A Young Draughtsman in a Broad-Brimmed Hat

On the didactic context of such studies, see the Getty exhibition catalogue Drawings by Rembrandt and His Pupils: Telling the Difference (Getty, 2009–10 ↗) and, most recently, Jochem van Eijsden's study of teaching practice in Rembrandt's workshop (Kroniek van het Rembrandthuis, 2024/25 ↗). The sheet entered this research programme in July 2026 through its custodian, Michael, who had begun submitting questions about it to the relevant museums.

Documented source RKD records Zittende jongeman aan een tafel(Seated young man at a table), no. 116036, as an anonymous work from the Northern Netherlands, seventeenth century, in black and white chalk on blue paper, with sheet dimensions of 352 × 251 mm. These are the comparator's catalogue details; they do not establish the maker, date or materials of the present drawing. Sotheby's 2005 catalogue, discussed in the next section, uses the title “filling his pipe” for this sheet and records a pencil inscription “Terburg”. That is a revision of the working description of the shared activity. It is not an authenticated signature, and it does not change the attribution of either drawing. RKD object record ↗

Comparison RKD 116036 has the same composition as the present sheet. The seated draughtsman, the broad hat, and the arrangement of the head, arms, clothing and legs correspond, as do several contour routes and boundaries within the drapery. Dresden C 1421 is a different design of a related subject and is discussed afterwards. The same composition does not establish that every individual stroke corresponds, or that the drawings were made by the same person.

Direct observation In the available RKD reproduction, a light outer strip and more than one boundary articulate the rear of the torso. In the present sheet, the same region is more simply bounded by a strong brown contour. With the same whole-image alignment used throughout the comparison, that contour lies near part of the inner modelling of the RKD back.

Working hypothesis This may represent a selective translation of the design, in which an internal modelling boundary becomes part of the outer silhouette. Identifying the exact line correspondence requires better reproductions. The visible boundary should not be described as the sitter's spine, and the current comparison does not demonstrate an anatomical error or a mechanical tracing.

Comparison The route taken by the back boundary in the present sheet makes the figure narrower there without producing an obvious anatomical incoherence. That observation is kept separate from a different question: what the RKD line itself is, whether an interior contour, an abandoned first line or something else, and when it was drawn. Both remain unresolved.

Comparison The breeches provide a further correspondence in the arrangement of folds and light areas, with a more explicitly linear treatment in the present sheet. The distinction between a shared contour route and the individual marks used to draw it matters when comparing different media.

Limit of the comparison The shadow beside the working arm shows clear horizontal, stepped hatching in the present sheet. An equivalent sequence is not clearly legible in the RKD reproduction, where this area appears more diffuse. Similarities should therefore be assessed alongside local differences, rather than described as an exact match of all strokes.

Choose a colour for each drawing, adjust their visibility, or return to the images without recolouring. The alignment is fixed: changing colours, opacity or zoom does not alter the relationship between the drawings. Use the back, breeches and shadow controls to inspect the same areas discussed above.

The images are aligned by uniform scaling, rotation and translation, without local warping. This is a comparison at a common display scale, not a display of the sheets at their physical sizes. Their reported sheet sizes differ. That difference does not rule out transfer through a resized reproduction or another copying process.

Viewing limit Colours follow image darkness and include paper tone and reproduction artefacts; they do not isolate drawing material. The small photograph of the present sheet and the available RKD reproduction do not allow a reliable comparison of every stroke's length, interruptions or order of execution.

Supported observation: a close relationship in the organisation of the designs, with both local correspondences and differences that can be inspected directly. On the present evidence, a shared design or a derivation is better supported than coincidence between two wholly independent studies. The identity of every stroke, a mechanical tracing, the direction and date of any transfer, and common authorship remain unproved.

Evidence-backed question: whether one drawing derives from the other, directly or through a reproduction, or whether both use an earlier design. A repeated design by one artist and a copy by another remain distinct possibilities.

Unresolved: the direction, date and means of any transfer, and the authorship and sitter of the present sheet. The overlay does not demonstrate a mechanical tracing or establish a common drawing session. Dresden remains relevant to the broader subject and attribution history, but differences from Dresden alone cannot establish independence from the same composition recorded as RKD 116036.

