Chapter 12

After the Original: Hollar and Leonardo’s Salvator

The Bohemian artist Wenceslaus Hollar’s etching Salvator Mundi, dated 1650 in the plate,¹ is of signal importance, since it provides the first clear association of the name of Leonardo with the canonic typology.² It is inscribed in the bottom margin of the print: ‘Leonardus da Vinci pinxit, Wenceslaus Hollar fecit Aqua forti, secundum Originale, A^o 1650’ [‘Leonardo painted it. Wenceslaus Hollar made it in etching, from the original, in the year 1650’]³ (Fig. 1.3). The implication that, in 1650, Hollar directly copied a painting he considered to be Leonardo’s original, raises a number of questions. First, why did he attribute this composition to Leonardo? Second, where did he copy the model for his etching, published in Antwerp? Third, which of the known versions of this composition does Hollar’s etching most resemble? Finally, can we pinpoint the location of any of these versions in 1650?

Hollar was in the vicinity of the Caroline collections before, during, and after the Civil War and Interregnum. John Evelyn signalled his appreciation of Hollar’s etchings after Arundel’s collection of drawings by Leonardo.⁴ It is a tantalizing thought that Hollar may have had contact with the two preparatory drawings of the sleeve and the chemise with the crossed bands, then still bound into the large folio of drawings (the Windsor Volume) we believe was in Arundel’s hands during the 1630s. We know very little of how paintings were attributed to painters during this period in England, as also elsewhere. Were there labels, hand-lists? How important was oral tradition? How did paintings retain their identities as collections were deaccessioned and transferred across regional and national boundaries? What was thought to be a Leonardo in the age before mechanical reproduction and during the second major dispersal of his writings after Leoni’s death? In London, Arundel’s collection of Leonardo’s drawings and writings surely fostered and informed the understanding of the artist’s holograph, as did the, admittedly limited, first-hand contact with his works that the king, Arundel, Jones, Hamilton, Buckingham, Evelyn, and Gerbier experienced on the Continent.

This brings us to the question of the versions and variants of Leonardo’s canonic Salvator Mundi. Robert Simon has unearthed around twenty-six versions to date, an increase on Trutty-Coohill, who in 1982 identified nine versions, following Heydenreich, who in 1964 published seven.⁵ There has been a tendency to assume that, because we now know of numerous extant versions, there must have been many in circulation at the beginning of the seventeenth century and the composition must have been well known. This is not the case, and until a comprehensive comparative scientific analysis of these versions is undertaken, we know little of their age, origin, or status.

Hollar’s known movements before 1650 suggest contact with a model in either England or the Low Countries; Snow-Smith hypothesized that he encountered the painting in France. We shall review the historical context of Hollar’s generation of the etched copy after Leonardo’s composition, focusing on the paintings recorded in the Royal Collection, the Hamilton collection, and the painting in the collection of the Marquis de Ganay at Paris during the twentieth century. We will consider whether Hollar was characteristically faithful to a model, and which of the many versions most closely correspond to his etching. Visual comparison between the strongest candidates and the Hollar etching will focus on (1) the iconographical props of the composition, that is, the orb, halo, orphreys or bands, drapery, beard, and the ‘omega tuck’ evoking the wound of Christ; (2) the integral formal qualities of execution and spatial organization; (3) the material presentation of each painting, the history of its condition; and (4) congruence or divergence between the putative prototypes and Hollar’s etching.

Consonances and Dissonances

Hollar’s etching demonstrates a close affinity with only three of the twenty-six known versions: the painting formerly in the collection of the Marquis de Ganay (Plate 8), the Cook painting, and a panel documented since 1666 in the Capella Muscettola in San Domenico Maggiore, Naples, largely inaccessible in an elevated marble niche until its first museum exhibition in 2017, but since 1987 attributed to Girolamo Alibrandi (Fig. 5.5).⁶ This painting is highly consonant with the etching, except that it lacks a beard and prominently exhibits the so-called ‘omega tuck’. Since it has apparently remained in the same position since the mid seventeenth century, it is unlikely to have been Hollar’s prototype. Apart from the red colour of the chemise, the position of the thumb of the blessing hand, and a corkscrew curl resting, somewhat inelegantly, on Christ’s proper left shoulder, the Naples version, which is on walnut, corresponds very closely in every detail—the knot work of the crossed bands, the crystal orb, the lack of a full beard—to the Cook painting. Dianne Modestini points out that ‘this makes the genesis of the Naples panel particularly interesting.’⁷ Before closer scrutiny facilitated by the 2017 exhibition, it was not possible to confirm its features. In its present state, the handling of key passages, such as the curls of hair and white drapery around the wrist, appears formulaic. At the time of going to press, scientific analysis is in progress. This version apparently lacks the central jewels of the neckband and stole, but since the state of conservation and the presence of any underpainting is not ascertained, it is possible that they are currently obscured. As such, this study focuses on a comparison between the Ganay and the Cook paintings as the two most forensically examined panels.

