on which the Madonna and Child rest, is also very close to the original concept, but this time shows it in its final stage. The third group consists of just three paintings and includes, rather unusually, cherries and an apple. A fourth group combines16. MANY MADONNAS Detail of the Madonna of the Yarnwinder (The Buccleuch Madonna), plate 2. Courtesy National Galleries of Scotland and Buccleuch Living Heritage Trust. 204 THE INFLUENCE OF AN IMAGE the Madonna and Child composition with an entirely different composition. The final group comprises paintings by Spanish artists and reflects the presence in Leonardo's workshop of at least two Spanish pupils.THE BABY-WALKER GROUP There are five paintings in this group, which include in the left middle-ground a group comprising Joseph with a baby-walker, the baby Jesus and two other figures. As both the Lansdowne and Buccleuch Madonnas have very similar groups in the underdrawings, it is safe to assume that this group of paintings is based on the Madonna of the Yarnwinder in its early stages. A version at the National Gallery of Scotland, Edinburgh, possibly by Fernando Yáñez de la Almedina (fig. 80), and another previously in the Wood Prince Collection, Chicago (fig. 81), are most comparable and perhaps of higher quality than the others in the group. The baby-walker 'family' of variants all contain a spectacular dam through which a river flows down a steep incline, with further hills stretching beyond. Two of the versions (Edinburgh and ex-Wood Prince) show yarnwinders decorated with a vine, symbolising the wine of the Eucharist. Below the base of the yarnwinder is a basket of wool scattered with flowers, and in the Edinburgh and ex-Wood Prince works there appears a spindle. In all the paintings, the Madonna, Child and basket are placed on a stone wall. The presence of a basket is clearly described in Fra Pietro da Novellara's letter of 1501, but it does not feature in the two prime versions. As with the figure group, it seems to be the case that a basket was included in a very early stage of the painting which has been picked up by the works in this group. The Madonna's left hand closely holds the Child, with a distinctive upward curve to her thumb, unlike the Buccleuch and Lansdowne paintings where her thumb is tucked under the Christ's right arm. Equally distinctive is the positioning of the fingers in the Madonna's right hand, which are held at varying angles rather than the softer, more 'cupped' arrangement in the prime pictures. They are distinct from the other groups of variants in not depicting the Virgin with a headdress, perhaps suggesting that the early version of the original did not include one. This hypothesis is supported by the x-ray that Suida published in 1931 and which was discussed earlier (fig. 1). Instead, her dark curled hair hangs loosely about her shoulders. The face of the Virgin, with its smiling lips, is not unlike the Madonna and Child with the Infant St John by Fernando Yáñez de la Almedina at the National Gallery, Washington, discussed in the Spanish group below (fig. 87). The figures in these paintings are larger and may have been copied by eye. A drawing of the Madonna and Child only, with an indication of drapery and a wall, now at the Uffizi (fig. 85), shows close similarities to the 'family' that includes the baby-walker group in the tightness of composition and positioning of hands. It also bears similarities with a painting by Yáñez, discussed below (fig. 86). A sub-group of two paintings shows the Madonna and Child in reverse and without a basket (figs. 88 and 89 Fig. 90 After90distinctive, Leonardesque rocky outcrop on which the Christ Child is kneeling in the Lansdowne and Buccleuch Madonnas is a prominent feature in a group of eight paintings, which are close descendants of the orig-inal composition. This group includes works by artists from Italy, Northern Europe and possibly Spain, and is a test-ament to the popularity of the original painting. One of the group, possibly by the Lombard painter and close follower of Leonardo, Cesare da Cesto (1477-1523), and now in a private collection, appears to be the basis for the other paintings (fig. 90). Two further pictures copy closely the configuration of rocks seen in the first picture figs. 91 and 92). The Northern European style of the landscape does not argue against a Lombard origin for some of these paintings, since a number of artists from this region (such as Marco d'Oggiono, Giampetrino, Francesco Melzi and Cesare Magni) employed a Netherlandish style in their backgrounds.A pair of paintings from this group -one now in the Duke of Wellington's collection, and another in the Cathedral of Granada, Spain -are possibly by the same artist (figs. 93 and 94). The provenance of the Apsley House painting goes back to the Spanish Royal Collection, suggesting a Spanish origin which is strengthened by its similarities with the Granada picture. The painting in Granada has not been closely measured but the dimensions of the figures in the Apsley House version are so close to the Lansdowne Madonna as to indicate that the figure group was drawn by a reproductive process, either from a common cartoon or by a process of tracing from a Leonardo painting or replica. The facial features of the Virgin are similar to those in the Louvre painting and in drawings of the head of the Virgin in the Metropolitan Museum of Art, New York and the Uffizi (figs. 97 and 98). Given the Spanish provenance it is tempting to place a tentative attribution to followers of Fernando Yáñez de la Almedina, mentioned earlier, or of Fernando de Llanos, both of whom are discussed below. Most definitely a Northern painting is the Madonna of the Yarnwinder attributed to Cornelius van Cleve (1520-1567) (fig. 95). Set in a landscape most unlike the prime versions of the painting, but typical of Northern paintings, this version has the enamel-like quality of figures typical of Cornelius van Cleve and his followers. A painting which is rather like a Madonna of the Yarnwinder in a Virgin of the Rocks setting is the version found at Christ Church Picture Gallery, Oxford (fig. 96). Lombard in origin, this work is considerably larger in size than the prime versions. Virtually all the paintings in this group closely follow the core of the original composition. Leonardo da Vinci (Cesare da Cesto?), Madonna of the Yarnwinder, oil on canvas, private collection, Photo © Christie's Images/ The Bridgeman Art Library (Leonardo e il Leonardismo 40). Fig. 91 After91Leonardo da Vinci, Madonna of the Yarnwinder, Paris, Musée du Louvre (Leonardo e il Leonardismo 42). Fig. 9292Cesare da Cesto?, Madonna of the Yarnwinder, oil on panel, 44 × 34.7 cm. Private collection.features include the Virgin's headdress, the shape and positioning of her hands (with bizarrely long fingers on her left hand), the pose and placement of the Christ, and the formation of rocks in the foreground. Fig. 98 After98Leonardo da Vinci, Study of the head of the Madonna of the Yarnwinder, brown and white wash on white paper, 22 × 88 cm, Florence, Uffizi Gallery, G.D.S. 429e (Leonardo e il Leonardismo 65). Three versions of the Madonna of the Yarnwinder are distinct in their inclusion of an apple and three cherries (figs. 99, 100 and 101). Very close similarities in this distinctive composition indicate that the works were executed by a closely knit group of artists, or perhaps by the same artist. The fruits are placed where the basket is located in other versions. The cherries are representative of paradise and the apple of the Fall. The Child kneels on part of the Virgin's drapery, protecting him from the stony surface on which they rest. One can detect the distant origins of the stonework in the Buccleuch and Lansdowne Madonnas. Signs of a closer connection to them in these paintings include the Madonna's headdress, which follows in detail the intricate twists and folds of the cloth and the translucent veil which covers her forehead. The Madonna's clothing is comparably similar to the prime versions in the way it is layered and draped across her body. Finally, both hands of the Virgin, with the tucked-in left thumb and extended left fingers gripping Christ's side (no longer visible in the Lansdowne painting), and the softly curved shape of the right hand, show strong similarities with the prime versions. 212 THE INFLUENCE OF AN IMAGE Fig. 99 After Leonardo da Vinci, Madonna of the Yarnwinder with cherries and apple, oil on poplar (later cradled), 42 × 33 cm, private collection. Fig. 100 After100Leonardo da Vinci, Madonna of the Yarnwinder with cherries and apple, private collection, oil on panel. Photo © Christie's Images/The Bridgeman Art Library. Fig. 101 After101Leonardo da Vinci, Madonna of the Yarnwinder with cherries and apple, oil on canvas, private collection. Photo © Christie's Images/The Bridgeman Art Library. 99 100 MANY MADONNAS 213 MADONNA OF THE YARNWINDER WITH ST JOHN, TWO FIGURES AND FRUIT BOWL One of the stranger groups of variants combines figures from two separate paintings. In five paintings the Madonna of the Yarnwinder is conflated with the Madonna and Child with St John and Two Figures attributed to the Florentine artist Franciabigio (1482-1525), who worked with Andrea del Sarto (fig. 102). The group from the Franciabigio painting is placed in the left foreground in a rather awkwardly tight space, and the two figures behind St John become the Angel and Tobias. Two of the works depict an arched gallery of a building looming above the scene to the right (figs. 103 and 104). Instead of a basket, four paintings include in the right foreground a table with a shallow silver bowl filled with fruit (figs. 103-106). The resulting composition is rather crowded and the spatial handling of figures somewhat unrealistic. One example from this group adds a further figure of St Joseph to the right, leaning forward to gaze intently at the yarnwinder which the Child holds (fig. 105). One version attributed to Martino Piazza (active 1513-1522) includes a fruit bowl, but the figures on the left are now the infant St John and Elizabeth, and there is a sheep -a 'sacrificial animal' -placed prominently in the foreground (fig. 106). A final picture in this group is distinctive for its architectural setting (fig. 108). This Madonna of the Yarnwinder is unlike any other of the compositions because it places the Madonna and Child within a large loggia or gallery. Two large columns rest upon a wall with relief work showing scenes of the Annunciation and Adoration. Its location is unknown but earlier studies of the painting by art historians Hans Bodmer and Wilhelm Suida attribute the work to Bernardino Luini (c.1480-1532), with Suida effusively complimenting the 'delicious grey reliefs' Franciabigio ?, Madonna and Child with St John, Two figures, and fruit bowl, 118 × 88 cm, location unknown (Leonardo e il Leonardismo 49). Fig. 103103Follower of Leonardo da Vinci, Madonna and Child with St John, Tobias and the angel, and fruit bowl, Museo de Bellas Artes, Córdoba (Leonardo e il Leonardismo 47). Fig. 104104Unknown artist, Madonna and Child with St John, Tobias and the angel, and fruit bowl, location unknown (Leonardo e il Leonardismo 45). Fig. 105105Follower of Bernardino Luini?, Madonna and Child with Joseph, St John, Tobias and the angel, and fruit bowl, location unknown (Leonardo e il Leonardismo 46). Fig. 106 Martino106Piazza ?, Madonna and Child with St Joseph, St Elizabeth, a sheep and fruit bowl, Galleria Nazionale d'Arte Antica, Rome (Leonardo e il Leonardismo 59). Fig. 107107Follower of Leonardo da Vinci, Madonna and Child with St John, Tobias and the angel, location unknown (Leonardo e il Leonardismo 48). Fig. 108108Follower of Bernardino Luini?, Madonna of the Yarnwinder, oil on panel, 59.1 × 47.6 cm, formerly in the collection of Godfrey Locker-Lampson (Leonardo e il Leonardismo 62). 216 THE INFLUENCE OF AN IMAGE THE SPANISH PICTURES The last and one of the most interesting groups of variants includes the paintings made by or attributable to Spanish artists, most notably Fernando Yáñez de la Almedina, mentioned earlier, and Fernando de Llanos. Both artists came from La Mancha but had early training in Florence during the first years of the sixteenth century. Either artist could be the assistant referred to as 'Ferrando spagnolo' in two 1505 documents which describe him as an assistant of Leonardo da Vinci during his work on the Battle of Anghiari fresco in the Sala del Consiglio of the Palazzo Vecchio in Florence.3 There is confusion between these two artists not only because they shared similar names but also because they collaborated in their work and both absorbed influences from Leonardo and other artists inFlorence. They went to Valencia in 1506 where they undertook a number of commissions, most notably a major project comprising twelve large altarpieces for the Cathedral of Valencia, carried out in the period 1507-10. There has been much debate as to who painted which picture. By around 1513 the artists began to work independently. One work in the group, the Madonna and Child with the Infant St John by Fernando Yáñez de la Almedina (fig. 87), shows traces of the composition of the Madonna of the Yarnwinder, but in reverse, like the two paintings mentioned earlier in the baby-walker group (figs. 88 and 89). The same reverse feature is found in one of the Valencia Cathedral panels (not illustrated); this time the Madonna and Child are integrated into a Rest on the Flight into Egypt, sometimes attributed to Yáñez and sometimes to Llanos. Two further works show strong similarities in style, composition and iconography (figs. 109 and 110). They both feature the basket of wool, Christ resting on a ledge with an almost identical blanket beneath him, Madonnas with almost identical hand positions, a headdress unlike the Buccleuch and Lansdowne Madonnas but very like each other and with the Madonna's hair tucked away, and a building in the left middle-ground. This final feature is interesting, as the underdrawings of both the Buccleuch and Lansdowne paintings show slight sketches of arched architectural structures in the same location. A final painting in the group by Yáñez includes the head of Joseph appearing over the left shoulder of the Virgin (fig. 86). An inscription on the back of the panel appears to give the name of the artist, 'Hernandiañes' or Hernando Yáñez, and the year 1523 in Arabic. 4 Another Spanish artist taken by the theme of the Madonna of the Yarnwinder is Luis de Morales (1510-1586), working later than the 'Ferrandos' and who executed at least five versions of the painting, four of which are shown here (figs. 111-114). He is known for his many devotional paintings, which focused mainly on the Madonna and Child and the Passion. Among the Madonna and Child works, Morales's own version of the Madonna of the Yarnwinder displays a pious intensity typical of the artist. The close-up depiction with dark background is common with Morales, as is the striking use of bold colours. His earlier interest in Leonardo and the Lombard school was later augmented by influences from Netherlandish art. Painted mainly during the 1560s, the Morales Madonna of the Yarnwinder versions show the reverse composition Fig. 109 Fernando109Yáñez de la Almedina ?, The Virgin and Child in a Landscape (Madonna of the Yarnwinder), oil on panel, 36.2 × 28.5 cm. Courtesy Sotheby's Picture Library. MANY MADONNAS 217 found in other works, but are unique in depicting the Christ Child clutching a spindle of wool. For the most part the paintings show only slight variations in their composition. A final painting in the Spanish group depicts an altogether different Madonna of the Yarnwinder, with a blonde Virgin in a rather stage-like setting with Joseph behind her (fig. 115). It is one of the most distant relations of the original painting, being painted during a later period, and has fully absorbed its Spanish origins. Fig. 110 Fernando110Yáñez de la Almedina ?, Virgen con Niño, Museo de Bellas Artes, Murcia, Fig. 112112Luis de Morales, Madonna of the Yarnwinder, 1560s, oil on canvas, 71.5 × 52 cm, St Petersburg, The State Hermitage Museum. Fig. 113113Luis de Morales, Madonna of the Yarnwinder, 1560s, oil on panel, 48.5 × 33.3 cm, Berlin, ©bpk, Gemäldegalerie, Staatliche Museen zu Berlin. Fig. 114114Luis de Morales, Madonna of the Yarnwinder, 1560s. Courtesy of the Hispanic Society of America, New York. Fig. 115 Unknown115Spanish artist, Holy Family (Madonna of the Yarnwinder with Joseph), oil on panel, Museo Provincial de Bellas Artes, Guadalajara. 220 THE INFLUENCE OF AN IMAGE A BRIEF CONCLUSION TO A COMPLEX BUSINESS Nothing with Leonardo is ever simple. The extraordinary trajectory of his career across his varied fields of activity, his endless inventiveness and the erratic progress of his paintings from conception to completion, confront the historian with a formidable obstacle course. If we add the intricate uncertainties in the histories of our two Madonnas, we can understand why the research has been very challenging over a number of years and why the reader has faced not inconsiderable demands in following the stories. That Renaissance masters should have created copies or close variants of popular compositions is not surprising. This process of replication is particularly noticeable with small-scale devotional paintings by such famous masters as Sandro Botticelli in Florence and Giovanni Bellini in Venice. Leonardo himself, in the absence of a regular supply of autograph paintings, clearly sanctioned the production of 'Leonardos' by painters in his circle, making use of the resources of his workshop, such as cartoons or part-cartoons, drawings of various kinds and unfinished pictures. The variants were often compiled in a 'cut-and-paste' manner. of how the two pictures evolved is not totally resolved in all respects but the general outline of their parallel development is clear. The same clarity cannot be claimed for their provenances. Clearly, if one of the paintings could be traced seamlessly back to Florimond Robertet, it would achieve some kind of public priority in what the press has seen as a kind of 'battle of the babies' . Even so, in the light of what we have discovered, the closely similar status of the two A BRIEF CONCLUSION TO A COMPLEX BUSINESS 221 pictures would not be affected, since they are both 'originals' . The evidence as to which is the 'Robertet version' points in two directions. The extensive French provenance of the 'Buccleuch Madonna' , both real and hypothetical, clearly favours it as the one actually delivered to Blois. However, we know nothing certain about the provenance of the Lansdowne Madonna pre-1800, and it is as likely that it was in France as in Italy. It could be one of those mentioned in the French collections. The landscape in the Buccleuch version, which is less overtly Leonardesque, fits better with the idea that it remained in the master's workshop after 1507, awaiting completion. This would identify it as the painting in the Salaì list in 1525. We have set out the arguments as fully as we can -aware that it is difficult to assign relative weights to the evidence of provenance and of the visual judgements, neither of which is decisive in itself. The question remains open, and readers are invited to make their own judgements. should ideally be matched by another examination of the Buccleuch Madonna. We are happy that we have evidence to paint a remarkable picture of the evolution of the two Madonnas. The leapfrogging of alternate technical examinations could continue for an indefinite period. All the complexities should not be allowed to obscure the significance of Leonardo's two little pictures. They were as revolutionary in their own way as anything he produced in larger-scale painting, most notably the Last Supper, or in portraiture. He took a stock subject, the Madonna and Child with attendant symbols, and transformed it into a visual and emotional drama within the intense compass of a small panel. Its dimensions are close to those of a fully opened lap-top computer. That such subtle grandeur and psychological complexity could become manifest on this modest scale is incredible. Everything is triggered by the yarnwinder, the symbol of Christ's crucifixion, which is woven into the emotional fabric of the picture in an entirely new way. Raphael was most directly affected, but the whole trajectory of Florentine Madonnas, including those by Michelangelo, was diverted into a new channel. Via Raphael and his contemporaries, and courtesy of the many variants by Leonardesque painters, the new mode of characterisation raced across Europe. Madonna painting was never to be the same again. 