Issues of Attribution in the Rembrandt Workshop

Arthur K. Wheelock Jr.

As an introduction to the entries on paintings by Rembrandt and his school it seems appropriate to discuss briefly the history of Rembrandt connoisseurship as it pertains to this collection. This overview is intended to provide a framework for understanding the approach to attributions that is taken in this collection.

That Rembrandt was a genius has never been questioned, yet one of the ironies of Rembrandt scholarship is that each generation has developed its own perception of his genius.[1] This phenomenon is most clearly seen in the different character of the artist’s oeuvre described in the literature since the early nineteenth century. In 1836 the English art dealer John Smith, who compiled the first catalog of Rembrandt’s paintings, accepted 614 works as by the master, although many of these he knew only from written descriptions or from reproductive prints.[2] Still, Smith was convinced that the qualities of Rembrandt’s genius are evident to an “experienced eye.” According to Smith, these qualities span beyond manual dexterity to “beauties which emanate from a higher source, such as expression, delicacy of gradation, and harmony of tints, [and] are beyond the reach of all who are inferior to the master himself.”[3] Smith, however, knew neither the very early nor the very late paintings by Rembrandt; the latest dated work he knew was Rembrandt’s 1664 depiction of Lucretia (fig. 1). As was often the case, Smith saw in Rembrandt’s choice of subject reflections of the artist’s own life. Lucretia’s tragedy, he felt, was painted in reaction to the trouble Rembrandt had experienced during the painful close of his life.[4]

In reality, the nature of Rembrandt’s oeuvre and its relationship to events in the artist’s life, has proved to be far more complicated than Smith anticipated. By the beginning of the twentieth century the great Rembrandt scholar Wilhelm von Bode had rejected a large number of works included in Smith’s catalog and added an almost equal number of new discoveries in his eight-volume corpus of Rembrandt’s works.[5] Many of the fifty to sixty newly discovered works included in the last volume of Bode’s corpus were oil sketches, unfinished preparatory works, or boldly executed paintings from the end of Rembrandt’s career. One of Bode’s major discoveries, for example, was the Man with a Gilded Helmet (fig. 2) that he bought for the Berlin Museum.[6] The broad brushwork of paintings like this one, epitomized for Bode Rembrandt’s stylistic independence and creative genius. Bode, whose outlook was deeply influenced by the romantic movement and its ideals of creative genius, greatly admired Rembrandt’s chiaroscuro effects, which suppressed surface details to such an extent that the artist could render “souls rather than existences.”[7] For Bode, Rembrandt’s art marked “a climax in the development of universal art.”

As great an artist as Rembrandt was, Bode maintained that the “thinker and poet in him were still greater than the painter; they even worked occasionally to the detriment of the artist, seducing him into a fantastic handling of simple motives that demanded a purely realistic treatment.”[8] Bode noted that not all of Rembrandt’s works were of equal quality, attributing any failings to “merely the defects incidental to his great and original genius.”[9] Bode believed that Rembrandt had to be “studied as a whole, only thus is he comprehensible and unsurpassable.” According to him, it was necessary to study the totality of the master’s oeuvre, which includes drawings, prints, and paintings, as well as the coarse, the refined, the elegant, and the harsh elements of his art.

Bode’s inclination to incorporate a wide range of styles into his Rembrandt corpus was expanded upon by his protégé Wilhelm Valentiner. Through his numerous publications on Rembrandt in the first decades of the twentieth century, Valentiner also introduced a large number of paintings into Rembrandt’s oeuvre.[10] At the turn of the twentieth century, with the inclusive approach of Bode, Valentiner, and Cornelius Hofstede de Groot, who had assisted Bode in producing his Rembrandt catalog, and who later compiled his own catalog of Rembrandt paintings,[11] the corpus of Rembrandt’s work grew to more than seven hundred paintings. Yet many of Bode’s discoveries and most of Valentiner’s have not withstood the test of time.[12] Later twentieth-century scholars, in particular Abraham Bredius, Jakob Rosenberg, and Kurt Bauch, rejected a number of attributions made by Bode and Valentiner.[13] Until Horst Gerson’s 1969 revised edition of Bredius’s 1935 catalog, art historians had not recognized the degree to which Rembrandt’s oeuvre had become bloated with wrongly attributed paintings.[14] Gerson’s publication, which accepted only 435 paintings as authentic, shocked the art world because he rejected a number of paintings that had long been considered among Rembrandt’s major works, including The Descent from the Cross (fig. 3) and the 1650 Self-Portrait (now called Portrait of Rembrandt) in the National Gallery of Art (fig. 4). Gerson attributed a large number of his rejected paintings to Rembrandt’s pupils but also maintained that many had been painted by later imitators.