Ganay

Until its publication in 2011, the whereabouts of the Cook version were unknown, the only witness of its existence the c. 1911 greyscale photograph in the Witt Collection at the Courtauld Institute, which shows the painting in a grotesquely overpainted state (Fig. 1.13). The recent conservation and process of attribution of the Cook painting, its acceptance as an autograph work by Leonardo by a majority of the principal specialists, and its subsequent exhibition at the National Gallery, London, in 2011 has produced renewed comparison with the Ganay painting, whose claim to authenticity, at least as a studio version involving Leonardo, long championed by eminent Leonardo scholar the late Prof. Carlo Pedretti and Joanne Snow-Smith, has not met with academic acceptance. Authentication of works of art is an opaque matter, usually contentious, often bad-tempered, and sometimes litigious. In the interests of transparency, there follows a summary of the established provenance, technical analysis, attribution history, and exhibition history of the Ganay painting.

The Comtesse Martine Marie Pol de Béhague (1870–1939) owned four drapery studies attributed to Leonardo and two important manuscripts related to his writings—MS Ganay, Nicolas Poussin’s apograph of the 1651 treatise on painting, and an album of sketches attributed to Rubens for a study of the ‘Theory of the Human Figure’, published posthumously in Paris in 1733, containing copies of Leonardo’s drawings.⁸ In 1902 the countess purchased a Salvator Mundi painting from the Parisian dealer Bourdariat.⁹ He had acquired it from Clément Gustave Henri de Baillardel, baron de Lareinty (1824–1901), who purchased it from the Convent of the Cordeliers in Nantes, according to Snow-Smith and Pedretti, following a tradition of the De Ganay family.¹⁰ Snow-Smith hypothesized that the Salvator Mundi was placed in the ‘Convent of the Sisters of St Claire’ by Louis XII on the death of Anne of Brittany; therefore the provenance of the Ganay painting could be traced back to this location, and before that, to the French royal collection.¹¹ She proposed that Hollar copied the Salvator Mundi in the convent in Nantes in 1650 while in the train of the exiled Queen Henrietta-Maria. Since no Cordelier convent existed in Nantes in 1650, her conjecture that the Salvator Mundi was located there from 1514 until the nineteenth century is improbable, as is the proposal that Hollar was in the employment of Henrietta-Maria in 1650.¹² Nor is there any evidence that Hollar was ever in France; he left no drawings of French topography or subject matter.¹³

It is feasible that the painting was indeed purchased in Nantes—the baron was general councillor of the town and could have acquired the painting from the art market, since, as the principal creditor of Pommeraye et C^ie, he became in 1849 the sole proprietor of the famous covered arcade of antique shops, the Passage Pommeraye, at Nantes; it was sold by his widow in 1901.¹⁴ Given the proximity of dates, it is almost certain that Bourdariat acquired the painting at the sale of Lareinty’s estate.¹⁵

In 1939 the Marquis Hubert de Ganay inherited the painting from the comtesse, his aunt. In a letter of 1972, he recounted to Snow-Smith the opinion of the prominent American art historian Bernard Berenson, who, while a house guest of the marquis, ‘glanced at the painting on the wall’, rejecting it as a ‘poor copy by an inferior pupil of Leonardo’.¹⁶ The countess and Berenson were well acquainted, so it was presumably not the first time he had expressed this view, although the family maintained that she never had the painting authenticated.¹⁷ Kenneth Clark was also a frequent visitor to the Château de Courances at Fontainebleau, where the painting hung in the library. He was first (in 1935) to connect the two Windsor drapery studies (Plates 6 and 7) with this composition,¹⁸ but his opinion of the Ganay painting is not recorded. It is a fascinating irony that, as Robert Simon observed in Chapter 1, Lord Clark also disregarded the Cook painting, which he would have seen at the auction of the residue of the collection in 1958. It seems that the only supporters of the Ganay painting during the mid twentieth century were the marquis himself and an important Leonardo collector who saw it in the spring of 1952: Elmer Belt, a surgeon whose notable Leonardo library is now at the University of California in Los Angeles.