222 MADONNA OF THE YARNWINDER: A HISTORICAL & SCIENTIFIC DETECTIVE STORY APPENDIX CHRONOLOGY OF THE LANSDOWNE AND BUCCLEUCH MADONNAS 223 1501 1651 Raphael1651Trichet du Fresne (1611-1661) writes the Life of Leonardo (Trattato della Pittura di Lionardo da Vinci, 1651 edn), which describes a Madonna and Child in the collection of another great-grandson of Florimond Robertet, the Marquis de Sourdis (Charles d'Escoubleau, (c. 1588-Dec. 1666). Full page image -PART I Chapter openingPART 1 FORENSICS Infrared reflectogram detail of the Madonna of the Yarnwinder (Buccleuch Madonna), plate 2. Courtesy National Galleries of Scotland and Buccleuch Living Heritage Trust. Christ 'at the age he disputed in the Temple' or of any other subject that showed the 'sweetness and suavity' of his innovatory style.Since he came to Florence he has done the drawing of a cartoon. He is portraying a Christ Child of about a year old who is almost slipping out of his Mother's arms to take hold of a lamb, which he then appearsto embrace. His Mother, half rising from the lap of St Anne, takes hold of the Child to separate him from the lamb (a sacrificial animal) signifying the Passion. ing the Child from the lamb. She is perhaps intended to represent the Church, which would not have the Passion of Christ impeded. These figures are all lifesized but can fit into a small cartoon because they are all either seated or bending over and each one is positioned a little in front of each other and to the left hand side. He is hard at work on geometry and has no time for the brush. I write this only to let Your Excellency know that I have received your letters.thing. The upshot was that if he could discharge himself without dishonour from his obligations to HisMajesty the King of France as he hoped to do within a month at most, then he would rather serve Your Excellency than anyone else in the world. But that in any event, once he had finished a little picture that he is doing for one Robertet, a favourite of the King of France, he will immediately do the portrait and send it to Your Excellency. I leave him well entreated. The little picture he is doing is of a Madonna seated as if she were about to spin yarn. The Child has placed his foot on the basket of yarns and has grasped the yarnwinder and gazes attentively at the four spokes that are in the form of a cross. As if desirous of the cross he smiles and holds it firm, and is unwilling to yield it to his Mother who seems to want to take it away from him. This is as far as I could get with him.10 Treasurer to three successive French kings, and who had first visited Italy with Charles VIII's invading French troops in 1494.Robertet was able to use his considerable diplomatic leverage to obtain works from those authorities, with whom he negotiated on behalf of his various royal masters. He even persuaded the Florentines that he was the right recipient of a bronze David by Michelangelo, originally promised to Pierre de Rohan, once Marshal of France but since fallen out of favour. 12 told. The Child is embracing the lamb and his mother tries to restrain him, and she in turn is restrained by St Anne. The Carmelite's reading of the meaning, in terms of the Church and Christ's destiny as Saviour, seems well conceived. Although the Virgin's gestures with respect to the yarnwinder are more ambiguous than his account might suggest, the apparent eagerness with which Christ embraces his forthcoming crucifixion is fully borne out in the surviving versions. sion of them between his two sisters. One of the items in a list of paintings is a 'Madonna with a Child in her arms' . The list features, among other notable items, the Leda, the Mona Lisa and Madonna, Child, St Anne and a Lamb. The values were high -the Leda was top of the list at 200 ducats, which indicates that they were almost certainly the major works by Leonardo himself. The Mona Lisa is estimated at 100 scudi. The value accorded to the first Madonna above, at 25 scudi, symbolism' , but it does serve as a neat label for what Leonardo was accomplishing in these two innovatory compositions on his return to Florence. Effectively, the way the subject was to be portrayed was radically reformed. Neither Michelangelo nor Raphael failed to tion when a prospective purchaser appeared. One Bolognese painter in the fourteenth century specialised in crucifixes on various scales and with varying degrees of rich materials at different price levels. He became known as Simone dei Crocefissi. The precious workshop book of Neri di Bicci, one of the older generation of painters in Florence, gives an insight into paintings as a form of merchandise.13 The workshop resources of drawings and cartoons provided the basis for variations on stock themes by the master and his assistants, and by both in collaboration. 14 If we should think that Leonardo was above such things, or had no patience with such routine matters, we only need read Fra Pietro's account, and look at the products of his workshop. The two versions of It made sense to assemble small-scale paintings in a cut-and-paste manner or by a process of repetition, since they were sold on a different basis from major commissions for one-off pictures on a large scale or mural paintings.16 As Leonardo himself said, paint-Fig. 11 Leonardo da Vinci, The Virgin of the Rocks, c.1483-1486, oil on wood transferred to canvas, 199 x 122 cm, Paris, Musée du Louvre.Fig. 12 Leonardo da Vinci, The Virgin of the Rocks (The Virgin with the Infant Saint John adoring the Infant Christ accompanied by an Angel), c.1491-1508, oil on wood, 189.5 x 120 cm, London, The National Gallery.We return now to May 1990 in Edinburgh, where I suggest that the way to take matters forward, retaining the research impetus and drawing the other versions of the painting into our net, is to plan a small but high-quality exhibition in the National Gallery's fine classical building by James Playfair at the Mound in Edinburgh. This is where Scotland's select and beautiful collection of old master paintings is displayed and the Duke's picture had once been on loan.It is also a happy coincidence that the Gallery holds in its storerooms below the public galleries a version of the Madonna (fig. 2), which is not perhaps either the closest to Leonardo or of the very highest quality, but which provides testimony to the wide-ranging variant types that the original composition had inspired. What we were not to know is that the little figure group in the left middle-ground was to play a key role as the research unfolded. It shows the Virgin and her inquisitive child, with another woman, watching Joseph (a carpenter or joiner by trade) making a frame on wheels, which is recognisable as a 'baby-walker' of the kind that both my children enjoyed before they could walk upright without assistance or support. A wooden baby-walker of this type was later made for me by Gordon Brown, who undertook various jobs for me, and road-tested by my grandson Etienne. 1 The idea for the show meets with a good reception from Timothy Clifford, the imaginative, ambitious and irrepressible Director of the National Galleries, and from Michael Clarke, the Keeper.The proposal for the exhibition in the Gallery's files is dated November 1989.Michael Clarke indicates that the best date for the show would be 15 May to12 July 1992, and this is agreed in principle. The Duke expresses his enthusiasm for the scheme and a more developed proposal is produced, with a preliminary list of loans. The Duke's support for the research, whatever its results, is less than that of the works that head the list, indicating that it was on a considerably smaller scale and that it belonged to a different category. The list notes which paintings were unfinished, and from this we can assume that it was finished.elled within Italy and finally to France. 20 This seems to leave one of the two prime versions of the Madonna of the Yarnwinder as the painting retained by Salaì after Leonardo's death. connoisseurship. Even before examining the picture in the National Galleries studio in Edinburgh, and having previously seen it hanging in a room at Drumlanrig Castle in the Scottish borders, I was reasonably sure on grounds of quality that the Buccleuch picture directly involved Leonardo's participation. There were passages of painting, such as the headdress of the Virgin, that seemed beyond the reach of even his most able assistants. I was particularly attracted to the painting of the rocks on the right, which showed a fascination with the way that the horizontal strata in a thin-bed-ded limestone cut across the mineral veins that created a rhythmic pattern across its exposed face (fig. 13). None of his entourage would have bothered with such details, with the possible exception of Francesco Melzi, the young Milanese nobleman who later became the master's most faithful companion. But Melzi could not paint like that. Now we also had pentimenti, not conclusive, to be sure, but certainly supportive of the idea that the composition was designed and laid in by Leonardo and at least in part painted by him. The reception of the paper, not least by Pietro Marani, the leading younger Leonardo scholar in Milan and coeditor of the resulting publication, who delivered an ingenious paper on a little-known version of the Virgin of the Rocks, was decently encouraging. of California in Los Angeles. Like any publication by Pedretti, it is a mine of information, not least about the painted copies and variants, together with related drawings. Carlo's knowledge of Leonardo's legacy and versions of his compositions is of a depth and breadth that no one is likely ever to emulate. He expresses the painting given pride of place in Vinci was bright and fresh, perhaps disconcertingly so. The soft and rounded modelling of the forms in the Scottish painting, bearing witness to Leonardo's renowned sfumato (smoked) effects of blended contours, was less evident, and there was a certain sense of flatness in the brilliant colours. It was evident that it had been cleaned fairly recently. No pictures of this age have entirely escaped the intervention of past restorers concerned either to preserve the picture or more often to 'improve' its appearance. The extraordinary conservation history of the 'New York' version that emerged during the course We have become so used to seeing Leonardo paintings dirty, above all in the Louvre, that we somehow do not expect them to seem fresh and bright, as they must have been when first executed. The relatively recent restoration of the Last Supper and its freeing of the remaining fragments of Leonardo's paint layers from later over-paint, like the cleaning of Michelangelo's Sistine Ceiling, has been met with unease and worse. Magazine. 4 I recognise that 'it is not hard to marshal evidence against the Reford version as Leonardo's original' , but stress that 'the credits on the balance sheet between New York and Scotland are not all on Any exhibition, whether in Vinci or Edinburgh, remains a constant source of pleasure. He might, after all, find that his painting did not stand up to the more sustained scrutiny that we were proposing. There had been one previous exhibition centred on the Madonna of the Yarnwinder. In 1981-2 the Lansdowne version had been the focus of an exhibition, Leonardo dopo Milano. La Madonna dei Fusi, in the Castello dei Conti Guidi in Leonardo's home town of Vinci. Leonardo material, the Biblioteca Vinciana, had been set up in an adjacent building. Although only a small hill town, its centre no larger than in Leonardo's day, and the infant Leonardo was raised in the family home of his grandfather. In 1457 Leonardo is recorded by Antonio in his tax declaration as one of the 'mouths' he had to feed. 3 Ser Piero, the errant father, was a notary of more than provincial importance. He acted for Cosimo de' Medici in Florence in 1462, and was appointed in 1469 as lawyer in the Palazzo del Podestà, the seat of the chief law and police officer in Florence. It is possible that he took Leonardo to Florence at that time to enter the studio of the leading master Andrea appearance of the picture. Finally the facing cloth is removed, and we have a painting on canvas -but at a price. This hazardous procedure was invented by French restorers working for the Royal Collection before the Revolution in 1789. The inventor of the method in the mid-eighteenth century was Robert Picault. One of his set-piece transfers was the Raphael St Michael, rot could set in, and woodworm was endemic in old wood. Even if the panel did not fall apart, the paint and priming surfaces tended to become detached and flake off. We should recall that there were no such things as air-conditioned galleries until comparatively recently. Francis I was known to have housed Leonardo's Leda and Mona Lisa in his Apartment des The National Gallery via Pedretti enquires about the potential availability of the picture, and is asked to provide more detail about the rationale for the show and how the painting is to be catalogued. I respond drilled holes and old labels, the New York version was boxed in so that the rear surface was not accessible. In any event it was apparent from what I was told and what could be seen that no original support had survived. At some point the pigment layers and perhaps some or all of the priming had been stripped off its wooden support and laid down on canvas. (The technical details are discussed in Part II, 'Conservation and attribution' .) Apart from being reasonably sure that Leonardo did not paint on canvas, I could also see how the pattern of cracking on the paint surface exhibited characteristics showing that at one time it had been on a wooden panel which had cracked slightly along the grain. For the layman, and even for the specialist, the idea that a painting should be stripped off its wooden support sounds frightening. It involves 'facing up' the paint surface -that is to say the pasting of a layer of cloth limbs in a very un-Leonardesque manner. Looking at the painting later with William Griswold of the Getty Museum confirms the impression that some of the outlines have indeed been reinforced by a later restorer. more probably from the splitting of a single panel. At the top, in the sky above the Virgin's head, this central axis of instability has resulted in the flaking of a roughly triangular-shaped fragment of sky. A series of smaller losses has occurred as the fissure progresses across the Child's waist, genitals and bent leg. It was no doubt such signs of problems with the wooden panel that encouraged its transfer to canvas. Other signs of deterioration are characteristic of Renaissance panel paintings; the pattern of cracking in the Virgin's blue garment, in places quite conspicuous and deep, has been observed when a thick layer of azurite blue of those who wished to preserve it. Over the years it has also been subject to the effects of exposure to light (more destructive on some pigments than others), the remorseless yellowing and cracking of successive layers of varnish, some of which will have been removed and replaced over the years, and accumulations of dirt, some more pernicious, adhesive and corrosive than others. That we have anything worth admiring at all is a tribute to the long tradition of craftsmanship on which Renaissance masters depended (and sometimes even to the skills of earlier restorers). The New York painting was not the best I had seen, but it was certainly a long way from being the worst. On the other hand, the puckering and fine wrinkling of the paint surface in the Virgin's cheek and chin is recognisably a Leonardo quirk, shared with the Madonna and Child with a Carnation in Munich (fig. 14). 6 This effect resulted from Leonardo's application of layers of oil-rich pigment over underlayers that were not completely set. It is apparent that some areas of the New York painting have suffered more than others. A particularly conspicuous patch of retouching, presumably covering a damaged area, is apparent in Christ's cheek. The area around his genitals is undefined and difficult to read, for reasons that subsequently became all too clear. There are little areas of discoloured All the modelling is inevitably dimmed and blurred, and the colder colours, such as the blue of the sky, are transformed by the yellowish overlay. Should they be cleaned? I advised the Duke not, when he enquired about the possible cleaning of his Madonna. I am a minimalist in matters of cleaning, favouring intervention only if the paint surface is threatened structurally or the overall appearance is grossly disfigured. How In any event, the New York picture had been cleaned, no doubt more than once. The ravages of its history are clear, but so are its painterly and colouristic delights, which survive in good shape. The magical setting of typically Leonardesque mountains, painted with swift and certain strokes of a small brush, far surpasses the pedestrian landscape of the Buccleuch version. There are some brilliant passages of painting, especially involving the kinds of animated details in which Leonardo took particular delight. Eyes are of special import in Leonardo's paintings. The impact of the Mona Lisa, for example, depends utterly on the way that Leonardo exploits her eye contact with us. For him, the eyes are the 'windows of the soul' , permitting us to see the luminous wonders of nature from our bodily prison, and enabling others to search deep into the inner life of our thoughts. The Virgin's eyes are elusively lowered as she looks anxiously at her son, while Christ's right eye is fully visible, portrayed withThe obedient Fra Pietro duly investigated what Leonardo was doing. On 3 April 1501 he reported that: Your most Illustrious, Excellent and singular Lady I have just received Your Excellency's letter and will carry out with all speed and diligence that which you instruct me to do. From what I gather, the life that Leonardo leads is haphazard and extremely unpre- dictable, so that he only seems to live from day to day. St Anne, rising slightly from her seated position, ap- pears to want to restrain her daughter from separat- I shall carry out your Excellency's commission and keep Your Excellency informed. … 9 Eleven days later he had gained direct access to Leonardo via his entourage (which was not seemingly an easy thing to do), writing to Isabella that: During this Holy Week I have learned the intention of Leonardo the painter through Salaì his pupil and some other friends of his who, in order that I might obtain more information, brought him to me on Holy Wednesday. In short, his mathematical experiments have so greatly distracted him from painting that he cannot bear the brush. However, I tactfully made sure he understood Your Excellency's wishes, see- ing that he was most eagerly inclined to please Your Excellency by reason of the kindness you showed to him in Mantua, I spoke to him freely about every- This second letter (fig. 10) contains what is the first notice of what we now call The Madonna of the Yarnwinder. Although the essence of his description corresponds well to the two prime versions, the basket ISABELLA'S AGENT 31 (with spindles?) is nowhere to be seen. In the variants that do show a basket with spindles, including that in Edinburgh (fig. 2), it is not close to the Child's feet but is placed either on a rocky ledge or table at the other side of the painting, near his left knee. 11 Although we look hard in Edinburgh at the Buccleuch Madonna for some trace of the basket and spindles in the lower left of the painting, the infrared rays have difficulty pen- etrating a particularly dense area of pigment and do not reveal anything that we can read. Fra Pietro proves to be a good witness. He gives us a date by which the painting was under way, and he tells us the name of its commissioner, Florimond Robertet, a notable diplomat who was Secretary of State and Fra Pietro also gives a highly perceptive account of how Leonardo's new kind of Madonna and Child com- positions were conveying meaning. The devotional subject was being endowed with something close to a narrative. The symbols of Christ's passion -the lamb in the case of the cartoon, and the cruciform yarnwinder -were woven into a fabric of motion and psychological interplay. In essence, a 'story' was being I cannot remember whether it was John Shearman (who was responsible for my postgrad-uate studies at the Courtauld Institute of Art in London) or myself (or even someone else) who coined the phrase 'dynamic Fig. 10 Letter from Fra Pietro da Novellara to Isabella d'Este, 14 April 1501. Private collection. gories of 'Leonardos' . 19 Following the death of Salaì in 1524, a notary listed and valued his not inconsiderable possessions, with a view to making an equitable divi- take note. What Fra Pietro's account of Leonardo's working prac- tice also indicates is that the painter's attention was characteristically diverted away from painting and on to other matters, most notably geometry, and that there was something of a small 'production line' in the workshop, which involved assistants making copies in which Leonardo intervened. There was nothing par- ticularly surprising about such a procedure. Although the popular image is of the great genius-artists cre- ating masterpieces in transports of creative energy, journeying artistically where none had gone before, Renaissance masters, however grand and innovative, had workman-like ways of making saleable prod- ucts, particularly of small-scale devotional items for domestic settings. Some of these might be described as off-the-peg works, ready for sale or for comple- the Virgin of the Rocks (figs. 11 and 12) testify to the procedures of replication or near-replication that he was willing to follow. And, as it happens, we have a drawing by Gian Antonio Boltraffio that collages a Madonna and Child from the heads of the Virgin and Jesus, probably using the cartoons for the heads in the altarpiece. 15 Boltraffio was working in Leonardo's studio in the 1490s, and on one occasion had a valu- able silver-point drawing instrument stolen from him by Salaì, the rascally but beguiling young apprentice who was to remain with Leonardo throughout his life. ers should 'keep aside some good work saying, this is worth a good price, this is medium-priced and this is run-of-the-mill, thus showing that they have works in all price ranges' . 