Christ Crucified between the Two Thieves ("The Three Crosses")

In 1653 Rembrandt finished, signed and dated a large drypoint of Christ's crucifixion built almost entirely from scratched, unworn burr rather than etched line - a technical risk almost no printmaker before him had taken at this scale. He considered it done: the plate carries his signature and the date at the third state. Then, at a point scholars still cannot agree on, he went back into the same copper and tore it apart. Figures were scraped and burnished into oblivion, a horseman and a man in an improbable tall hat appeared where a crowd had stood, and thick, driving diagonal strokes turned an already sombre scene into something close to total darkness. States I to III and states IV and V are so different from each other that this record treats them, deliberately, the way it would treat two related but distinct objects - because for more than a century after Rembrandt's death, print connoisseurs were not sure the two halves of this sequence came from the same plate at all.

This record anchors on the Rijksmuseum's impression of the fourth state, printed on Japanese paper, one of two Three Crosses impressions the museum received in successive years from the Amsterdam collectors Isaac de Bruijn and Johanna Geertruida de Bruijn-van der Leeuw. The museum also holds an impression of the clean, signed third state and a first-state impression pulled, unusually, on vellum rather than paper - so the whole arc from finished image to its own destruction and remaking sits, physically, in the same print room. The reasons scholars give for why Rembrandt did this range from the practical (drypoint burr wears out fast) to the deliberately experimental (a documented practice, across several of his late plates, of manufacturing collectible variation) to the theological (a literal staging of the Gospel of Luke's darkness at the ninth hour). This record states plainly what it could and could not confirm about each explanation, and about when, exactly, the second half of this story happened.

What this record anchors on

The impression at the centre of this record is Rijksmuseum object RP-P-1962-40: a drypoint printed on Japanese paper, sheet 375 by 450 millimetres, catalogued by the museum itself under Bartsch 78, Hollstein (Dutch) 78 and New Hollstein (Rembrandt) 274, state 4 of 5. It came to the museum by bequest in 1962 from Mr and Mrs De Bruijn-van der Leeuw of Muri, Switzerland, and carries an as-yet-unidentified collector's stamp catalogued as Lugt 228. This is not a lifetime-clean, gallery-friendly impression of a settled composition. It is a sheet from the plate's second life: the state Rembrandt produced only after scraping away large parts of the image he had already signed and dated as finished three years, or perhaps as many as seven or eight years, earlier.

The print as a whole shows Christ on the central cross between the two thieves, with Mary supported by the Evangelist John, Roman soldiers on foot and on horseback, and a crowd of bystanders arranged around the base of the crosses, worked in a scene the Rijksmuseum's own catalogue text describes simply as Christ hanging between two thieves, surrounded by apostles, his mother, followers and soldiers. What separates this print from almost anything else in Rembrandt's graphic output is that it was executed, in its first three states, almost entirely in drypoint - scratching a needle directly into the copper rather than biting a line with acid - a technique that, at this scale and level of ambition, the museum's own text states had almost never been done before. Drypoint gives a line a soft, velvety burr, but that burr is fragile: it flattens and disappears after a small number of pulls through a printing press. Rembrandt built his single most ambitious composition of the early 1650s on the medium least suited to being printed more than a few dozen times, and that fact turns out to be central to everything that happened to the plate afterward.

The subject, and the darkness the Gospel itself specifies

The scene is drawn from the Passion narrative shared, with variations, across all four Gospels: Christ crucified between two convicted thieves, one traditionally read as penitent and one not, with the Virgin, the beloved disciple, Roman soldiers and a crowd gathered at the foot of the crosses. The Rijksmuseum's own cataloguing cites Matthew 27:34-44, Mark 15:23-32, Luke 23:33-43 and John 19:18-24 for the print as a whole. One detail specific to Luke's account becomes unavoidable once the plate's fourth state is in front of you: Luke 23:44-45 states that from the sixth hour until the ninth, darkness came over the whole land, because the sun's light failed. States I to III show that darkness only as an atmospheric backdrop to a legible, populated scene. State IV does something closer to literalizing the text: thick, driving diagonal drypoint strokes fall across both sides of the composition like curtains closing in on Christ, obscuring most of the crowd that had been clearly visible moments (in the plate's working life) before.

