Among the most important 15th-century Flemish paintings remaining in private hands, this altarpiece has been a highlight of the Metropolitan Museum of Art’s painting galleries since 1998. In its current state, it reveals one of only a few surviving Renaissance preparatory underdrawings visible to the naked eye. In the following pages, Peter van den Brink presents a compelling argument that the altarpiece is by the great Netherlandish painter, Hugo van der Goes, an attribution first proposed by Claus Grimm. Van den Brink opens by positioning the composition in its historical context, tracing its roots to the innovations of Jan van Eyck and Rogier van der Weyden and exploring its relationship to the work of Hugo van der Goes’ contemporaries in Bruges such as Hans Memling. He next reviews the painting’s provenance, going back to the early 18th century, when the Virgin and Child in the center of the composition were carefully stripped away and repainted with an architectural view of a church interior. At that time, the altarpiece’s four saints were modified to become participants in a new scene representing the Marriage of Henry VII and Elizabeth of York (the figure of John the Baptist was also stripped away). Effectively disguised in this manner, the painting passed through several distinguished collections, including that of the famous writer and connoisseur, Horace Walpole, at Strawberry Hill, England. These alterations were removed when the altarpiece was conserved between 1977 and 1987, thereby revealing the artist’s original composition.
Van den Brink next analyzes several paintings and drawings that other scholars have used to suggest alternative attributions for our altarpiece. He demonstrates that the overall handling of the painting and underdrawing is stylistically distinct from that of a Crucifixion in a Spanish private collection, which is by a follower of Hugo van der Goes, although an alternative attribution to Justus of Ghent has also been suggested. Comparison of the altarpiece with the work of Jean Hey, particularly the early Nativity with Cardinal Rolin in Autun, also reveals that our painting is by a more accomplished hand. The underdrawing of the altarpiece is likewise shown to be more sophisticated than the more mechanical hatching system used in a drawing of the Virgin in Dresden, again attributed to a follower of Hugo van der Goes. Van den Brink dates our picture to the first half of the 1470s, comparing it to Hugo van der Goes’ Adoration of the Magi (the Monforte Altarpiece) in Berlin and to some of his unquestioned drawings. He concludes his discussion by revealing how several aspects of the painted surface, such as the modeling of the faces, the anatomical treatment of the legs and feet, the handling of the drapery, and the use of color and light, are especially close to Hugo’s Portinari Altarpiece in Florence, thus cementing his attribution.
* * *
This painting is remarkable, for more reasons than one. What catches the eye immediately is the absence of paint in two parts of the composition. The central area confronts us with a drawing of the seated Virgin and Child on a throne in a church interior. On her right hand side another figure is partly painted; only his bare legs and the lower part of his cloak are left in paint. The rest of the painted figure, as with the Virgin and Child, was apparently scratched away, leaving only parts of the underdrawing visible. Unlike the Virgin and, to a somewhat lesser extent, the Christ Child, the face of this figure cannot be determined at all; hardly any underdrawing remains. Nevertheless it is very likely that the painter had presented John the Baptist here, probably holding a book. The rest of the composition was left mercifully intact and it represents a so-called Sacra Conversazione, a composition type that was especially popular in Italian painting of the Quattrocento and Cinquecento, but is known in northern painting, too, especially with Van der Goes’ contemporary colleague Hans Memling and with Gerard David, a generation later. As we know, the most famous predecessor in the Netherlands is Jan van Eyck’s Virgin and Child with the Canon Van der Paele in Bruges. As Claus Grimm quite correctly states, both central thrones are placed upon a low stone base, but there is another important link between the two paintings. The play of hands by Mother and Son is intriguingly comparable. In Van Eyck’s masterpiece the Christ Child has just grabbed a small bouquet of flowers from his mother’s fingers, whereas his other hand supports a green parrot on the Virgin’s lap (fig. 1). What exactly takes place in our Sacra Conversazione is less clear-cut (fig. 2). Based on the position of her left thumb and index finger the Virgin Mary seems to hold a single flower or a small bouquet as does Van Eyck’s Virgin, whereas the hands of the young Child enthusiastically reach for this treasure in His mother’s hand.