17 The best-quality pre-made paintings could be offered to the most prestigious clients. In 1506-7, when Leonardo was shuttling between Florence and Milan before formally entering the service of the French king, Louis XII, he promised to bring with him two pictures of 'Our Lady of different sizes' for the King himself or someone else deemed appropriate. 18 A document discovered in Milan by Graziosi Sironi and Janice Shell had also shed light on different cate- 4. EDINBURGH, VINCI, NEW YORK: LOOKING, PLANNING AND RESEARCHING The Milan paper referred to in Chapter 2, and out- lined here, failed to emphasise that only the Madonna Litta in the Hermitage and the Madonna of the Yarnwinder match the description in the Salaì list. The former appears to have been substantially exe- cuted by Boltraffio in the first Milanese period, and is unlikely to have been one of the cherished paintings to which Leonardo clung in his later years as he trav- It was such issues of categories of painting and work- shop production that were discussed in the Milan conference paper, bringing together documentation about patronage, technical evidence and traditional Fig. 13 Detail of rock formation, Madonna of the Yarnwinder (The Buccleuch Madonna), plate 2. Courtesy National Galleries of Scotland and Buccleuch Living Heritage Trust. his view that 'the Reford [Lansdowne] version has the greater probability of being his [Leonardo's] own pro- totype' , but avoids drawing dogmatic conclusions. Alongside the two prime versions, he lists and illus- trates six painted variants. The Vinci show presented the first opportunity for me to see the 'New York' version. At first sight, it looks very different in appear- ance from the Buccleuch picture. Compared to the mellow tones of the latter, muted by darkened varnish, of the subsequent research will be detailed later, and will do much to explain why it looked superficially (literally superficially) so unlike and so much brighter than the Buccleuch version. We simply do not expect to see our old masters radi- ant with fluorescent hues. And, of course, we now harshly flood the cleaned work with light of a qual- ity and quantity unavailable in Renaissance interiors. Leonardo and Michelangelo knew, as did their prede- cessors, that enhanced luminosity of the painting aided the visibility of mural paintings in dimly lit interiors. I review the Vinci exhibition for The Burlington one side' . Already I am warning that 'in this case, more than most, the dividing lines between autograph, partly autograph and non-autograph, between master, studio and follower, between original, copy and version are likely to elude close analysis' . I conclude, on little more than a hunch, that 'neither the Reford nor Buccleuch versions should be definitely excluded as direct prod- ucts of Leonardo's studio' . At this stage, unaware of what I would subsequently learn, I consider that the Buccleuch picture involved more of Leonardo himself than did the Lansdowne version. WORKING TOWARDS THE EXHIBITION IN EDINBURGH As it happens, one of my graduate students at this time, Thereza Crowe, is already undertaking doctoral researches on Leonardo's French patrons and is, inevi- tably, looking into Florimond de Robertet and the Madonna he commissioned. It is agreed that I should select and catalogue the exhibition, with Thereza as research assistant. She was to re-enter the story a number of years later. AN EXHIBITION IN VINCI The castle, dominating the centre of the town, houses the Museo Leonardo da Vinci, and contains a series of small-scale models of uneven quality based upon some of his engineering drawings. A notable library of Vinci is dominated by the posthumous presence of its gigantic son. In case anyone should miss the link, a massive wooden sculpture of the famous 'Vitruvian Man' , made in 1987 by Mario Ceroli, stands toweringly on the terrace of the castle. The man's orbiting hands and feet trace his inscription in a circle and a square, as first described by the ancient Roman architect Vitruvius, but now within a spherical frame reminis- cent of an armillary sphere -the ancient model of the cosmos according to Ptolemy. Tourism and Leonardo merchandise in the town below underscore his endur- ing presence. A small house at Anchiano in the olive groves just above the town, where it is said Leonardo was born, is a place of pilgrimage for the reverent and the curious. Leonardo's family were prominent local land and property owners and notaries. Leonardo's grandfather, Antonio di Ser Piero da Vinci, recorded in a memo- randum that Caterina, a farmer's daughter, had on 15 April 1452 given birth to an illegitimate son fathered by his own son, Ser Piero. 2 Leonardo's mother was sub- sequently married to a local man the following year, del Verrocchio, versatile sculptor, painter, engineer and master of courtly designs for armour, banners, and so on. Ser Piero's third and fourth wives produced as many as seventeen half-brothers and half-sisters to Leonardo. The catalogue of the Vinci exhibition was mainly writ- ten by Alessandro Vezzosi with Carlo Pedretti, the Italian Leonardo scholar then based at the University on to the front of the picture with a soluble adhesive to secure it -placing the picture face down on a flat surface, and then progressively softening and shaving the wood away from the rear. Once the white gesso ground is reached, this can be left or stripped away to varying degrees. The resulting layer, thin and shiveringly fragile, is then fixed to a canvas using adhesive. To ensure a tight fit, the sandwich is ironed using the kind of old flat-iron once used for laundry. More recently, a vacuum table would have been used to 'suck' the surfaces together. The older use of an iron inevitably squashes the pigment lay- ers, with the characteristic result that any impasto resulting from more thickly applied pigments, such as white highlights, is flattened, to the detriment of the which he described as taking eight months, during which time I passed three months practically without sleep, and for many nights I was not permitted to abandon the work for a minute of slumber.… The workmen are constantly in danger because of their being forced to inhale nitrous gases given off by the material involved. 5 By 1767, the favoured restorer was Jean-Louis Hacquin, working in the Louvre, who assumed responsibility for the wholesale transfer of important paintings in the Royal Collection. The transfer process became the technological miracle of the eighteenth-century restorer's art, though, like any technique of radical intervention (then as now) it attracted committed opposition. The rationale of its proponents was clear. Wood panels were demonstra- bly subject to severe deterioration. They warped and cracked, continually moving with changes in tempera- ture and above all humidity. Joins opened up under the strain, as glues and joints failed. In damp conditions, Bains in Fontainebleau, where humidity fluctuations were likely to have been extreme, with apparently disastrous consequences for the former, which was described as deteriorating in 1624 and is no longer in existence. By contrast, paintings on canvas, a support increas- ingly used during the sixteenth century, presented fewer problems. A canvas that had become slack over the years could be tightened by tapping in the wedges at the corner of the wooden framework over which the canvas was originally bent and tacked. One that was showing signs of failing could be relined with another canvas, attached with ad-hesive and flat-iron. Paintings on canvas could be rolled for involving the gathering together of precious material from various geographical locations involves care- ful diplomacy before any loan requests are made. The next step in developing the exhibition for the National Gallery in Edinburgh is to ascertain if the other frater- nal twin might be available. A close juxtaposition of the two prime versions lies at the centre of the show, and provides the chief rationale for subjecting such treasured items to the hazards of air and surface trans- port. Carlo Pedretti acts as intermediary between me and the owner of the Lansdowne Madonna, which, at that stage, I had seen only in Vinci. It was, as its desig- nation in the Vinci catalogue had suggested, to be seen in New York. on the Gallery's behalf on 31 May 1991, explaining that: 'I do not expect the juxtaposition to show that there is a simple answer to the question as to which is Leonardo's "original" painting (if either). Indeed, I do not think this is quite the right question with respect to Leonardo's smaller devotional images. ' I express the hope that I can make 'the same kind of close study of it as I have made of the Buccleuch version' . I say, I hope reassuringly, that 'from the exami- nation of the Buccleuch Madonna, it is my opinion that it is not wholly autograph, but does involve Leonardo's participation. I have no reason at present for regarding the ex-Reford painting less or more favourably. ' The reassurance seems to do the trick. VIEWING THE LANSDOWNE MADONNA In early December 1991 I am in New York for vari- ous reasons, including a lecture at the Metropolitan Museum, and arrange to see the Lansdowne painting on Monday 9 th . I am granted the opportunity to give it a prolonged inspection, under different lights and using a magnifying glass. It is virtually the same size as the Buccleuch picture and the figures are clearly on the same scale. I look at the back of the picture. If such things as inscriptions, labels and seals survive, the backs of panels and canvases can be eloquent about the construction of the picture, and give valuable infor- mation about the collections and sales it has passed through. Whereas the back of the Buccleuch picture revealed the original panel, peppered with shallow The varnish seems uneven in places. There are also signs of cracking and paint loss that had occurred when the painting was still on its panel. Along with a set of smaller vertical cracks along the grain, par- ticularly near the top margin of the panel, there are clear indications of a more continuous central fissure. It seems likely that the central crack resulted from the panel being made of two separate small planks, but this was not subsequently confirmed, and it resulted has been overlaid by lapis lazuli (ultramarine), a pig- ment of considerable expense that was prized for its scintillating richness. The same quality is apparent in the Buccleuch Madonna. The brown 'greens' are also familiar enough in Renaissance paintings, the result of the instability of the pigments in the copper resinate family. Around the edges of the Lansdowne painting, especially to the left, movement of the frame against the paint surface has resulted in some marginal abrasions, with the white ground showing through. Although the natural edge of the panel has been irredeemably lost, the margin of the surviving paint surface, surrounded by a later fillet of wood, appears to be natural, indicating that the picture has not been trimmed or cut down. To a non-specialist all this may sound like a description of a ruin. But in reality, we are dealing with a picture that is more than half a millennium old, which has, like its contemporaries, somehow survived through hot and cold, humid and dry, been subject to vex- ing transport to far-away countries, and suffered the well-meaning but often ill-considered interventions The Mona Lisa and the Buccleuch painting are in nota- bly good condition for paintings of their age, still on their original panels, with little paint loss and pro- tected under layers of old varnish. The old varnish brings another problem however, that of legibility. transport and storage, and even cut more readily to size when required. The transfer method became fashionable internation- ally as the state-of-the-art in conservation methods for Renaissance panel paintings. There is not a major gal- lery or collection in the world containing Renaissance or later pictures that does not possess its share of transferred paintings. There are tell-tale signs: any pic- ture that exhibits a cracking pattern characteristic of wood but now shows the imprint of a canvas support has been transferred. What we learnt, as we contin- ued to investigate the conservation history (outlined in Chapter 13), was that the Madonna was a late victim of this severe process, the transfer having occurred at some time around 1910, and that it was subsequently re-mounted on board in 1976. Looking at the Lansdowne Madonna in New York, particularly under raking light and with low magnifi- cation, the impress of canvas is all too apparent. Some of the thinner areas of paint, such as the refined glaz- ing in the flesh tones, disconcertingly look as if the pigments had been applied quite sparingly to linen. retouching in the sky and landscape. Retouching pigments invariably deteriorate differently from the originals, becoming more discordant over the years. Some of the thin brown brushstrokes used to give definition to the drawing of the forms in some areas raise suspicions, including a few of the curls in the child's hair and the emphatic outlining of some of the Fig. 14 Leonardo da Vinci, Madonna and Child with a Carnation, c.1475, oil on wood, 62 x 47.5 cm, Munich, Alte Pinakothek. gross is gross is obviously a matter of subjective judgement. We have a notable ability to compensate perceptually for veiling varnish, and good lighting can work wonders, as can skilful photo-graphy. It is now possible to undertake virtual restoration digi- tally, manipulating an image on computer rather than intervening physically on the actual work. When there is even the slightest hint that the Louvre curators might be considering the cleaning of their holiest of holy cows, the press clamour is such as to deter any director who has to work with the political realities of the post. EDINBURGH, VINCI, NEW YORK: LOOKING, PLANNING AND RESEARCHING 43 42 FORENSICS I explain in New York that the infrared examination of the Buccleuch Madonna had produced some very interesting results, and the owner agrees that a comparable technical examination is desirable in the context of research on the exhibition, which is proceeding well. Thereza Crowe and I are developing a list of items for the exhibition, together with draft justifications for each loan to form the basis of requests sent by the Gallery to the various owners, private and public. The items include relevant Leonardo and Leonardo school drawings from the Royal Collection at Windsor, where Jane Roberts is very supportive, and from Venice and Turin. A number of items connected with Robertet are also selected, including a portrait medal (fig. 19), which is the most reliable likeness of Leonardo's distinguished French patron. ). It testifies to Christ's insight. Its vivid portrayal is completely worthy of an artist who spent so much time investigating the vitreous and aqueous humours of this most perfect of human organs. He knew, as had Leon Battista Alberti, the Renaissance author of a book entitled On Painting, that it was the direct axis of our vision -what Alberti called the 'prince of rays'-that 'certifies' sight and gives us the clearest vision of things. There is no doubting the way that the Child's line of sight is focused on the cross, at the point marked by the tip of his index finger, with all the short-range engagement that is typical of a young child who has just learned to locate forms visually in space, striving to coordinate sight and touch. 7 Leonardo, like other Renaissance masters, is not aspiring to portray a 'real' baby at a particular stage in its development, with head, body and limbs naturalistically proportioned both in themselves and with respect to his mother, but he does incorporate unusually sharp observations of infant behaviour in his drawings and paintings of the baby Jesus. The hair of the Child, curling in vortices that swirl like turbulent water (fig. 15), is readily recognisable from his drawings, and was a notable invention of Leonardo's. The pupils and followers imitated the form of such hair but never captured the inner vitality of his vision and touch. Similar qualities can be observed in the rhythmic interplay between the twisted fabric of the Virgin's headdress and the cascading curls of her hair, at first partly veiled and then emerging in full glory (fig. 16). The diaphanous material of her veil undergoes cleverly characterised optical transitions. My inclination at the time of the exhibition is to regard it as 'the work of a close imitator' . I now regard this formulation as too pusillanimous. Executed with great skill in red chalk and metalpoint on pink prepared paper, like the Windsor drawing, but with additional touches of white heightening, it now seems to me to lie beyond any of his assistants. It is not utterly characteristic of his style in all respects, but I am now prepared to allow greater elasticity in how we draw boundaries around what Leonardo might do at any one time. A drawing for a woman's head and shoulders in Turin (fig. 22) poses many of the same problems, but in this case I still do not favour its attribution to Leonardo, and its connection with the composition is less evident. 11 Also in Venice was a drawing related to the Virgin's foreshortened hand, though this time in reverse for a left hand (fig. 23). 12 This motif had first appeared at, say, 30 o to the surface will be a third less bright. 9 This rule is not to be applied literally in every instance, as we now might do using ray tracing on a computer to render a form, but provides a substratum of understanding when Leonardo portrays a subject such as the woman's torso. The level of light from the source is also to be plotted carefully. He noted that faces of women, in particular, gain extra 'grace' in subtle light, like that in his portrait of Cecilia Gallerani. Although outdoors, the light that plays across the features of the Virgin and the Child in the painting is softer and less assertive in the shadows than might be expected in such a setting. He also insisted that the painter observe the way that light rebounds from strongly lit surfaces, casting a reflected glimmer into adjacent shadows. The light that visibly strikes the underside of the Virgin's hand does just this, having bounced off the Child's thigh. Other drawings presented trickier problems of attribution. A very nice drawing in Venice (fig. 21) clearly relates to the preparatory stages for the composition,Fig. 22 Follower of Leonardo da Vinci, Drawing of a woman's head and shoulders, c.1501, red chalk and metalpoint with white highlights on pink prepared paper, 22 x 17.5 cm, Turin, Biblioteca Reale.Fig. 23 Attributed to Cesare da Cesto or Fernando Yáñez de la Almedina, Study for a foreshortened left hand, c.1507-10, red chalk on pink prepared paper, 22.3 x 16.4 cm, Venice, Gallerie dell'Accademia.Fig. 20 Leonardo da Vinci, The Bust of the Madonna, c.1500, red chalk over metalpoint on pink prepared paper, 22.1 × 15.9 cm, The Royal Collection © 2011 Her Majesty Queen Elizabeth II (RL 12514). The atmospheric effects, the fading into soft blue with increasing distance and with hints of mists clustering in the intervening valleys, are just what he recommended in his writings.As the planning of the exhibition proceeds, Thereza Crowe continues to researchRobertet and his background, alongside Leonardo's other French patrons, including successive kings and Charles d'Amboise, who administered Milan on behalf of Louis XII. Leonardo was involved in various jobs for Charles, including the designing of a residence. 21 Robertet himself proved to be a notable collector and builder, aspiring ambitiously to adopt the new Italian Renaissance style, saturated with references to ancient Rome. His father, Jean, resident in Lyons, had been involved in the avant-garde, Italianate literary culture of the city, writing Petrarchian poetry on his own account, and Florimond would have received a Humanist education, rich in classical learning. 22 Florimond was to build a substantial palace-château at Bury, in the countryside near the royal city of Blois, long since demolished but known from drawings and engravings (fig. 31). This is where the Michelangelo bronze David, prised from the hands of the Florentine Republic, eventually found its home, standing proudly on a column at the centre of the main courtyard. It has vanished along with the château. On the other hand, Robertet's Hôtel d'Alluye, his town palace at Blois, has survived. Its modification of the standard French townhouse in the direction of an Italian palazzo testifies to his international taste (figs. 32 and 33). 23 The decorative motifs, including the Roman emperor heads in the courtyard, speak of his ambitions to emulate the all'antica manner that he had admired on his travels with the invading French kings. That Leonardo should have attracted the interest of internationally minded Frenchmen was not surprising. He was still in Milan when Louis invaded in 1499, and he had apparently entered into some kind of arrangement with the King, as Fra Pietro testified. It was said that the King was so impressed by the Last Supper that he instigated a scheme to have it transported to France from its wall in the Milanese refectory of S. Maria delle Grazie, a plan that was not realised. After his return to Florence, Leonardo seemingly intended to remain there, and it is likely that he set aside his commitments to the French invaders. Then five years later, under heavy diplomatic pressure, the Florentines consented to Leonardo leaving Florence for Milan to work for the French rulers, temporarily as they hoped. In the event, although he returned to Florence for two periods of some months during 1506-8, his residence in Florence was effectively coming to an end. Robertet played a conspicuous role in persuading the Florentines to release Leonardo from his obligations, reporting that the King wanted 'our dear and well-loved Leonardo da Vinci painter and engineer of our trust' to serve him in Milan. 