Curators at more than one institution holding a fourth-state impression read this as a genuine shift in what the image is doing, not merely how it looks. The Cleveland Museum of Art's own catalogue text for its impression describes the intensified blackness as an active force that threatens to extinguish the light of Christ, explicitly tying the change to Luke's account of darkness covering the earth. A description associated with the Smith College Museum of Art's impression goes further, arguing that where state III let the viewer experience the crucifixion's trauma through the bystanders' visible reactions, state IV removes nearly all of them and concentrates attention on Christ alone in the dark, a shift severe enough, in that reading, to cast doubt on the redemption narrative the earlier states more straightforwardly told. This record treats both readings as exactly what they are: documented curatorial interpretations of a real, physical change to the plate, not a claim about what Rembrandt privately believed. No source consulted here supplies a document in which Rembrandt explains his own religious intention for this print, and this record does not supply one either.

States I to III: the print Rembrandt considered finished

Three states of the plate are documented before the radical reworking, and by the account given across the institutional sources consulted, the changes between them are changes of degree rather than of kind: further shading and tonal deepening added state by state, with the overall cast of figures and the composition's basic structure staying legible and consistent throughout. It was only at the third state that Rembrandt added his signature and the date, Rembrandt f 1653, at the bottom centre of the plate - an act one museum's own catalogue text reads plainly as evidence that Rembrandt regarded this state as finished and ready for the market. Sources consulted for this record differ on exactly how many impressions of the earliest states survive; one secondary account puts the figure at roughly twenty-one known impressions of the first state and roughly ten of the second, though this record could not verify that count against a primary catalogue directly and states it here as reported rather than confirmed.

The Rijksmuseum holds impressions from two points in this clean sequence: a first-state impression pulled, unusually, on vellum rather than paper (RP-P-1961-1196A, from the same De Bruijn-van der Leeuw bequest, one year before the museum's Japanese-paper fourth-state sheet arrived), and a third-state impression on paper (RP-P-OB-617), carrying the recurring, unidentified Lugt 240 mark discussed further below. Both show the print as Rembrandt left it before he went back into the plate: a crowded, legible crucifixion scene, tonally dark in places but with every figure the composition requires still visible.

State IV: the plate remade

What happened between the third state and the fourth is not a further refinement. It is, in the words used by more than one institution holding an impression of it, a re-invention of the print - compositionally, tonally, and in what the image means. Rembrandt scraped and burnished away large portions of the figures he had drawn at the sides of the composition, replacing much of the crowd with heavy, driving parallel drypoint lines that read, in a well-inked impression, as an almost solid field of shadow. New elements were drawn into the reworked space: two mounted soldiers introduced at the left, a redrawn group around Mary and John at the right, and, most discussed in the literature consulted for this record, a mounted figure in a tall, non-European hat where a differently dressed, kneeling centurion had stood in the earlier states. The Roman soldiers' medieval-style armour and feathered plumes, present in states I to III, were replaced in the reworking by the more exotic hat, a costume change more than one source treats as a deliberate revision of Rembrandt's own earlier, historically inconsistent staging rather than an accident of the burnishing process. The second thief's cross, at the right, is thrown into a shadow deep enough that the figure is far harder to read than it was before.

Every institutional description of this transition consulted for this record agrees on the scale of the change; the Fitzwilliam Museum's own text for its own third-state impression calls the resulting fourth state much more sombre, much more bleak, and describes the narrative as having shifted backward in time, from the moment just after Christ's death to the moment of his own cry of abandonment. This is the single most dramatic plate reworking documented for any print in Rembrandt's output, and it is the reason this record exists in the form it does: an ordinary catalogue entry for a signed, dated, finished print from 1653 would not prepare a reader for what the same copper produced afterward.

