The Lost Tudor Princess

by Alison Weir

  This may also have been a marriage portrait. The costume is French in style; thanks to the “Auld Alliance” and cultural links with France, French Renaissance influence was strong in Scotland, where French fashions were worn at court. Margaret would have grown up wearing such fashions, and her husband, Lennox, would have been accustomed to seeing women dressed similarly in France and Scotland. The style of dress in the portrait was still worn in France in the mid-1540s, and it would have been a nod to Margaret’s birthright, and to her new husband, for her to have been painted wearing French attire, especially if this was a marriage portrait.

  It has been suggested that three Holbein portraits that almost certainly depict Katherine Howard are of Margaret.4 There is no certain portrait of Katherine Howard, although a good case can be made for identifying these three—two miniatures and a half-length panel portrait—as her. The existence of two versions of the same miniature suggests that the sitter was a lady of some prominence,5 and her cloth-of-gold bodice, rich jewels and fur sleeves show her to have been of high rank. One version, dating from ca.1540, is in the Royal Collection. It was first said to be Katherine Howard around 1837, but may perhaps be identified with one of a group of miniatures at Lee Priory, Kent, that were described by George Vertue in the 1730s as being of Anne of Cleves,6 Jane Seymour, and Katherine Howard. On the back of the miniature is engraved an inscription, probably dating from the nineteenth century: “Catherine Howard, Queen of Henry 8th by Hans Holbein.”

  The other version of the miniature, signed by Holbein, was in the collection of Katherine’s relative, Thomas Howard, Earl of Arundel, in the early seventeenth century, and was engraved in 1645 by Wenceslaus Hollar without any identifying inscription. In 1743 it was engraved by Jacobus Houbraken as being Katherine Howard, which, together with Vertue’s identification of the first miniature as Katherine in the 1730s, suggests that this portrait type was accepted as her likeness by 1730. The second miniature was acquired by Horace Walpole in the eighteenth century, although he believed that it probably portrayed Mary Tudor, Queen of France. It was sold to Walter Montagu-Douglas Scott, 5th Duke of Buccleuch, in 1842, and is still owned by the present Duke. The original miniature has been trimmed, as the hands are cropped.

  Both miniatures show a young woman with dark auburn hair wearing a tawny-gold gown with a deep jeweled border at the neckline, a French hood, and an ouch and a pearl necklace that can be seen in portraits of Jane Seymour and Katherine Parr. Only recently have several portraits once thought to be of Lady Jane Grey been identified, on the evidence of jewelry, as Katherine Parr, and prior to that there were theories that Jane Seymour may have given away those jewels, perhaps to Mary Brandon, Lady Monteagle, who was possibly the sitter in these miniatures.7 But Katherine Parr is wearing the ouch in two portraits in the National Portrait Gallery, and the necklace in a portrait at Seaton Delaval Hall. Since the Queen’s jewels were handed down from consort to consort, the sitter in the miniatures is almost certainly Katherine Howard; furthermore the “square of jewels” edging the neckline of her bodice and the rich habiliments in the hood have been identified with wedding gifts given to Katherine by Henry VIII, the border being described in an inventory of her jewels as a “square containing xxiii diamonds and lx rubies with an edge of pearl containing xxiii.” As has been observed, the identification as Margaret cannot explain why she is wearing the royal jewels.8 It is hard to understand why, given the evidence of jewelry, historians have been reluctant to accept these miniatures as Katherine Howard. David Starkey, Alasdair Hawkyard and Bendor Grosvenor have collectively stated that “a comparison between the sitter’s jewelry and the inventory of jewels presented to Katherine on her marriage demonstrates beyond doubt that the sitter is Henry’s fifth queen.”9

  Holbein’s portrait of a young woman “in her twenty-first year” (according to the Latin inscription), painted around 1540, almost certainly features the same sitter. The original hangs in the Toledo Museum of Fine Arts, Ohio, and there are copies in the National Portrait Gallery and at Hever Castle (formerly at Trentham Hall); the latter has been tree-ring-dated to the middle of the sixteenth century.10 The rich clothing and jewelry show that she is of high rank. The Toledo portrait was not said to be Katherine Howard until 1898, so historians have long questioned the identification. It descended in the Cromwell family, and in 1969, Roy Strong suggested that the lady was Jane Seymour’s sister Elizabeth, who married Gregory, Thomas Cromwell’s only son. But the costume does seem rather lavish for the daughter of a knight and wife of a gentleman, and more recently, Dr. David Starkey has identified the sleeves and jewels in the Hever Castle version with items in Katherine Howard’s inventory. That evidence, the existence of three versions, and the resemblance of the lady to the sitters in the miniatures and engravings all suggest that the Toledo portrait is of Katherine Howard, whose image was at one time in demand. Lisle doubts that people would have kept portraits of a disgraced queen—none survive from Anne Boleyn’s lifetime, for example—but these particular paintings have emerged from an obscurity in which they were kept for centuries.