24 It was probably Leonardo's return to Milan that enabled Robertet to press successfully for the completion and delivery of his Madonna. In January the Florentine ambassador at Blois, Francesco Pandolfini, reported that 'a little picture by his hand has recently been brought here and is held to be an excellent thing' . 25 Pandolfini tells how the arrival of the picture had again whetted Louis's appetite for Leonardo's work. The King expressed a desire to have 'certain small pictures of Our Lady or other things according to what I might devise in my imagination [fantasia]' . His renewed interest explains Leonardo's reported promise to Charles d'Amboise to bring two paintings with him from Florence: I am sending Salaì to you to explain to your Lordship that I am almost at an end of my litigation with my brothers, and that I expect to find myself with you this Easter, and bring 2 pictures of … Our Lady of different sizes … for our most Christian King or for whomsoever your Lordship pleases. 26 It seems likely that the larger Madonna would have been one of the variants on his favourite themes of the Virgin, Child and St Anne, while the smaller might have been a second version of the Madonna of the Yarnwinder. It might seem odd to us that Leonardo would offer the King a close variant of a paint- October 1991, the Daily Record in Glasgow announces a 'Testing time for the duke' , telling us that 'the richest man in Scotland is at the centre of a £ 50 million art wrangle. For one of the Duke of Buccleuch's treasured paintings is to have rigorous scientific testing. An American collector claims he owns the REAL Leonardo da Vinci masterpiece. ' On 26 October, the Duke is said to be 'awaiting with trepidation a verdict on one of his most prized paintings … by a group of art boffins' assembled by St Andrews University. The following year, Martin Bailey in the Observer reported on 8 March that 'The Duke of Buccleuch has just been told that he is the lucky owner of the original version of Leonardo da Vinci's most popular painting. ' The story quotes the Duke himself, and cites my 'opinion' to the effect that 'evidence suggests that the Duke of Buccleuch owns the original version' . It is unlikely that I said this, because I am arguing against the concept of a single 'original' . The most extensive notice is provided by Richard Cork in the 'Life and Times' supplement to The Times on 24 March. Cork opens with the dramatic idea of a combat between the pictures, with the prospect of a 'clear winner' , but then reviews the complex evidence with considerable acumen. The National Gallery of Scotland's own brochure for the months of April and May announces that 'this summer, visitors to the National Gallery will be able to consider one of the most intriguing Leonardo mysteries: which is the real Madonna of the Yarnwinder ?' The 'focal point of the exhibition' is to be the version 'from the collection of the Duke of Buccleuch' . The brochure for June and July has the Buccleuch Madonna on its front page, and reports inside that 'experts are increasingly confident that a portrait of the Madonna of the Yarnwinder, owned by the Duke of Buccleuch, is the work of Leonardo da Vinci' . We are told that 'the version which resembles the Buccleuch painting most closely has been lent by a private collector in New York' . I do not remember if the brochure copy has passed through my hands. I hope not. Virtually the same wording appears in a Scotsman feature on 17 April 1992.62 FORENSICS COMMERCE AND CLAMOUR 63 Leonardo would have recognised the kind of things that were going on. Leonardo was a hired hand, and readily accepted his role as a stipendiato of Ludovico's or Louis's court. He had himself invented diverting things for smart dinners, such as a trick that turned white wine into red -whether still or spumante is not recorded. Robertet knew that art blended enjoyment, money and power. The social contexts for the leading players are now very different, but the scripts remain recognisable. Speeches are made, and gratitude is expressed. The Chairman of the Trustees is flying back from New York overnight and cannot attend, so the Director is briefed to thank the right people, including British Airways, who have provided support in kind for the transporting of people and things. Remaining pangs of conscience are lubricated by good wine, neither transformed from white to red nor from water. The presence of Martini Brut is evident, obviously -as it is at the private views set up for various groups and privileged categories of individuals in the normal way. Then, early in October 2007, a detective constable from the Dumfries and Galloway Constabulary calls the Department in Oxford asking if I will contact him. Even before calling back, I feel a rising sense of optimism. The police indicate that the painting may have been recovered, and ask for confirmation about tests that would confirm that any recovered painting is indeed the one that was stolen and not a clever replacement. I explain that the unpublished IR reflectograms would provide about as good forensic evidence as they might wish. No forger would know about the underdrawing in the detail that we did. They ask if I will travel to Edinburgh to look at the picture they have in their hands. I am very sorry to say that I cannot go on the date they suggest, which falls during the frantic first days of term. It also coincides with the run-up to the opening of the exhibition Seduced. Art and Sex from Antiquity to Now that Marina Wallace and I are curating, with Joanne Bernstein, at the Barbican Art Gallery. Michael Clarke, Director of the National Gallery in Edinburgh, looks at it in my stead. He knows his way around old master paintings and can of course make an informed judgement. I arrange to visit Edinburgh on 24 October, which is my first free day. 1 .1PROLOGUE1 Edoardo Villata, ed., Leonardo da Vinci. I documenti e le testimonianze contemporanee, Ente RaccoltaVinciana, Castello Sforzesco, Milan, 1999, p. 208, no. 240. 2. Villata, 1999, pp. 287-8, no. 333. 3. Bruno Latour and Steve Woolgar, Labo- ratory Life: The Construction of Scientific Facts, Princeton, 1986. 2. THE STORY BEGINS: THE SCOTTISH PICTURE 1. Martin Kemp, 'The Hunterian Chardins X-rayed' , The Burlington Magazine, cxviii, 1976, pp. 144-8. 2. Martin Kemp, ' A Date for Chardin's Lady Taking Tea' , The Burlington Magazine, cxx, 1978, pp. 22-5. 3. Wilhelm Suida, 'Leonardos Madonna mit dem Kreuzstab (oder dem Garnwinder)' , in Miscellanea di studi lombardi in onore di Ettore Verga, Milan, 1931, pp. 333-9. 4. Wilhelm Suida, 'La Scuola di Leonardo' , in Leonardo da Vinci (1939), ed. Costantino Baroni and Sandro Piantanida, London, 1967, p. 322. 5. This and other citations of John Dick are taken from his emailed response to the first draft of the present text. 6. Martin Kemp, 'The Madonna of the Yarn-winder in the Buccleuch Collection Reconsidered in the Context of Leonardo's Studio Practice' , in Il Leonardeschi a Milano: fortuna e collezionismo, Acts of an international symposium held at Milan (25-26 September 1990), eds. Pietro Marani and Maria Thereza Fiorio, Milan, 1991, pp. 35-48. 3. ISABELLA'S AGENT 1. Edoardo Villata, ed., Leonardo da Vinci: I documenti e le testimonianze contemporanee, Ente Raccolta Vinciana, Milan, 1999, p. 113, no. 129. 2. Villata, 1999, p. 114, no. 130. 3. François Viatte and Varena Forcione, eds., Léonard de Vinci: Dessins et manuscrits, exhibition catalogue, Musée du Louvre, Paris, 2003, pp. 185-9, no. 61. 4. The status of the Ashmolean drawing is reviewed by Martin Kemp and Juliana Barone, eds. Leonardo da Vinci: I disegni di Leonardo da Vinci e della sua cerchia, Collezione in Gran Bretagna, Florence, 2010, no. 115. 5. Villata, 1999, p. 131, no. 144. 6. Ibid., p 134, no. 149. 7. For the Isabella-Perugino correspondence, see David Chambers, Patrons and Artists in the Italian Renaissance,London, 1970, p. 139. 8. Ibid., p. 131. 9. Villata, 1999, pp. 134-5, no. 150. 10. Ibid., p. 136, no. 136. 11. Alessandro Vezzosi, et al., Leonardo dopo Milano: La Madonna dei Fusi, exhibition catalogue, Vinci, Castello, 1982, for a selection of copies; and Martin Kemp and Thereza Crowe, Leonardo da Vinci: The Mystery of the Madonna of the Yarnwinder, exhibition catalogue, National Gallery ofScotland, Edinburgh, 1992, pp. 46-7, no. 5. 12. For Robertet as a patron, see The Mystery, 1992 and Claude Albert Mayer and Dana Bentley-Cranch, 'Florimond Robertet: Italianisme et Renaissance française' , in Melanges à la mémoire de Franco Simone (IV, Tradition et originalité dans la création littéraire), Geneva, 1983, pp. 140-59; Dana Bentley-Cranch, ' A Sixteenth-Century Patron of the Arts, Florimond Robertet, Baron Alluye and his "Vierge Ouvrant"' , Biblothèque d'humanisme et renaissance, xlix, 1988, pp. 317-33; and Florimond Robertet (? -1527): Homme d'Etat Français, Paris, 1994; for the lost Michelangelo David, see Gert Rudolph Flick, Missing Master-pieces: Lost Works ofArt, 1450- 1900, London, 2003. AnnThomas, The Painter's Practice in Renaissance Tuscany, Cambridge, 1995. EDINBURGH, VINCI, NEW YORK: LOOKING, PLANNING AND RESEARCHING 1. For an ingenious but inaccurate reconstruction of the baby-walker with a round frame and 'ball-bearing' wheels, seeAlessandro Vezzosi, ed., 'Parleransi li omini…' . Harry Hahn, The Rape of La Belle, Kansas City, 1946, pp. 170-3. 6. Cornelia Syre, Jan Smidt and Heike Stege, Leonardo da Vinci Die Madonna mit der Nelke, exhibition catalogue, Alte Pinakothek, Munich, 2006. 7. See Larry Feinberg, 'Sight Unseen: Vision and Perception in Leonardo's Madonnas' and 'Visual Puns and Variable Perception: Leonardo's Madonna of the Yarnwinder' , Apollo, July and August 2004, pp. 28-34 and pp. 38-41. 8. Kenneth Clark, A Catalogue of the Draw-ings of Leonardo da Vinci in the collection of His Majesty the King at Windsor Castle, Cambridge, 1935; 2 nd rev. ed., with Carlo Pedretti, London, 1968-9, no. 12514. 9. Clark and Pedretti, 1968, no. 12604r. 10. Luisa Cogliati Arano, Disegni di Leonardo da Vinci e della sua cerchia alle Gallerie dell'Accademia, Venice, 1996, no. 42, attributed to Cesare da Cesto. 11. Carlo Pedretti, Disegni di Leonardo da Vinci e della sua scuola alla Biblioteca Reale di Torino, Florence, 1975, no. 17, attributed to a pupil, probably Cesare da Cesto; and Martin Kemp and Thereza Crowe, Leonardo da Vinci: The Mystery of the Madonna of the Yarnwinder, exhibition catalogue, National Gallery ofScotland, Edinburgh, 1992, pp. 72-3, no. 18. 12. Cogliati Arano, 1996 and The Mystery, 1992, pp. 78-9, no. 21, attributed to Fernando Yáñez. 13. For some of the many versions and variants of this popular composition, which must have reached at least the stage of a highly developed design, see Nicoletta Baldini in Gigetta dalli Regoli, Romano Natali and Antonio Natali, Leonardo e il mito diLeda, exhibition catalogue, Vinci, 2001, pp. 122-5. The most authoritative versions are those owned by the Galleria Palatina (Pitti deposit), and the Ashmolean Museum in Oxford (although badly damaged).14. Arthur Popham, The Drawings of Leonardo da Vinci, ed. MartinKemp, London, 1994, nos. 8b-14. 15. Martin Kemp and Margaret Walker, Leonardo on Painting: An anthology of writings by Leonardo da Vinci with a selection of documents relating to his career as an artist, ed. Martin. 333. 20. David Allan Brown, Madonna Litta, xxix Lettura Vinciana (Vinci, 1989), Florence, 1990. NOTES TO PART I 122 FORENSICS 4. Leonardo e l'Europa, dal disegno delle idee alla profezia telematica, with Carlo Pedretti, exhibition catalogue, Museo Ideale, Vinci, 2000, p. 86. 2. Edoardo Villata, ed., Leonardo da Vinci: I documenti e le testimonianze contempora- nee, Ente Raccolta Vinciana, Milan, 1999, p. 3, no. 1. 3. Villata, 1999, p. 6, no. 2. 4. Martin Kemp, 'Vinci: Leonardo dopo Milano' , The Burlington Magazine, cxxiv, no. 957, 1982, p. 788. 5. . 17.19. Especially ibid., nos. 12405-8 and 12410, 12411, 12413 and 12414, dateable across the range c.1500-1513 and ibid., pp. 68-9, no. 11. 20. Carlo Starnazzi, Leonardo ad Arezzo A.D. 2000: La 'Madonna dei fusi' di Leonardo da Vinci e il paesaggio del Valdarno Superiore, exhibition catalogue,Arezzo, 2000, figs. 26-30. 21. Villata, 1999, p. 204, no. 237. 22. Jean Robertet, Oeuvres, ed. Margaret Zsuppán, Geneva, 1970 and Claude Albert Mayer and Dana Bentley-Cranch, 'François Robertet: French sixteenth-century civil servant, poet, and artist' , Renaissance Studies, xi, 1997, pp. 208-22. 23. Martin Kemp, 'From Scientific Examination to the Renaissance Market: The Case of Leonardo da Vinci's Madonna of the Yarnwinder, Journal of Medieval and Renaissance , no. 619. 27. Henri Chesneau, Bury Rostaing, 1650; see 'Inventaire des objets d'art composant la succession de Florimond Robertet, Ministre de François I, dressé par sa veuve, le 4 jour d'Août 1532' , Mémoires de la Société Impériale des Antiquaires de France, 3 rd series, x, 1868, pp. 1-66. 29. The Mystery, 1992, nos. 22-8. 5. LATE SURPRISES FROM NEW YORK 1. Kenneth Clark, A Catalogue of the Drawings of Leonardo da Vinci in the collection of His Majesty the King at Windsor Castle, Cambridge, 1935; 2 nd rev. ed., with Carlo Pedretti, London, 1968-9. 2. Kenneth Clark, Leonardo da Vinci: An Account of his Development as an Artist (1939), ed. Martin Kemp, London, 1989. 3. The New York World, 18 June 1920. NOTES TO PART I 123 4. John Brewer, The American Leonardo: A 20 th Century Tale of Obsession, Art and Money, London, 2009. 5. Harry Hahn, The Rape of La Belle, Kansas City, 1946, p. 74. Edward Gage, Scotsman, 1 June 1992. 2. Martin Gayford, The Telegraph, 2 June 1992. 3. James Hall, The Independent, 23 June 1992. 4. John McEwen, Sunday Telegraph, 31 May 1992. 5. Among the specialist reviews were: Michael Bury, 'Leonardo/non-Leonardo: The Madonna of the Yarnwinder' , Apollo, cxxxvi, No. 367, 1992, pp. 187-9; Cecil Gould, 'Leonardo's Madonna of the Yarnwinder: Revelations of reflectogram photography' , Apollo, cxxxvi, no. 365, 1992, pp. 12-16; Carlo Pedretti, 'The Mysteries of a Leonardo Madonna, Mostly Unsolved' , in Achademia Leonardi Vinci: Journal of Leonardo Studies & Bibliography ofVinciana, vol. 5, 1992b, pp. 169-75, and Il Giornale dell'Arte, no. 101, June 1992a, p. 18; and Nicholas Penny, 'Leonardo's Madonna of the Yarnwinder: Edinburgh, National Gallery of Scotland' , The Burlington Magazine, cxxxvi, no. 1073. Carlo Pedretti, 'Il tema del profilo, o quasi' , in Il Leonardeschi a Milano: fortuna e collezionismo, Acts of an international symposium held at Milan(25-26 Septem- ber 1990), eds. PietroMarani and Maria Thereza Fiorio, Milan, 1991, pp. 14-24; and Paul Joannides, 'Creative Distortion in the Renaissance: Lippi, Leonardo and Parmigianino' , Apollo, cxxxxvi, 1992, pp. 239-46. 7. Cecil Gould, Leonardo: The Artist and Non-Artist, 28. Dana Bentley-Cranch, Florimond Robertet (? -1527): Homme d'Etat Français, Paris, 1994. 6. Important Old Master Paintings and Sculpture, Sotheby's, New York, 28 January 2010, lot 181. 7. Martin Kemp and Thereza Crowe, Leon- ardo da Vinci: The Mystery of the Madonna of the Yarnwinder, exhibition catalogue, National Gallery of Scotland, Edinburgh, 1992, pp. 22-3. 6. COMMERCE AND CLAMOUR 1. 1 .1Kenneth Clark, A Catalogue of the Drawings of Leonardo da Vinci in the collection of His Majesty the King at Windsor Castle, Cambridge, 1935; 2 nd rev. ed., with Carlo Pedretti, London, 1968-9, no. 12579. 2. Martin Kemp and Thereza Crowe, Leonardo da Vinci: The Mystery of the Madonna of the Yarnwinder, exhibition catalogue, National Gallery of Scotland, Edinburgh, 1992, p. 46, no. 5. 3. Martin Kemp, 'From Scientific Examination to the Renaissance Market: The Case of Leonardo da Vinci's Madonnna of the Yarnwinder' , Journal of Medieval and Renaissance Studies, xxiv, no. 2, 1994, pp. 260-1. 8. ICONOGRAPHIES 1. For the Bigallo Madonna, see Hanna Kiel, Il Museo del Bigallo, Florence, 1977, pp. 4-5; and Luigi Passerini, Curiosità storico-artistiche fiorentine: la Loggietta del Bigallo, Florence, 1866, pp. 91-8. 2. Edoardo Villata, ed., Leonardo da Vinci: I documenti e le testimonianze contemporanee, Ente Raccolta Vinciana, Milan, 1999, p. 135, no. 150. 3. Michael Bury, 'Leonardo/non-Leonardo: The Madonna of the Yarnwinder' ,Apollo, cxxxvi, No. 367, 1992, pp. 269-70. 4. Carmen Bambach, ed., Leonardo da Vinci: Master Draftsman, exhibition catalogue, Metropolitan Museum ofArt, New York, 2003, pp. 366-70, no. 45; and Martin Kemp, ibid., pp. 145-6; for the underdrawing of the National Gallery picture, see Luke Syson and Rachel Billinge, 'Leonardo da Vinci's use of underdrawing in the Virgin of the Rocks in the National Gallery and St. Jerome in the Vatican' , The BurlingtonMagazine, cxlvii, no. 1228Magazine, cxlvii, no. , 2005. The infrared scanner employed is described in L.Pezzati, M. Materazzi and P. Poggi, 'Infrared Reflectography and the INOA High Resolution Scanner' , in C. B. Strehlke and C. Frosinini, eds., The Panel Paintings of Masolino and Masaccio. The Role of Technique, Milan 2002, p. 267. The IR scanning work in London was undertaken by C. Frosinini, R. Bellucci and P. Poggi (all of the OPD in Florence), attended by R. Billinge, A. Roy and L. Keith from the National Gallery. 6. Codex Leicester 34r (translation by Domenico Laurenza and Martin Kemp for a forthcoming internet edition). 7. Leo Steinberg, The Sexuality of Christ in Renaissance Art and in Modern Oblivion, London, 1984. 8. See Steinberg, 1984, pp. 174-83, for examples of the later concealment of penises, and, sometimes, their reinstatement in modern restorations. 9. THE UNIVERSAL LEONARDO AND THE LABORATORY 1. For examinations taken under the auspices of the Opificio delle Pietre Dure in Florence, discussed in this section, see Roberto Bellucci and Cecilia Frosinini, 'The Madonna of the Yarnwinder: a case study of non-invasive technique for art historians and conservators'; R. Fontana, et al., '2d and 3d optical diagnostic techniques applied to Madonna dei Fusi by Leonardo da Vinci'; A. Casini, et al., 'Fiber optic reflectance spectroscopy and hyper-spectral image spectroscopy: two integrated techniques for the study of the Madonna dei Fusi'; N. Grassi, 'Composition and stratigraphy of the paint layers: investigation on the Madonna dei Fusi by ion beam analysis techniques'; and M. Callieri, et al., 'Tools for inverse mapping and visualization of multi-spectral image data on 3d scanned representations of drawings' , all in Optical Methods for Arts and Archaeology, Proceedings of SPIE Volume, eds. Renzo Salimbeni and Luca Pezzati, Munich, 2005. 2. Martin Kemp and Thereza Crowe, Leonardo da Vinci: The Mystery of the Madonna of the Yarnwinder, exhibition catalogue, National Gallery of Scotland, Edinburgh, 1992, pp. 78-9. 3. For the Munich Madonna, see Cornelia Syre, Jan Smidt and Heike Stege, Leonardo da Vinci Die Madonna mit der Nelke, exhibition catalogue, Alte Pinakothek, Munich, 2006; and for the Mona Lisa, see Laurence de Viguerie and Phillipe Walter, et al., 'Revealing the sfumato Technique of Leonardo da Vinci by X-Ray Fluorescence Spectroscopy' , Angewandte Chemie, International Edition, 49, 2010, pp. 6125-8. 4. Madeleine Hours, 'Etudes analytiques Tableaux de Léonard de Vinci au Laboratoire du Musée du Louvre' , in Leonardo: Saggi e ricerche, ed. Giorgio Castelfranco, Rome, 1954, pp. 13-26; and Kazimierz Kwiatkowski, La Dame à l'hermine de Léonard da Vinci: Etude technique, Warsaw, 1955. 10. DAYLIGHT ROBBERY AND MORE FORENSIC WORK 1. Martin Kemp, Leonardo, Oxford, 2004. 2. Martin Kemp and Pascal Cotte, La Bella Principessa. The Story of the New Masterpiece by Leonardo da Vinci, London, 2010. 3. Godfrey Barker, 'Rich Pickings' , Sunday Times, Focus, 31 August 2003, p. 19. 4. Codex Urbinas, 148v-9r; Martin Kemp and Margaret Walker, Leonardo on Painting: An anthology of writings by Leonardo da Vinci with a selection of documents relating to his career as an artist, ed. MartinKemp, New Haven, 1989, p. 76, no. 187. For other such effects, seeMartin Kemp, 'In the Beholder's Eye: Leonardo and the "Errors of Sight" in Theory and Practice' (Hammer Prize Lecture), Achademia LeonardiVinci, v, 1992, pp. 153-62. 126 MADONNA OF THE YARNWINDER: A HISTORICAL & SCIENTIFIC DETECTIVE STORY PART II CONSERVATION AND ATTRIBUTION THE LANSDOWNE AND BUCCLEUCH MADONNAS: CONSERVATION AND CONNOISSEURSHIP 127 ). Not only have articles written many years ago produced13. THE LANSDOWNE AND BUCCLEUCHMADONNAS: CONSERVATION AND CONNOISSEURSHIP Previous page: Infrared reflectogram detail of the Madonna of the Yarnwinder (The Lansdowne Madonna), plate 1. .One of the contributors to the 1898 exhibition was the Italian art historian Dr Gustave Frizzoni, who also wrote a review of the show in which he specifically discussed the question of attribution of the two ers of Leonardo. Cook believed the two paintings were by two different hands but shared a common source, such as a drawing by Leonardo. He then went on to suggest the painting was by Salaì based on its resemblance to the Virgin of the Balances, then attributed to Salaì, at the Louvre. The confusing result is that the public was faced with two paintings, three scholars and three different attributions. now houses Harvard's Center for Italian Renaissance Studies and attracts scholars from all over the world.Madonnas. 