Traditionally regarded by connoisseurs as a masterpiece by the greatest of all Dutch artists, in recent years, along with many other Rembrandt paintings from the years 1643-45, the status of Man with a Sword has been disputed, with attributions made to various pupils of Rembrandt rather than to the master himself. Recently the subject of a thorough re-appraisal, this painting sheds fascinating new light on Rembrandt’s studio practice during one of the most enigmatic and least well-documented phases of his career. Through a process of scientific investigation, which had never before been conducted on the picture, the removal of an old obscuring varnish, and fresh scholarly analysis, Man with a Sword has now been acknowledged as a reliably signed and dated portrait which was conceived by Rembrandt and then fashioned into a historical portrait or tronie by another artist active in the Rembrandt workshop.
Man with a Sword has a long and, for the most part, illustrious history within the Rembrandt canon. First documented in an English collection in 1765, the picture had returned to Holland by 1777, where it entered two notable Amsterdam collections in quick succession, that of Everard van Tindinghorste and then Pieter Locquet. It was most probably acquired in Amsterdam by Sir James Carnegie, 8th Earl of Southesk (1799-1849), who was documented there in 1819 on the return journey from his Grand Tour. The picture was subsequently acquired from his son by Robert Stayner Holford (1808-1894), one of the greatest collectors of Old Master pictures in England during the Victorian era. In 1898 the picture was included in the seminal Rembrandt Tentoonstelling exhibition at the Rijksmuseum, the first major monographic show of Rembrandt paintings and one of the first blockbuster art exhibitions of the modern era.
The great German connoisseur Gustav Waagen had seen the picture in Holford’s collection in 1854, remarking on its ‘extraordinary power and depth of tone’ (loc. cit.) and for the next century Man with a Sword was unanimously heralded as a Rembrandt by all of the key scholars on the artist, including Hofstede de Groot, Von Bode, Valentiner, Rosenburg, Bredius and Kurt Bauch, the latter describing it as a possible self-portrait. Only in 1968 was an element of doubt about its authorship first raised when Horst Gerson remarked: ‘I do not know the painting itself. To judge from the photograph, the attribution to Rembrandt is not very convincing’ (loc. cit.). In 1970, while on loan to the Bowdoin College Museum of Fine Arts, in Maine, from the estate of Sir Harry Oakes (1873-1943), two founder members of the Rembrandt Research Project, Bob Haak and Pieter van Thiel, examined the picture in person and rejected the Rembrandt attribution altogether, even suggesting that it was an 18th century pastiche. By the time the picture appeared on the art market, at auction in 1989, it had been given to one of the Rembrandt’s most accomplished pupils Govaert Flinck (1615-1660), puzzling since if the date of 1644 is to be believed, Flinck was no longer active in Rembrandt’s workshop. In 1994 Werner Sumowski suggested another Rembrandt pupil, Ferdinand Bol (1616-1680), as the possible author, comparing it to his Self-Portrait in the Springfield Museum of Art, Massachusetts, and the Portrait of a Man in Gouda. In the same year (1994), the picture reappeared at Sotheby’s, again as by Flinck, this time linked inextricably to the only other Rembrandt painting of a man dated 1644, the Man with a Steel Gorget in the Metropolitan Museum of Art, New York (fig. 1). At that time the Metropolitan Museum painting, which was soon to feature in the 1995 exhibition Rembrandt/not Rembrandt, was given to yet another of the master’s pupils Gerbrand van den Eeckhout (1621-1674) by Walter Liedtke, while Hubert van Sonnenburg, the co-author of the exhibition catalogue, still considered it to be by Flinck. In the Sotheby’s catalogue entry, the conclusion was reached that ‘Although the authorship of the Metropolitan painting may be in dispute, there is no doubt that that painting and the present work were executed by the same artist’. This is interesting in itself because the Man with a steel Gorget has only recently been re-attributed to Rembrandt by Ernst van de Wetering.