Why Rembrandt did it: several explanations, and one this record could not confirm

The sources consulted for this record offer three genuinely distinct explanations for the reworking, and none of them, on the evidence available here, is a story about rescuing a print that had failed to sell.

The first is practical. Drypoint burr wears down under the mechanical pressure of a printing press far faster than an etched or engraved line does; the Cleveland Museum of Art's own catalogue text states that the plate's delicate burr had already worn after roughly forty impressions of the first three states, and that Rembrandt reworked the plate, at least in part, simply to keep using a copper that could no longer print the image as originally drawn. On this account the dramatic result is a byproduct of a mundane maintenance problem: once large passages had to be re-cut or replaced anyway, Rembrandt chose to replace them with something new rather than a repair of the old.

The second explanation is deliberately artistic. Robert Fucci, whose 2015 exhibition and catalogue Rembrandt's Changing Impressions at Columbia University's Wallach Art Gallery examined roughly eighteen of Rembrandt's most dramatically altered late plates as a group, argues that Rembrandt thematized the notion of change itself in his printmaking practice, using successive states of the same plate to create multiple variations of a single image, both as an artistic experiment in what drypoint and burnishing could do and as a way of giving serious collectors comparative material worth owning in more than one form. On this reading the Three Crosses is not an isolated case of a plate going wrong; it is one of Rembrandt's clearest demonstrations of a working method he pursued deliberately across his late career.

The third explanation is theological, and is really a description of the reworking's effect rather than a claim about Rembrandt's stated intent: whatever the technical or commercial reasons behind it, the resulting image visualizes Luke's account of darkness at the crucifixion more literally and more completely than the earlier states did, shifting the emotional register described above from public spectacle to isolation. This record found no source that frames the reworking as an attempt to rescue a commercially unsuccessful print - the market evidence discussed below, including a documented culture of collectors actively seeking out variant states, points the other way, toward Rembrandt's plates as objects whose value the artist himself had reason to multiply, not repair.

When state IV was actually made: a dating dispute the sources do not agree on

Exactly when Rembrandt carried out the reworking is a genuinely unsettled question, and this record states the disagreement rather than picking a side it cannot support. Three distinct dates appear across the institutional sources consulted here, and they are not close together. The Fitzwilliam Museum's own text states that Rembrandt reworked the plate within a few months of signing the third state in 1653, putting the whole sequence inside a single year or two. A Sotheby's London cataloguing of a fourth-state impression, offered as lot 49 in its 17 March 2015 Prints and Multiples sale, instead dates the reworking to circa 1658. A third position, long the standard art-historical dating and still the one used in the title of the Mildred Lane Kemper Art Museum's own impression (catalogued as The Three Crosses, 1660-61), places the reworking as late as 1660 or 1661, on the strength of a specific stylistic comparison: the tall hat worn by the mounted figure introduced in state IV closely resembles the hat worn by the central figure of Claudius Civilis in Rembrandt's 1661 painting The Oath of the Batavians, now in the Nationalmuseum, Stockholm. The circa 1660-61 date, if correct, would place the reworking in the same narrow span of years as Rembrandt's 1661 painted Self-Portrait as the Apostle Paul, now in the Rijksmuseum, a work this record includes as an artist portrait precisely because it falls inside that same disputed window.

That stylistic argument has since been challenged on harder evidence. According to a summary of Erik Hinterding's research consulted for this record, watermark analysis - identifying the paper mill and production period of the specific sheets on which early fourth-state impressions survive, independent of any argument from costume or brushwork - strongly suggests otherwise, and does not support the circa 1660-61 date the stylistic comparison to the Oath of the Batavians had implied. Hinterding is a co-author, with Jaco Rutgers, of the New Hollstein Rembrandt catalogue (2013), the modern reference this record cites throughout for state numbers, and watermark evidence of exactly this kind is the same physical method this archive's other Rembrandt print record, on the Hundred Guilder Print, describes being used to date that print's own lifetime impressions. This record was not able to obtain Hinterding's specific proposed date directly, only the fact that his watermark evidence displaces the stylistic one; it states that limitation here rather than picking a number to fill the gap, and records the state's true reworking date as an open question below.