Another important model is Rogier van der Weyden’s so-called Medici-Madonna in the Frankfurt Städel Museum, where the semi-circular arrangement of the four saints must have held a strong appeal for the painter of our Sacra Conversazione (fig. 3). However, the closest parallels are to be found in the oeuvre of Hans Memling, and these are most certainly contemporary in time, such as the Altarpiece of the two Saint Johns in Bruges. As in our painting, the four saints surrounding the Virgin enthroned – apart from the two protagonists, these are St. Catherine and St. Barbara – are placed in a loggia with a view to the outside world, a construction that was repeated by Memling in the slightly later Triptych of John Donne in London.
Apart from the scratched out John the Baptist, the other saints represented in the present altarpiece are – at left – Saint Thomas, dressed in startling green and purple and holding his attribute, a spear. At right we can make out Saint Jerome, identifiable by his cardinal’s robe and hat as well as the lion. Finally, to the far right, Saint Louis is painted, easily identifiable from the fleurs-de-lis pattern on both his crown and robe. These four saints are placed around the Virgin’s throne, standing on a tiled floor in an open loggia, giving way to a dream-like landscape with trees and hills, beautifully lit. They are separated from the seated Virgin and Child by two columns, while the Virgin’s throne is placed on a carpet on a slightly higher platform. Moreover, the Virgin and Child are placed in an altogether different reality. Unlike the standing saints in their loggia, the divine Mother and Child are presented in a gothic church, a construction that is unique in the history of early Netherlandish painting. In Memling’s Altarpiece of Jacob Floreins in the Louvre from c. 1490, all figures are united before the church interior.
The Virgin Mary of the present altarpiece is wearing a crown that identifies her as the Queen of Heaven, whereas her position on a throne before a church interior – the Temple of God – points towards her position as the Mother of God. The glass vase with Columbines - the area is abraded, partly compromised by restoration – refers to Mary’s Virginity and the presence of the Holy Ghost. The two capitals on the Virgin’s left and right illustrate Esther before Ahasuerus and the Finding of Moses; Esther symbolizes God’s grace towards the Virgin, whereas Moses as God’s chosen one is the precursor of Christ. It is fascinating, as Claus Grimm already stated, that all four saints carry a book in their hands  – even John the Baptist, in a similar way as in Rogier’s Medici Madonna (fig. 3) – marking them as witnesses to a religious event, deeply contemplating the birth of Christianity and the word of God.
The Virgin and Child with Saints Thomas, John the Baptist, Jerome and Louis has had a remarkable history as already can be judged from its present appearance. Before 1977, when the long process of cleaning and restoration by David Bull started, the painting looked wholly different (fig. 4). When the picture was acquired by the London based art dealer Edward Speelman the year before, its curious composition represented The Marriage of Henry VII with Elizabeth of York. The Virgin and Child were covered by a perspective view of a Gothic church, Saint John the Baptist had been transformed into Elisabeth of York, whereas both St. Jerome and St. Louis were adapted to play their new roles as the Archbishop of Canterbury and Lord Chancellor, Cardinal Thomas Bourchier, and King Henry VII. Nevertheless, over the course of the eighteenth- and nineteenth century, the identification of the subject matter was much debated, and so were the attribution and the odd stylistic disparities. Between 1842 and 1977 the painting remained at Sudeley Castle in Gloucestershire, in the collection of John Dent and his heirs. During that long period of ownership the painting was displayed only once at a public exhibition, in 1890, in the New Gallery in London. The exhibition was entitled The Royal House of Tudor and the picture’s formal identification was the reason for its presence in that show.