  The Toledo portrait dates from ca.1540; Margaret was in her twenty-first year in 1536. The sitter in this and the miniature bears no resemblance to her as she appears in the Darnley Memorial. The nose, again, is too long. Comparing portraits on appearance alone is a subjective approach, however, and one must look for other evidence. In regard to all these Holbein portraits, there is none to link any of them with Margaret Douglas, and much to connect them with Katherine Howard.

  A drawing by Holbein at Windsor, which has no inscription but has sometimes been said to portray Katherine Howard, does not appear to be the same sitter as the lady in the portraits just discussed. If the Somerley portrait is not Margaret, this may just possibly be her, as it bears a resemblance to her in later years. Although this sitter is fuller-faced, that may be down to her being about thirty years younger.

  A sitter who is perhaps an older Margaret may appear in a little-known but fine portrait of a lady with a black hood and a fox fur around her neck, painted between 1544 and 1555. In 1956–57, when it was displayed at the Royal Academy’s British Portraits exhibition, it was attributed to Hans Eworth. Prior to that it was thought to be by Hans Holbein. In 2009, Ludwig Meyer attributed it to Guillaume Scrots, who worked at the English court from 1545 until 1550.11 The sitter is clearly a lady of high rank, as denoted by her embroidered cuffs and furs, and her loose black gown suggests that she is pregnant, as Margaret was so often in this period. The headdress resembles the ones worn by Elizabeth Stonor, Lady Hoby and Margaret Wyatt, Lady Lee, in portraits executed by Holbein before 1544. There is a resemblance to the later portrait of Margaret, and to her tomb effigy. The portrait, formerly in a Welsh private collection, was sold at Christie’s, London, in 2011.

  A full-length portrait of the 1550s of a lady in black, formerly at Sawston Hall, Cambridgshire, and attributed to Guillaume Scrots, has traditionally been identified as Mary I, who took refuge from Northumberland’s pursuers at Sawston after Edward VI’s death. However, in recent years the identity of the sitter has been the subject of much debate, as she bears little resemblance to other portraits of Mary. Mary’s biographer Linda Porter was so convinced that this is Mary that she chose the painting for her book jacket. But Sir Roy Strong has never been convinced of this identification, and Dr. Tarnya Cooper, sixteenth-century curator at the National Portrait Gallery, has stated: “We concluded that, while it is undoubtedly a very interesting and important painting, it cannot represent Mary I mainly because of facial dissimilarity with other authentic portraits of her. It is more likely to be a member of the nobility, possibly from within Princess Mary’s circle.” Professor John Scarisbrick asserts: “There was nobody outside the royal family important enough for such a lavish full-length painting—and if it isn’t Mary, who is it?”12

  Professor John Guy has argued that it is Margaret Douglas.13 However, the painting is far more likely to portray Margaret Howard (ca.1515�
�71), sister of Katherine Howard and wife of Katherine Parr’s chancellor, Sir Thomas Arundell, who was executed in 1552. Before it came to Sawston Hall in 1918, the portrait was at Sutton Place, Surrey, where it was recorded by George Vertue in 1733.14 Lady Arundell’s daughter, Dorothy, married Henry Weston of Sutton Place in 1559.