Next evidence: a better photographic source for RKD 116036, a perpendicular photograph of the present sheet, and the reproduction history of the RKD drawing. These would allow the back boundary and other particular contour routes to be tested more closely.

Present-sheet photograph supplied by the custodian. Comparator: RKD 116036, colour reproduction, digital object 9141481, used here for non-commercial research. Alignment and display derivatives prepared for this dossier. The colour file checked on 23 September 2026 is the same reproduction used to build the alignment. Catalogue details were checked against the RKD record on 19–20 September 2026 and belong to the comparator, not to the present sheet.

Documented source RKD attaches two image records to 116036: a colour reproduction, digital object 9141481, and a black-and-white photograph, 9158962. The available files differ in framing, margins, lettering and pixel dimensions, so the monochrome file is not simply a greyscale conversion of the colour one. That does not prove two independent photographic captures, and the source and date of either reproduction remain unknown. On 26 September 2026 the record, last updated on 9 April 2024, gave as provenance only the Sotheby's sale of 6 July 2005, lot 141, and listed no literature or exhibitions. RKD 116036 ↗

The sheets below are different poses. They are shown side by side for comparison. They are not layers in the present sheet–RKD viewer, and Dresden C 1421 is not returned to that selector. Bodies are not warped to force a match. No bibliography consulted for this section specifically attributes RKD 116036, or the present sheet, to Gerard or Moses.

Documented source Sotheby's, Old Master Drawings, London, 6 July 2005, lot 141, printed page 159, describes a young man in a hat seated at a table, filling his pipe. It gives black and white chalk on blue paper, 352 × 251 mm, and retains the attribution Dutch School, seventeenth century. The RKD record already cited in section 02 agrees on an anonymous seventeenth-century sheet of that medium and size, catalogued as Northern Netherlands. The sale reference and the matching illustration identify lot 141 with RKD 116036. They do not prove who owned the sheet before the sale. The printed reproduction is halftone and limits stroke-level comparison. RKD 116036 ↗

Limit The catalogue reports the pencil inscription “Terburg”. It does not describe that inscription as an artist's signature. “Mentioned on object”, elsewhere in this section, is the same kind of limit: a recorded name is not an authenticated signature. A dated family example, in which the father annotates a drawing by Harmen, is RP-T-1887-A-887 below. It does not identify the writer of “Terburg”. The published title revises the working description of the shared composition. The present sheet is not silently retitled, and the object in the hands of this drawing remains unresolved in the photographs.

Documented source The catalogue's own words for the mark are: bears pencil inscription, lower left: Terburg. It does not say signed. The crop below is that corner of the illustration printed with lot 141. It is a halftone. The letters cannot be read from it, so this page does not offer a new reading and does not treat the mark as an authenticated signature. The reading “Terburg” remains the catalogue's.

Documented source The custodian supplied a two-page PDF: the front cover and printed page 159. It is published here unchanged (PDF, 2 pages, 3.2 MB), with the two pages and the drawing as printed on page 159 below. The cover names the sale Old Master Drawings, including property from the Arthur Feldmann Collection, London, 6 July 2005. That title does not show that lot 141 was Feldmann property. The lot is catalogued as Dutch School, 17th Century, black and white chalk on blue paper, 352 × 251 mm, and reproduced as a halftone.

Documented source The same page carries a catalogue note, repeated in Sotheby's online entry for sale L05040: “One can easily imagine finding a figure based on this study from life in a painting by one of the Bamboccianti. Little is, however, known of the figure drawing style of most of those artists, so until this figure is actually found in a painting, this highly accomplished study must remain unattributed. The clue to its authorship may perhaps lie in the fact that although the figure type and approach to lighting seem strongly Italianate, there is also a certain stylistic affinity with figure drawings by Rembrandt school artists such as Backer and Flinck.” The printed estimate is £5,000–7,000 (€7,300–10,200; US$9,200–12,900). Neither the page nor the online entry records a result. Sotheby's, sale L05040, lot 141 ↗