The catalogue entry is extremely informative on the provenance history of the picture, at the time attributed to Jan de Mabuse (Jan Gossart), quoting Horace Walpole (1717 - 1797) from the first volume of his 1762 Anecdotes of Painting in England (see Literature). Indeed, Walpole owned the picture and had it displayed in his collection at Strawberry Hill in Twickenham, until 1842, when it was sold with the rest of the collection. Walpole was extremely proud of the painting, calling it “[...] a celebrated picture in my possession” and stating that “the Earl of Oxford once offered £ 500 for it”. In addition Walpole noted that the picture was owned previously by Henrietta Louisa, Countess of Pomfret († 1761), acquired for £ 200 and “hung for some years at their seat at Easton Neston in Northamptonshire, whence it was sold after the late Earl’s death”. As is known from Walpole’s published correspondence, he was well acquainted with the collection of Thomas Fermor, the 1st Earl (1698 - 1753) and the Countess of Pomfret and he must have seen the painting during one of his visits to Easton Neston. In effect, Walpole acquired the painting at the estate sale of Lord Pomfret, in 1754. The picture was offered on the second day of the auction, as lot 53, curiously attributed to Zucchero , as The Union of the Houses of York and Lancaster. Walpole bought it for £ 88.4.0. At the immense Strawberry Hill sale in 1842 the painting fetched a much higher price. On the 21st day of the sale it was bought by Bevan, a dealer, for John Dent, for £ 178.10.0.
However, the 1890 catalogue entry does not mention how the painting became the topic of an agitated discussion between Horace Walpole and Andrew Coltée Ducarel in three letters, written on February 23, 24 and 27, 1762, shortly after the publication of Walpole’s first volume of Anecdotes of Painting in England, in which, as stated before, the painting was referred to in detail. The content of these letters is of the utmost importance for the history of the painting in its disguised, overpainted condition, because for the first time doubts were being expressed on its visual integrity.
Ducarel (1713 - 1785) was born in Paris from Huguenot parents from Normandy and was trained as a “civilian”, a lawyer practicing civil law, but he was more interested in books and antiquities. Ducarel was regarded as a parvenu by some, including Horace Walpole, who held a very low opinion on him. Nevertheless Walpole had sent the first volume of the Anecdotes of Painting in England as a gift to Ducarel, who showed his gratitude for it in the opening sentence of his first letter, although he makes clear that he had prepared two papers, one of them containing some critical remarks concerning Walpole’s important publication. One remark concerns the painting with the Marriage of Henry VII and Elizabeth of York. Ducarel calls the remark an anecdote himself, communicated to him by George Vertue on January 28, 1754. Vertue (1684 – 1756) was an English antiquary and he was famous as an engraver, who received many commissions from noble families. In addition, he was a connoisseur and member of the famous Rose and Crown Club. Walpole acquired Vertue’s notebooks after the latter had passed away and these notebooks were the proper foundation for Walpole’s Anecdotes of Painting in England.
As is clear from Walpole’s published correspondence, he and George Vertue corresponded very often and did not always share the same opinions. In fact, Vertue’s posthumous ‘anecdote’ on Walpole’s picture was nothing less than a devastating blow to the painting’s quality and integrity: “That Lord Pomfret bought this picture of one Old Sykes above 30 years ago, which Sykes dealt in pictures and was a noted tricker—that he (Sykes) gave it that name, well knowing how to give names to pictures to make them sell [...] that upon the strictest examination Vertue could never be convinced that the man was Henry VII, the face not appearing to him like any of the pictures he had seen of that king—that as to the woman, she had pomegranates upon her clothes, which certainly did not belong to her—that the church in which they are married, as represented in the picture, did not appear to be any English church, and that, upon the whole, it was suspected, at the time that Lord Pomfret bought it, that Old Sykes, who was a rogue, had caused the figures and representation of the marriage, to be added to the representation of the inside of a church, Old Sykes having before been guilty of many pranks of that sort.”