In recent years the issues of Rembrandt attributions have been kept very much alive by the publications of the Rembrandt Research Project (RRP), hereafter referred to as theCorpus. This project was organized in 1968 with the intent of studying Rembrandt paintings within a scientific framework.[15] Working first under the leadership of Josua Bruyn and then under Ernst van de Wetering, the RRP has applied even more stringent standards than those of Gerson. Rembrandt’s oeuvre, as a result, has once more been radically trimmed: the RRP now accepts around 250 paintings as authentic works by the master.

The RRP has provided a new foundation for subsequent Rembrandt research. Aside from its careful descriptions of the works themselves, which also include the publication of much new technical information, particularly about materials and supports, the project has analyzed anew contemporary documents, including inventories, and examined issues of Rembrandt’s workshop. However, the structure of the RRP, the personalities involved, and the art historical approach evident in the volumes of the Corpus have changed dramatically over the years, demonstrating, yet again, the continually evolving character of research on Rembrandt’s life and work. To begin with, the first three volumes of the Corpus, which were written under the guidance of Josua Bruyn, were organized chronologically; the last two volumes, under the guidance of Ernst van de Wetering, are organized thematically, with volume 4 devoted to Rembrandt’s self-portraits and the projected volume 5 devoted to small figure history paintings and landscapes.[16]  

The most significant changes between the first three volumes and the latter volumes of the Corpus, however, are in approach. In the later volumes, the authors are more willing to acknowledge uncertainties in our understanding of the past than was evident at the beginning of the project. In my assessment, the first three volumes of the Corpus the RRP made too many precise judgments of attribution given our partial understanding of various aspects of Rembrandt’s life and work, whether they be his relations to his patrons or the running of his workshop.[17] A number of critics rightly argued, moreover, that the RRP’s view of Rembrandt’s range of style and technique had been too narrow and had been premised on the idea that his work had progressed in a totally logical manner.[18] As many have noted, the evidence gained from Rembrandt’s etchings and drawings indicates that he often radically altered both style and technique to create different effects, sometimes within the same image.

The RRP, in general, has been extremely critical of Rembrandt scholarship that grew out of the romantic era. It has argued that the range of styles of painting that Bode, among others, embraced is too broad. To substantiate its belief, the RRP has given great weight to evidence gained through scientific investigations, particularly in matters of painting technique, including the types and sizes of the supports, composition of the grounds, buildup of paint layers, and individual characteristics of brushwork. While such technical examinations are important, I differ from the approach taken by the RRP in that I believe that nineteenth- and early twentieth-century interpretations of Rembrandt’s paintings can provide important insights into understanding the past. I tend to believe that Rembrandt’s personal circumstances may well have affected the mood and subject matter of his works. While his career has certainly been over-romanticized, there are usually germs of truth in the myths surrounding his life and art.

One assumption made in the first three volumes of the Corpus that is consistent with the tradition of Smith and Bode is that Rembrandt executed paintings without the assistance of members of his workshop.[19] While the discussions of some paintings in the Corpus raise the possibility of workshop collaboration in the master’s works, such instances have been treated as the exception rather than the rule.[20] Unfortunately, the designations A, B, and C that the RRP has used in its first three volumes to differentiate between “accepted,” “do not know,” and “rejected” paintings do not provide a framework for works produced collaboratively in the workshop.[21] The precise distinctions the RRP made in these volumes between autograph and non-autograph works, however, does not accord with workshop practice in the seventeenth century, and this approach has been discontinued in volume 4.[22]  Despite the opinions of Bruyn, it seems most likely that Rembrandt, like Rubens in Antwerp and Van Dyck in England, used studio assistants to help him produce paintings for the market, especially during the 1630s when his work was in great demand.[23]

Despite the research of Van de Wetering, Bruyn, and others, many questions still remain about the practices of the Rembrandt workshop.[24] In 2005, Van de Wetering raised a fundamental question about the nature of workshop production that was not seriously considered by earlier volumes of the Corpus: “does the work betray a genesis which would suggest that the maker was also the person who developed the conception of the work? If the answer . . . is yes, can it be the work of Rembrandt himself or a pupil or an assistant, or was it executed by several people?”[25] Documentary evidence indicates that during the mid-1630s in particular, Rembrandt worked together with pupils and assistants on the same compositions. In other instances he may have encouraged students or assistants to devise compositions based on a conceptual framework he had developed in his own paintings, drawings, or quick oil sketches.