Belt’s eloquent and affecting description of his encounter with the Ganay painting was published in 1980, on the occasion of its first public exhibition, in Vinci.¹⁹ He saw it in Paris, en route to Vinci, where he was to be made an honorary citizen on the occasion of the quincentenary of Leonardo’s birth. It was an experience that, for him, clearly reverberated down the years, culminating in the enterprise of 1980. His recollection, although lengthy, is worth recounting in full, because it illuminates a key moment in the candidacy of the Ganay painting and the chronology of its promotion.

After viewing the works of art in various rooms, we came to the library—a memorable experience. I remember even the smell of the precious wood of the furniture, the smell of the leather bindings. Here was the famous library of the Countess de Béhague, the library that she had entrusted to a young librarian who was to become one of the greatest poets of France, Paul Valery, whose 1894 essay on Leonardo’s method marked the beginning of his career.

In a corner of the library, in a sort of Leonardo light filtered through the curtains of the window, there suddenly appeared to us the glowing image of the ‘Salvator Mundi’ a painting I did not even know—in fact, a painting that I did not even know existed. A version of the same subject in the Institute of Fine Arts at Detroit was brought by Valentiner to Los Angeles in that very year, for an exhibition at our County Museum on the occasion of the Leonardo Centennial. The several known versions of that painting were later studied by Professor Heydenreich, to whom the de Ganay version, superior to all the others, remained unknown. The photograph that the Marquis gave me on that occasion attracted the attention of Carlo Pedretti, who published it in his 1973 book on Leonardo together with Leonardo’s preliminary studies at Windsor. He dated those studies to about 1510–15, and attributed the painting to a pupil, possibly the author of a rather repulsive effigy of an old man, also at Windsor, which shows the same type of hair as the Christ in the painting. [RL,12501] Later his student Joanne Snow-Smith returned to the subject with Professor Heydenreich, approaching it from the point of view of documentation, iconography, and laboratory examinations in order to propose an attribution of the painting to Leonardo, identifying it with the one recorded in the Fontainebleau inventory of 1642 and engraved by Hollar in 1650.

In front of that painting the Marquis confessed to me that the whole family had always entertained a certain scepticism towards the enthusiasms of his aunt, who had acquired it in 1902 and had always been convinced, to the very end of her life, that it was Leonardo’s lost original. Perhaps one of the few disappointments in her life, otherwise so full of glamour and success, was brought to her by her friend Bernard Berenson who, when a house guest in her house, hardly glanced at the painting to dismiss it quickly as a pupil’s work.

Here the Marquis smiled, as if to intimate that he had already enough Leonardo material without having to force the hand of scholars who, sooner or later, would have to reconsider that painting.²⁰

Carlo Pedretti, according to his publication of 2017, considered the painting to be by Leonardo himself, with the assistance of a pupil.²¹ According to Snow-Smith’s account, it was Pedretti who, in January 1972, first brought it to her attention.²² In 1974 she shared her research with Prof. Ludwig Heydenreich, who had published a pioneering article about Leonardo’s presumed-lost original Salvator Mundi a decade earlier.²³ He congratulated Snow-Smith on her conclusions, and two years later examined the painting at Paris, when he ‘evaluated [the Louvre’s] evidence definitively linking the de Ganay Salvator Mundi with the Saint John the Baptist in that museum’, confirming her conclusion that it was by the hand of Leonardo.²⁴