Japanese paper, and why this impression is on it

The task this record was set requires an explicit answer, not an inference: is the Rijksmuseum impression at the centre of this record printed on Japanese paper? Yes. The museum's own catalogue entry for RP-P-1962-40 states the medium as drypoint, printed on Japanese paper, and the file's own Wikimedia Commons record confirms the same description. This is not the only paper Rembrandt used for this print; the Rijksmuseum's own first-state impression (RP-P-1961-1196A) is printed on vellum, not paper at all, described in the museum's own text as pulled on costly vellum, and its third-state impression (RP-P-OB-617) is on ordinary European paper. Three different supports, for three different states, in the same print room, from the same broad decade - a genuine demonstration of how deliberately Rembrandt varied the physical material his prints were pulled on.

Rembrandt is documented as one of the first Western artists to print etchings on Japanese paper, beginning around 1647 and continuing to his last etching in 1665, using it regularly for both plates and drawings. The Japanese sheets he favoured were gampi-fibre papers, light-brown to light-yellow or ivory in colour, with a smooth, shiny surface markedly different from the rough, matte European laid paper otherwise available to him; that surface did full justice to drypoint work specifically, since fine, absorbent paper of this kind picks up the subtle, velvety tonal range a burred drypoint line can produce far more faithfully than a coarser sheet does. How the paper reached Rembrandt's workshop is not established in the sources consulted; gampi papers arrived in the Netherlands via the Dutch East India Company from around the 1640s, but the exact channel by which Rembrandt himself acquired sheets, and how much he paid for them, is not documented in any source this record could verify. What is documented is the effect on this specific state and impression: the same driving, near-total darkness that defines state IV reads, on the warm, luminous surface of Japanese paper, with a tonal richness the coarser European sheet used for the Rijksmuseum's own third-state impression does not reproduce.

This impression's own collectors: the De Bruijn-van der Leeuw bequest

The Rijksmuseum's credit line for this impression names Mr and Mrs De Bruijn-van der Leeuw Bequest, Muri, Switzerland, with the object entering the museum's collection by bequest in 1962. The couple were the Amsterdam-connected collectors Isaac de Bruijn (1872-1953) and his wife, Johanna Geertruida de Bruijn-van der Leeuw (1877-1960), who lived in Spiez and later Muri, near Bern, and who assembled one of the more significant private holdings of Rembrandt's graphic work in the first half of the twentieth century. Rather than a single clean gift, the couple's arrangement with the museum was structured in stages: they donated works in 1949 while retaining usufruct - the right to keep and enjoy the objects during their own lifetimes - with the collection physically transferring to the Rijksmuseum only in 1960, and further individual objects, including the two Three Crosses impressions discussed in this record, arriving by bequest in 1961 and 1962 respectively, in the years immediately following Johanna's death.

The couple were active, informed buyers in the international print market of their day, not passive inheritors of a family collection. One documented purchase places them at a Rembrandt sale from the Kinnaird collection on 28 May 1936, where they paid 47,500 pounds sterling, and at the Henry Oppenheimer sale in London of 10-13 July 1936, where they acquired a further work for 700 guineas through the London dealership P. and D. Colnaghi and Co. Those purchase prices and that named dealer establish the couple as serious participants in the mid-1930s London old-master print trade, buying at the same level and through the same channels as the era's major institutional and private collectors, in the years immediately before the Second World War disrupted that market. This record was not able to establish, from the sources it could access, exactly when or from whom the couple acquired either of their two Three Crosses impressions specifically, as distinct from the general fact of their participation in that market; that gap is recorded as an open question below.

The Rijksmuseum's other impressions, and a mark this archive keeps meeting

Beyond the two De Bruijn-van der Leeuw sheets, the Rijksmuseum holds at least three further impressions of this print: RP-P-OB-616 and RP-P-OB-617, both cited to Bartsch 78 and New Hollstein 274 as states 1 and 3 of 5 respectively, and RP-P-OB-618, a further impression of state 4. All three carry the same collector's stamp, catalogued as Lugt 240, and none of the museum's own online catalogue entries consulted for this record identifies the collector behind that mark.