  Another portrait that has tentatively been identified as Margaret is a three-quarter-length of a lady by Hans Eworth in the Fitzwilliam Museum, Cambridge, painted around 1556. Again, she was once thought to be Mary I, but that was questioned by Strong as long ago as 1965. The history of the picture is unknown before 1856. She has the snub nose, wide cheekbones, upturned mouth and pointed chin seen in the portrait of Margaret in the Darnley Memorial; she has red hair and she wears a black gown with a stand-up collar over a rose-pink kirtle and undersleeves, a square French hood and jewelry rich enough to mark her out as someone of high rank. The ornamental prayer book hanging from her girdle is embossed with the letter “D.” However, she is unlikely to be Margaret Douglas, because the jewel at her breast depicts the maiden Esther kneeling before King Ahasuerus, which almost certainly indicates that the sitter is unmarried.15

  The Victoria and Albert Museum holds in storage a miniature of a lady by the court artist Levina Teerlinc, which was formerly in the collection of the Dukes of Beaufort. The sitter wears a square-shaped French hood and bonnet with a jeweled brooch and a black feather, a black gown with a high furred collar, a small ruff and a crucifix, which suggests that she was Catholic. There is a marked resemblance between her features and Margaret Douglas’s in the Darnley Memorial and her tomb effigy, and although she looks older than in the memorial, that may be because Teerlinc, who came to England in 1545 and was appointed painter to Henry VIII, was not the most accomplished of artists. She served as a gentlewoman in the households of Mary I and Elizabeth I, and painted miniatures of Elizabeth I, Lady Katherine Grey and Katherine Parr, among others. Her style is exemplified by female sitters with stick-thin arms.

  This miniature has been dated to ca.1560. The square-shaped version of the French hood is first seen in the mid-1540s, when it appears in two portraits of Katherine Parr, who was also painted wearing a feathered bonnet on top of a circular pearl-edged pleated coif, not unlike the one the sitter wears in the Teerlinc miniature, although the latter is a squarer shape. Such bonnets also feature in portraits of the 1530s. The square hood gained popularity and is familiar in portraits of Mary I (reigned 1553–58); it remained in fashion until around 1567. It is extremely rare to find examples of it surmounted by a cap such as the Teerlinc sitter wears. The furred collar and the high-necked ruff appear in portraits of the 1550s–60s, although there are examples of the latter in Holbein’s drawings of the 1530s. Allowing for the fact that some portraits show older women wearing the outdated fashions of their youth, the miniature probably dates from around 1550–65. The fact that Teerlinc painted this sitter at all strongly suggests that the lady was at court and of high rank. Given that, the jewelry and the facial similarities, it may portray Margaret, but there is also a facial resemblance to Bess of Hardwick as she appears in two portraits at Hardwick Hall.16

  In the National Portrait Gallery there is a half-length portrait by an unknown artist, dating from ca.1560–65, of an unknown woman who was identified in 1866 as Margaret solely on the grounds of similarity with the portrait of her in the Darnley Memorial. In fact there is little resemblance at all, the sitter in this portrait being older, more angular and thinner-faced than Margaret appears in the memorial. The painting’s history before 1866 is unknown.17

  There are several portraits said to portray Margaret in her widowhood. The most important is a full-length by an unknown artist in the Royal Collection at the Palace of Holyroodhouse, which depicts a woman in black mourning wearing a white cap. It bears a framed inscription: “THE LADY MARGARET. HIR GRACE / LATE WIFE TO MATHEW ERLLE / OF LENNOX REGENT OF SCOTLANDE / AND MOTHER TO HENRY KINGE / OF SCOTLAND / Aetatis 55 Aí Dni. 1572.” It was first recorded in the Royal Collection in 1639, in the Queen’s Gallery at Greenwich,18 when it was described merely as “a Scottish lady at length in mourning habit and a clock upon the table in a gilded frame,”19 but ten years later it was sold as “The Lady Margaret at length, wife to ye Regent of Scotland.” It would appear that the inscription was added sometime between 1639 and 1649, otherwise the sitter would surely have been named in the earlier inventory. Doubts have been expressed about her identity: The portrait was not known to be of Margaret in 1639; the features do not tally with Margaret’s in the Darnley Memorial, and the age is given incorrectly:20 Margaret had reached fifty-five in October 1570. After 1660 the portrait was returned to the collection of Charles II.21 It may well have been the one recorded as hanging in the Queen’s privy chamber at Hampton Court in the reign of George I.22 The panel has been extended on all four sides.

