Peter Paul Rubens

Head of an Old Woman c. 1615
Oil on oak, 52 × 43.3 cm
Purchased 1925, no. 3339

The panel was bought at auction in 1925, catalogued as Head of an Old Woman, and “said to be the mother of the artist.” The identification was an historic inaccuracy, and can be traced back through the late eighteenth century. The sitter is in fact unknown. She has a distinctive physiognomy, and was painted and drawn by several artists working in Antwerp in the 1610s, including Jordaens.1 While from the life, the work was not intended as a portrait; rather it was painted by Rubens as a tronie or study head. It served as a model, a character head, to be copied and inserted into other works.2 The Ottawa tronie was painted in c. 1615, and employed in at least three paintings dating to the 1610s.3

The use of head studies to provide models for larger paintings was an established practice – a good number of tronies survive from the workshop of Frans Floris, the leading Antwerp painter of the mid-sixteenth century. Rubens built on the efficiencies these studies offered, but rethought the tradition: he painted them from the life, unlike Floris, and he seemed to value the direct connection to reality, which would make the final painting the more convincing by providing a trusted structure on which the artist could build. Rubens was also influenced by contemporary Italian painters who strove for a new degree of verisimilitude – painting directly from life had a new importance at this time, both celebrated and condemned by contemporaries.4 Tronies also filled a teaching role, allowing students to study and understand the master’s technique. Rubens, when he first returned to Antwerp from Italy, built up a stock of these heads – types who would come to people his works – and so providing a foundation for his workshop. He sought models, posed them and often painted them from different perspectives. It seems certain that Rubens also made studies of this woman from different aspects, although none have survived.5 One is documented in the so-called Drawing Book, which records material from Rubens’ studio (illus. 1).6

This sitter’s physiognomy must have fascinated Rubens, looking for a useful and characteristic female type, who could serve as, say, an old maid or elderly saint. This type was in use across all Europe, part of a well-established repertoire. Examples had appeared in Rubens’ earlier paintings, their features exaggeratedly pronounced and time-worn as a kind of compensation for their otherwise generic quality. The artist must have been amused to find a living, real example on the streets of Antwerp.7

Rubens seems more focused on recording her facial features and character, rather than working toward any specific pose. The head-on view makes her incorporation into any painting awkward, and she remains somewhat isolated within the figure groups in which she has been inserted. So too do her slightly averted eyes – as if she could not meet Rubens’ scrutiny of her during the painting session. The Ottawa tronie stands out among similar objects in its blunt frontality and lack of affect, and appears to be an almost an objective record of the subject, seeming to capture something of her character. Other heads show the sitters more artfully posed; often adopting a theatrical attitude.

Features of the carpentry of the Ottawa panel suggest it was originally wider, and once may have held another view of the same woman’s head. A comparable example depicting the same woman shows her in three-quarter and full profile (illus. 2). This has been traditionally attributed to Jordaens – based on style and technique, it cannot be by Rubens.8 However, closely related profile views, featuring the same headdress, appear in at least three paintings from Rubens’ workshop, dating to the late 1610s and early 1620s.9 The simplest explanations are that either the Jordaens profile was available for use in Rubens’ workshop, or possibly that Jordaens copied a tronie by Rubens that is now lost. Not all Rubens’ tronies were necessarily painted by him: his stock also held examples by Van Dyck.

If there was another head on the Ottawa panel, it too may have been a profile and a good candidate for use is in the Holy Family with St. Elizabeth and the Infant John the Baptist, now in the Art Institute of Chicago (c. 1615) [ illus. 3 ]. St. Anne’s head shows similarities to the Ottawa head in lighting and in the handling of physiognomy, the result albeit idealized for the purpose of the role played.10 A drawing by Artus Wolffort shows the same woman’s physiognomy in profile, and a good comparison can be made (illus. 4).