The evidence for such an integrated approach to workshop production comes from many sources: inscriptions on paintings and etchings, reworkings found on drawings, collaborative works listed in his inventory, and, finally, the visual and technical evidence of the works themselves. Indications of workshop collaboration include the series of four etched oriental heads from 1635 that are based on prototypes by Jan Lievens and are signed “Rembrandt geretuckert” (Rembrandt retouched)[26]; the large Abraham’s Sacrifice in Munich, a version of Rembrandt’s painting in the Hermitage that is inscribed“Rembrandt, verandert. En over geschildert. 1636” (Rembrandt, changed. And painted over. 1636); and a number of paintings in the 1656 inventory of Rembrandt’s possessions that are described as having been “geretukeert” (retouched) by Rembrandt, and one work that is said to have been “gemodelt” (designed by) Rembrandt.[27]

Much debate exists over the meaning of these inscriptions and terms, but it seems quite clear that they all indicate a form of collaborative undertaking. In some circumstances Rembrandt's intent may have been to correct compositional or design elements in paintings, much as he sometimes did with students’ drawings. It also seems probable, however, that he developed various ways to utilize his workshop to increase his own production, particularly during the 1630s, when he received a large number of commissions.

Although the size of Rembrandt’s workshop at various stages of his career is uncertain, a number of artists, including Govert Flinck (1615-1660), came to Rembrandt only after they had received preliminary training with another master. They must have come to learn his style, but presumably they were proficient enough to assist Rembrandt with his own productions. The collaboration in Rembrandt’s workshop thus could have taken many forms: Rembrandt could have provided a model in the form of a drawing or oil sketch for an assistant to use as a basis for a painting executed in his style; he could have blocked in the composition on the canvas or panel before it was completed by an assistant;[28] he could have had assistants paint costumes and even hands on commissioned portraits;[29] and finally, he almost certainly retouched and improved upon works produced by students and assistants. One can also imagine that two or more students could have worked on the same painting. Moreover, a frequently overlooked complicating factor in determining attribution questions is that many paintings that were probably worked on over a period of time. Whether or not the nature of workshop productions appreciably altered from one decade to the next is difficult to determine, particularly since little documentation about the workshop exists after Rembrandt’s financial difficulties of the 1650s.[30]

The hypothesis that Rembrandt worked closely with members of his workshop at various stages of his career accounts for the many Rembrandtesque paintings that are difficult to attribute to a specific member of his workshop. It is also consistent with his documented practice of signing works made by members of his workshop. It helps explain why paintings not executed primarily by Rembrandt were inscribed with the master’s name, and, finally, why so many works listed in Rembrandt’s inventory as being by his hand appear to have been workshop productions.[31]

This synopsis of the history of Rembrandt connoisseurship in the last century and a half has specific implications for the National Gallery of Art collection. The Rembrandt paintings here, with one exception, came from two major collections: that of the Widener family and that of Andrew Mellon.[32] Both collections were formed at a time when Bode’s influence on Rembrandt connoisseurship was at its height. The Wideners collected their Rembrandt paintings from 1894 until the 1920s. Peter A. B. Widener probably met Bode when the German art historian came to Philadelphia in 1893, during his tour of American collections, after viewing the Chicago World’s Fair (see The History of the Dutch Paintings Collection).[33] Widener’s later advisor, and the advisor of his son Joseph Widener, was none other than Wilhelm Valentiner. Mellon collected slightly after Widener, and he acquired four Rembrandt paintings from the Hermitage during the 1930s. Nevertheless, his generation had also learned to love and appreciate Rembrandt through the eyes of Bode. Thus, as a collection, the Gallery’s paintings tend to be the types of paintings Bode and Valentiner most admired, works from the end of Rembrandt’s life, when his brushwork is bold and evocative. Following romantic inclinations, these collectors were drawn to paintings that were intimately associated with the artist’s life.

Widener’s first painting by Rembrandt was a portrait of Saskia van Uylenburgh, the Wife of the Artist (fig. 5) and his most famous work was thought to represent the mill of Rembrandt’s father (see The Mill, fig. 6); each collector owned a self-portrait (see fig. 4 and fig. 7, Self-Portrait ). Virtually all the Rembrandt paintings in these collections were covered with discolored and even tinted varnishes to give them that “golden glow” so admired in the nineteenth and early twentieth centuries.

