A 2021 technical and provenance study established that the 'Portrait of an Old Man' in the J. A. van Tilburg Collection of the University of Pretoria Museums — attributed to Rembrandt van Rijn since the 1890s — is a deliberate nineteenth-century fabrication: micro-XRF mapping found zinc white and barium sulphate with no lead-based ground, materials incompatible with a seventeenth-century creation, even though the dendrochronologically dated oak panel (earliest creation date 1646), a cancelled 'HVU' stamp and a purported Rembrandt invoice had supported the attribution. The study left a precise and still-open agenda: eight further versions of the same composition survive (New York/Leiden Collection, Haarlem/Teylers Museum, Paris/Louvre, London/Ranger's House, Brussels/Geneva, and three of unknown whereabouts), none of which has undergone comparable non-destructive technical analysis, so their material histories and authorship remain undetermined; the maker, workshop and market pathway of the Pretoria forgery remain unidentified; and the reliability of the early twentieth-century chemical authentication protocols used by A. M. de Wild has never been systematically reviewed. The objective is a complete, evidence-weighted authentication record for this group of nine paintings.

Background

Rembrandt van Rijn (1606–1669) is the most intensively studied painter of the Dutch Golden Age, and deciding which paintings in his style are actually by his hand has occupied scholars for centuries. The Rembrandt Research Project (RRP), founded in 1968, spent roughly fifty years re-examining the traditionally accepted oeuvre and publishing the multivolume 'A Corpus of Rembrandt Paintings'; its final volume, prepared by Ernst van de Wetering (2014), again changed the status of many works. The RRP's method combined connoisseurship — attribution by expert judgment of style and quality — with the physical study of paintings, yet critics such as Gary Schwartz (2018) have documented that even the Corpus's final judgments were sometimes delivered as expert opinion without published physical analysis. A 'tronie' (Dutch for 'face') is a study of a head or character type rather than a commissioned portrait, a common genre in Rembrandt's workshop; these small panels are especially hard to attribute because they were freely painted and copied within the studio. One such composition, an elderly bearded man in profile with a fur cap, gazing toward a book, exists in multiple versions. The version now in the Leiden Collection, New York (inventory RR-109, 'Portrait of an Old Man (Possibly a Rabbi)') has been attributed variously to Rembrandt, his pupil Gerbrand van den Eeckhout, and Samuel van Hoogstraten; van de Wetering judged it autograph and the primary version, and the Leiden Collection catalogue (4th edition, 2023–) still carries it as Rembrandt, ca. 1645. The catalogue's list of further versions and copies is longer than the group studied here: de Kamper and McGinn reconcile the Leiden Gallery's 2017 report, which lists eleven artworks, to the nine distinct versions counted in this problem, believing that three of the entries refer to the same painting. Provenance research — the reconstruction of chains of ownership through auction catalogues, stock books and archives — by de Kamper and McGinn (2021) traced one of those versions, the 'Pretoria painting' in the University of Pretoria's J. A. van Tilburg Collection, through London sales from 1899 to a 1947 Dutch auction of wartime-confiscated goods, before it reached South Africa. Their technical analysis illustrates the standard tools: dendrochronology (dating a panel by its wood's annual growth rings, which gave an earliest creation date of 1646, consistent with Rembrandt's lifetime), and X-ray fluorescence (XRF, an elemental analysis of paint that distinguishes pigments; scanning micro-XRF produces spatial maps of the element distribution). Whereas Rembrandt's practice relied on lead white (confirmed by X-ray diffraction studies of his impasto), the Pretoria painting showed no lead-based ground and instead large quantities of zinc and barium sulphate — pigments (zinc white, available only after 1834; blanc fixe from the 1820s; their mixture lithopone from 1874) that cannot have been used in the 1650s. The painting is therefore a nineteenth-century fake whose deceptively 'authentic' features — a scavenged old panel, a cancelled stamp possibly evoking Hendrik van Uylenburg (a dealer closely associated with Rembrandt), and a purported Rembrandt invoice — were assembled deliberately to deceive. The same study notes that the painting had been chemically 'authenticated' in the 1940s by A. M. de Wild, a leading scientific restorer of the era whose method sampled paint with a steel awl for microscopic examination, and that none of the eight sibling versions had been tested for authentication.