  The sitter wears a heavy wide-skirted black overgown over a black gown trimmed with fur at the neckline and down the front, and carries a pair of leather gloves trimmed with the same black bows that adorn the dress; a thin scarf knotted at the ends hangs almost to the length of her hem. The dark background is plain, but she stands on a rich Turkey carpet, proclaiming to the world her high status, as such carpets were costly and were usually draped on tables to preserve them from wear and tear. At her feet prances a griffon, a toy dog from Flanders of a breed that was popular among the English aristocracy. It was believed that griffons mated for life and never sought a new partner after one had died. Behind the sitter is a finely turned side table covered with a black cloth with a fringe or border, on which the sitter rests her hand. Upon the table stands a gold clock in the form of a tempietto—a small round chapel—with a dial surmounted by a hound and a shield bearing what may be the royal arms of Scotland. The sitter is clearly of high status, and if the arms are royal, then she probably is Margaret. That being so, the portrait was either appropriated by Elizabeth after her death, or came by descent to her grandson James.23 A nineteenth-century copy by Rhoda Sullivan hangs in the Darnley Room at Temple Newsam, a companion full-length to one of Mary, Queen of Scots. Both are in the collection of Leeds Museums and Art Galleries.24

  The Royal Collection Trust has another portrait that might be Margaret. It was bought by Queen Victoria in 1843 from John Ponsonby, 4th Earl of Bessborough, as a portrait of Mary I. Later it was identified as Margaret Douglas, but that was disputed by Oliver Millar. It is said to date from ca.1575, and is probably by an English artist, but has been damaged and overpainted.25 Given the distinctive nose and deep-set eyes, it is possible that the identification was correct, but without further investigation it is impossible to tell. If it is Margaret, it must date from early in her widowhood, as her hair has not yet gone gray and she is wearing deeper mourning than in her other portraits from the 1570s.

  A half-length portrait of Margaret, similar to the full-length in the Royal Collection, was owned by the Earls of Morton and once hung at Dalmahoy House near Edinburgh. Its present location is uncertain, but it was described by Strickland, and is known from a nineteenth-century engraving by Gourlay Steell.26 The portrait was inscribed “The Lady Margaret, her Grace, The Countess of Levenax, Aetatis sue 50.” The sitter wears black damask, which was not necessarily mourning, as deep black dye was costly and therefore black was a high-status color; she has a white ruff, and the same thin scarf and white hood with a cornette as in the Royal Collection portrait. This headdress may be similar to the one provided for Margaret in the Tower in 1565. The only jewelry is a balas ruby brooch with a pendant pearl. The frame is enameled in black and white on gold and adorned with marguerites.27 If this was painted when Margaret was in her fiftieth year, it must have been done between October 1564 and April 1565, before she was placed under arrest. But she looks older and plumper than in the Darnley Memorial of 1568, so the Dalmahoy portrait was probably painted later, in which case the age given in the inscription is incorrect.

  A miniature of 1575 by Nicholas Hilliard, also showing an a
ging woman in black, in the collection of the Rijksmuseum, Amsterdam, is also said to portray Margaret. The sitter has gray hair; in the Darnley Memorial, painted seven years earlier, Margaret’s was dark brown, so if this is her she had aged considerably during those sad years. Yet this sitter has a distinctive hooked nose, unlike Margaret’s in the Darnley Memorial or her tomb effigy, and one eye partially closed, suggesting some facial paralysis or possibly a stroke. This miniature was perhaps the one recorded in Charles I’s inventory in 1639: “The picture…being upon blue grounded-card written about the year of Our Lord with gold letters 1575, and also her age, 53, being the Lady Margaret Douglas, aunt to Queen Mary of Scotland, in a black and white mourning widow’s habit [and] a little plain ruff.” The miniature has since been cut down, so the age is missing, but in October 1575 Margaret was fifty, not fifty-three. There is another version, damaged, with the face entirely repainted, dated 1576 and called “Anne [Morgan], Lady Hunsdon,” in the Fitzwilliam Museum, Cambridge, but it is impossible to say for certain that it is by Hilliard.28

  A painted genealogy dating from ca.1603 is at Parham Park, Sussex, and shows the ancestry of James I. The royal figures in the genealogy are based on known portraits, but those from which the images of Margaret and Lennox originate are unknown, and probably lost. An engraving of Margaret dating from ca.1603 is in Benjamin Wright’s The Royal Progeny of King James, and again is based on an unknown portrait, although the snub nose and pointed chin are evident.29

  Leanda de Lisle suggests on her website that a portrait of a woman of ca.1560 in the Royal Collection may be Margaret Douglas. The sitter has traditionally been identified as her cousin, Frances Brandon, Duchess of Suffolk, but bears no resemblance to Frances’s tomb effigy, or to Margaret in the Darnley Memorial, and there is no evidence to support the identification as either of them.