Comparison The 2005 cataloguer recorded the “Terburg” inscription but did not discuss it. The stylistic suggestion pointed instead to the Bamboccianti, painters of Roman street life, and to Rembrandt-school figure drawings by Jacob Backer and Govert Flinck. Museum records confirm that both made figure studies in this medium: the Albertina holds Backer's knee-length study of a man with hat and cloak seated before a table, inv. 10212, black and white chalk on blue paper, once attributed to Jürgen Ovens; the Rijksmuseum's Flinck, RP-T-1977-40, is in the same medium and lists Backer as a rejected attribution; the Ackland Art Museum's Backer, 2017.1.1, is on blue-grey paper. These are cited without images. Albertina 10212 ↗ · Rijksmuseum RP-T-1977-40 ↗ · Ackland 2017.1.1 ↗

Limit The note is a 2005 opinion about lot 141, not evidence of authorship, and it bears on the present sheet only through the shared composition. The medium does not identify a circle: Ter Borch family drawings in black and white chalk on blue paper are also documented below. None of the three museum records mentions RKD 116036, and all three sheets are smaller than lot 141. The Ter Borch lead stands beside this older suggestion; neither decides the attribution.

Documented source Sotheby's online entry for lot 141 has no provenance section, whereas Feldmann lots in the same sale, such as lot 33, record Feldmann, the Gestapo seizure and the 2003 restitution to his heirs. ARTnews reported the Feldmann group as 55 drawings in a sale of 244 lots. RKD's provenance field for 116036 names only this sale. A lot printed without provenance has an undocumented earlier history, not a known one; no source consulted supplies a Feldmann provenance for lot 141. Lot 33 ↗ · ARTnews, 19 July 2005 ↗

Direct observation The dossier editor supplied three crops of the lower-left corner: from the RKD monochrome reproduction, 281 × 134 pixels; from the RKD colour reproduction, 249 × 135 pixels; and from the Sotheby's PDF, 381 × 208 pixels. They are shown at their full captured size. Faint cursive marks are visible, with a reproduction pattern and limited tonal separation. Enlargement makes the marks easier to inspect but cannot restore detail the reproductions do not contain.

Limit These are three reproductions of one drawing, not three signed works. They support examination of the reported reading “Terburg”; they do not authenticate a signature, identify its writer, date the inscription, or distinguish Gerard from Moses or a later annotator. Its autograph status remains unverified.

Comparison The comparison can now go beyond the broad-brimmed hat. In the Cleveland Museum of Art's A Man Standing Seen from the Back, catalogued as Gerard ter Borch, c. 1630, longer contours coexist with short fold accents, and soft shading and stronger dark marks distinguish the clothing's volumes. Hair, the transitions between jacket and breeches, and folded footwear are useful points of comparison. These are visual affinities, not a demonstrated match of individual drawing habits. Cleveland records black chalk with traces of white on brown laid paper, sheet 289 × 150 mm, and a graphite name on the verso. That name is recorded as an inscription, not authenticated as a signature. A Google Arts & Culture page repeats Cleveland's record; it is not an independent second attribution. Cleveland 1941.603 ↗ · Cleveland API ↗

Related seated and standing figures at Boijmans and the Städel, and a Sotheby's workshop drawing, broaden the comparison. The attributions differ. Some records name Gerard II, others keep a question mark or a Moses alternative from a note on the mount, and Sotheby's uses Studio of Gerard ter Borch II. Repeated folds and selectively strengthened outlines deserve comparison. This mixed group is not a uniform sample of one artist's hand.

Comparison The regional observations below summarise what the photographs support. Continuity of design, already discussed for RKD 116036, and the behaviour of the execution are different questions. Nothing here asserts the same hand, the same speed, the same tool, or the same hat.

The four plates below set cropped regions side by side: hair and contour, drapery and volume, footwear, and broad-brimmed hats. Scale, pose and medium differ between the crops. Each plate illustrates one row above and carries the same limit; the constituent images are credited in the comparison register at the end of this section.