The Ottawa tronie was first employed in the Discovery of Erichthonius (c. 1616), in the collection of the Princes of Liechtenstein, Vienna, where it was used, appropriately, for the character of the old maid (illus. 5). The preparatory sketch for this work shows a figure more integrated into the action and the group.11 The resolution is less satisfying with the eventual insertion of the Ottawa head, demonstrating the limitations of the frontal view. Successful integration required more adaptation, as seen in the workshop picture Drunken Silenus Supported by Satyrs, c. 1620, at the National Gallery, London, for which the Ottawa tronie provides a blueprint for the old bacchante (illus. 6).12 Details, the headdress, the highlights on the flesh, and the general physiognomy are loosely transposed to make an otherwise unstructured figure type more direct and convincing. The Ottawa head is also used in the Prado Achilles Discovered among the Daughters of Lycomedes, c. 1617–18, another workshop painting featuring a bland, characterless rendition as another old maid (illus. 7).13

Such heads were also used in training: in copying, the assistant would come to understand and ideally replicate the master’s notation and touch. Two close copies survive. The best, and clearly from the workshop, is now in Oldenburg (illus. 8). The copyist is clearly focused on Rubens’ method of building up form, the relationship between transparent, thin and opaque paint, and on mimicking touch.

While the Ottawa head is not directly copied into any paintings after the early 1620s – an exact copy would have been ungainly insertion, as Rubens’ style had changed – it seems to have stayed in the studio. It appears, for instance, in a workshop painting, another version of the Discovery of Erichthonius from the early 1630s, as the only element to reappear from the first version, suggesting that she remained, for Rubens, the archetypal old maid.14

The painting offers a virtuoso demonstration of Rubens’ accomplished technique, which can be fully appreciated for the first time in two hundred years following its recent restoration. Over the chalk and glue preparatory layers – called a “ground” – Rubens cast a streaky, warm grey paint layer, termed an “imprimatura.” This layer makes the underlying ground less absorbent, tones down the strong white of the ground, and provides a mid-tone onto which the artist would begin to drawn and paint. As often in Rubens’ work, this toning layer remains visible where the overlying paint is thinner or absent. It can be most clearly seen in the bonnet as a brushy, diagonally applied wash. When applied, it was also trapped in striations in the underlayed ground, adding further unevenness – this was Rubens’ preferred way of working, and he must have welcomed the visual vibration it provided as he worked into it, using thin, semi-opaque applications to build up form.

No underdrawing could be detected either by infrared reflectography, or through close examination under a microscope. Rubens would not have needed any, since this is a relatively informal work, and painted directly – in fact, the face and bonnet were finished in one session. The background and dark clothing seem to have been painted just afterwards, and the white of her undershirt, visible at her neck, was painted only after the black had dried fully. There is no retouching or correction in the face whatsoever. The last descriptive touch on the painting is the orange scumble that re-establishes the shape of her neck, which had been somewhat lost when the white collar was painted.

Rubens’ mastery is best observed, of course, in the flesh and in the flurried, yet considered highlights on the bonnet. These strokes show his control over the paint’s consistency: the stiffer, bodied paint could be pushed out to a very thin film, or diluted to almost liquid to achieve a direct, yet soft rendering of form. The paint of the flesh shows Rubens modulating warm and cool tones to achieve volume, as was typical of him. In this case, he uses browns and grays, the tints worked wet-in-wet. On paintings built in layers, this was traditionally achieved in stages of underpainting in warm and cool colours – sometimes a strong red and blue – which were then scumbled over. The control shown in allowing the imprimatura to modify the thinnest applications of semi-opaque paint and in working tint into tint is masterful. Rubens shows great foresight and economy, for good reason: painting like this could quickly become muddy if there is too much blending. The final touches in the face are the soft, yet firm application of highlights in paler tints, and an agitated, deepening red lake-based glaze beneath the nose to define lips and mouth, and around the eyes. Rubens would have used several brushes during this session, to keep his tints as clean as possible. To judge from the tracking in the paint, all were likely made of soft hair rather than bristle. The widest was 7 mm, with a flat tip; the so-called pencil brush, smaller and pointed, was used for the darker strokes in her hair on the right, blending at her left eyebrow, and in strokes of emphasis elsewhere.