The questions surrounding the Metropolitan picture serve to highlight the uncertainties and special complexities involved with attributions for Rembrandt paintings executed in the years immediately after 1642. This uncertainty has been exacerbated by the fact that the chronological progress of the Rembrandt Research Project’s Corpus stopped at 1642, meaning that these pictures have yet to be submitted to the same degree of rigorous scrutiny as the earlier works. What is certain is that Rembrandt’s output of paintings diminished dramatically after the completion of his spectacularly ambitious masterpiece the Nightwatch in 1642. Having apparently exhausted the artistic possibilities of portraiture during more than twelve years of intensive graft, culminating with the Nightwatch, and having made himself extremely wealthy in the process, Rembrandt appears to have abandoned portrait commissions altogether, thereby freeing himself to pursue other personal and artistic goals. For almost a decade, from 1643, he produced virtually no portraits at all, returning to the genre again only in 1652/3 as financial pressures began to weigh down on him. This shift in focus after 1642 must have had a considerable impact on practice within the Rembrandt workshop, and may to some extent explain the unevenness in quality of the known pictures from 1643-44, this picture and the Man with a steel Gorget included. The same applies to the two portraits of men that were painted in the Rembrandt workshop in 1643 - the Falconer (private collection, England), which is now considered by Ernst van de Wetering to be by Rembrandt with studio participation; and the so-called Admiral (New York, Metropolitan Museum of Art) which he believes was partly or entirely executed by someone other than Rembrandt in the master’s workshop.
The scientific examination of the Man with a Sword, conducted by Dr. Jilleen Nadolny and Dr. Nicholas Eastaugh in 2012/13, has established first and foremost that the picture definitely emanated from Rembrandt’s workshop. Pigment testing has revealed that the materials in the painting, including a quartz ground that was particular to Rembrandt’s studio, are entirely consistent with those seen in other works by Rembrandt. Eastaugh’s study ‘Found no material evidence in odds with the date and attribution suggested by the inscription, and indeed, a good degree of consistency to what is known of Rembrandt’s working methods and materials’. Furthermore, examination of the signature and date found it to be contemporary with the rest of the picture - ‘very thoroughly and cleanly transversed by the craquelure of the painting’, where they observed ‘no physical anomalies that would introduce any particular uncertainty’.
The most striking revelation made by technical examination is the presence of extensive alterations throughout the composition, which demonstrate clearly how the initial concept for the picture was substantially changed. Adjustments to the sitter’s pose, substantial repainting of the costume, and changes to the background, all become visible in the x-radiograph and infrared images (figs. 2 and 3). The original scheme appears to have been for a more frontally posed portrait with a higher outline of the sitter’s left shoulder. A curtain across the upper right corner, originally red in hue, was painted out and the column extended up to the top edge. The sitter was originally depicted bare-headed, with long hair (much of which is still visible), to which a hat was added, originally with a large plume, which was subsequently painted out (visible in the infrared scan). The man’s fringe seen protruding from under the hat has also been added. He was originally shown in a straight-topped white shirt, which was shown protruding in a triangular shape in his midriff. The round-necked shirt, the chain and much of the purple coloured tunic visible now, all appear to be part of a second campaign of painting. Furthermore, the sitter’s right hand, originally shown further to the left with the thumb protruding upwards, and without a lace cuff, has been painted out and replaced with another hand, now clutching a sword, also an addition to the original scheme.
Seen in this new light, Van de Wetering has reached the conclusion that the original concept for this picture was by Rembrandt, showing a formal, commissioned male portrait. The raw, bravura brushwork seen in the black cape, particularly across the left shoulder, is indicative both of Rembrandt’s handling and of its unfinished state. So too the rendering of the shaded part of the left hand reveals qualities that point strongly to Rembrandt’s technique. It is, above all, the outstanding quality of the head and hair, largely unaltered, but possibly for some pink highlights added in the cheeks and in the eyes, that suggests unmistakably Rembrandt’s authorship and, at the same time, rules out any of his pupils. This portrait was either never sold (the sitter could have died or reneged on the payment for instance), or it might simply have been abandoned by the artist, perhaps as part of his move away from portraiture after 1642. The significant conceptual changes that were subsequently made to it were most likely added by a pupil who, by superimposing elements over an unfinished picture, effectively transformed Rembrandt’s portrait into a more generic historical figure or tronie. As pointed out by Van de Wetering, there are several other instances within Rembrandt’s oeuvre where portraits were adapted into tronies. For instance, the Bust of a woman (Los Angeles, Wright Art Gallery, University of California), whereby an unfinished female portrait was changed by a pupil into a depiction of a Sybil; and the Girl at a picture frame, of 1641 (Royal Castle, Warsaw), which was also originally conceived as a portrait.
We are grateful to Professor Ernst van de Wetering for his analysis of this painting and his endorsement of the present cataloguing. We are also grateful to Dr. Jilleen Nadolny and Dr. Nicholas Eastaugh of Art Access & Research (UK Ltd.) for their scientific analysis of the picture and to Martin Bijl who cleaned it and submitted it to further scientific testing.