Comparison The Rijksmuseum's Moses ter Borch, Drawingis a closer comparison of pen execution than the chalk figure studies. The sitter is Moses. The author, in the museum's record, is Gerard ter Borch II. Those roles are not reversed. The graphite or chalk beginnings and the variable ink reinforcement discussed by Korevaar and Tauber concern documented Ter Borch works, not an established sequence in the present sheet. RP-T-1887-A-1202 ↗ · Korevaar and Tauber 2014, p. 360, fig. 8 ↗

Documented source The Rijksmuseum's Girl Reading, also titled Portrait of a Girl Reading a Book, RP-T-1887-A-764, is by Gerard ter Borch the Elder, the father of Gerard II and Moses. The museum dates it c. 1630–1635, records it as made in Zwolle and signed by the artist, with the inscription “Ger Sr; G Ter Borch Sr”, and cites Kettering GSr 107, vol. I, pp. 78–79. Pen in brown, brush in brown and grey, red chalk, 97 × 88 mm. The Wikimedia Commons description calls the girl probably the artist's eldest daughter Anna; the current museum record does not name her. RP-T-1887-A-764 ↗

Comparison Like the present sheet and Gerard II's drawing of Moses, it records a family member absorbed in an occupation: the head bowed, the eyes lowered to the hands, drawn at close range. Its medium, brown pen with brush, is nearer to the present sheet than the chalk drawings in this section. In the reproduction, the grey and brown washes stay within the figure on an untouched ground. That is comparable to the local, contour-bound wash that section 07 identifies in the present sheet as a divergence from the Rembrandt-school exemplars. The modelling differs: here it is carried by wash, with parallel pen strokes only in the hair, where the present sheet relies on serial hatching.

Documented source Repetition and transfer are documented in works by and associated with Ter Borch. The Rijksmuseum's three studies of a horse and sleigh show one motif repeated at different sizes. Korevaar and Tauber discuss that sheet alongside evidence of transfer from blackened drawing versos. The Städel counterproof 2916 Z, above, is a separate documented transfer, limited to its pair with 2917 Z. Neither example establishes how the present sheet and RKD 116036 became so closely related, and neither explains their difference of scale. The sleigh is not claimed to be geometrically exact, and it is not offered as the transfer mechanism, or as proof of tracing, for this case. RP-T-1887-A-789 ↗ · Korevaar and Tauber 2014, pp. 352–353 ↗

Documented source The full article was read on 26 September 2026. It is a technical study of two pairs of painted portraits: Godard van Reede, in the Rijksmuseum and at Slot Zuylen, and Jacob de Graeff, in the Rijksmuseum and in Saint Louis. It concludes that Ter Borch painted both versions of each pair (pp. 349, 376–377). It does not study drawings, and it does not mention the present sheet or RKD 116036. What it offers this case is a method.

It orders two versions by their changes. The Oud-Zuilen portrait follows the final painted hairline of the Rijksmuseum portrait and lacks that portrait's one pentimento (p. 361). In the De Graeff pair, the eyes were moved early in the Amsterdam version and the corrected position was copied to Saint Louis (p. 376). Short, faint parallel lines, shifted by 2–3 mm, are read as the double register of a mechanical transfer (p. 362). Three transfer methods are defined: tracing through a blackened or whitened verso, pouncing through pinpricks, and the counterproof, which gives a mirror image (note 11, p. 378). Blackened versos survive in the Ter Borch family drawings (pp. 352–353, fig. 3). The sheet in those figures is RP-T-1887-A-810, given in the article to Gerard II or Harmen. It is not RP-T-1887-A-887. No blackened verso, transfer or watermark is established for A-887. Gerard II could also copy free hand, enlarging or reducing a motif while keeping its parts in proportion (p. 353, fig. 4).

The authors also set a limit on handling. The Saint Louis portrait is so much freer that “one might suspect another hand”. They still give it to Ter Borch, because the two versions share pentimenti and construction, and they note that the replica has minor pentimenti of its own (p. 376). Their list of Deventer pupils (Caspar Netscher, Antoni Jordens and Bartholt Berentsen, with Pieter van Anraedt and Roelof Koets accepted as working in the studio) is the documented circle for paintings (note 4, p. 377). None of them is proposed here for either drawing.

Working hypothesis Applied to this case, the method narrows the questions without answering them. The superposition above uses no reflection, so the two figures face the same way; a counterproof, as defined in the article, is not their direct relation. The reported sheet sizes differ, about 206 × 153 mm against 352 × 251 mm, but the figures have not been measured on the sheets. Tracing, pouncing and counterproofs transfer a design at the same size. If the figures prove to differ in size as the sheets do, none of the three methods linked the two sheets directly; a resized intermediary or proportional free-hand copying would remain possible. Proportional copying is documented practice (p. 353), not an observation on these sheets. The ordering test asks which drawing carries first placements that the other lacks. The RKD back shows more than one boundary; the present sheet shows one, and a possible first placement at the hat crown is recorded as a possibility only (section 05). Neither can be read as a change of mind, or put in order, from the present reproductions. Differences of handling and medium do not decide the hand, in either direction.