Problem Statement

Determine the material histories and authorship of the nine known versions of the composition 'Portrait of an Old Man (Possibly a Rabbi)' — a small tronie of a bearded man in profile wearing a fur cap — and reconstruct the market pathway of the fabricated version, as left open by the 2021 University of Pretoria study. Concretely: (1) provide, for each of the eight versions not yet subjected to authentication-grade technical analysis — the Leiden Collection New York version (RR-109, currently carried as an autograph Rembrandt, ca. 1645, whose published technical record comprises x-radiography, infrared reflectography and a dendrochronological terminus post quem of 1631, without authentication-grade elemental analysis), the Teylers Museum Haarlem chalk drawing, the Louvre version (RF 2384), the Ranger's House London version, the ex-van Gelder Brussels/Geneva version, and the three versions whose current whereabouts are unknown — an evidence-weighted assessment, from the published technical, documentary and archival record, of whether its materials, construction, support and any dendrochronological evidence are consistent with seventeenth-century manufacture (by Rembrandt, his workshop or his circle), with a copy in period-correct materials of undetermined authorship or date, or with later fabrication, specifying the non-destructive analyses still required where that record leaves the question open, appropriate to each version's materials and support, and what each such analysis would decide; the assessment returns, for each of the eight versions, either its best-supported determination or an explicit cannot-be-determined finding, rests for the three versions whose current whereabouts are unknown on the available documentary, published and archival record, and, among the versions found consistent with seventeenth-century manufacture, identifies which one, if any, is the autograph prime version, while any version found to be of later fabrication is distinguished as deliberate forgery or non-deceptive copying by the criterion applied to the Pretoria painting, applied, for a work without a panel such as the Teylers chalk drawing, to its drawing materials and support; (2) identify, if the evidence permits, who produced the nineteenth-century Pretoria fabrication, in what workshop or network, and by what techniques the deception cues — the scavenged oak panel with a 1646-eligible ring sequence, the cancelled 'HVU' stamp on the frame, and the purported Rembrandt invoice of 1899 — were assembled; (3) pursue further provenance studies of how the painting moved through the 1899–1947 art market, addressing the links the 2021 study could document only by presumption — the private sale presumed to have followed the restorer E. Markey's purchase at the 11 June 1917 Christie's sale, the unrecorded interval between Walter Dowdeswell's December 1919 purchase and W. N. Ferguson's May 1927 sale, and the dating and authorship of the panel cradle, presumed to have been added by Markey; and (4) undertake further research into the methodology of the early twentieth-century chemical analyses used to authenticate the painting — exemplified by A. M. de Wild, who sampled paint with a small steel awl for microscopic examination (his 1928 book) and on whose advice the painting was purchased in 1941 — addressing how such tests, which the source calls 'not quantitative or definitive', could produce a false indication of age or authenticity, as happened with this painting, together with possible re-authentication of works tested using such methodologies. The single overarching objective is a complete, evidence-weighted authentication record for this group of nine paintings, in which expert opinion is weighed as one evidence class alongside connoisseurial, documentary and technical findings, and each version is assessed — against the published technical, documentary and archival record — as consistent with seventeenth-century manufacture (by Rembrandt, his workshop or his circle), with a copy in period-correct materials of undetermined authorship or date, or with later fabrication, or is explicitly reported as not determinable from the record, and the record identifies which version, if any, is the autograph prime version, answering the question the source study leaves open: which version is the original. Where later manufacture is established, deliberate forgery is distinguished from non-deceptive copying by the criterion applied to the Pretoria painting: anachronistic materials that survive the restoration hypothesis, combined with assembled authenticity cues (an evocative stamp, a fabricated invoice, a scavenged old panel) evidencing an intention to deceive from first creation.

References

  1. Gerard de Kamper and Isabelle McGinn, 'Of Unknown Men: Rembrandt or Not? A South African Provenance Story', de arte 56(1) (2021) 5–22, DOI 10.1080/00043389.2021.1882768, https://doi.org/10.1080/00043389.2021.1882768.
  2. David De Witt, 'Portrait of an Old Man (Possibly a Rabbi)', in The Leiden Collection Catalogue, 4th ed., ed. Arthur K. Wheelock Jr. and Elizabeth Nogrady with Caroline Van Cauwenberge (New York, 2023–; entry dated 2017), https://theleidencollection.com/artwork/a-portrait-of-a-rabbi/ (accessed 26 August 2026).
  3. Ernst van de Wetering, A Corpus of Rembrandt Paintings VI: Rembrandt's Paintings Revisited – A Complete Survey (Dordrecht: Springer, 2014), DOI 10.1007/978-94-017-9240-0, https://doi.org/10.1007/978-94-017-9240-0.
  4. Joconde/POP database notice 000PE008574, 'Tête de vieillard', Musée du Louvre, inv. RF 2384 (record updated 28 February 2023), https://pop.culture.gouv.fr/notice/joconde/000PE008574 (accessed 26 August 2026).
  5. Jilleen Nadolny, 'A history of early scientific examination and analysis of painting materials ca. 1780 to the mid-twentieth century', in Conservation of Easel Paintings, ed. Rebecca Rushfield and Joyce Hill Stoner, 2nd ed. (London: Routledge, 2020) 354–358, DOI 10.4324/9780429399916-24, https://doi.org/10.4324/9780429399916-24.
  6. Victor Gonzalez, Marine Cotte, Gilles Wallez, Annelies van Loon, Wout de Nolf, Myriam Eveno, Katrien Keune, Petria Noble and Joris Dik, 'Unraveling the Composition of Rembrandt's Impasto through the Identification of Unusual Plumbonacrite by Multimodal X-ray Diffraction Analysis', Angewandte Chemie International Edition 58(17) (2019) 5619–5622, DOI 10.1002/anie.201813105, https://doi.org/10.1002/anie.201813105.
  7. Gary Schwartz, 'The Transparent Connoisseur 5: Keeping the Rembrandt Research Project to Its Word', garyschwartzarthistorian.nl, 8 May 2018, http://www.garyschwartzarthistorian.nl/364-the-transparent-connoisseur-5-keeping-the-rembrandt-research-project-to-its-word/.