The painting was included in a touring exhibition of the Comtesse de Béhague collection of Leonardiana curated by Carlo Pedretti, staged in Italy and the United States, 1980–81. The countess’s collection was auctioned in Monaco in 1989;²⁵ the Salvator Mundi, attributed to ‘circle of Leonardo’, was sold at Sotheby’s New York on 28 May 1999 as lot 20 for a $332,500 hammer price, against an estimate of $80,000–$120,000. In 2013 the painting was exhibited at Sao Paolo, where Pedretti catalogued it under the name of Leonardo.²⁶ In 2017 it was displayed at the Museo Diocesano, Naples, in an exhibition curated by Pedretti, where the painting was presented as a masterpiece of Leonardo and his immediate workshop.²⁷ Display panels explicated aspects of the history and research into the Ganay painting. The Cook painting was illustrated in the exhibition catalogue by the early-twentieth-century black-and-white photograph and not according to its modern appearance. Pedretti’s article in L’Osservatore Romano, 12 July 2011, ‘Se Leonardo é una chimera’, denounced the new attribution of the Cook painting, which had leaked into the public domain the previous month. An illustration of the Ganay painting was captioned: ‘The best version of the School (probably Giampietrino) of the same subject’.²⁸ In the 2017 exhibition catalogue Pedretti criticized ‘the alarming clouding of judgment which has led to an attribution to Leonardo of one of the many versions of the canonic Salvator Mundi, and this, after a restoration that is far from reliable, and what’s more, based on undisclosed presumed laboratory analyses.’²⁹ He did not mention Dianne Dwyer Modestini’s 2014 article, which presents the conservation treatment and scientific analysis.³⁰ Pedretti, citing Snow-Smith, repeatedly linked the Ganay painting with a 1642 Fontainebleau ‘inventory’.³¹ As we have seen, no such royal inventory exists; the reference is to Pierre Dan’s ‘Christ à demy corps’, attributed to Leonardo, a painting connected to the Leonardesque Young Christ at Nancy.

Condition and Congruence

Condition is a critical concern for any visual comparison of the Cook and Ganay paintings with the Hollar etching. In 1982 Snow-Smith pointed out that the embroidered pattern on the crossed bands and neckband is broadly consonant between all three items. However—and understandably—given, as she pointed out, that she had not been able to inspect the Cook painting, she contended ‘that the Cook painting could not have been the prototype of Hollar’s etching is evidenced, however, by the fact that there is no jewel in the center of the stole nor an accommodation in the orphrey for the insertion of one; the embroidery merely continues the hexagonal pattern without interruption whereas Hollar includes a change in the pattern around a mounted jewel. Moreover, the Cook copy includes only cross-hatching in the neckband while Hollar makes the embroidery of continuous filaments, as does the Ganay Salvator Mundi.’³²

During the cleaning of the Cook painting undertaken by Dianne Modestini, the jewel at the centre of the crossed stole was revealed, and the elegantly constructed three-dimensional knot work emerged from beneath layers of overpaint. Comparison of infrared reflectography of the two paintings conclusively demonstrates the pre-eminence of the Cook painting.³³ The tracery of the interlocked filaments in the Cook orphrey is refined, carefully modelled, and complex in design. The composition appears in places to be underpinned by a scaffold of spolveri and incisions, indicating transfer from a cartoon.³⁴ There are revisions in the positioning of both of the hands, signalling that the composition was developed in progress. The embroidered bands, the yoke of the chemise, and the hair are underpinned by fully realized monochrome (grisaille) brush drawings.³⁵ Light and shade are confidently modelled, and minor revisions are apparent here and there. In contrast, the underdrawing of the Ganay painting is a summary exercise, made by means of transfer of a tracing of a line drawing, with no attempt to model the form; the hallmarks of the reproduction of an extant prototype (Fig. 6.2).

Visual comparison confirms several divergences between the Hollar etching and the Ganay and Cook panels, some of which are common to both and some of which occur only in one painting. Hollar’s etching does not reproduce the ‘omega tuck’ in the drapery of the chemise, apparent in both paintings but much less obvious in the Cook painting and invisible until the recent cleaning. The Ganay painting exhibits the ghostly outlines of a globus cruciger and the crescent form of a halo surmounting Christ’s head. These components were probably gilded; the spectral halo is incised with lines and traces of decorative diapering. It is not known exactly when these elements were introduced, whether at the time of the picture’s generation or afterwards. Some white highlights, in imitation of reflections, are painted over the remnant of the globus cruciger, perhaps in emulation of the Hollar etching. The Cook painting does not feature a cross-bearing orb; the viewer’s focus is directed to the forensic portrayal of the inclusions and refracted lights within the rock crystal sphere, a wondrous object fashioned from a material of sacred status in the Renaissance.³⁶ Hollar imitated the unadorned rock crystal orb of the Cook painting, rendering it with granular attention, taking care to distinguish the interior lights of the gem from the sheen of glass; the surface reflections are muted.