3 The short catalogue entries by Herbert Cook for the two paintings also dwell on the issue of attribution; the Buccleuch Madonna entry notes that Frizzoni attributed the painting to Sodoma and cited a painting with a similar style by that art- ist found at the Brera Gallery in Milan. On the other hand, the great Italian art historian Adolfo Venturi resisted the Sodoma claim. The Lansdowne entry by Cook describes the subject of the Madonna of the Yarnwinder as popular among the immediate follow- Ten years later, in February 1908, the Lansdowne Madonna was purchased from Constance Rothschild, Lady Battersea (1843-1931), the widow of Lord Battersea (who had died the previous year), by the Paris-based dealers Nathan Wildenstein (1852-1934) and René Gimpel (1881-1945), who will both be discussed later. They were keen to settle the issue of attribution; after all, a painting by Sodoma and a painting by Leonardo carried two very different price tags. They contacted the eminent art historian Bernard Berenson to ask his opin- ion. He wrote to Gimpel on 27 October 1909: I have looked carefully at the 'Madonna and Child' which you ascribe to Sodoma. I entirely agree with your attribution, but I seriously believe that Sodoma only executed your exquisitely beautiful picture. I think that it was designed, and probably that its actu- al cartoon was probably made by Leonardo da Vinci himself. Indeed, excepting the four or five pictures that Leonardo himself painted with his own hands, I know none that approaches the great master so closely as your Madonna does. 4 It seems that Wildenstein and Gimpel believed the Lansdowne Madonna was executed by both Leonardo and Sodoma but Berenson was not convinced that Leonardo had played a part other than to have been the original inspiration. THE BERENSONS, DUVEEN AND THE BUCCLEUCH MADONNA Berenson's opinion on the Lansdowne Madonna was to be taken seriously -he was recognised as the lead- ing authority on Italian Renaissance painting. Born in Lithuania, but brought up in Boston, Berenson (fig. 62) went on to study at Harvard. His wife, Mary, whom he married in 1900, was also an art historian in her own right. In 1894 she had published a guide to Italian paintings at Hampton Court Palace, which established her name as an authority. She contributed much to Berenson's own work with writing and editing as early as his 1894 publication Venetian Painters. With Mary's help, Berenson published a succession of books and monographs on Italian art. They lived at Villa I Tatti outside Florence, which Berenson had bought the year they married. Mary died in 1944 and when Berenson died in 1959, he left I Tatti to Harvard University. It 134 CONSERVATION AND ATTRIBUTION The Berensons' work extended beyond libraries and museums. As noted earlier, he was in touch with the dealer Wildenstein. He was also deeply involved with other dealers, most notably the Duveen brothers. The Duveen family were originally from Holland and had a history that included collectors as well as deal- ers in furniture, porcelain, tapestries and Old Master paintings. 5 In 1879 the firm of Duveen Brothers was established with Joseph Joel and Henry presiding. Both the London and New York branches thrived, with the brothers forging close relationships with some of the wealthiest clients in the business, including the future King George V, who would later knight Joseph Joel. Joseph Joel's son, Joseph, became involved at the early age of seventeen when he travelled to New York to learn from his uncle's business. It was Joseph who saw the opportunities in moving on to painting and Fig. 62 Berenson standing in front of Domenico Veneziano's Madonna at Villa I Tatti, 1903. The Berenson Archive, The Harvard University Center for Italian Renaissance Studies, Villa I Tatti, courtesy of the President and Fellows of Harvard College. The Duveen firm was no stranger to scandal. In 1910 the New York Gallery was accused of false declarations on the import of objects from Europe and was forced to pay the highest settlement fine up to that time. InBerenson an offer to become his paid adviser on Italianpictures. Duveen would award him an annual retaining fee and a commission on sales. He regularly sent photographs of works to Berenson who noted his attribution on the photos and signed them. Duveen kept get a copy of this photograph and let me have it for careful study. You would be doing me a great favour if you could manage this and I shall not fail to let you know my conclusions about it. I have not myself seen the picture for over twenty years. 8 ment between Duveen and Berenson: if Berenson did not publicly attribute the painting to Leonardo, then it could be had for a low sum, and a subsequent sale of the painting as a Leonardo would ensure an enormous profit for both the dealer and for Berenson. The other It is not known whether Duveen ever made an offer for the Leonardo that he had appeared so keen to secure. Perhaps ultimately the Duke could not bear to be parted from it. By good fortune it has remained in Buccleuch hands to this day, with two exceptions. trade, but it appears that it took some time for him to acknowledge this. The dealer Gimpel recalls in his diaries an exchange between Berenson and the antique When I saw it in 1911 [in Paris] at Wildenstein's, it had been relined, cleaned, restored and altered in several details. The loin cloth had been removed, the restored two fingers of the Madonna's left hand THE LANSDOWNE AND BUCCLEUCH MADONNAS: CONSERVATION AND CONNOISSEURSHIP 139 It shows that Reford was active in supporting research into his paintings, agreeing to send the work to Vienna for the examinations. They revealed that the underdrawing is, in some respects, different from the final composition. The examinations were reported by Suida in a 1931 publication.1921 Joseph became embroiled in a legal battle involv- ing the Hahn La Belle Ferronnière discussed in Chapter 5 (fig. 64). Among the experts called in by Duveen was Berenson, whose connoisseur's testimony unravelled under cross-examination. Eventually Duveen settled out of court for $ 60,000. Duveen's reputation had been avidly examined by the press and the case had been a platform on which to scrutinise the practice and legiti- macy of the connoisseur. Duveen and Berenson's relationship was close. 6 They first met in Paris in 1906 and Duveen soon made these 'attestation' photographs and presented them to his clients as certificates of authenticity. If Duveen was successful in selling the work, Berenson received twenty-five per cent of the sale price. Although Berenson would sometimes authenticate paintings for other dealers, this practice irritated Duveen, and Berenson's main business was with him. Their close relationship, private and professional, lasted until the early 1930s when they fell out over the attribution of a painting that Duveen was about to sell to the eminent American collector Andrew Mellon as a Giorgione, but which Berenson believed to be by Titian. When Berenson refused to change his attri- bution to accommodate Duveen, their relationship was severed. Fig. 64 The Leonardo da Vinci Dispute 'La Belle Ferronnière Case'-Joseph Duveen 2 March 1929, © Illustrated London News Ltd/Mary Evans. 136 CONSERVATION AND ATTRIBUTION Before the ties were broken, however, Berenson and Duveen became involved with the Duke of Buccleuch and his Madonna of the Yarnwinder during a period that spanned ten years. The connection was initiated by Berenson in 1926 after he had read Emil Möller's article on the Madonna of the Yarnwinder in The Burlington Magazine, where Möller accorded a Leonardo attribu- tion to the Buccleuch painting (as discussed below). 7 Berenson wrote to Duveen on 10 August explain- ing that he had read the article and wished to see the photograph that had been taken of it in 1914. He went on, As the firm has been in relation with his Grace, per- haps you could, without exciting too much comment, A month later, on 10 September 1926, a short Duveen memorandum describes a visit by Joseph Duveen, Mary Berenson, and Arthur Ruck -a London dealer and Duveen associate -to the Duke of Buccleuch's home in London, where they examined the paintings in the collection. The memo mentions only one detail: 'Mrs Berenson does not think the Leonardo is right. ' 9 The following day, Mary Berenson wrote to Duveen and repeated that although she did not think the paint- ing was by Leonardo, and believed her husband would agree, they still wished to see photographs and could Duveen send them. Although Mary Berenson did not have faith in the painting it seems that Duveen saw something. On 21 September a Duveen memo states that Joseph Duveen was to visit the Duke of Buccleuch to see the painting. Some months passed without any further develop- ments, but it appears that Duveen's interest in the Buccleuch Madonna had not diminished. A cable let- ter between the Duveen brothers dated 18 January 1927 describes how Ruck, who acted for Duveen, had been in negotiations with the Duke of Buccleuch regarding the Madonna as well as his Rembrandt: Ruck says if you would let him know most you would give for Rembrandt and Leonardo da Vinci he would make deal with two pictures and furniture together, of course not settling anything without obtaining your instructions. 10 On 9 February 1927 a coded cable was sent. Using coded names for clients, paintings and artists was a Duveen practice when confidentiality was required. Verging on a schoolboy plot, the following cable reveals their nefarious intentions: Have you photograph Cleughbuc [Buccleuch] Greck- gem [Leonardo] stop Seems to me both Dorises [Ber- ensons] believe this topaz [painting] right therefore I am keen on Ernest personally following this without delay stop Generally speaking when Dorises ask pho- tographs they feel pretty sure and personally I believe topaz right. 11 The following day Duveen was keen to secure the Madonna from Buccleuch. A memo from that day explains that Ruck was planning to visit the Duke to ask for a price for both the Rembrandt and the Leonardo before asking for photos of the Leonardo, 'so as not to let him think about the Leonardo in particular'. 12 Things become curious when a Duveen cable dated THE LANSDOWNE AND BUCCLEUCH MADONNAS: CONSERVATION AND CONNOISSEURSHIP 137 11 February reveals a completely different Berenson opinion from that in the 9 February coded cable: 'Regarding Buccleuch Leonardo da Vinci, Berenson asked for photograph for his record. He says Leonardo da Vinci composition but both Berensons are con- vinced picture is not painted by master himself. ' Another cable concerning the same planned visit to Buccleuch and dated 18 February concludes with, 'I see that Berenson still persists that the picture is not by the master. ' 13 Why does the Duveen correspondence record two different attributions by Berenson in such a short space of time? That Duveen continued to pursue the Madonna from Buccleuch seems to indicate two possibilities. The first is that there was a secret agree- possibility is that Berenson had changed his mind and concluded the painting was not by Leonardo, and that Duveen did not agree and continued to pursue it. In any event, the purchase of neither the 'Leonardo' nor the Rembrandt went ahead. Several years passed before the Buccleuch Madonna appearsed once again in the Duveen correspond- ence. In April 1932, a cable from London to New York describes a meeting with a Mrs Blois, a well-connected lady who advised Duveen that Buccleuch was looking to sell his Holbein and his Leonardo for £ 150,000 each, adding that the Duke was in need of some money. When the painting was sent to the Leonardo exhi- bition in Milan in 1938, it became caught up in the events of World War II. The family did not know of its whereabouts until it reappeared in 1946 when the Italian Embassy in London made contact to let them know that the painting was with them. The second time the painting was away was during the period it was held by thieves from 2003 until 2007. By 1932 Berenson and Duveen's relationship had been severed. Berenson's reputation in the academic world suffered because of his association with the art dealer Bauer: Once, in the days when he [Berenson] was not so pow- er-ful, he said to Bauer 'A man as scholarly as your- self shouldn't be a dealer, it's horrible to be a dealer. ' To which Bauer replied: 'Between you and me there's no great difference; I'm an intellectual dealer and you're a dealing intellectual. ' Berenson never forgave him that. 14 Berenson himself eventually regretted the work he did for dealers, writing that, 'I took the wrong turn when I swerved from more purely intellectual pursuits to one like the archaeological study of art, gaining thereby a troublesome reputation as an "expert". My only excuse is … I too needed a means of livelihood. … The spiritual loss was great and in consequence I have never regarded myself as other than a failure. ' 15 138 CONSERVATION AND ATTRIBUTION WILDENSTEIN'S CONSERVATION OF THE LANSDOWNE MADONNA In comparison with the Buccleuch painting, the Lans-downe Madonna's relationship with the art trade has been considerably more intimate and pro- longed. As we saw earlier, it became the property of Wildenstein and Gimpel, and it was in their hands that the dramatic transfer of the painting from panel to canvas took place. This practice, now rarely car- ried out, was still quite common and was considered a suitable precaution to stabilise the paint layers. The results of the transfer were described by the Leonardo scholar Emil Möller in the 1926 Burlington Magazine art-icle, which Berenson had read: had been cleaned off, the folds smoothed out, the veil on the bosom altered, and the piece of rock under the Child's foot made visible. The total impres- sion was pleasant but overtrim, so that this impor- tant copy had largely been deprived of its original value. I am indebted to M. Wildenstein for kind information and for a photo of the picture in its pre- sent state. 16 Although Möller thanked Wildenstein for a pho- tograph of the painting, the image reproduced in the article was taken before the transfer and the dramatic changes described by the scholar cannot yet be discerned. There are no known records detail- ing the transfer process that took place in Paris under the Wildensteins. ATTRIBUTION AGAIN: THE LANSDOWNE MADONNA AS A COPY Möller's article was the first detailed comparative study of the two Madonnas. In it he listed the numerous scholars who had given their judgements, the major- ity of whom had been persuaded by the Lansdowne picture. Möller's own opinion differed, however. He believed the Lansdowne Madonna to be a copy of the Buccleuch painting, executed by a pupil, citing the ' disturbed forms' in the landscape and 'decidedly vulgar' Madonna. He went on to describe clumsy limbs and a 'heavy and uncertain' design. Finally he com- pared the area of crumbling stones below the base of the yarnwinder in both versions, observing that the passage in the Buccleuch painting was far superior in its detailed execution. 17 In discussing the Buccleuch painting in detail, Möller could readily see the landscape was inferior to the rest of the composition and not by Leonardo. He believed incorrectly that beneath the landscape was another landscape, like the one in the Lansdowne painting, but better. He backed this theory by citing other early versions of the painting which had been painted with the Lansdowne landscape and then a later version, painted in the early eighteenth cen- tury with a Buccleuch landscape. He concluded that the Buccleuch landscape had therefore been painted over around the same time as the early eight- eenth-century copy. Möller's theory is ingenious, but subsequent analysis of the Buccleuch painting shows no signs of a Lansdowne-like mountainous landscape. In his closing remark, Möller refers to the Buccleuch painting as having recently been 'cleaned from repaint- ings' . Comparing the reproductions from the 1898 exhibition, the 1931 image and today (figs. 52, 65 and plate 2), there appears to be very little if any change in the state of the painting, so any conservation work seems to have been conservative. In the 1920s the Lansdowne painting was taken to Wilden-stein's in New York and in May 1928 Wildenstein sold it as a Sodoma to Robert Reford, a Canadian industrialist and shipping tycoon. The first visual evidence of the painting after the transfer appears in Hans Bodmer's 1931 publication, Leonardo (figs. 55 and 66). 18 Employing a game of spot the difference, one can see a number of changes. Those described by Möller are clearly apparent. The most noticeable are that the drapery over Christ's lower Fig. 65 1931 reproduction of Leonardo da Vinci, Madonna of the Yarnwinder (The Buccleuch Madonna). Reproduced in Giorgio Nicodemi, ed., Miscellanea di studi lombardi in onore di Ettore Verga, Milan, 1931. Fig. 66 c.1931 reproduction of Leonardo da Vinci, Madonna of the Yarnwinder (The Lansdowne Madonna). Reproduced in Bodmer, Leonardo, 1931. Image from the Harvard University Center for Italian Renaissance Studies, Villa I Tatti, courtesy of the President and Fellows of Harvard College. Fig. 67 1931 ultraviolet image of Leonardo da Vinci, Madonna of the Yarnwinder (The Lansdowne Madonna). Reproduced in Giorgio Nicodemi, ed., Miscellanea di studi lombardi in onore di Ettore Verga, Milan, 1931. THE LANSDOWNE AND BUCCLEUCH MADONNAS: CONSERVATION AND CONNOISSEURSHIP 141 abdomen has been largely or completely removed, so that his genitals are visible, and lower parts of the fingers of the Virgin's left hand have been taken away. However, the veil around the Virgin's neckline appears not to have been substantially altered. A LEONARDO UNDER THE SKIN? In about 1930, ten years after the Lansdowne Madonna's transfer from panel to canvas and two years after it was bought by Robert Reford, the painting underwent x-ray and ultraviolet examination under the guidance of Wilhelm Suida, Professor Dr Robert Bigenberger, and the artist R. Maurer (figs. 1, 56 and 67). This relatively new form of examination enabled researchers to see beneath the surface of the painting for the first time. It is clear that Reford was active in engaging the interest of scholars in his painting. He had a close rela-On the first visit he[Valentiner] was obviously greatly impressed with its beauty and remarked that it was finer than the other picture said to be identical in the collection of the Duke of Buccleuch. He observed it tion work, one of which is recorded in the Reford letter. Whichever is the case, the painting had again been subject to considerable intervention.Reford's letter and to which the Duke of Buccleuch agreed to loan his painting, was also an opportunity for the Buccleuch Madonna to be examined. The exhibition, which took place at the Palazzo dell'Arte, was a major Leonardo show. As described and reproduced such as the Virgin of the Rocks and Adoration of the Magi. Its presence over the surface of a support makes it difficult to read x-rays, as lead white obstructs the rays. Unwisely, Marchig dissolved the white lead layer during his restoration work. Marchig noted that THE LANSDOWNE AND BUCCLEUCH MADONNAS: CONSERVATION AND CONNOISSEURSHIP 145 picture are as follows: 1) in 1893 there was an elaborate veil of translucent drapery running across the Virgin's breast above the neckline of her dress, spiralling across her right shoulder and upper arm. It may also have passed over her wrist and hand, which are clearly surrounded by drapery, in a way no longer visible;tionship with Wildenstein and they no doubt helped and supported him in establishing contacts with experts. It is not known whether Reford himself insti- gated the examinations carried out by Suida, but he had a clear interest in the undertaking. In February 1932 another scholar was invited to inspect the paint- ing. This time William R. Valentiner, then director of the Art Institute of Chicago, visited Reford in Canada to see the Madonna. The visit is recorded in the Reford archives: very carefully, commented again and again on its wonderful drawing and beauty in every way. Dr. Val- entiner then asked leave to study it again which he did this forenoon and emerged from it even more enthusiastic than after seeing it for the first time. He re-confirmed his opinion that it is beyond doubt by Leonardo and stated that he personally found it in several respects more beautiful than the Mona Lisa primarily on account of the glorious blue colouring in the picture and the spiritual expression and beauty of the Madonna's face. 21 This enthusiastic support for a Leonardo attribution no doubt pleased Reford a great deal. Some months later, in September 1932, Joseph Stransky, a collector who also worked for Wildenstein, sent Suida's published report on the examination to Reford, accompanied by his own support of the Leonardo attribution: Most of the Leonardos in the world are just ruins, badly restored, including the Mona Lisa. There is practically nothing left of the Master's brushwork. You and Mrs. Reford have a picture in the best of states and are therefore the proud possessors of an original work by the Master. I am glad that I was in- strumental in bringing it to your attention. 