This is not the first time this mark has surfaced in this archive's own research. This record's sister record on Rembrandt's Hundred Guilder Print documents the identical Lugt 240 stamp on that print's own Rijksmuseum impression (RP-P-OB-602), and this archive's record on Durer's Rhinoceros documents a Lugt 240 stamp on that sheet's Rijksmuseum impression as well - three separate prints, by two different artists, in three different centuries of manufacture, all carrying the same numbered mark inside the same print room. The Hundred Guilder Print record's own research already raised, without confirming, the possibility that Lugt 240 functions as an internal Rijksmuseum print-room stamp rather than a private collector's mark, given how differently dated and differently subjected the objects that carry it are. This record cannot resolve that question any more than its predecessor could, but records the pattern's third occurrence here as a finding in its own right: whatever Lugt 240 turns out to be, it is not confined to one collection, one artist, or one moment in this print room's history, and a future editor with direct access to the Lugt reference catalogue should settle the question once, for all three records at once.

The pendant: Christ Presented to the People

The Rijksmuseum's own catalogue text for its vellum first-state impression of the Three Crosses describes the print as the pendant to Christ Presented to the People, on the opposite partition wall - the museum's own curatorial framing for how the two prints are displayed and understood as a pair, rather than a claim about a documented joint commission. Christ Presented to the People, Rembrandt's other large-scale drypoint of the Passion, showing Pilate presenting Christ to the crowd before the crucifixion (Ecce Homo, John 19:4-6), is close in size to the Three Crosses - both prints among the two largest Rembrandt ever made - and exists in its own sequence of eight documented states, all dated to around 1655, two years after the Three Crosses was first signed. The Rijksmuseum holds its own impression of that print as well, RP-P-OB-611, an impression of a later state (the museum's own record for this sheet notes an annotation reading 6de druk extra raar (sixth impression, extra rare), a sixth-state impression marked by an early owner as exceptionally rare).

Read together, the two prints bracket the Passion narrative - Christ shown to the crowd before his death, and Christ dying on the cross - in the same technique, at nearly the same scale, within a few years of each other, at the height of Rembrandt's ambition as a printmaker working almost exclusively in drypoint. Christ Presented to the People was itself extensively reworked across its eight states, including, in its most discussed alteration, the removal of a foreground crowd Rembrandt had originally etched at the bottom of the composition and replaced with an empty platform - a different kind of dramatic revision on a different plate, made by the same hand within the same few years, that this record notes as context without asserting a documented, contractual pairing between the two prints beyond the museum's own descriptive framing.

1653: what else was on Rembrandt's ledger the year he signed this plate

Rembrandt dated the plate's third state 1653 in the same year his own finances were, on the documentary record, under acute strain. His ownership of the large house he had bought on the Sint Antoniesbreestraat in 1639 took full legal effect only after a judicial delivery on 8 January 1653, a transfer that removed any further possibility of delaying the outstanding purchase payment. Forced into an immediate cash need, Rembrandt took out a series of loans in what one scholarly account of his finances describes as panic; among them, the Amsterdam patrician and later burgomaster Cornelis Witsen lent Rembrandt 4,180 guilders that same January, recorded in an aldermen's letter that stipulated a general hypothec, a lien over his general assets rather than a specific one. His financial position did not recover: the documentary record shows it deteriorating steadily through the first half of the 1650s, brought on, by the same account, considerably by Rembrandt's own financial mismanagement, and culminating in the formal declaration of insolvency in 1656, with debts eventually totalling roughly 13,000 guilders.

This record does not claim a documented causal link between the Witsen loan and the decision, whenever exactly it was taken, to rework this specific plate; no source consulted draws that connection explicitly, and this record will not manufacture one. What can be stated plainly is that the print exists, was signed, and entered active commercial and collecting circulation in the same calendar year Rembrandt was scrambling for cash to keep the roof over his own workshop - a coincidence of timing worth recording precisely because Rembrandt's own market for prints, discussed below, was already functioning as a source of income he had every incentive to protect and extend.