The panel currently comprises five separate pieces of oak from the Baltic region or Poland. Dendrochronology indicates the two principal members were available for use in the early seventeenth century, certainly by c. 1610; the two remaining datable members were felled later, and were available for use by the early 1640s.15 The join between the principal members and the others runs vertically, approximately 3 cm from the right edge of her bonnet. While many of Rubens’ panels demonstrate unusual construction, the case here is that the original panel was cut down and then extended. The two original members run horizontally with the join at the level of the woman’s eyes and would have extended to the right, permitting another head study, as discussed above. The three other members were added to centre the head on the newly resized panel. The additional wood, also with horizontal grain, was joined to the original panel with a pegged lap joint; this type of joinery is unusual for professional panel makers at this time. It was also unusual to extend planks along the grain, however we know of many examples within Rubens’ production that does exactly this. This means of enlarging the panel may seem unnecessarily complex, but it was economic and understands how wood behaves – it avoids the dangers of cross-grained joinery, and the join remains intact to this day. The newly added surface was prepared in a similar way to the original panel, but with a more opaque imprimatura. The addition was then painted to match, with new paint extending over the original surface, presumably to disguise the transition, although it can be clearly seen. This is due to the fact that the newer paint consists of a cruder, coarser mixture of lead white and charcoal black than lies beneath. A previous restoration suppressed the disparity.

Given the dates of the different elements and the location of the first cut, a likely scenario is that the larger panel showing two heads was cut in half and then adapted to produce two saleable paintings shortly after Rubens’ death. The contents of his studio sale of 1640 list many such heads, and there are well-known incidences of this material being adapted for the market. Its later seventeenth-century history is unknown.

By the early eighteenth century, the Ottawa tronie was sufficiently valued that an unidentified collector commissioned an expensive custom frame, which it retains (illus. 9). Its quality and design place it in Paris, where Rubens’ art was deeply appreciated by both collectors and artists.16 The value of such a frame testifies to the remarkable transformation of a piece of studio furniture into a prized art object.

This old woman enjoyed some celebrity in Paris during the eighteenth century. This panel – or possibly a copy of it – was in the collection of the Marquis de Marigny, an important collector who was responsible for much royal patronage of the arts; he was well-placed, the brother of Madame de Pompadour, official mistress to Louis XV. In his collection the painting was identified as a depiction of Rubens’ mother – perhaps inspired by another tronie of another old woman in a princely German collection, well-known through reproduction.17 And yet another tronie showing the subject of the Ottawa work her was also in Paris at this time.18

By the early nineteenth century the panel had been bought by the Earl of Darnley, who had assembled a remarkable collection of Old Masters, buying voraciously at a time when much material became available through the breaking up of artistocratic collections following the French Revolution.19 Here the works was attributed to Rubens and – although with some doubt – claimed to represent his mother.20 The tronie would remain with this family until bought by the Gallery in 1925. During the second half of the twentieth century, scholars at times showed doubt as to the attribution, sometimes arguing it was by Rubens’ workshop or perhaps even by Jordaens – doubt likely due to the effects of the old varnish, and to a lack of knowledge of Rubens’ tronies. The Gallery retained the attribution to Rubens, which its recent cleaning has justified.

Physical History and Technical Notes

At the time of purchase, the Gallery engaged F.W. Colley, an independent restorer, to improve the painting’s appearance; Colley, who may have also partially cleaned the painting, revarnished it. A year later, he undertook structural work, mounting the panel to a brass plate with a lead white-based oil putty, likely after partially thinning it. No photograph of the panel’s verso exists. The painting remained largely untouched until 2013, when it was given a comprehensive restoration. During that treatment, we were able to determine that Colley’s efforts at cleaning had been, thankfully, superficial and that the painting still bore overpaint and varnish from before his treatment.

Provenance

Presumably among “une quantité des visages au vif, sur toile, & fonds de bois, tant de Mons. Rubens, que de Mons. van Dyck” listed in Rubens’ 1640 studio sale (Muller 1989, p. 145). Unidentified French collection by the 1720s. John Bligh, 4th Earl of Darnley (1767–1831), when recorded at Cobham Hall, Kent by c. 1806; by descent; sold Christie’s, London, 1 May 1925, lot no. 74 as “Sir P.P. Rubens,” “Head of an Old Woman (said to be the mother of the artist)”; bought £2,100 by the Gallery through Agnew, London.

Notes

The frame (Paris, 2nd quarter 18th century) would seem to have been made for the work, implying that it was in an unidentified French collection at this time. Its later eigthteenth-century provenance is unclear: Edgar Harden (private communication) notes that French collectors did not always choose to replace frames from this period, seeing them as compatible with contemporary taste.