Walpole reacted as if he was bitten by a snake: “I am much amazed at Vertue's blunders about my marriage of Henry VII. His account is a heap of ridiculous contradictions. He said, Sykes knowing how to give names to pictures to make them sell, called this the marriage of Henry VII and afterwards, he said, Sykes had the figures inserted in an old picture of a church. He must have known little indeed, Sir, if he had not known how to name a picture that he had painted on purpose that he might call it so! That Vertue on the strictest examination could not be convinced that the man was Henry VII not being like any of his pictures. Unluckily he is extremely like the shilling, which is much more authentic than any picture of Henry VII—but here Sykes seems to have been extremely deficient in his tricks: did he order the figure to be painted like Henry VII and yet could not get it painted like him, which was the easiest part of the task? Yet how came he to get the Queen painted like, whose representations are much scarcer than those of her husband? And how came Sykes to have pomegranates painted on her robe, only to puzzle the cause? It is not worth adding, that I should much sooner believe the church was painted to the figures than the figures to the church. They are hard and antique; the church in a better style, and at least more fresh. If Vertue had made no better criticisms than these, I would never have taken so much trouble with his MSS. Adieu! Sir.”
Suffice to say that poor Ducarel, being only the messenger of bad tidings, made clear in his reply of 27 February 1762 that he wholeheartedly agreed with Walpole’s view of the matter: “to the remarks contained in my last, if any of them have given you the least uneasiness, I am very sorry for it—Vertue's note about your picture of Henry VII I sent you just as Vertue gave it to me—for I was so far from laying any stress upon it, and from believing it not to be Henry VII's marriage that I went twice to Easton on purpose to see that picture, and was long since convinced that it is not only what you say, but likewise one of the finest English historical picture[s] I ever yet beheld.”
Vertue’s posthumous critical remark and Walpole’s fascinating reaction on it provides us with a lot of new information of great importance. First and foremost, both Vertue and Walpole felt that the picture was not the result of a single creative process. Whereas Vertue thought that the figures were added to the church interior, Walpole on the contrary expressed his feeling that the gothic interior was of a younger date. Now we know that both men were right (and wrong). The second bit of information Vertue brought into play, apparently did not surprise Walpole in the least. His overly angry reaction shows that he was well aware of the doubtful role William Sykes had played in the process, but decided not to mention him in the heroic tale of his picture in the first volume of his Anecdotes of Painting in England, hence his ridicule with regard to Sykes in his letter to Ducarel.
Following Ducarel’s message of Vertue’s ‘anecdote’, the Earl and Countess of Pomfret acquired the painting directly from the painter and art dealer William Sykes (c. 1659 – 1724) and therefore must have kept the picture in their possession for at least thirty years. Sykes may have had a doubtful reputation, as Vertue claims, since he was known to have committed comparable treacheries and dealt in forgeries. Vertue most certainly had personal knowledge of William Sykes’ activities as a painter and a dealer, since they both held a membership of the Virtuosi of Saint Luke, a rather distinguished society of gentlemen, artists and art lovers. In addition, Sykes was known as a portrait painter; a (posthumous) portrait of Mary, Queen of Scots, in Kedleston Hall, Derbyshire shows his abilities in that field. That painting is a copy after the original portrait, by Federigo Zuccari, now in the collection of the Duke of Devonshire. When the painted portrait is compared to the painted figure of Elizabeth of York (fig. 4), it seems highly likely that Sykes was indeed responsible for adding that figure in the painting and for adding the gothic church interior in the center of the composition, no doubt based on seventeenth-century Flemish examples, such as those from Pieter Neeffs or Hendrick van Steenwijck. Moreover, the lead structure of the windows that he had painted over the landscape behind Elizabeth of York on the left and behind the transformed Saint Jerome on the right are almost identical to the pattern seen in his portrait in Kedleston Hall, and is directly based on the model of Zuccari.
In the only critical publication on the painting during its disguise as the Marriage of Henry VII with Elizabeth of York – the review of the 1890 exhibition in the Gazette des Beaux-Arts (see bibliography) – Claude Phillips criticizes the central part of the picture as “ […] une perspective d’église nue et froide, comme en peignaient les Steenwyck et les Saanredam, page curieuse dont le ton blafard et l’exécution sèche jurent cependant avec le coloris puissant et magnifique de l’oeuvre primitive”. Moreover, Philips had discovered something else; he was able to make out the appearance of the Virgin and Child underneath this church interior: “Au milieu, se voyait, sans aucun doute, le groupe traditionnel la Vierge avec l’enfant”. It would last almost a full century, before the picture became the subject of serious research and by then its outlook had changed dramatically.