The market for impressions of this print

Three documented sales of impressions of this print, spanning more than a decade, are directly traceable in the sources consulted for this record, and they illustrate different tiers of the same market rather than a single price for a single object. At Christie's New York, on 23 October 2024, an impression of the Three Crosses was the star lot of Exceptional Impressions: The Alan and Marianne Schwartz Collection, an auction that totalled 9.89 million dollars; the print itself sold for a total of 1,380,000 dollars, 38 percent above its 1,000,000-dollar low estimate. The source consulted for this figure states only the combined total and does not separately report a hammer price and buyer's premium, and this record does not back-calculate one from the other.

A second sale, identified through search-engine-indexed reporting rather than a direct fetch of Christie's own sale records (the primary listing returned an access error on every attempt made during this research), places a third-state impression from the celebrated Rembrandt print collection of Sam Josefowitz (1921-2015) at Christie's London on 7 December 2023, as lot 17, with a reported total of 870,000 pounds sterling against a 300,000-pound low estimate; press coverage around the wider Josefowitz dispersal describes this specific lot as the most expensive of five highlighted impressions drawn from more than seventy Rembrandt prints in that collector's holdings. This record flags that figure as resting on secondary, search-indexed reporting rather than a source it could read directly, consistent with this archive's practice of stating that distinction plainly rather than presenting an unverified figure with false confidence.

A third data point, independently and directly confirmed, is an estimate rather than a result: Sotheby's London offered a fourth-state impression as lot 49 in its Prints and Multiples sale (L15160) of 17 March 2015, with an estimate of 200,000 to 300,000 pounds, described in the catalogue as a fine impression with rich burr and dramatic tonal contrast, in good condition with a repaired tear, foxing and minor creasing, and with a distinguished provenance: ex-collection Adalbert von Lanna, the great Prague old-master print collector, whose mark is catalogued as Lugt 2773. This record could not establish the lot's price realized from the sources it could access and states that gap here rather than guessing at it. Taken together, these three data points place documented impressions of this print, across three different states and three different auction houses, consistently in the low hundreds of thousands to low millions of dollars, a market that rewards early or dramatically reworked, well-provenanced sheets rather than a single settled price for the print as an undifferentiated object.

Exhibition history, and this archive's other Rembrandt print

The Three Crosses, in its third state, was included in Rembrandt: The Late Works, the joint National Gallery, London and Rijksmuseum, Amsterdam exhibition that ran in London from 15 October 2014 to 18 January 2015, drawing roughly 40 paintings, 20 drawings and 30 prints organized thematically around self-scrutiny, experimental technique and the observation of everyday life in Rembrandt's final years - an exhibition the National Gallery's own reporting states was seen by more than a quarter of a million visitors, one of the most attended in the gallery's history. Its inclusion in a show explicitly built around Rembrandt's experimental technique is consistent with, though this record does not treat it as proof of, the dating that would place the plate's dramatic reworking in that same late period of his career; as set out above, the sources consulted disagree on whether that late dating is actually correct.

This archive already holds a record on Rembrandt's other great mid-century drypoint, the Hundred Guilder Print (Christ Preaching, Healing the Sick, and Blessing the Children), and the two objects are worth distinguishing clearly rather than folding together. The Hundred Guilder Print's own dramatic later alteration was posthumous and physical: the London collector William Baillie bought the worn plate in 1775, reworked it by his own hand, printed roughly a hundred impressions, and then cut the single copper into four separate pieces, each further reworked and printed independently - an alteration made by a different, named eighteenth-century hand more than a century after Rembrandt's death. The Three Crosses tells a different story: every state discussed in this record, from the signed, finished third state to the darkened fourth and fifth, was made by Rembrandt himself, within his own working lifetime, whatever the correct date for that second campaign of work turns out to be. Both prints show a Rembrandt plate transformed almost beyond recognition; only one of the two transformations belongs to Rembrandt.