The etching diverges from both paintings in the treatment of the halo; a nimbus of light emanates from behind Christ’s head. Leonardo sometimes depicted halos as floating semi-transparent discs in perspective, the ellipse delicately circumscribed by a golden line.³⁷ The Cook painting, in its restored condition, does not present any type of halo, whether the type occasionally favoured by Leonardo or a subtle emanation of supernatural light, so it is not possible to ascertain whether Hollar depicted the nimbus in this way in fidelity to a prototype or for some other reason. We should be mindful of the condition of Hollar’s model, which may have been around 150 years old when he copied it. In conservation, a layer of dark green paint was removed from the background of the Cook painting, so it is not possible to determine whether it once presented a penumbra corresponding to the features of the 1650 etching.³⁸ The date of the addition of the globus cruciger and patterned halo to the Ganay painting has not been determined, although modern methodologies of technical analysis could provide information, including a reconstruction of the incised patterns of the halo, which may offer insights to the painting’s early provenance.

That said, Hollar’s understated nimbus agrees in sensibility with the eloquent materiality of the crystal sphere. It chimes with Leonardo’s characteristic play on the properties of light, since it comprises an exploration of the natural light falling from Christ’s frontal right, the emanation of supernatural light from Christ, confounding the physical laws governing optics, and an analysis of refraction and reflection in the crystal orb. If the Hollar etching does indeed reproduce Leonardo’s original conception, it provides insights into the artist’s instinctive manipulation of the standard apparatus of Renaissance iconography, or perhaps a particular theological context governing the painting’s generation.

This returns us to the correspondences between the respective paintings and Hollar’s etching. Although the abrupt termination of Christ’s sleeve at the left margin might suggest that the painting has been cut down on this side, this is not the case. The lower left margin of the unpainted wood is at a slight angle—it is half a centimetre narrower at the lower end of the panel—but the trimmed edge nowhere intrudes on the paint surface.³⁹ Nor has the bottom of the paint surface been trimmed where the edge slices across Christ’s knuckles. The unexpected severing of forms at the edges of a picture is a strategy deployed in the near-contemporary Mona Lisa, where slivers of framing columns defy orthodox pictorial conventions, perplexing generations of copyists.⁴⁰ Hollar carefully respects the margins of the Cook picture. By contrast, the Ganay version, which is 2.9 cm wider and 2.6 cm taller, includes more of the sleeve on the left and displays a small area of drapery below the knuckles on the bottom right. This provides the strongest argument that the Hollar was working from the Cook panel.

The thumb of the blessing hand presents another interesting comparison. During cleaning of the Cook painting, a significant pentimento was found in the underdrawing. An alternative position of the digit had been partially sketched, in which the pad of the thumb faces the viewer, contrasting the more realized version, where the thumb is rotated through 90 degrees to show the thumbnail in profile (Plate 4). This presented a dilemma for the conservator, Dianne Modestini. Eventually, in consultation with National Gallery curatorial staff and the owners, she chose to suppress the pentimento; the thumb was presented in profile.⁴¹ The Ganay painting depicts the other position, where the thumb pad is shown. The Hollar etching depicts an awkward sausage-like thumb in profile, with the nail visible.

The beard presents the most obvious point of divergence between the two compositions. In Hollar’s etching and the Ganay painting, Christ wears a full beard and moustache, whereas the Cook painting presents at most a light, almost imperceptible growth. As noted above, only the Cook and the Naples panels lack full beards. The etching depicts a full beard that divides into two spirals of curls at the cleft of the chin. The Ganay panel exhibits a mass of small curls, forming a neat, but abundant, beard, surmounted by branching tendrils on the upper lip, individually delineated, as are the lower and upper eyelashes, eyebrows, and wisps of hair framing the forehead. Dianne Modestini considers it is doubtful whether the Cook painting ever had a fully realized beard. This sets up a dissonance between Hollar’s fulsomely hirsute Christ and the Cook panel’s five-o’clock-shadow.

In conclusion, while Hollar’s etching is very similar, it is not an exact reproduction of either the Cook or the Ganay painting. It presents congruence of only one or other of the principal compositional elements. To summarize: spatial organization (neither, exactly, but closer to Cook), sphere (Cook), crossed bands (Cook + Ganay), halo/aura (neither), omega tuck (neither), thumb (neither, exactly), and beard (Ganay, inexactly). On balance, Hollar’s etching more closely reproduces the Cook painting. Above all, the correspondence of the margins of the Hollar etching to the intact paint surface of the Cook painting strongly supports its dependence on the latter. We should not look for perfect alignment of details, since this is not a characteristic of Hollar’s copies after Leonardo. Richard Godfrey cautions that ‘it is a mistake to regard [Hollar’s] work as wholly faithful to the originals. This dictum may be applied to his work in general. … this most rational-minded of artists was moved to improve designs to remedy defects, to tidy up, complete and elucidate areas he found unclear. A distressing number of paintings and drawings copied by Hollar are now lost; if any should surface, their authenticity should not necessarily be disputed because of slight discrepancies between them and the Bohemian’s print—in fact, the discrepancies may work to confirm the drawings’ authenticity. … in copies of drawings by Leonardo and Holbein he could show fidelity to his models, yet he would sometimes combine elements from several sheets to make a single etched design.’⁴²