Shortly after he had acquired the painting, Edward Speelman asked David Bull to take up the task of removing all the overpaint and to restore the picture. In the first stage of the process, in 1977-78, Bull removed a later board that had been added by William Sykes in the central portion of the picture and rejoined the original panel. In addition, he removed huge areas of overpaint with a scalpel under a microscope, revealing a completely different composition underneath the eighteenth-century overpaint. Only the faces of Saints Thomas, Jerome and Louis, their hands and attributes, as well as the red and white cloak of Saint Jerome, the beautifully painted Gothic decoration on the upper left and right and large parts of the landscape had remained untouched by William Sykes.
Only several years later, shortly after David Bull stepped down as director of the Norton Simon Museum in Pasadena in 1981, he finished retouching the painting, leaving the unpainted parts of the picture untouched, along with the green cloak of Saint Thomas that apparently had suffered from the wrong use of solvents, as was stated by Claus Grimm in 1988. After he had finished retouching, the picture returned to London, where it was bought by the present owner in April 1986. The last stage of the restoration project took place in 1986-1987 and included, among other things, the removal of the overpaint on the cloak and the crown of Saint Louis, since these were still there when the picture was acquired in 1986, as can be judged from the only known color image from the pre-1986 condition (fig. 5).
Two years later, in 1988, Claus Grimm was the first to publish the altarpiece in full. Grimm stated that the painting should be regarded as a work by Hugo van der Goes, placing it shortly after his earliest paintings, such as the Virgin and Child in Frankfurt, the Fall of Man in Vienna (fig. 6) and the Virgin and Child with Saint Anne in Brussels (fig. 7), but before the Monforte Altarpiece in Berlin (fig. 8). A date around 1470, as suggested by Claus Grimm, does accord with the dendrochronological analysis that was carried out by Peter Klein, establishing a terminus post quem of 1466. Grimm’s attribution was criticized by Jochen Sander in his 1992 monograph on Hugo van der Goes. Although Sander stated that the painting must have been made shortly after 1475, he argued that it could not be attributed to the master himself, but that it must have been painted by Jean Hey, the Master of Moulins, who must have been active in Hugo van der Goes’ workshop during the second half of the 1470s. However, Sander’s case for Jean Hey was severely criticized by Lorne Campbell in 1994: “Though Sander has the reputation of being a reliable connoisseur, his assertion that a Virgin and Child with saints, published in 1988 by Claus Grimm as the work of Hugo, is an early work by the Master of Moulins undermines faith in his judgment. This interesting picture, partially stripped to its underdrawing by an early eighteenth-century dealer, may be related not to Hugo or to the Master of Moulins but to a Crucifixion with saints and donors which is in a Spanish collection and which has been attributed to Justus of Ghent or, more plausibly, to a gifted follower of van der Goes.” Dirk de Vos re-attributed the present altarpiece to Hugo van der Goes, but contrary to Claus Grimm, regarded the painting as a late work, in which he firmly recognized influences from Hans Memling. Neither Elizabeth Dhaenens nor Molly Faries would discuss the matter of attribution and Maryan Ainsworth, in her publications of 1998 and 2003, avoided the subject of attribution as well, although in her 2003 review of the exhibition of Early Netherlandish drawings in the Rubenshuis in Antwerp, she called the painter “a close follower of Van der Goes”. More recently, however, Maryan Ainsworth spoke out more clearly on the topic of attribution in a lecture at Symposium XIX for the Study of Underdrawing and Technology in Painting, in Bruges, 11-13 September 2014, giving the painting to a close follower of Hugo van der Goes. Although the lecture was transformed into an article, it has not yet been published.