Possibly the work listed in the post-mortem inventory of the Marquis de Marigny (1727–1781) : “Le portrait de la Mère de Rubens Coiffée d’un bonnet de Toile prisé Trente Six Liv.” (Gordon 2003, p. 296, no. 821); his sale, Paris, 18 March – 6 April 1782 (this lot 25 March), lot no. 149 “Tête de vieille Femme, habillée & coiffée dans l’ancien costume Flamand; elle est vue de face, & porte un habillement noir: par un ancien Maître,” 16 pouces 6 lignes by 13 pouces 6 lignes (45 × 36.5 cm); bought 38 livres by Caupris (unidentified). The work was not highly valued, and this may argue against identifying it with the Ottawa painting. Its measurements accord better with the copy in Oldenburg (see below).

The painting is first documented with the Earl of Darnley as Rubens’ “Old Woman” in a plan of the picture gallery at Cobham Hall, his country seat. (Darnley papers, Medway Archives and Local Studies Centre, Strood, Kent, U565.P67.) Undated, but before c. 1806/09, when the gallery was remodelled by James Wyatt. (For Wyatt’s alterations see J. Newman, The Buildings of England: West Kent and the Weald. 2nd ed. (New Haven, 2002), pp. 234–235.) There is a note of doubt running through the inventories of the Darnley collection as to the sitter’s identity: the 1831 probate inventory taken at the death of the 4th Earl describes it as by Rubens, “The head of his Mother when very old,” again listed in the picture gallery (National Arts Library, London, Ms. 86.OO.9, fol. 138), yet Douglas Guest’s 1833 catalogue of the pictures at Cobham Hall, which makes reference to a “Catalogue shewn to Visitors,” describes it under no. 22, “An old woman said to be his Mother,” by Rubens, valued at £40 (Darnley papers, U565.F27, vol. 1). A typescript list dated 1899 again catalogues it as “Rubens’ mother (supposed to be)” (Darnley papers, U565.H052, fol. 12, no. 79).

R. Arnold, Cobham Hall, Kent. (Rochester, n.d.), pp. 20–21, notes that Bligh bought two studies of heads by Rubens along with three other paintings for £300 in Brussels, December 1793, citing a letter from Bligh to his wife; the letter could not be found in the Darnley papers. It is unclear if this refers to the Ottawa painting – it is just possible that Bligh later sold the two studies, as he sold certain of the other works bought with them. Unfortunately, the entries in Bligh’s bank books in general lack detail, and few records of his art purchases survive.

Exhibition History

  • 1949 Montreal, Montreal Museum of Fine Arts, Masterpieces from the National Gallery of Canada (no catalogue)

Selected References

  • Neale 1823 (unpaginated) as Rubens, “Old Woman, said to be his Mother, Head”
  • Rooses 1886–1892, v. 4, no. 1111 was only aware of the copy in the Oldenburg museum (see below).
  • Goris and Held 1947, no. 35, autograph, c. 1615–16
  • Held 1980, no. 444, autograph, c. 1615–16; the panel structure is described inaccurately.
  • Baumstark in New York 1985, under no. 205, as autograph, the model for the maid in the Liechtenstein Discovery of Erichthonius painted c. 1616
  • McNairn in Laskin and Pantazzi 1987, pp. 261–262 as autograph, incorporated in paintings dated to c. 1615–17
  • Jaffé 1989, listed under no. 320 (The Discovery of Erichthonius,) as accepted by Held 1980 and Baumstark in New York 1985

Copies

I – Landesmuseum für Kunst und Kulturgeschichte Oldenburg, no. 15615, oil on oak. 43 × 36 cm. Marks and Inscriptions: verso, Antwerp brand.
Provenance: Possibly the Marquis de Marigny by 1781 (as above). Bought 1869 from Bartels (Berlin). Presumably Karl Bartels (1794–1868), although no 1869 sale is recorded.
Selected References: Rooses 1886–1892, vol. 4, no. 1111; H.W. Keiser, Gemäldegalerie Oldenburg, Landesmuseum Oldenburg. (Munich, 1967), p. 73 as workshop

II – Ex-Christie’s. Oil on wood. 48.9 × 39 cm. Provenance: according to the 1919 auction catalogue, the Prince de Polignac and then Jules Porgès (1839–1921); Brussels, 3–4 July, 1919, lot no. 16 as Rubens, “La Mère de Rubens.” With the Art Collectors Association, London 1921? Private collection, France; by bequest, and then by descent; sold Christie’s, London, 22 April 2005, lot no. 18, as Circle of Rubens, “Portrait of an Old Woman,” private collection.