Proximity

We now turn to the second important question: can we pinpoint the location in 1650 of any of these paintings, and if so, can Hollar be placed in the proximity? We cannot determine the location of the Ganay painting in 1650; Snow-Smith’s assertion that it was placed in an apocryphal convent of the Cordeliers in Nantes from 1514 until the nineteenth century is clearly untenable. However, the detailed paper trail given in Chapter 7 strongly suggests one, possibly two, paintings of a Salvator Mundi type attributed to Leonardo da Vinci which were on public view between 1649 and 1651 at the Commonwealth Sale in Somerset House, London. Then there is the Hamilton painting, vividly described in c. 1638–41 at Chelsea, but dropping from sight thereafter, despite the sale of the 1st Duke of Hamilton’s collection c. 1649 in the Low Countries, although the precise circumstances are still to be fully established.⁴³ What do we know of Hollar’s whereabouts in 1649–50; can we locate him in London?

Hollar: Dealer at the Commonwealth Sale?

Little is known of Hollar’s activities between 1644 and 1651/2, a period of intense productivity to judge by the c. 350 etchings he executed. These were produced in collaboration with at least eleven publishers and covered a variety of subjects; some were after old masters. He maintained a working relationship with his former colleague Hendrick van der Borcht II, concentrating on subjects relating to the Arundel collection; their provenance acknowledged as ‘ex Collectione Arundeliana’. He also worked with Francis van den Wyngaerde, who specialized in reproducing old masters, especially those of the Italian Renaissance, and also Holbein. Hollar drew source material from fine collections in Antwerp, including that of Jan and Jacob van Veerle.⁴⁴ It is possible that works of art on the Antwerp art market also served as his models. The Civil War in England was bringing to market paintings from the famed collections of Stuart courtiers such as Hamilton and Buckingham, as well as Arundel. For instance, Hollar’s 1650 etching after Titian’s Christ before Pilate [Ecce Homo], a painting formerly in the Buckingham collection, was auctioned in Antwerp in 1648.⁴⁵

A good starting place is Hollar’s own witness. In its bottom margin, he says he made the etching, after Leonardo’s original, in 1650. He does not state the location of publication, the name of the publisher, the name of the owner of the painting, or of the plate. This is unusual in the context of Hollar’s output. Yet this work is universally classified under Hollar’s period in Antwerp, because a number of his etchings were published in this location in 1650. The standard chronology of Hollar’s life, according to the principal specialists—Richard Pennington (1982), Richard Godfrey (1995), and Simon Turner (2009)—places him in Antwerp between 1644 and 1652. This timeline has been constructed according to the place and dates of publication of etchings (although this does not necessarily denote uninterrupted residency, since Hollar endured a peripatetic career throughout his life, and since he may have sold plates to an Antwerp publisher at any point during this period). An important contemporary witness, John Aubrey, says that Hollar stayed in the Low Countries until about 1649.⁴⁶ Aubrey, an acquaintance, credits Hollar himself as the source of this information.

Hollar had arrived in England in 1636, in the train of Thomas Howard, Earl of Arundel, by whom he was employed to make drawings and etchings after the nobleman’s remarkable collection of old master paintings and drawings, antique sculpture, decorative arts, and natural curiosities. Many of these, inscribed ‘ex collectione Arundeliana’, were published in Antwerp after 1644. It is believed that Hollar had some employment in the royal household around 1639–40, some sources say he acted as a drawing tutor to the young Prince Charles.⁴⁷

The second state of Hollar’s etching after Jan Meyssens’s grisaille portrait of the artist, c. 1649, adds an inscription in the margin of biographical details, presumably provided by Hollar himself, claiming that he was a serviteur to the Duke of York.⁴⁸ If so, he would perhaps have had access to the royal works of art, although he seems to have etched very few of them. We understand little of Hollar’s contact with the Royal Collection before 1644; whether he copied paintings and drawings belonging to the king and queen, from which he made the etchings published in Antwerp, c. 1649–50, is not known. It is possible that the drawing on which Hollar based his etching of Leonardo’s Salvator Mundi was made in the Royal Collection before the outbreak of civil war.