22 LATER RESTORATION WORK While changes were being detected underneath the Lansdowne painting, further alterations were to take place on its surface. There are two images which appear to date from the late 1930s to early 1940s. In the first, taken c. 1939, the image is not sharp, but a most significant change is clear, namely the disap- pearance of Christ's genitals (fig. 57). It appears that the painting had undergone some kind of restoration work. Another image, found at the Witt Library as a black-and-white photograph, is much clearer (fig. 58). The missing genitals are confirmed but there is also a dramatic change in the appearance of the Madonna's right wrist and sleeve. The Witt photograph shows a more detailed wrist, foreshortened with a sleeve around it. Furthermore, there are indications of a fuller veil rather awkwardly painted along the front of the Madonna's dress. How and why did the changes evident in the two images take place? Restoration is carried out on a painting if the stabil- ity of the work is in jeopardy, it has been damaged, or in order to make aesthetic improvements. The Reford family archives give us the answer to our question. They THE LANSDOWNE AND BUCCLEUCH MADONNAS: CONSERVATION AND CONNOISSEURSHIP 143 include a letter from Robert Reford to Wildenstein, dated 20 January 1940 referring to restoration work following damage incurred during the loan of the painting to the New York World's Fair exhibition: 'My object in visiting New York was to arrange for the restoring of the Leonardo da Vinci and a Bronzino, at the New York Exhibition owing to dampness and bad packing by the Exhibition Committee, regardless of my understanding that they were to be sent to your Gallery for examination and packing. ' 23 There is no record in the family papers detailing the exact nature of the damage or who might have car- ried out the restoration work. It is possible that the two images reveal two separate instances of restora- THE ATTRIBUTION DISPUTE CONTINUES Before Reford agreed to loan his painting to the New York exhibition, he was keen that the Leonardo attri- bution was given in the catalogue and on the label. He made contact with Felix Wildenstein in April 1939 in order to ask his advice. He explained in his letter that Valentiner, who was Director General of the exhibi- tion committee, and who had supported the Leonardo attribution in 1932, now wanted the painting to be catalogued as 'workshop of ' or 'ascribed to' Leonardo, neither of which Reford was pleased with. He went on to explain why: 'I refused to send the picture to the Milan Exhibi- tion [1938], though they specifically asked for it as Leonardo. The Duke of Buccleuch, who has the other version of this picture, has sent his, and my feeling is that if it is catalogued as 'Leonardo' and mine as ' As- cribed to Leonardo' it will be something that the crit- ics will use for all time as derogatory to the picture' . 24 The reply from Wildenstein shows his continued sup- port of the Leonardo attribution: I fully agree with you as to the inadvisability of loan- ing the painting unless it is catalogued properly. I certainly would not allow it to be listed as 'Workshop of Leonardo' and hardly as 'ascribed to' … I, therefore, think that your painting should be catalogued as 'Leon- ardo' or perhaps as 'Leonardo aided by his pupils. ' 25 It is worth noting here in relation to the Buccleuch Madonna that the Milan exhibition, referred to in by Sandro Piantanida in 1939, an X-ray was taken of the painting (fig. 3). 26 The most striking feature is the series of small white dots covering the surface of the painting. These are the drill holes that we discussed earlier in Chapter 2. The Lansdowne painting appears to have undergone a well-deserved period of rest during which no further interventions took place. It remained in the Reford family for several decades. Then, following the death of Reford's widow Elsie in 1967, the Reford family again made contact with Wildenstein's in 1971. The Refords were planning to auction their painting and had written to Louis Goldenberg, President of Wildenstein, to ask for his support in a Leonardo attribution. Goldenberg's response summarises secondary research carried out by his colleagues on publications which addressed the issue of attribution. He concludes, 'Our opinion is that the painting of the Madonna and Child is a beautiful and important work. We do not think that the attribu- tion to Leonardo da Vinci can be supported. ' 27 The Reford family went ahead with the auction and on 8 December 1972 the Lansdowne Madonna was sold at Christie's in London as a Sodoma. It fetched £ 17,850. The buyer was Wildenstein's. The price was of course less than might have been expected with a Leonardo attribution. A SECOND TRANSFER Within five years, the Lansdowne Madonna was to undergo a second transfer -this time from canvas on to a composite panel. Wildenstein's contracted the conservator Gianino Marchig, a Geneva-based restorer and painter to carry out the work. Marchig had undertaken work on paintings by other major artists such as Goya and Filippino Lippi. He was well known to Wildenstein's and carried out work on a number of their paintings. In a brief document dated 1 April 1976, Marchig outlined the transfer process and discoveries he made. 28 First, the relining canvas had to be removed. During the removal, it was discov- ered that three canvases were present. Reading from the outermost layer towards the paint surface, these were: (i) a canvas of very rough texture, which had left its imprint on some areas of the surface of the paint- ing; (ii) a canvas with a less rough texture; (iii) a very fine canvas. The third canvas had been bonded to the original preparation with lead white. The first of these (i.e. the one with the coarsest texture) had clearly been added during the first process of transfer from panel to canvas. The innermost, fine canvas had been used by the artist, as was not unusual, to help stabilise the ground on the surface of the panel. The middle layer of canvas is more indeterminate in origin, but is unlikely to have been added by Leonardo. The lead white layer that had been used by the artist as a preparation of the fine canvas indicates that the original wood panel had been prepared using a technique that differed from the simple gesso sotile (gypsum and glue size) which was standard in Italian studios of the late fifteenth and early sixteenth centuries. The use of lead white by Leonardo is evident in other paintings by the artist, the experimental nature of the preparation could be compared with Leonardo's own experimental nature, which was well known. He went on to suggest that the painting was prepared in Leonardo's studio. He also described the presence of a pentimento in the position of Christ's left leg which, he noted, was not generally found in copies. This second transfer from canvas to a composite board was the last radical intervention to take place. It is clear that between 1893, when the painting was still on its original panel, and today, conservation and restoration work has dramatically changed the paint- ing, and although the background landscape appears unchanged, some key differences in other parts of the tor of the present Lord having repaired thither, with the same views as Sir William Petty [1 st Marquess of Lansdowne], like him too, became an agent for the Cromwellians, and carved out an excellent estate for himself; his grandson was ennobled by George I. The clipping goes on to describe the 4 th Earl: The present peer, born June 30 1767, educated at Eton, whence he removed to Christ Church Oxford; and having succeeded his father, the late Earl, in 1781, was created Doctor of Civil Law in 1793. Anterior to this (in 1791), he married Elizabeth, daughter of the late William Brownlow, of Lurgen in Ireland, by whom he has issue, both male and female. 2 Succeeding his father at the early age of fourteen, Darnley was a youthful traveller to the Continent, making a journey to The Hague in 1787, before his coming of age, and possibly visiting Amsterdam as well. 3 The Grand Tour was in full swing during this time, and it was common for young members of the upper classes to travel to the Continent as part of their education. Various hotspots were always on the itinerary, but Italy was the most prestigious in terms of art. In 1788 Darnley went to St Petersburg and was in Berlin by October, following a visit to Warsaw. 4 Back in London the following year, he took his seat in the Irish House of Lords, before returning to the Continent in 1790.This time he is known to have visited Italy, appearing in Rome in March 1790 and Venice later that month.5 April 1798, Dublin). 78 In another letter to Holland (dated April 1798) he repeats the compliment but adds an observation that is not altogether complimentary: 'Lady Gifford [sic] … is so desirable, that as I politely insinuated to her -'pour faire l'agreable' -her bubbies would furnish an ordinary woman with a smart pair of buttocks -making truth and flattery go hand in hand by the aid of a happy illustration. ' 79 In the meantime, Wycombe's affair with Mrs Wyndham had become more complicated, with her abandoning her husband and children in order to flee to Ireland. Wycombe arranged for a house at Sandy Mount, outside Dublin, where he stayed with her. It also happened to be neighbouring one of the Giffards' homes. Writing in June Wycombe complained to Holland that '[Mrs Wyndham] takes umbrage at everything particularly at Lady Giffard and my wench Rosy. The 168 PROVENANCE former is certainly a prodigious fine sow, and the latter a very immaculate good girl I believe. My grey hairs to say nothing of my Morals should screen me from suspicion' (Wycombe to Holland, 23 June 1798, Dublin). 80 Despite Wycombe's complicated personal life, which had also become known in England, he did manage to find time for the arts. During a stay in Wales (where he had gone with Mrs Wyndham), he writes, 'I shall beg your acceptance of a picture of Lord E. Fitzgerald executed by a very decent painter' (Wycombe to Holland, 16 August 1798, Aberystwith). 81 Back at Sandy Mount, he makes his first reference to 'Hamilton [not Gavin] the portrait painter' (Wycombe to Holland, 10 October 1798), 82 whom he would mention in later letters, and who seems to have been based in Ireland for a time. Earlier that summer, in a letter to Lady Holland, Wycombe refers to a collection of paintings of his which had gone astray: 'I believe I have lost all my pictures in Switzerland by the roguery of a Banker[?]' (Wycombe to Lady Holland, 24 June 1798). 83 It is possible that these pictures were recovered and are the same group of paintings he referred to in the 1799 letter to his brother, mentioned earlier. Maria Giffard too was a patron of the arts in Dublin. During the absence of Sir Duke Giffard, Wycombe notes: 'I have promised to go to the play with her [Lady Giffard] tonight, and want her to be drawn into the character of Penelope, as Chinnery [George Chinnery, 1774-1852] is to do an oil picture of her' (Wycombe to Holland, 15 March 1799, Dublin). In the same letter he also writes, 'do not forget that you owe me a picture by Hop[p]ner' . 84 The Hoppner portrait is mentioned again, this time after Wycombe had been in London, where he planned to have a base: you have not fulfilled your promise of giving me your picture by Hop[p]ner. It is much wanted, Great alterations are about to take place in Duke St. [London]. Dennison [?] is going now to decorate his principal room and if I lodge with him again as I suspect, it is my intention to set it off with a few select portraits Back in Ireland, Wycombe's personal tribulations continued. In discussing Sir Duke Giffard, he wrote, 'he cannot succeed in governing the Sow [Lady Giffard], who looks wanton and stays in Dublin constantly in spite of all he can do' (Wycombe to Holland, 26 July 1799, Sandy Mount). 86 At least by this time, Wycombe had convinced Mrs Wyndham to leave Ireland, though he would continue to see her, and the following year he would have to fight a lawsuit against him by her husband. By early September Wycombe had returned to England, this time staying in Margate. At some point he had his portrait made and offered it to Lord Holland: You will hardly acquit me of my vanity when you find that I have accompanied this Envoi by a portrait of your humble … It is not to the likeness of an individual -but to the evidence of Irish excellence, that I am anxious to draw to your attention …' (Wycombe to Holland, 11 December 1799, Margate). 87 Writing again about a present for the Hollands, this time from London, Wycombe wrote, 'Have the goodwill to tell Lady Holland that she shall shortly receive THE LANSDOWNE MADONNA: A PICTURESQUE PROVENANCE 169 an engraving of Buonaparte taken with more success than might have been expected, from a very indifferent picture she may have seen in my possession' (Wycombe to Holland, December 1799,London).88 of men of Genius only (Wycombe to Holland, 20 July 1799, Bangor ferry). 85 Flower came from a family that had made its fortune in Australia's merchant trading business. He was educated at Harrow and Trinity College, Cambridge, before being called to the bar in 1870. He was a Liberal MP from 1880 to 1892, and Lord of the Treasury under Gladstone. In 1892 he was made Baron Battersea of London and Overstrand. THE LANSDOWNE MADONNA: A PICTURESQUE PROVENANCE 173 In 1877 at the age of thirty-four he married Constance de Rothschild, the eldest daughter of Sir Anthony de Rothschild. This marriage greatly elevated his fortune and political connections. It also enabled him to pursue an active interest in the arts. In 1867 he had been elected a member of the Burlington Fine Arts Club and two years later he commissioned Sir Edwin Landseer Lutyens to construct a cliff-top seaside villa in Overstrand which he named Pleasance. In 1879, the same year he bought the Madonna, he and Constance set up their London home at Surrey House, a former home of the Dukes of Norfolk near Hyde Park. The following year, the first known public showing of the Lansdowne Madonna took place in London during the Royal Academy's winter exhibition of old masters. Battersea's taste for things Italian reached beyond paintings. In her memoirs, Constance described how 'Italy … allured and enchanted him from the first. He would have loved to own a villa on the heights overlooking the lovely city of Florence. … He did his best at any cost to transplant Italian colouring and Italian design into his London and Norfolk homes. ' 106 She also described him as a 'devoted student of Italian art' , which would indicate that the Madonna would have been a star item in his collection. 107 It seems that collecting was an obsession with Battersea. Constance noted how 'he unearthed treasures hidden away in little shops and dark streets, both at home and abroad; for wherever he went he brought back an extraordinary assortment of objects'. 108 It seems that the collection was ever-expanding because, as Constance again describes, 'wherever we chose to live he added and built, to accommodate furniture, statuary, pictures and books, of which we had a goodly collection'. 109 It seems that the Madonna was a favourite of Battersea's because he exhibited it on at least two other occasions, in 1893/4 at the New Gallery, London, and again in 1898 at the Burlington Fine Arts Club. Although he clearly had a strong interest in Italian art, Battersea was also keen to encourage the artists of his day. As his friend Lady Frances Balfour wrote, 'Christie's was familiar with his presence, and his love for the old Masters Sourdis Madonna: 'There is by him [Leonardo], in the 184 PROVENANCE THE BUCCLEUCH MADONNA AND A FRENCH CONNECTION 185 cabinet of M. the marquis de Sourdis, a Virgin holding a baby Jesus in her arms. ' 18 It is noteworthy that both du Fresne and Félibien made a point of highlighting this particular painting. Du Fresne mentions only a few paintings by Leonardo in France, including a small painting of the Battle of Anghiari then at the Palais Royal. Félibien describes four paintings in the King's collection: 'un Saint Jean au desert [Bacchus], une Vierge et une Sainte Anne [Virgin and Child with St Anne], et autre Vierge à genoux [Virgin of the Rocks]' and later, 'la Gioconde [Mona Lisa]'. 19 The only other privately owned Leonardo painting besides that of the marquis de Sourdis mentioned is a head of St John the Baptist. 20 The de Sourdis painting was therefore a precious one. In 1651 the marquis of Sourdis was Charles d'Escoubleau. He came from a powerful and esteemed family that was well connected with the Crown and Church. Charles was known for his large collection of art, which he inherited and increased. Most importantly, Charles was the great-grandson of Florimond Robertet, the French minister who first commissioned the Madonna of the Yarnwinder from Leonardo. It seems reasonable that the painting described by du Fresne and Félibien is that which Robertet commissioned. By looking more closely at the Robertet family and its descendants, it can be seen how this painting could have stayed in the family for at least three generations. Florimond Robertet and his wife, Michelle Gaillard, had four children, the youngest of whom was a daughter, Françoise Robertet, dame d'Alluye (d. 1580). In 1539 at Blois, Françoise married Jean Babou de la Bourdaisière (1511-1569), Seigneur de Thuisseau, Comte de Sagonne et de Voilhon, and Grand Master of the royal artillery. He was a prominent figure at the court of King Henri II, becoming Conseiller d'Etat in 1568 and Knight of the Order of Saint-Michel. The couple had four children. The eldest, Françoise, married François Antoine IV d'Estrées, marquis de Coeuvres (1526-1609). One of their daughters was Gabrielle d'Estrées, who became famous for being the mistress of King Henri IV with whom she had three children, all legitimised by him. During her time with Henri IV, Gabrielle lived in the Louvre. Their marriage was never to take place, however. In April 1599 she died after suffering eclampsia followed by the birth of a stillborn baby. The marriage of Gabrielle's parents disintegrated when her mother, Françoise, ran away with her lover, the marquis d'Allègre. This scandal would make it highly unlikely that Françoise would have been favoured with any kind of inheritance, let alone a painting by Leonardo. A second daughter of Françoise and Jean Babou and a more likely candidate to inherit was Isabeau Babou de la Bourdaisière (d. 1626), who in 1572 married François d'Escoubleau, the marquis de Sourdis (d. 1602). The d'Escoubleau family was an eminent one, and like the Robertets, had served the Crown for several generations. François's grandfather, Jean d'Escoubleau, was a page to François I, had been with him at the Battle of Pavia in 1525 and had stayed with the King when he was imprisoned in Spain following the French defeat at Pavia. François, who gained the Robertet title of marquis d'Alluye when he married Isabeau, was First Equerry of the Grande Ecurie (Royal Stable), knight of the Crown and the King's lieutenant-general in Chartres. Through their cousin Gabrielle's alliance with Henri IV, the families were almost related. Indeed, in the early years of their marriage, François and Isabeau lived next to the mother of Henri IV in the Enclos des Tuileries. Of the eight children produced by Isabeau and François, three sons, François II, Henri and Charles, became important military and Church figures. As will be seen, they in turn inherited the wealth and assets of the family. We will discuss each in turn, to review the likelihood of one of them owning the Madonna of the Yarnwinder. 21 FRANÇOIS II DE SOURDIS: RELIGION AND ROME 22 Born in 1574 at the family château de Sourdis in the village of Châtillon-sur-Sèvres, François II was the eldest child of Isabeau Babou and François d'Escoubleau. As an eldest child he was expected to follow in his father's footsteps and pursue a military career. He first studied at the college of Navarre in Paris in 1589 and in 1591 fought alongside his father in the Siege of Chartres. He became engaged to Catherine Hurault de Cheverny, daughter of Philippe Hurault, the royal chancellor. However, neither this engagement nor his military career was to last. François II's first visit to Rome took place in November 1593 as part of the duc de Nevers's embassy to Pope Clement VIII following Henri IV's rejection of Protestantism to embrace the Catholic faith. The Pope wanted Henri IV to persuade his people to become Catholic also. When the Pope rejected the French King's absolution granted by his clergy, a great tension was created. To appeal to the Pope, Henri IV asked Louis de Gonzaga, the duc de Nevers (1539-1595), to go to Rome on his behalf. The entourage included François II d'Escoubleau. Although the mission was a failure and returned in January 1594, François II d'Escoubleau was much taken by Rome as well as the Church and stayed on. There he became friends with Federico Borromeo (1564-1631), who became archbishop of Milan in 1595, and with the ecclesiastical historian Cardinal Caesar Baronius (1538-1607). When François returned to France, he broke off his engagement and joined the seminary in Chartres, thereby abandoning his rights as eldest son and heir. When in February 1599 Henri IV became engaged to Gabrielle d'Estrées, duchesse de Beaufort (1571-1599), the d'Escoubleau family became ever closer to the Crown. Gabrielle was François's first cousin, and a close friend of his mother Isabeau. As a sign of the alliance, Henri IV gave François clerical titles. The reconciliation between the Church and French Crown in 1595 had created a need for French cardinals and on 3 March 1599 at the age of twenty-five François II was duly elevated to a cardinalship. By July of the same year he was also consecrated archbishop of Bordeaux. 