1In addition to Rubens and Jordaens, she was also painted by Artus Wolffort, an artist outside of Rubens’ immediate circle, although influenced by him. He painted her in profile view in a tronie now in the Stift of St.-Florian (near Linz, Austria) and in a drawing after this now in the Louvre (inv. 20019); [illus. 4] See Vlieghe 1977, p. 117. Our thanks to Dr. Friedrich Buchmayr of the Stiftsbibliothek, St. Florian for his generous help.

2See Held 1980 vol. 1, pp. 597–599 for a concise introduction to Rubens’ study heads. The historic context and functions of these study heads is set out in Gottwald 2011, chap. IV.

3Assigning a date to it is difficult in the absence of a catalogue of Rubens’ tronies.

4For instance, works by the Carracci and their circle. They did not serve the same function, being exercises done to train the eye and hand rather than for use in future production.

5 At least one other is inferred through paintings that incorporate what must be the same woman’s features recorded from a different view and lighting, e.g. The Old Woman with the Candle, c. 1616–17 (Jaffé 1989, no. 430).

6 A set of prints reproducing study heads, anatomy studies, motifs, etc. that were in Rubens’ workshop. Likely published after his death, it made this material available for study by other painters, in the tradition of so-called “drawing books,” meant as exercises for young artists and as compilations of useful or characteristic motifs. See Bolten 1985, p. 103 and following, which sets this publication in context. This tronie may have been used in the Judith and Holofernes, c. 1617, in Brunswick (Jaffé 1989, no. 453) and in an addition made c. 1625 to the earlier The Drunken Silenus in Munich (Jaffé 1989, no. 428). It is likely not the “pair” to the Ottawa tronie, as the woman wears a different bonnet.

7See Mirimonde 1958 for a discussion of the characteristic types and uses of female tronies in Rubens’ work.

8 Jordaens also copied the profile view in a drawing; this is marked up showing proportions – suggesting a didactic or theoretical function. D’Hulst 1974, no. A52.

9 For example, The Descent from the Cross, Lille, Musée des beaux-arts, c. 1617–18; The Adoration of the Shepherds, Munich, Alte Pinakothek, c. 1619; The Adoration of the Shepherds, Soissons Cathedral, c. 1620 and various studies and paintings connected with this. Nico van Hout attributes the first to Jordaens, working for Rubens; see Nico van Hout, p. 41. Jordaens’ tronie of the old woman would require adaptation to be successfully integrated into a Rubens piece. Rubens did, at times, rely on other painters, sub-contracting with them if outside the workshop or, in the case of Van Dyck when he worked in the studio, allowing him considerable responsibility. Rubens’ 1640 studio sale includes tronies described as painted by Van Dyck, suggesting he may have been entrusted with this task when working with Rubens in the mid to late teens. See Nora de Poorter’s insightful comments in Barnes et al. 2004, p. 17. For an example of a tronie by van Dyck employed in a Rubens’ workshop picture, see Barnes et al. 2004, no. I.90. The work shows two views of a man’s head, one of which is used in the National Gallery, London, Drunken Silenus discussed here; the other appears in Van Dyck’s own Let the Children Come to Me in this exhibition.

10Miramonde 1958, p. 228 argues that St. Elizabeth is rather based on another tronie now in the Musée des beaux-arts, Besançon; Held 1980, vol. 1, no. A41 accepts this, but argues that the attribution of the tronie to Rubens is uncertain, and it is perhaps by Van Dyck. The saint’s physiognomy would seem to be closer to the Ottawa panel.

11Two painted sketches for it are known, Courtauld Institute and ex-Dorotheum (sold Vienna, 18 April 2012, lot no. 606). The attribution and their exact relationship to the painting are both unclear, although the latter is the better candidate for a preparatory sketch by Rubens himself. Neither shows the maid figure in the pose seen in the painting.

12See Keith 1999 for a detailed analysis of the picture and its place in Rubens’ workshop practice; he suggests Van Dyck was responsible for much of the work, including the figure of the old bacchante.

13Vergara in Madrid 2012–13, no. 60, which sets in it terms of Rubens’ workshop practice. Traditionally seen as partly by van Dyck working for Rubens; Vergara rejects this.