Physical checks Four checks follow and are added to the conservation session: the verso for blackening or chalk rubbing; raking light for stylus indentation along the contours; transmitted light for pinpricks; and magnification for doubled, slightly shifted lines. A negative result on the present sheet would not exclude free-hand copying. Korevaar and Tauber 2014, pp. 348–381 ↗

Comparison The hat in Gerard II's A Boy Picking Fleas off a Dog is a further costume comparison, and a secondary one. Its crown and folded brim belong to a comparable visual repertoire. The painting does not establish the same physical hat, the same sitter, or the same drawing hand. The museum record consulted for this dossier does not publish a signature. Image dimensions are 344 × 271 mm; the support is canvas laid on wood. This painting does not lead the question of authorship. Alte Pinakothek, inv. 589 ↗

Comparison The Moses group is kept separate from Gerard's signature control. The two chalk self-portraits are material comparisons for the catalogue description of RKD 116036, which is also black and white chalk. They are not evidence that Moses was the sitter or the maker of the present sheet, whose visible medium is brown. The horse is an inscription control: the Rijksmuseum describes it as signed by the artist. That institutional statement does not authenticate the pencil inscription on lot 141.

Documented source The Rijksmuseum's Portrait of Gerard ter Borch the Elder, RP-T-1887-A-1042, is by Moses ter Borch and is dated by its inscription. Along the lower edge, in pen and brown ink and in the father's handwriting according to the museum, it reads “Mosus ter Borch. nae het leven getijckent den 23. Januarij 1660”: drawn from life on 23 January 1660. The museum records it as made in Zwolle, when Moses was fifteen and his father nearly eighty. Black chalk heightened with white, some black brush, on blue paper, 287 × 230 mm; Kettering M 54, vol. I, pp. 316–317, as cited by the museum. RP-T-1887-A-1042 ↗

Comparison The resemblance to the present sheet lies in the hat and its handling. Both show a very broad brim and a raised, round-topped crown, with hair falling below the brim, and in both the brim is shaded in parallel hatching. The medium is the black and white chalk on blue paper recorded for RKD 116036, as in Moses's two self-portraits above. The differences are as plain: an old man in a frontal bust rather than a youth seated in profile, and chalk rather than the brown medium of the present sheet.

Documented source The Rijksmuseum records RP-T-1887-A-887, Zittende, tekenende man (Gerard de Jonge?), translated here as Seated man drawing (Gerard the Younger?), under Harmen ter Borch (mentioned on object). The English title is a translation, not a second official identification. The question mark in the title stays. The date is 13 January 1650, and the place is Zwolle (possibly). The drawing is black chalk on paper; the sheet measures 154 × 139 mm. These are the museum’s particulars, not findings from the original. Rijksmuseum record ↗ · Dutch record ↗

The brown-pen annotation along the lower edge reads “Her[man].t.Borch. naet leve[n]. geteickent de[n] 13.Januarij Anno 1650”: “Her[man] t. Borch, drawn from life on 13 January 1650.” Geteickent means drawn, not a signature. The museum assigns the handwriting to Gerard ter Borch the Elder. The annotation is in pen; the figure is catalogued as black chalk. This is a control for the reported “Terburg” inscription on RKD 116036: a name on a drawing can be written by someone else. It does not show who wrote “Terburg”, whether that inscription is contemporary, or whether either sheet belongs to the family.

Comparison In the museum image: a broad hat with an uneven brim, curled locks, short curved marks at the sleeves, angular and curved boundaries in loose breeches, grouped hatching, and stronger dark accents at selected folds. The chalk handling is relevant to the RKD chalk drawing, and the sheet adds a documented seated figure in Harmen’s recorded work. The figure is more frontal, with different arms, legs and furniture. It is not the same design as the present sheet and RKD 116036. It is shown in its own orientation, not in the registered overlay, and no geometric agreement is claimed. Short or interrupted marks can come from medium, paper and reproduction. They do not establish a hand or a speed. The hat and clothing are period features, not exclusive family traits.