Although the present altarpiece has been on the wall of the Metropolitan Museum of Art since 1998 and the underdrawing of the Virgin and Child as well as Saint John the Baptist has been accessible to every visitor and scholar interested in that topic, it was only recently examined and documented with the aid of infrared reflectography (fig. 9). The visible underdrawing of the Virgin and Child and Saint John the Baptist is equaled by the rest of the underdrawing. Indeed, it is a very detailed and worked-up underdrawing in pen and brush that is followed in general in the covering paint layers, indicating that the painter did not depart from the original lay-out of the composition, at least in those areas that are covered with their original paint layers. Certainly, there are various details that have ultimately been executed differently than in the first set-up. The feet of Saint Thomas and John the Baptist were somewhat smaller at first; in the final paint stage they were elongated and enlarged, whereas one of the poulaines of Saint Louis was repositioned, as can be seen in the IRR mosaic (figs. 10-11). The position of Saint Jerome’s adorable lion was shifted as well, whereas the incense burner in the foreground was somewhat smaller in the first lay-out, and positioned differently. The same is true for the flower vase on the left foreground, but this area, as can be judged from both the IRR mosaic and the X-radiograph, is partly compromised and is in part the result of David Bull’s reconstruction. This extremely detailed, refined and very skilled drawing seems to imply the use of a model, probably a model drawing, perhaps even a presentation drawing. The underdrawing of the landscape, however, shows a much freer set-up.
A comparable setup can be witnessed in the Crucifixion, now in a private collection in Spain. As stated before, Lorne Campbell observed a close resemblance between this painting and the present altarpiece. On the surface there are certainly similarities, such as the facial types of Saint James or the Magdalen, which do coincide with those of Saint Thomas and the Virgin in the present altarpiece. However, the execution of the painting in Spain is quite distinct from our altarpiece, not only in the subtlety of the faces, but in the brocades and decorative elements as well, such as the cross of Saint Jerome. The underdrawing of the Crucifixion is at first glance very similar to that of the present altarpiece. A very detailed lay-out with carefully placed parallel- and cross-hatchings, combined with a very free underdrawing of the landscape, follows a similar path. Like the execution in paint, however, a comparison of the drawing of the Virgin’s face in the present altarpiece and the underdrawn face of Saint Catherine in the Crucifixion, or any other of the faces for that matter reveals such a difference in style that it is impossible to view the Crucifixion as a youthful work of the same artist; it can only be regarded as the work of a different, yet still very talented painter. However, because of the superficial affinities with the present altarpiece, both in facial types, color of the palette and underdrawing method, I would regard the painter of the Crucifixion as a follower of the other, more accomplished artist.
Jean Hey, or the Master of Moulins, as Lorne Campbell pointed out, cannot be the painter of the present altarpiece. The earliest known painting by Jean Hey, the Nativity with Cardinal Rolin from c. 1475-80 in the Musée Rolin in Autun (fig. 12), diverges from the present altarpiece in its overall cool tonality, especially in the typical pale, almost wax-like, flesh colors. The brightly lit landscape of the Autun painting is reminiscent of both the present work and the Berlin Monforte Altarpiece, though. The underdrawing of this early work by Jean Hey, however, published recently by Martha Wolff , is characterized by carefully rendered strokes of parallel- and cross- hatching, done with the brush, applied with a lot of variation, to indicate folds or shadows. As Wolff states: “For this range of strokes and emphasis on light and shade, Hugo van der Goes’ early underdrawing process as discernible in the Monforte Altarpiece may have provided a model.” This may be so, but the hatching system in the underdrawing underneath the figures of the Berlin Adoration of the Magi is much more powerful and complicated than that of the young Jean Hey, with hatching strokes, going in every possible direction, as can be seen in the underdrawing of Joseph’s cloak, for example (fig. 13). The underdrawing of the present altarpiece, too, is not dissimilar to that of the Autun painting, especially in the softly applied, refined hatchings in the faces of the figures, or the great variety of hatching strokes to define light and shade in the clothing. They differ, however, as well. The hatching system in the underdrawing of the present altarpiece, as in the underdrawing of the Adoration of the Magi in Berlin, is much more complicated and developed, with an emphasis on islands of parallel hatching going in every possible direction, molding the various drapery folds and thereby creating a greater plasticity.