In 1641, Lord Arundel spent £850 escorting the queen’s mother, Marie de Médicis, who was fast becoming a political liability, from England to Cologne, where he proceeded to sell part of his collection of old master drawings to the marchand-banquier Everard Jabach, in an attempt to cover his personal debts. He raised the astronomical sum of £54,000, which served to finance the armies of Charles I.⁴⁹ Clearly, Hollar’s position was imperilled; by 1642 Arundel House and Tart Hall were shut up, and those parts of the art collection that could be transported were crated and hurriedly shipped to Holland against the looming sequestration of royalists’ assets.⁵⁰ As the earl and countess disposed of parts of their vast collection in Antwerp, Cologne, and Amsterdam, Hollar published prints after its former glories. His actions were no doubt a consequence of sudden redundancy, but his prints may have served the additional purpose of advertising for sale the Arundel collection, and thereby, his loyalty to Arundel and the royalist cause.⁵¹ Little is known of this type of activity, but there is one similar case: the artist-dealer Adam Bierling (1625–75) used prints to advertise stock. In 1649 Bierling published Hollar’s etchings after Holbein’s portraits of Sir William and Lady Margaret Butt, indicating a commercial relationship.⁵²

Seen in this light, Hollar’s sudden reappearance in England in 1649, coinciding with the start of the Commonwealth Sale, appears strategic. If he had indeed been employed in the royal household before the outbreak of war, Hollar may have been a Crown creditor, returning to claim his due. Alternatively, he may have been sent to London by Antwerp dealers, publishers, and collectors to make copies of the sale catalogue, or perhaps to delineate certain items of interest as a sort of illustrated inventory. This idea is plausible; on 23 March 1653, John Evelyn recorded in his diary a visit from the French engraver Louis Richer (fl. 1640–59): ‘Came to see me that rare graver in Tallie douce Monsieur Richett, he was sent over by Card: Mazzarini, to make Collections of Pictures’.⁵³ This enterprise was surely annexed to the list of paintings and tapestries earmarked for Mazarin by Croullé in May 1650. Richer provided a few plates for John Ogilvy’s translation of Virgil—The Works of Publius Virgilius Maro, Translated, Adorn’d with Sculpture and Illustrated with Annotations, published in London in 1654. Hollar etched forty-three plates for this publication; the majority of the other plates were supplied by Pierre Lombart (1613–82), who arrived in London in 1650 and stayed for around a decade, leaving shortly after the Restoration of Charles II. He too paid Evelyn a visit in 1653: ‘Monsieur Lombart a famous Graver came to see my Collections.’⁵⁴ Antony Griffiths considers he may have come to London partly to deal in pictures at the Commonwealth Sale.⁵⁵

Hollar may have been able to rely on the protection of his former patrons in England during that time. The presence of the Countess of Arundel in London, in 1652, has been overlooked by scholars.⁵⁶ In his journal entry for Sunday, 11 February 1651 (Dutch Old Style), Lodewijk Huygens recorded a meeting with the countess, who was a family friend. ‘ I went with [Peter] Du Moulin to see Lady Pye and also for about an hour I talked with her and Lady Arundel, the daughter of the Earl of Shrewsbury, who is staying in the same street and had asked me to visit her.’⁵⁷ It is surprising to find the countess in London at the beginning of the Interregnum, at a time when she has been thought to be exiled in Amsterdam. The French agent Bordeaux wrote to Cardinal Mazarin from London several times during October and November of 1653: the heirs of the Earl of Arundel wished to sell many paintings, ‘needing money, being Catholic and indebted’.⁵⁸

What about the practicalities of making a copy of a painting in Somerset House? It is clear from the accuracy of the etching that Hollar had direct contact with a model. That is, the copy was not drawn by eye while hanging on a wall, but brought down, laid flat, and a reduced drawing made. Would he have been permitted to do this? Were artists able to remove items from the Commonwealth Sale in order to make copies? It seems so, to judge from an annotation to a memorandum of over forty paintings in the sale drawn up by the Spanish diplomat Alonso de Cárdenas for Luis de Haro, Marquis del Carpio, principal minister to Philip IV of Spain and financier of purchases made on the king’s behalf. Cárdenas told Haro that a Titian had not been viewed ‘because it has been taken out of the sale to be copied’.⁵⁹ In principle, Hollar should have not experienced any difficulty in making a direct copy of a painting at the sale; if he copied Leonardo’s Salvator Mundi, he probably did so between the date of the shopping list sent by Croullé to Mazarin in May 1650 and the end of the year, perhaps at the cardinal’s behest.