Limit This sheet does not establish Harmen, Gerard or Moses as the maker of the present drawing or of lot 141. It supplies no common sitter, paper or date for those sheets. No watermark description was found in the inspected public record for A-887. That is a limit of the record, not evidence that the sheet has no watermark. The same date on RP-T-1887-A-888, Seated Child and a Child Playing the Violin, also under Harmen, with an annotation the museum likewise gives to the father, documents a family habit of annotation. It does not prove a shared sitting. RP-T-1887-A-888 ↗

Documented source The museum cites Alison McNeil Kettering, Drawings from the Ter Borch Studio Estate (The Hague, 1988), vol. I, H 69, p. 226, illustration p. 227, displayed as “Kettering 69-H”. That printed entry has not been read for this update. H 68, the adjacent entry for A-888, is cited by the museum on the same pages and is useful for the father’s annotations. It is not an independent reassessment of H 69. Rijksmuseum publication record ↗

Working hypothesis The evidence supports a focused research question: does the RKD design have a connection to the Ter Borch family or workshop? Even a future attribution of that sheet would not identify the person who made the present drawing. Same-hand repetition, workshop copying, another intermediary, and later derivation remain distinct possibilities. A later copyist is not ruled out. These examples do not date the present sheet, and they do not make Titus, Moses, or Gerard the identity of the youth.

Documented source Consulted for this lead, on 25 September 2026, were the specified pages and records, not every volume in full. Kettering 1988 was not read directly for those stylistic comparisons; references were checked through institutional records and the 2022 Sotheby's note. Part I was later consulted for the seven watermark entries and the correction of M 18 in section 06. The watermark plates in part II, pp. 776–802, are still outstanding and have not been inspected. That paper reading does not reassess the stylistic comparators on this page. The 2022 note is not used to assign the other comparators to the unidentified-pupil appendix. Rijksmuseum bibliographic record ↗. Korevaar and Tauber 2014 was first read at pp. 352–353 and 360, and read in full on 26 September 2026 from a copy supplied by the dossier editor; it does not study these two drawings. DOI ↗. John Smith's 1842 supplement, Gerard Terburg, pp. 529–537, and Émile Michel, Gérard Terburg (Ter Borch) et sa famille (1887), are historiography. They mention young men, hats, study scenes, a pipe, family resemblances, and the Munich boy with a dog. No object-specific link to RKD 116036 or to the present sheet was identified, so those generic subjects are not used as proof of authorship. Michel is not a current authority for reassigning either drawing.

Luijten's 2005 article on Moses learning to draw, the 1974 and 2004/2005 exhibition catalogues, Hofstede de Groot 1913, and Plietzsch 1944 were not used as read sources for this section. The immediate priorities are a better photograph of the RKD original and of its inscription, and direct examination of the relevant catalogue entries and attribution history.

Repeated unusual details or spatial ambiguities can help investigate design transmission. Their absence does not by itself exclude copying, because a copy may select, simplify or alter its source. The documented observations about C 1421 below concern that drawing; they must be distinguished from this dossier's interpretation of the present sheet. C 1421 presents two documented anomalies, both noted by Hofstede de Groot in 1890 and reported in the 2004 catalogue entry: the desk plane lacks ground anchoring, and the thighs are drawn overlong, with the desktop appearing to rest upon them. The annotated comparison below records how the present sheet treats the corresponding passages.

The research notes are published in readable form, each with a Markdown download: research notes. Notes 00–14 record the September 2026 research. Where they say that no Ter Borch Strasbourg-lily record had been found, or direct a watermark enquiry only to RP-T-1887-A-1327, that result is superseded by note 15 and note 16, prepared 5 October 2026. The watermark evidence table is separate. The public research map (CSV) lists each source note with its status; private material is withheld. The 2005 catalogue pages (PDF, 3.2 MB) are the unchanged scan.

Submitted, not endorsed The appendix below reproduces the 22 presentation images received with the case, as documents of what was submitted. They are not ConfirmArt findings. Where an image asserts an attribution, a sitter, a provenance, a paper match or a named “modello”, that claim is not supported by the evidence on this page. The responses are in the notes on evidence status, paper and watermark, provenance and two-hands claims, and in section 06 of this page.