Robert Stayner Holford (1808-1894)
By virtue of two successive inheritances, first from his bachelor uncle in 1838 and then from his father a year later, Robert Stayner Holford found himself, at the age of 30, the heir to an immense fortune. His family’s wealth owed to his grandfather Peter Holford (d. 1804) who had created a canal system to supply London with fresh water. Holford started collecting on a grand scale straight away, using the legendary Scottish dealer William Buchanan as his advisor. His tastes were eclectic and extended beyond that of his contemporaries in that, in addition to paintings from all schools, he was also interested in manuscripts, prints, maiolica, metalwork, tapestries, porcelain and furniture. A central figure in the London art scene of the day, he was a founding member of the Burlington Fine Arts Club in 1856. Holford was also a builder, but his architectural projects were mainly motivated by his desire to showcase his enormous collection. To this end, he commissioned the architect Lewis Vulliamy not only to remodel his country house at Westonbirt, Gloucestershire, but also to erect a magnificent Italianate palace on Park Lane - Dorchester House, now home to the Dorchester Hotel.
When the great German connoisseur Gustav Waagen visited Holford in 1854, he was astonished by the collection, especially given ‘the youth of the proprietor’ and the fact that it had ‘been formed in a comparatively short period’. Holford had already amassed more than 100 pictures by this time, and Waagen considered the collection an ‘example of how much may be done where great wealth is combined with excellent powers of judgment’ (op. cit., pp. 193-4).
Holford held a particular fondness for Italian art which numerically outweighed all other schools in his collection. In this respect he may well have been influenced by the taste of his brother-in-law Alexander Linday, 25th Earl of Crawford and 8th Earl of Balcarres, or Lord Lindsay (1812-1880), who was married to his wife’s sister, and who was also their second cousin. Highlights from his Italian collection included Lorenzo Lotto’s Lucrezia (London, National Gallery), Pesellino’s Madonna and Child with Saints (New York, Metropolitan Museum of Art), and Giovanni Bellini’s Portrait of a Boy (Birmingham, Barber Institute of Fine Arts). Holford owned a remarkable assortment of portraits, including such masterpieces as Petrus Christus’s Portrait of a Man (Los Angeles, Los Angeles County Museum of Art), Velázquez’s Philip IV of Spain (Sarasota, Ringling Museum of Art) and Gaspar de Guzmán, Conde-duque de Olivares (New York, Hispanic Society), and two exceptional full-lengths by van Dyck - Marchesa Balbi (Washington, National Gallery of Art) and Abbé Scaglia (London, National Gallery). Other stars included Aelbert Cuyp’s View of Dordrecht (Ascott, Anthony de Rothschild collection, National Trust), Claude’s Temple of Bacchus (Toronto, National Gallery of Canada), and Ruisdael’s View of Alkmaar (Upton House, National Trust).
Most of all though, Holford had a special predilection for Rembrandt, as might be discerned from his own portrait of 1862. He owned three other Rembrandt paintings: the magisterial Marten Looten of 1632 (Los Angeles, Los Angeles County Museum of Art; fig. 4), the Man in a Beret, whose attribution, previously dismissed, is now being reconsidered (New York, Metropolitan Museum); and the Lady with a handkerchief, currently attributed to Rembrandt’s prodigious pupil Carel Fabritius (Toronto, Art Gallery of Ontario).
The entire picture collection went under the hammer at Christie’s during three days of sales in July 1927 and May 1928, after the death of Holford’s childless son in 1926. This series of auctions constituted one of the most significant single-owner sales of Old Master pictures in the 20th century and the results provide a fascinating insight into where taste and market was aligned in the late 1920s. Demand for the Italian pictures was relatively restrained with a few notable exceptions, such as the Pesellino which sold for 16,500 guineas and Lotto’s Lucretia which fetched 22,000 guineas. Velázquez was apparently treated with some indifference, the full-length of Philip IV fetching just 3,300 guineas, while the Claude sold for a paltry 950 guineas. The highest prices were largely reserved for the Northern pictures: 14,000 guineas for the Petrus Christus, 21,000 guineas for the Cuyp, and 30,000 guineas for Abbé Scaglia are some striking examples. It was the four Rembrandt paintings, above all else, that commanded the most attention and the prices reveal a preference for the artist’s later style. Marten Looten was sold for 26,000 guineas (the lowest price of the four), the Portrait of a Lady for 30,000 guineas and the Man with a Beret for 44,000 guineas. It is testament to the extraordinary appeal of the Man with a Sword that it became not only the most expensive of the four Holford Rembrandts, but also the top lot of the entire sale, selling for 48,000 guineas.