The sense we have of Hollar as a Royalist sympathizer, aloof in bearing and apparently remote from common concerns, sits uneasily with the idea that he may have capitalized on the late king’s misfortune, but we should remember he had little means of support for his family during this time. His prolific output in Antwerp stands testimony to his efforts to earn a living. Further, he had as exemplar the behaviour of the majority of the former royal servants associated with the art collection—Jan van Belcamp, the de Critz family, Sir Balthasar Gerbier, Remigius van Leemput, the Lanier family—who quickly switched their allegiances during the Interregnum, and switched them back again at the Restoration, having profiteered in picture dealing and making copies in the intervening ten years.⁶⁰ Antony Griffiths reminds of the ‘financial stringency that forced [Hollar] to work anonymously on propaganda sheets on behalf of the King’s enemies’.⁶¹

Weighing the Evidence

A number of paintings in the Hamilton collection seem to have found their way into the hands of Flemish marchand-amateurs, including Jan and Jacob van Veerle.⁶² An archducal passport of April 1649 records twenty-eight paintings imported into Brussels, ten of which may be identifiable with items from the della Nave and Hamilton collections. Hollar etched six of the paintings recorded in the passport.⁶³ Two—Palma Vecchio’s lost Portrait of Caterina Cornara⁶⁴ and Titian’s Portrait of his daughter [Lavinia]⁶⁵—state that their prototypes were owned by ‘Van Verle’, and were published in 1650 by Frans van den Wyngaerde. The Portrait of Caterina Cornara was said, in the second state of the etching, to belong to Diego Duarte (1612–91). The Portuguese merchant, resident in Antwerp, was formerly ‘Jeweller in Ordinary’ to Charles I, and dealt in art.

While this intelligence supports the hunch that Hollar may have been more closely integrated in the Flemish trade in pictures from the great Caroline collections than has been understood, it does not shed any light on the model for Hollar’s etching of Leonardo’s Salvator. No record has yet been found of such a painting in the Hamilton collection after c. 1641. If the Salvator Mundi attributed by the first duke to Raphael in an inventory taken after February 1638 was packed up and sent to the Low Countries, it is not discernable from the generic descriptions in the inventories of the crated goods.⁶⁶ Hollar explicitly attributes his Salvator to Leonardo; it seems unlikely he would amend an existing attribution. No such painting reappears in the documentation, written or visual, of the ‘onward destinations’ of paintings from Hamilton’s collection, such as Archduke Leopold Wilhelm’s collection at Brussels or Vienna.

On the other hand, the two paintings of Christ, attributed to Leonardo, recorded in the former Royal Collection in 1649 and 1650, are comparatively tangible. If one of these paintings was the model for Hollar’s etching, why then, did he not acknowledge its royal provenance? The king dead, the heir to the throne exiled and disenfranchized, and the painting up for sale, effectively it had no owner. It is unlikely Hollar would wish to advertise his exploitation of the royal family’s downfall, especially if he hoped for a restoration of the Crown and future royal service. Snow-Smith’s assertion that Hollar sent the proofs of his etching to Queen Henrietta Maria in exile is absurd. Given what we know of the austerity of the queen’s new circumstances following her husband’s execution, an unsolicited and vivid reminder of the loss of her property and former status would have been an unconscionable insult for the courtly Hollar to inflict.⁶⁷ If he sent proofs to Paris, it is much more likely the intended recipient was the acquisitive Cardinal Mazarin, who, Henrietta Maria complained, ‘no longer recognised me a queen’.⁶⁸

Where and when Hollar made the intermediary drawing of Leonardo’s canonical Salvator Mundi, from which he would publish his etching dated 1650, has long been a mystery, not least because the location of the painting at this time was unknown. That two strong candidates were to be found in the collection of King Charles I and Queen Henrietta Maria, and were on public view between 1649 and 1651 at the Commonwealth Sale, may resolve this long-standing question, while also opening a view on Hollar’s activities during the Interregnum. He published the painting at this precise date because it was for sale, because of his demonstrable interest in Leonardo, and because potential foreign buyers were unlikely to know what the painting looked like. The broad physical agreement between Hollar’s etching and Leonardo da Vinci’s Salvator Mundi, formerly in the Cook Collection, in conjunction with the historical context of the other contenders, outlined above, suggests that this painting was Hollar’s model and that he copied it when it was for sale, 1649–50.