23 The c ardinal soon joined the religious intellectual elite in France, including the cercle d' Acarie, named after Madame Acarie, the founder of the Carmelite order in France. THE BUCCLEUCH MADONNA AND A FRENCH CONNECTION 187 In September 1600 the Cardinal François d'Escoubleau de Sourdis returned to Rome to receive his cardinal's hat. He was invited by the Pope to stay at the Palazzo Aldobrandini and resumed friendships he had made during his first visit. His third visit to Rome took place in 1605 as part of an embassy to influence the choice of successor to Pope Clement VII who had died in March. On his way to Rome, François stopped in Milan where he greatly admired his friend Cardinal Borromeo's institutional work, including the founding of the Ambrosiana library and work on the Duomo of Milan, where he was eventually buried. This trip most likely influenced François later in his own establishment of a convent, garden and museum project (unrealised) for Bordeaux. When he eventually arrived in Rome, François stayed at the Borromeo family palace. 24 Along with the wealthy collectors and patrons Cardinal Scipione Borghese (1576-1633), who established a great collection of antique sculpture in what is now the Borghese Palace, and Maffeo Barberini (1568-1644), later Pope Urban VIII, François visited the studio of the famous sculptor Gian Lorenzo Bernini (1598-1680). 25 It was around this time that François commissioned Bernini to create his portrait bust. 26 Between 1620 and 1622 he also commissioned The Resurrection of Christ186 PROVENANCE As both François and Henri had taken vows with the Church, the third son, Charles de Sourdis (1588-1666) was the family heir and therefore given the title of the marquis de Sourdis et d'Alluye. Indeed, this had happened when François joined the Church.It is therefore likely that while cash was given to Henri, the favoured son, family belongings, including estates, would largely have gone to Charles. Unfortunately, a post mortem inventory for Isabeau's estate is not known to survive. It would no doubt have provided rather more detailed information as to the contents of the various family properties.The year after Isabeau's death in 1626, her son François drew up his own will, from which it is clear that his greatest beneficiaries were to be the Church and town of Bordeaux rather than his family. Indeed, the will details that all the paintings in the chapel, gallery and cabinet should be left to his successors in the Church. 33The only mention of his younger brother, Charles, the marquis of Sourdis, is the bequest of a painting of St Cyrillus.34 This would mean that had the Madonna been in François's possession, it would have been left to the Church in Bordeaux. How, then, would it have returned to Charles, the youngest brother? HENRI D'ESCOUBLEAU DE SOURDIS Perhaps the answer lies in François's successor, his other brother, Henri d'Escoubleau de Sourdis (1594-1645). THE BUCCLEUCH MADONNA AND A FRENCH CONNECTION 189 When François died in 1628, Henri became archbishop of Bordeaux. Henri's patronage of the arts is not as well documented as that of his elder brother. He did, however, oversee the completion of François's building projects in Bordeaux and the ornate marble decoration and enlargement of the choir of the cathedral. 35 He also appears to have acted as a kind of artistic adviser to the powerful Cardinal Richelieu (1585-1642), who was seeking a suitable location for a recently obtained group of Italian paintings from Isabella d'Este's studiolo. In a letter dated 24 July 1632, Richelieu wrote to Henri, who approved of the idea 'to put the five paintings from Mantua in the great cabinet of the Corps de garde' . 36 Boniface Borilly (1564-1648) -a collector of medals, natural curiosities, and paintings, including works by Titian, Rubens and Michelangelo -wrote about Henri as a prospective buyer who was willing to pay a high price for works. Borilly wrote that he had 'three bodies [sculptures?] with different postures … which the Cardinal de Sourdis has offered one thousand livres to have'. 37 Henri apparently considered the collection left by his brother to the Church as part of his own personal inheritance. His stance became apparent during a trial in 1633. The church council, which had inherited from the Cardinal François furniture, silver and decorative pieces, sued Henri because he had allowed his chaplain to remove six silver candlesticks and brocades from the cathedral in order to use them for his parish of St Seurin at Christmas. 38 It seems reasonable, therefore, that had there been a painting precious to the family such as the Madonna of the Yarnwinder, Henri might have felt he had a right to keep it in the family. Indeed, in his will, Henri leaves the vast majority of this estate to his brother Charles, the marquis de Sourdis. Therefore, had the painting ever been in François's possession, it is likely to have gone from Henri to Charles. If Isabeau had given it to Henri, then it would have been passed on to Charles after Henri's death. Of course, the simplest answer could be that Charles received the painting directly after the death of Isabeau. In any case, after Henri's death in 1645, Charles was the owner of the vast majority of the d'Escoubleau fortune and estates.CHARLES D'ESCOUBLEAU, MARQUIS DE SOURDIS When in 1651 Raphael Trichet du Fresne described in his biography of Leonardo a Madonna painting in the collection of the marquis de Sourdis in Paris, he was referring to Charles, youngest son of the Babous. By 1651 Charles had become well known for his collection, much of which he inherited from his family. He was also an active buyer as well as seller of works of art and, like his brother Henri, Charles acted as a kind of art adviser to Richelieue as can be seen in a letter written by Cardinal Mazarin to Cardinal Antonio Barberini on 16 December 1634: I have sent on your behalf to the Cardinal four paintings which are the best and which were chosen by the Marquis de Sourdis, who is very competent to handle all of the sculpture and paintings of His Eminence: they are those by Titian and Pietro da Cortona, The Baptism of the Virgin on wood by Giulio Romano and the painting by Antonini which His Eminence expressly wished for after hearing the marquis de Sourdis describing its charm. 39 Charles cared for the family collection enough to clash with the Church in Bordeaux following the death of his brother Henri in 1645. He was present at Henri's death and carefully supervised the proceedings following it. On 16 January 1646, the Bordeaux Chapter opposed Charles's removal of his brother's paintings from the walls of the gallery, claiming that they were meant to remain with the Church. 40 Unfortunately, the archives do not give specific details of the paintings involved. Whatever they were, Charles managed to establish a fine collection, much of which was housed in a gallery of paintings in his Paris home. The nineteenth-century historian Edward Bonnaffé, who modelled his writings on Félibien, described the collection, which was visited by the Queen of Sweden: 'His gallery of paintings in Paris is one of the best; Queen Christina visited them at length in 1656, while in dicating to the marquis de Sourdis the values of the paintings. ' 41 Bonnaffé goes on to highlight in Charles's collection, Raphael's St George and the Dragon: '[from the] collection of the King of England, today in the Hermitage' . 42 This painting, now at the National Gallery, Washington, appears to have come from the collection of King Charles I and was likely bought by Charles de Sourdis from Edward Bass. 43 The only other painting specifically mentioned by Bonnaffé is, 'the Virgin holding the Infant Jesus by Leonardo da Vinci (n° 482 in the Louvre?)' . 44 Almost certainly the same painting described by du Fresne, the picture appears to have been not only a star in the collection, beingonly one of two paintings to be specifically mentioned, but also well known. Bonnaffé's guess that the paint-ing was the same as number 482 at the Louvre is as intriguing as it is confusing. It appears to be based onFrédéric Villot's 1849 catalogue of Italian paintings at the Louvre where the number 482 is given to Leonardo's Virgin of the Rocks. In outlining the provenance, Villot wrote that the Virgin of the Rocks was once in the collection of François I, but goes on to cite an 1803 biography of Leonardo by Gault de Saint-Germain: 'Gault de Saint-Germain, in his Life of Leonardo da Vinci, said that it came from Versailles, and had previously belonged to the Marquis de Sourdis. ' 45 Pierre Gault de Saint Germain's publication of Leonardo's Treatise on Painting does indeed describe the Virgin of the Rocks as coming from Versailles and de Sourdis: 'It previously belonged to the Marquis de Sourdis' , but unfortunately he gives no source for this information. 46 It is perhaps most likely that Gault de Saint-Germain was mistaken. The de Sourdis Leonardo was famous, as du Fresne and Félibien testify, but their description of the Virgin holding the Child does not match the Virgin of the Rocks. Charles lived between Paris and the family château in Jouy-en-Josas, several kilometres from Versailles. He was known for his intellectual pursuits and was part of one of the elite literary circles, which included Madeleine de Souvré, the marquise de Sablé (1599-1678), and Madame Cornuel (1604-1694). In the literary salons they would discuss human nature and the human condition, which would be distilled into brief descriptions of human conduct and wisdom called maximes. 47 Bonnaffé cites an important library held by Charles at his château. 48 190 PROVENANCE THE BUCCLEUCH MADONNA AND A FRENCH CONNECTION 191 In 1612 Charles married Jeanne de Montluc et de Foix, comtesse de Carmain, princesse de Chabanais. She was the daughter of the erudite and literary Adrien deMontluc-Montesquiou (1571-1646. They produced five sons and four daughters before Jeanne died in 1657, nearly ten years before her husband.188 PROVENANCE Coeffier deRuze (1612Ruze ( -1644. Of the sons, the eldest, Adrien François (1619-1638), died of his injuries at the siege of Renty at the age of nineteen. Virginal, abbé de Sourdis, also died young at twenty-two. Two further sons werePaul (d. 1690) andHenri (d. 1712). Charles's will, drawn up on 21 September 1666, is particularly interesting for the insight it gives into the relationship he had with his children, especially Paul, who at the time was the eldest son and heir. 50 The will reveals that Charles was less than happy about Paul's desire to marry the infamous beauty and confidante of King Louis XIV, Marie-Béigne de Meaux du Fouilloux. Indeed, he threatened to disinherit Paul and bequeath his entire estate to Henri, the Comte de Montluc, should Paul decide to marry Marie-Béigne. Paul was perhaps more of a womaniser than a soldier, despite some successful campaigns with Richelieu's armies in the Low Countries. 51 His affair with Marie-Béigne continued until the death of Paul's father in 1666.It was then that the couple married. Expecting that Paul would disobey his wishes, Charles had already partly taken from him an extensive source of income, that of the governance of Orléans. A further setback to Paul's fortunes took place in 1671 when a number of properties he had inherited were transferred to his brother, Henri. 52 These included the contents of the properties such as furniture and paintings, and might therefore have included the Madonna of the Yarnwinder. If this is the case, then Henri would have become the owner of the Leonardo. By 1673, Paul had also parted with the Château de Jouy, which was bought by Charles d'Albert, duc de Luynes (1620-1690), who will feature later in the Tallard story. The fortunes of Paul and Marie-Béigne were to fall further when Marie-Béigne became implicated in the infamous Affair of the Poisons, a murder scandal which involved poison and witchcraft and the inner court of King Louis XIV. Her close allegiance with one of the prime suspects, the Comtesse de Soissons, tarnished the couple. When Paul died in 1690, he still had not been forgiven by the King. Marie-Béigne outlived her husband by thirty years, living comfortably if not extravagantly at the Palais Royal in Paris. It seems evident that had Paul been in possession of the Madonna of the Yarnwinder, it did not stay with him for long. Bad judgement and ill-fortune caused him to lose most of his wealth and property. A final son of Charles de Sourdis was François (d. 1707). He had been given the Abbey of Aubagne, but left the clergy at the age of fifty apparently to marry Marie-Charlotte de Béziade d'Avaray (d. 1688), with whom he produced one child, Angélique (1684-1729). He was Lieutenant-General of the Armies of the King in 1682, Knight of the King's Order in 1689 and Governor of Orléans. Following the death of his father in 1666, François built the Château de Gaujacq in Guyenne in the style of the Italian Renaissance, which he finished around 1690. 53 A posthumous inventory from February 1708, following François's death in September 1707, lists 273 paintings, none of which can be identified as the Madonna of the Yarnwinder. All but eight of the paintings were sold. 54 It appears that had the Madonna of the Yarnwinder been in François's possession, it would have been at Gaujacq. A letter, written two months after his death, mentions that the majority of François's possessions were at Gaujacq and indeed this is where he was based. 55 The sale of the contents of the château brought in 23,000 livres, an impressive total. The contents included ornate furnishings one would expect from a house of this calibre, such as Flemish tapestries and valuable furniture. There are a number of portraits of popes and known French figures as well as paintings on panel from the sixteenth or early seventeenth centuries, therefore likely to have originated from his father Charles, uncles Henri and François, or his grandparents. Although the descriptions are brief, they are occasionally informative. A few were certainly Italian, such as 'Neapolitans playing instruments' and 'a small square painting representing the Castle of San Angelo in Rome' . 56 de Dreux describes in his will of 1680 thirteen paintings which he bought from 'Monsieur de Sourdis' . Dreux wished for his heir to keep the collection intact, but most of them were subsequently sold in 1681. His list reflects the strong presence of Italian paintings in the de Sourdis collection: x. The Apparition of the Angel … by Bassano, from monsieur de Sourdis 57 xi. Paul III by Titian, from monsieur de Sourdis xii. Julius II, by Giulio Romano, from monsieur de Sourdis 58 xiii. Saint Jerome by Palma Vecchio, [from] monsieur de Sourdis xxii. A sketch of the Apostles … by Caravaggio, [from] monsieur de Sourdis 59 192 PROVENANCE THE BUCCLEUCH MADONNA AND A FRENCH CONNECTION 193 xxv. A Pietà that I have from monsieur de Sourdis, by Antonio da Correggio 60 xxvi. The wife of Paul de Veronese painted by Paul de Veronese, from monsieur de Sourdis 61 xxvii. The daughter of Titian by Titian, from M. de Sourdis 62 xxviii. A Christ carrying the cross by Albrecht Dürer, monsieur de Sourdis 63 xxxii. A Titian painted for Charles V and given by him to the grandfather of monsieur de Sourdis 64 xxxiii. A picture by Paul Veronese of Christ nailed to the cross, monsieur de Sourdis 65 xxxvii. A Virgin and child by Annibale Carracci, M. de Sourdis 66 There does not appear to have been a sale dedicated to the de Sourdis collection. It is therefore likely that the collection was dispersed in small groups or individually through private sales and by inheritance. An example of such a possibility exists through Charles de Sourdis and his son François of the Château de Gaujacq. Indeed, the connection could lead directly to the duc de Tallard from whom the Buccleuch Madonna of the Yarnwinder was purchased. François and Marie-Charlotte's only daughter, Angélique d'Escoubleau, was married to François Gilbert Colbert (1676-1719), the marquis de Saint-Pouange and Chabannais, and maréchal de camp. They had two children, François Gilbert Colbert (1705-1765) and Antoine Colbert (1707-1788). While Antoine held the title of marquis de Sourdis and seigneur de Gaujacq Bastennes, François Gilbert became the marquis de Saint-Pouange and Chabannais. He was also a military man as maréchal des camps et armées of the King. When their mother died in 1729, the two brothers sold the Château de Gaujacq for 9,000 livres to Jean-Gabriel Cazenave, a major-general in the French navy, who no doubt would have been known to Tallard through military connections. The sale of the château, much beloved by their grandfather, must have been a desperate measure, and it seems likely that any paintings remaining in the family collection would have been considered for sale as well. There is a further and perhaps more reasonable link between the de Sourdis family and the Tallards. As previously mentioned, in 1673, Paul d'Escoubleau, the son of Charles de Sourdis, had fallen on hard times and sold the Château de Jouy, presumably with some of its contents, to Charles d' Albert, duc de Luynes. The duc de Luynes's descendant, Charles Philippe d' Albert de Luynes, duc de Luynes (1695-1758), was married to Marie Brulart, lady-in-waiting to the queen consort Marie Leszczynska at Versailles. It has been established that Marie-Isabelle de Tallard was part of the queen consort's court as lady-in-waiting and then governess to the royal children and it is therefore clear that the families would have known each other. Marie Leszczynska was known to have a close-knit circle of friends, which she called her honnêtes gentes and which included both the Tallard and Luynes families. This can be confirmed through memoirs written by the duc de Luynes. The memoirs, which cover the period 1735-58, describe events and the daily life at the court. In it, the duke mentions 'Madame de Tallard' at least ten times. 67 It is not unlikely, had the Madonna of the Yarnwinder come into the duc de Luynes's possession through the purchase of the de Sourdis château, that it could have been given or sold to that eager collector the duc de Tallard. A further possible connection between the duc de Tallard and another Madonna and Child by Leonardo da Vinci is worth mentioning. Again, the link is through Tallard's military career, and coincidentally, the Church in Bordeaux is featured as well. It has been established that Cardinal François de Sourdis created a great collection which he wished to leave to the Church in Bordeaux. When his brother Henri succeeded him, much of the collection remained in Bordeaux. Following the death of Henri, the next archbishop of Bordeaux was Henri de Béthune (1604-1680). Henry owned around 500 paintings which were kept in his palace in Bordeaux and his Château du Dormant in which he kept a library comprising around 1,000 volumes. His father, Philippe de Béthune(1565-1649), was an active patron and collector, having lived in Rome as ambassador where he became a collector of Caravaggio. When he died in 1649, Philippe's estate was divided between Henri de Béthune and his brother, Hippolyte (1603-1665). In Henri's will of 1680, he specifically mentioned a 'Nostre-Dame et le Petit-Jésus' located above the door of the bedroom, which he bequeathed to his sister-in-law, the duchesse de Béthune. 68 Although the artist is not named, the prominence of the painting and its location in the will indicate its importance. The possibility of the painting being by Leonardo is strengthened by the presence of a Madonna and Child by Leonardo in the possession of a descendant of Henri and Hippolyte. The 1725 will of Marguerite-Louise de Béthune, duchesse de Lude (1642-1726), 'bequeathes to the duc de Coislin, her nephew, one painting depicting the Holy Virgin holding the Child Jesus, which is an original painting by the hand of Leonardo da Vinci & which comes from Monsieur the Bishop of Verdun [Hippolyte de Béthune. The first cousin of the duchess' father]' . 69 The duc de Coislin was Henri Charles du Cambout de Coislin (1664-1732). Appointed bishop of Metz in 1697, Henri Charles was also known to Tallard's father, Camille d'Hostun, duc de Tallard. Although the connection is less firm than that between the de Luynes and the Tallards, it is conceivable that Camille Tallard bought his Leonardo through this channel. However, if this is the case, it would mean that the Tallard Madonna and Child could not be the de Sourdis painting and therefore less likely to have come from the collection of Florimond Robertet. A final link between a Leonardo Madonna and the Tallard Madonna should be mentioned. In 1741, an inventory of the estate of Cardinal Melchior de Polignac (1661-1741) was drawn up. Like Cardinal François de Sourdis, Cardinal Polignac had spent a great deal of time in Rome and was also a keen collector, especially of sculpture. A sale of his collection took place in 1742 comprising 137 lots containing Roman (modern and antique), Greek and some Egyptian bronze and marble sculpture. 70 Alongside a long list of paintings and artists in the 1741 inventory is the following notation: 'Virgin and Child by Leonardo da Vinci, one foot and a half high by thirteen inches wide. ' It was valued at 120 livres. 71 It is worth noting this inventory entry because the painting described here is exactly the same size as the Madonna of the Yarnwinder in the 1756 Tallard sale. Melchior de Polignac has another link with the Tallard sale. Lot 117 in the Tallard sale comprised two pastels by the female artist Rosalba Carriera (1675-1757), which are described in the catalogue as formerly 194 PROVENANCE in the collection of Cardinal de Polignac. 