14A fragment of which survives in the Allen Memorial Art Museum, Oberlin College, see www.oberlin.edu/amam/Rubens.htm [accessed 18 March 2013]. The old maid is clearly derived from the Ottawa tronie; the Oberlin painting is broadly treated, and the difference in style shows how artists would have to interpret the earlier study heads to adapt them to Rubens’ later production.

15Of the two principal members, the youngest heartwood ring dates to 1585; assuming a median of 15 sapwood rings were removed and a short period for seasoning, the panel could have been available for painting early in the 1600s. Only two of the other members are datable; the youngest heartwood ring dates to 1624, and following these assumptions, would have been available from 1641 forward. Our thanks to Dr. Peter Klein for the dendrochronology.

16Rubens’ work played a vital – if controversial – role in French art, both celebrated and condemned. It remained of real interest to many artists, e.g. Eidelberg 1997 which describes Watteau’s copies after Rubens’ tronies.

17Rome 1977, no. 417, for the 1775 print by M. Ernst after a painting then in the collection of the Elector of the Palatinate in Mannheim, perhaps now in the Alte Pinakothek in Munich. The subject is identified as Rubens’ mother.

18 Paris, 26 and following December 1797, lot no. 27, “Une vielle femme, vu en buste, la tête coîffé de ses cheveux, relevés & attachés avec un ruban blanc, la chemise ouverte, et vétue d’un vêtement brun … ” 18 pouces ½ lignes × 14 pouces ½ lignes, i.e. 50 × 39 cm. Later with Carl Robert Lamm (1856–1938), Näsby Castle, near Stockholm; his sale, American Art Assocation, New York, 21 and following February 1923, lot no. 563, Rubens, “Portrait of an Old Woman (Said to be the artist’s mother-in-law),” on panel, 19¾” × 16.” Judging from the photograph in the auction catalogue, it cannot be by Rubens, and it may be a copy of a lost tronie. She compares well to the figure of the old maid in a Rubens’ workshop piece, The Head of Cyrus brought to Queen Tomyris, c. 1622–23 (Museum of Fine Arts, Boston), also once owned by Lord Darnley.

19 Penny 2008, pp. 458–452.

20 The impulse to give identities to study heads was powerful: the same woman in the Ottawa tronie was also identified as Rubens’ old nurse when she appeared in another painting (a version of the Old Woman with a Candle, Jaffé 1989, no. 430) circulating in the London art market at this time.

illus. 1

Paulus Pontius, plate from The Drawing Book, after 1640, engraving

illus. 2

Jacob Jordaens, Two Views of an Old Woman’s Head, c. 1620, oil on panel, 58.5 × 65.6 cm. Musée des beaux-arts, Nancy, inv. 91. Photo: C. Philippot

illus. 3

Peter Paul Rubens, The Holy Family with Saints Elizabeth and John the Baptist, c. 1615, oil on panel, 114.5 × 91.5 cm. Art Institute of Chicago, Major Acquisitions Fund, 1967.229. Photo: The Art Institute of Chicago

illus. 4

Here attributed to Artus Wolffort, Head of Old Woman in Profile, black and red chalk with white chalk highlights, 36.3 × 27.6 cm. Musée du Louvre, Paris (inv. 20019 recto). © RMN-Grand Palais / Art Resource, NY

illus. 5

Peter Paul Rubens, The Discovery of the Infant Erichthonius, c. 1616, oil on canvas, 218 × 317 cm. Liechtenstein Museum, Vienna. (Inv.-No. GE111)

illus. 6

Attributed to Anthony van Dyck, Drunken Silenus Supported by Satyrs, c. 1620, oil on canvas, 133.5 × 197 cm. National Gallery, London, purchased 1871 (NG853). Photo: National Gallery, London / Art Resource, NY

illus. 7

Peter Paul Rubens and workshop, Achilles discovered by Ulysses among the Daughters of Lycomedes, 1617–18, oil on canvas, 107.5 × 145.5 cm. Museo Nacional del Prado, Madrid, Donación duquesa viuda de Pastrana, 1889

illus. 8

Workshop of Peter Paul Rubens, Head of an Old Woman, oil on oak, 43 × 36 cm. Landesmuseum für Kunst und Kulturgeschichte, Oldenburg. Photo: Sven Adelaide

illus. 9

Peter Paul Rubens, Head of an Old Woman shown in a frame made for the work. Frame: carved and water gilded oak, Paris, second quarter 18th century. National Gallery of Canada, purchased 1925, no. 3339