72 The pair fetched a very high price of 2,800 livres. The year before, the same pair of pastels had been bought by Tallard for 2,416 livres. 73 Although there are no known public sales which included a Leonardo Madonna and Child linked to Polignac, it is conceivable that Tallard acquired his painting through a private sale. Although the link is tempting, it is speculative and must remain so for the present. The persistent problem with the provenance of the Buccleuch Madonna is not that there are too few possibilities but that there are too many that are almost equally plausible. However, looking back at our journey through the various families, the most plausible chain is that the Robertet Madonna of the Yarnwinder passed from Florimond Robertet to his daughter, Françoise Robertet, and after to her daughter, Isabeau Babou de la Bourdaisière, wife of François d'Escoubleau. At this stage it came into the d'Escoubleau de Sourdis family. The painting could have been owned by the brothers François and Henri, but certainly became part of the collection of the third brother, Charles de Sourdis. Within a generation of Charles's death, the family sold most of its assets and the painting appears to have been part of this dispersal, perhaps ending in the collection of the duc de Tallard. On the other hand, the Buccleuch/ Tallard Madonna of the Yarnwinder also has connections with Pierre de Rohan, Robertet's contemporary in Italy who could have obtained the second version of the Madonna after the Robertet commission. Whichever scenario is preferred, the evidence for the provenance for the Buccleuch painting is consistently French. Indeed, potentially direct connections with Robertet can be adduced. On the face of it, this would seem to favour it as the one that was actually delivered to Florimond Robertet in Blois. However, the possible routes, incomplete as they are, leave plenty of gaps into which the Lansdowne Madonna may be inserted. After all, both Lords Darnley and Lansdowne were known to buy paintings from French collections and the de Sourdis Madonna could very well have been one of them. Furthermore, a number of the paintings owned by Salaì were obtained for the French Royal Collection, including of course the Mona Lisa. A French purchaser for the Madonna of the Yarnwinder is not unlikely. Getting the painting from France into the Darnley and Lansdowne collections is not difficult, given the huge numbers of works that were swept up by British purchasers following the French Revolution.We will return briefly to this issue in the conclusion.NOTESTO PART III 195 14. THE LANSDOWNE MADONNA: A PICTURESQUE PROVENANCE 1. Hugh Brigstocke, William Buchanan and the 19 th century art trade: 100 letters to his agents in London and Italy, London, 1982. 2. Medway archives, Strood, u565/f61. 3. Letter from Darnley to his mother written from The Hague, March 1787, and mentioning a possible visit to Amsterdam, see Esmé Wingfield-Stratford, The Lords of Cobham Hall, London, 1959, p. 281. Ligny in Rome and go with him to Naples. It is not known what the plan was or if it was carried out. See Codex Atlanticus 669 r. 46. These were Treatise on the Art of War and The Duty of a Prince or General towards a Conquered Country. 47. William Dunbar, 'Eulogy to Bernard Stewart, Lord of Aubigny [Withe glorie and honour]' , in John Conlee, ed., William Dunbar: The Complete Works, Kalamazoo, Medieval Institute, 2004. 48. He was his great, great, great, great, great grandson. 49. Brigstocke, Letter 50, 15 April 1804, 1982, p. 253. 50. Lansdowne House had been purchased, unfinished, from the Earl of Bute in late 1765. It was originally designed by Robert Adam and subsequently added to by George Dance the younger, François-Joseph Belanger and Robert Smirke. See Damie Stillman, 'The Gallery for Lansdowne House: International Neoclassical Architecture and Decoration in Microcosm' , The Art Bulletin, vol. 52, no. 1, 1970, pp. 75-80. 51. Ibid., pp. 76-7. 52. Adolf Michaelis and Arthur Hamilton Smith, A Catalogue of the Ancient Marbles at Lansdowne House, London, 1889. 53. Morrison, Hamilton and Nelson Papers, 58, no. 88; the Leonardo manuscript is perhaps Libro A. See Carlo Pedretti, Leonardo da Vinci on Painting: a Lost Book (Libro A): Reassembled from the Codex Vaticanus Urbinus 1270 and from the Codex Leicester, Berkeley and Los Angeles, 1964, p. 96, no. 5. 54. Martin Davies, National Gallery Catalogues: The Earlier Italian Schools, 2 nd ed., from Kings Mews, the same address as C.J. Greville. 58. Rutland, Charles Manners, Duke of, and John James Robert Manners, Duke of Rutland, Historical Manuscripts Commission, The Manuscripts of His Grace the Duke of Rutland … Preserved at Belvoir Castle, vol. III, London, 1894, p. 270. 59. Lot 47, ' An upright picture representing the Virgin, the Infant Saviour, St. John and Elizabeth. ' A Leonardo painting describing a 'Holy Family with St. Elizabeth' appeared in an 1810 Lansdowne sale, but as it sold for a low 30 guineas, it could not be the Virgin of the Rocks. See Davies, 1961, pp. 270 and 280, nos. 91 and 92. 60. Christian August Gottlieb Goede, England, Wales, Irland und Schottland: Erinnerungen an Natur und Kunst aus einer Reise in den Jahren 1802 und 1803, Vol. 4, 1805. 61. Ibid., p. 35. 62. Ibid., p. 40. 63. Ibid. 64. The 1806 Lansdowne auction lists a Pandora by Leonardo da Vinci which sold for £42. This is likely to be one of the many 'Leonardos' being sold during this period and well into the nineteenth century. 65. Bentham to George Wilson, 1781. In John Bowring, ed., The Works of Jeremy Bentham, 11 vols., 1838-1843. 66. Wycombe to Lansdowne, Paris, 21 September 1792. Bowood House Archives. 67. Bowood House archives. 68. Ibid. 69. Ibid. 70. Ibid. 71. Ibid. 72. Ibid. 73. Holland House Papers, British Library, ms 51318-52254. 74-88. Ibid. 89. The 1 st Marquess was the owner of a Salvator Rosa Incantation scene, which was put up for auction in 1806; Coxe, Burrell and Foster, Sale of paintings, March 1806, lot 43. 90. Wycombe left a portrait of Marat to Holland in his will; it was the only picture mentioned. A portrait of Marat appeared in the 1833 auction (lot 27) so perhaps Holland never received his painting. 91. Holland House Papers, British Library, ms 51318-52254. 92. Richard Arthur Austen-Leigh, Jane Austen at Southampton, 1949, p. 16. 93. The Observer, 17 September 1809. 94. Holland House Papers, British Library, ms 51318-52254. 95. James Edward Austen-Leigh, A Memoir of Jane Austen by her nephew, 1869, chapter 4. 96. I am grateful to Jean Watts for bringing this portrait to my attention. 97. Allan Cunningham and Peter Cunningham, The Life of Sir David Wilkie: With His Journals, Tours, and Critical Remarks on Works of Art; and a Selection from His Correspondence, vol. 1, 1843, pp. 191-6. A painting of The Marriage of the Dwarfs, with a portrait of the Marquess of Lansdowne by Chalon, was included in the 1879 sale in which the Lansdowne Madonna appeared. 98. Louis Simond, An American in NOTES TO PART III 197 Regency England, ed. Christopher Hibbert, London, 1968. 99. Holland House Papers, British Library, ms 51318-52254. 100. Ibid. 101. Will of 2 nd Marquess of Lansdowne, Bowood House archives. 102. Lord Granville Levenson Gower, Private Correspondence 1781-1821, see Austen-Leigh, 1949, [n.p.]. 103. 'Report of the Marriage of the Duke and Duchess of Cumberland' , The Times, 30 August 1815; cites the Dowager Marchioness of Lansdowne as one of the principal guests. 104. I am grateful to Jean Watts for bringing this to my attention. 105. Paintings which appeared in the 1834 and 1835 sales include a portrait of Napoleon and works by Wallis, Guido, Benvenuti, Van Wien, Jordaens, G. Dow and Salvator Rosa. 106. Constance Battersea, Reminiscences, A FRENCH CONNECTION 1. Catalogue raisonné des tableaux, sculptures, tant de marbre que de bronze, desseins et estampes des plus grands maîtres, porcelaines anciennes, meubles précieux, bijoux, et autres effets qui composent le cabinet de feu Monsieur le Duc de Tallard, Paris, Didot, 1756; and Fritz Lugt, Répertoire des Ventes, vol. I, Paris, 1983, no. 910; see Bur-ton Fredericksen, 'Leonardo and Mantegna in the Buccleuch Collection' , The Burlington Magazine, vol. 133, no. 1055, February 1991, pp. 116-18. 2. See Patrick Michel, Le commerce du tableau à Paris dans la seconde moitié du xviiie siècle, Paris, 2008, p. 230. 3. 'M Rémy peintre et négociant en tableaux et autres curiosités. ' From a manuscript list of prices at the end of the printed catalogue, bge la 85/4 (5), Bibliothèque Publique et Universitaire, Geneva. Several other copies, including a second in this library (bge la 85/11), provide the same information. 4. The four other paintings purchased by Rémy 'for England' ('pour L'Angleterre') are: no. 8, Francesco Vanni, Christ in the house of the Pharisee; no. 51, Lodovico Carracci, Dead Christ with the Virgin and two Angels; no. 98, Tintoretto, Adoration of the Shepherds; no. 125, Murillo, Madonna and Child. 5. There is a sale in 1779 in Paris of a painting resembling the Madonna of the Yarnwinder (Guilleaumon, Paris, 6 . '.La Sainte Vierge assise, aïant sur ses genoux l'Enfant Jesus qui tient une croix. Le fond du tableau est un païsage; sa hauteur est de 18 pouces de haut sur 13 pouces de large. Ce tableau est du meilleur tems de ce Maître & d'une conservation parfaite. Rien n'est si rare, ni en même tems si estimable, que les Tableaux de Léonard de Vinci. Cet excellent Artiste qui avoit la plus haute idée de la Peinture, n'a pas fait beaucoup de Tableaux, quoiqu'il ait poussé fort loin sa carrière. ' See chronology for full catalogue description. 7. Mariette wrote, 'Lettre sur Léonard de Vinci peintre florentine' , the introduction to The Count of Caylus, Recueil de testes, 1730, which reproduced Leonardo's grotesques from the Album de Caricatures (Paris, Louvre). Mariette's father had bought the grotesques from the Earl of Arundel. The drawings are now considered to be by a follower of Leonardo. 8. 'Les chairs sont d'une couleur de pain d'épices, qui n'est pas favorable au Tableau …. ' Copy 2, rc.ee.39b, Victoria and Albert Museum, London. 9. 'L'Enfant Jesus couché dans la crêche, aïant à ses côtés la Vierge & Saint Joseph à genoux: figures grandes comme demi nature. Dans le fond est une campagne où l'on apperçoit deux Bergers, dont un porte un agneau sur ses épaules. Plusieurs Anges, debout sur des nuages, semblent se réjouir & annoncer la venue du Messie. ' 10. 'Ce Tableau ne me paroit point de Leonard, il est trop mal dessiné, l'enfant est un enfant mort né. ' Copy 2, rc.ee.39b, Victoria and Albert Museum, London. 11. Lot 3, Christ on the Cross. The work brought in only 190 livres and was soundly criticised by Mariette who noted, 'How unjust to attribute to the greatest draughtsman a work which sins by its design?' ('Quelle injustice d'attribuer au plus grand dessinateur un ouvrage qui peche par le dessein'), ibid. 12. The paintings which went to the Montagu and then Buccleuch collections have already been mentioned. Another example is Lot 39 attributed to Correggio, which is now to be found at the Gemälde-16. Raphael Trichet du Fresne, Trattato della Pittura di Lionardo da Vinci, Paris, 1651. This first edition was in Italian. 17. 'Nel studio del Signor Marchese di Sourdis a Parigi vi è un altra Madonna di riputatione. ' 18. 'Il y a encore de lui [Leonardo] le cabinet de M. le marquis de Sourdis, une Vierge tenant un petit Jésus entre ses bras. ' André Félibien, Entretiens sur les vies et sur les ouvrages des plus excellens peintres anciens et modernes: Entretiens I et II (1666-1688), 1987, p. 257. 19. Félibien, 1987, pp. 257 and 260. These paintings are now all at the Louvre, Paris. 20. '… une tête de saint Jean-Baptiste qui est à present à l'hôtel de Condé dans le cabinet de M. le Prince. ' This is perhaps the Head of St John the Baptist by Andrea Solario, now in the Louvre, Paris. 21. Following the death of Jean Babou, Françoise Babou married Jean VI d'Aumont (1522-1595), but produced no children. 22. I am indebted to Thierry Morel for his crucial contributions to the de Sourdis family research. 23. Joseph Bergin, The making of the French episcopate: 1589-1661,London, 1996, pp. 704-5. 24. Arlene Quint Platt, Cardinal Federico Borromeo as a Patron and a Critic of the Arts and his Museum of 1625, Doctoral thesis, University of California Los Angeles (1974), New York, 1986. 25. A visit to Bernini's studio by François and Pope Clement VIII took place on 12 June 1622. See Oliver Laroza, Guide pour la Visite de la Cathédrale de Saint-André de Bordeaux, 1974, p. 8. 26. Church of Saint Bruno, Bordeaux. The de Sourdis family tomb was placed here in 1691. 27. Cathedral of Saint André, Bordeaux. 28. Church of Saint Bruno, Bordeaux. 29. Andrea Bacchi, Catherine Hess and Jennifer Montagu, Bernini and the Birth of Baroque Portrait Sculpture, exhibition catalogue, National Gallery of Canada, Ottawa, 2008. 30. See Robert Boutruche, ed., Histoire de Bordeaux de François de Sourdis, Archives Départementales de la Gironde, 1 December 1627, g. 259: 'Donnons aux archevesques nos successeurs tous les tableaux qui sont dans nostre Chappelle, gallerie & cabinet …. ' 34. Ibid., '… Et quand à nostre frère le marquis de Sourdis Tableau de St. Cyrille. '35. 'Ledit seigneur, estant à son château de Lormont, l'aurait envoyé pour faire advertir le chapitre qu'il avoit volonté de monter le choeur de cette église, faire le grand autel et de parer à neuf de marbre, fair une baluster de marbre, eslargir le choeur et icelluy parer de marbre quy priroit le chapitre d'y consentir …' , 14 January 1644, Archives Départementales de la Gironde, g. 297, f. 124. 36. 'de mettre les cinq tableaux de Mantoue dans le grand cabinet du Corps de garde. ' Cited in Antoine Schnapper, Curieux du Grand Siècle: Les collections d'art en France au xviie siècle, Paris, 1994, p. 142. 37. 'trois corps de différentes postures […] desquelles le cardinal de Sourdis en a fait offrir mille livres par les avoir. ' See Schnapper, 1994 p. 117. 38. Archives Départementales de la Gironde, g. 172. 39. J'ai envoyé de votre part à M. le Cardinal quatre tableaux qui sont les meilleurs et qu'a choisis le marquis de Sourdis, fort competent qui s'occupe de toutes les statues et peintures de Son Eminence: ce sont ceux du Titien et de Pierre de Cortone, Le Baptême de la Vierge sur bois de Jules Romain et le tableau d'Antonini que Son Eminence a voulu expressément, après avoir entendu le marquis de Sourdis en décrire le charme' , see ). Like the paintings discussed above, the Madonna 206 THE INFLUENCE OF AN IMAGE PREVIOUS PAGE Fig. 80 After Leonardo da Vinci (Fernando Yáñez de la Almedina?), Madonna of the Yarnwinder, oil on panel, 62 × 48.8 cm, National Galleries of Scotland, Edinburgh (Leonardo e il Leonardismo 53). Fig. 81 After Leonardo da Vinci, Madonna of the Yarnwinder, oil on panel, 62.2 × 48.6 cm, formerly in the Wood Prince Collection. Private Collection/Photo © Christie's Images/The Bridgeman Art Library (Leonardo e il Leonardismo 54).Fig. 82 After Leonardo da Vinci, The Madonna of the Spindle, © Musée des Beaux-Arts de Dijon (Leonardo e il Leonardismo 55).Fig. 83 After Leonardo da Vinci, Madonna of the Yarnwinder, location unknown (Leonardo e il Leonardismo 52). THIS PAGE Fig. 84 After Leonardo da Vinci, Madonna of the Yarnwinder, oil on panel, 61 × 51 cm, last known location: Paoletti Chelini Collection, Lucca, Italy (Leonardo e il Leonardismo 56).Fig. 85 After Leonardo da Vinci, drawing of the Madonna of the Yarnwinder composition, silver-point, brown and white wash on paper, 159 × 112 cm, Florence, Uffizi Gallery, inv 430 E (Leonardo e il Leonardismo 61). OPPOSITE Fig. 86 Fernando Yáñez de la Almedina, Variant of the Madonna of the Yarnwinder with Joseph, 1523, last known location: Carlos Grether Collection, Buenos Aires, Argentina. THIS PAGE Fig. 87 Attributed to Fernando Yáñez de la Almedina, Madonna and Child with the Infant Saint John, c.1505, oil and tempera on panel, 78.4 × 64.1 cm, Washington D.C., Samuel H. Kress Collection. Image courtesy of the Board of Trustees, National Gallery of Art Washington (Leonardo e il Leonardismo 60).Fig. 88 After Leonardo da Vinci, Madonna and Child (Madonna of the Yarnwinder), oil on wood, 67 × 50 cm, Wittelsbacher Ausgleichsfonds München, Munich (Leonardo e il Leonardismo 58).Fig. 89 After Leonardo da Vinci, Madonna of the Yarnwinder, location unknown (Leonardo e il Leonardismo 57). THE INFLUENCE OF AN IMAGE wears no headdress and her hands and fingers are depicted in the same way. The reverse composition isMANY MADONNAS 205 80 81 82 83 84 85 86 MANY MADONNAS 207 87 88 89 208 Fig. 93 After Leonardo da Vinci, Madonna of the Yarnwinder, c.1510-30, oil on panel, 48.5 × 34.5 cm, Collection of the Duke of Wellington (Leonardo e il Leonardismo 43).Fig. 94 After Leonardo da Vinci, Madonna of the Yarnwinder, Cathedral of Granada, Spain.Fig. 95 Cornelius van Cleve ?, Madonna of the Yarnwinder, oil on panel, 65.5 × 52.5 cm, private collection. © Christie's Images/ The Bridgeman Art Library.Fig. 96 After Leonardo da Vinci, Madonna of the Yarnwinder, oil on panel, 89.6 × 66 cm,Christ Church Picture Gallery, by permission of the Governing Body of Christ Church, Oxford (Leonardo e il Leonardismo 44).Fig. 97 After Leonardo da Vinci, Study of the head of the Madonna of the Yarnwinder, New York, The Metropolitan Museum of Art. Reproduced in Wilhelm Suida, Miscellanea di studi lombardi in onore di Ettore Verga, Milan, 1931 (Leonardo e il Leonardismo 43).MANY MADONNAS 209 210 THE INFLUENCE OF AN IMAGE 95 96 93 94 MANY MADONNAS 211 97 98 1 .1Alessandro Vezzosi, et al, Leonardo e il Leonardismo a Napoli e a Roma, Florence, 1983. 2. See Robert Langton Douglas, A Few Italian Pictures Collected by Godfrey Locker-Lampson, London, 1937, p. 30. 3. Edoardo Villata, ed., Leonardo da Vinci: I documenti e le testimonianze contemporanee, Ente Raccolta Vinciana, Milan, 1999, pp. 218 and 221. 4. Chandler Rathfon Post, A history of Spanish painting Vol. XI: The Valencian school in the early Renaissance, Cambridge, Massachusetts, 1953, pp. 241-4; see also Fernando Benito, Jose Gomez Frechina and Samper Embiz, Los Hernandos, Pintores Hispanos del Entorno de Leonardo, Valencia, 1998, pp. 144-7. NOTES TO PART IV 115 Fig. 111 Luis de Morales, Madonna of the Yarnwinder, 1560s, Madrid, Palacio Real, Copyright © Patrimonio Nacional. April -The head of the Florentine Carmelites, Fra Pietro da Novellara, gains access to Leonardo in Florence via Salaì on behalf of Isabella d'Este. He writes two letters to Mantua informing the patron Isabella d'Este about Leonardo's activities. In the second of these (dated 14 April) he describes the Madonna of the Yarnwinder which Leonardo is painting for Florimond Robertet (see p. 28). June -Pierre de Rohan, Marshal of France (Maréchal de Gié, 1451-1513), begins attempts to obtain from the Signoria in Florence a commission from Michelangelo to produce a copy of Donatello's bronze David. 12 January -the Florentine Ambassador to the French court at Blois reports 'a small picture by his [Leonardo's] hand which has recently arrived here' . Robertet acquires a St Veronica by Lorenzo Costa, now in the Louvre, as a gift from Fran-cesco II Gonzaga. Michelangelo's bronze David delivered to Robertet in France.1507 1508 Infrared reflectogram detail of the Madonna of the Yarnwinder (Buccleuch Madonna), plate 2. Courtesy National Galleries of Scotland and Buccleuch Living Heritage Trust. un caractère de vie, inconnu à presque tous les Peintres qui l'avoient devancé.)Figures .Van der Doort, Abraham, A Catalogue and Description of King Charles the First's Capital Collection of Pictures, Limnings, Statues, Bronzes, Medals and other Curiosities, London, 1757. Vezzosi, Alessandro, et al., Leonardo dopo Milano: La Madonna dei Fusi (1501), exhibition catalogue, Florence, 1982. Vezzosi, Alessandro, et al., Leonardo e il Leonardismo a Napoli e a Roma, Florence, 1983. Vezzosi, Alessandro, ed., 'Parleransi li omini …' Leonardo e l'Europa, dal disegno delle idee alla profezia telematica, with Carlo Pedretti, exhibition catalogue, Museo Ideale, Vinci, 2000. Viatte, François and Varena Forcione, eds., Léonard de Vinci: Dessins et manuscrits, exhibition catalogue, Musée du Louvre, Paris, 2003, pp. 185-9. de Viguerie, Laurence, et al., 'Revealing the sfumato technique of Leonardo da Vinci by X-Ray fluorescence spectroscopy' , Angewandte Chemie, International Edition, vol. 49, 2010, pp. 6125-8. Villata, Edoardo, ed., Leonardo da Vinci: I documenti e le testimonianze contemporanee, Ente Raccolta Vinciana, Milan, 1999. Villon, Frédéric, Notice des Tableaux Exposés dans les Galeries du Musée Impérial du Louvre, vol. 2, Paris, 1849. Waagen, Gustav Friedrich, Works of Art and Artists in England, 3 vols., London, 1838. Waagen, Gustav Friedrich, Treasures of Art in Great Britain: Being an account of the Chief Collectors of paintings, drawings, sculptures, illuminated MSS., &c. &c, 3 vols., London, 1854. White, Sale of Purling Collection, London, 16-17 February 1801. Whitley, William .Ashmolean Museum, University of Oxford: figs. 9, 24PHOTOGRAPHIC CREDITS 237 PHOTOGRAPHIC CREDITS CONSERVATION AND ATTRIBUTION PROVENANCE th by the Dean of Ch. Church who arrived in the nick of time, after we had almost despaired of him. '8 In the PROVENANCE galerie Alte Meister, Dresden, as school of Correggio. APPENDIX Gault de Saint-Germain, Pierre Marie, Traité de la Peinture, de Léonard de Vinci, Précédé de la vie de l'auteur et du catalogue de ses ouvrages, avec des notes et observations, Paris, 1803. BIBLIOGRAPHY McEwen, John, Sunday Telegraph,