A large number of paintings acquired by Widener and Mellon were extremely famous at the time of their purchase, and their acquisitions were highly acclaimed by both the leading experts and the popular press. Nevertheless, as scholars have reduced the size of Rembrandt’s oeuvre in the second half of the twentieth century, many of these paintings have been viewed critically; some have been attributed to pupils of Rembrandt, some have been called later imitations. Other paintings have entered into an undefined limbo, where vague doubts about the attribution are expressed verbally or in the literature, but where no specific arguments have been advanced to explain why they should not be accepted as authentic.

In trying to determine the attribution of 24 paintings to Rembrandt and his workshop for this edition, a wide range of material has been analyzed, from the provenances of the works, to old reproductive prints and descriptions in sales catalogs, to the opinions of other scholars. Iconographic issues have been explored by looking at both texts available to Rembrandt and probable visual sources. A concerted effort was made, moreover, to examine all of the paintings carefully in the laboratory. Paintings were looked at under the microscope, signatures were studied, X-radiographs were taken, infrared and ultraviolet photographs were made, and panel and canvas supports were analyzed. Most of these paintings have been restored so that discolored varnish and old overpaint could be removed. Other information has been acquired from further investigations undertaken in collaboration with members of our scientific department.[34]             

The results of the restorations and technical examinations, which are discussed in these objects’ entries, have often been dramatic, providing reassessments of the quality of a number of individual works, and, in some instances, leading to proposals for new attributions, as in the The Descent from the Cross. Nevertheless, as in the example ofJoseph Accused by Potiphar's Wife,  not all issues of attribution have been fully resolved, generally because a painting’s style and technique could not be identified with that of a specific artist from Rembrandt’s workshop. In an effort to reflect the complex nature of the Rembrandt School, attributions of these paintings have been designated in the following ways: Rembrandt van Rijn, Rembrandt Workshop, Follower of Rembrandt van Rijn, and Style of Rembrandt van Rijn. The distinctions are in basic accordance with National Gallery fo Art practice, as is outlined in the Notes to the User. Within the Rembrandt van Rijn section, however, are also works designated “Rembrandt van Rijn and Workshop,” a distinction here used when stylistic evidence indicates that Rembrandt collaborated in the execution of a painting. Collaboration between Rembrandt and members of his workshop may also have occurred in one form or another in paintings designated here simply as “Rembrandt Workshop,” but in these paintings, evidence of  Rembrandt’s own brushwork has not been found in the final image. When a specific artist can be associated with a painting executed in the workshop that was produced to be sold under Rembrandt’s name, the name of that artist is indicated in parentheses following the designations “Rembrandt van Rijn and Workshop” or “Rembrandt Workshop.”

Notes

[1] For overviews of earlier assessments of Rembrandt, see Seymour Slive, Rembrandt and His Critics, 1630–1730 (The Hague, 1953); R. W. Scheller, “Rembrandt’s reputatie van Houbraken tot Scheltema,” Nederlands Kunsthistorisch Jaarboek 12 (1961): 81–118; and Jan Ameling Emmens, Rembrandt en de Regels van de Kunst (Utrecht, 1968). For a general account of the history of Rembrandt criticism see Jeroen Broomgaard and Robert W. Scheller, “A Delicate Balance: A Brief Survey of Rembrandt Criticism,” inRembrandt: The Master and His Workshop, ed. Christopher Brown, Jan Kelch, and P. J. J. van Thiel (New Haven and London, 1991), 106–123.

[2] John Smith, A Catalogue Raisonné of the Most Eminent Dutch, Flemish, and French Painters, vol. 7 (London, 1836).

[3] John Smith, A Catalogue Raisonné of the Most Eminent Dutch, Flemish, and French Painters, 9 vols. (London, 1829–1842), 7:244.

[4] John Smith, A Catalogue Raisonné of the Most Eminent Dutch, Flemish, and French Painters, 9 vols. (London, 1829–1842), 7:xxiv.

[5] Wilhelm von Bode, with the assistance of Cornelis Hofstede de Groot, The Complete Works of Rembrandt, trans. Florence Simonds, 8 vols. (1897–1906).

[6] Despite the enormous fame of this work, the Man with a Gilded Helmet is no longer attributed to Rembrandt. See Jan Kelch et al., Der Mann mit dem Goldhelm (Berlin, 1986).

[7] Wilhelm von Bode, with the assistance of Cornelis Hofstede de Groot, The Complete Work of Rembrandt, trans. Florence Simonds, 8 vols (Paris, 1897–1906), 8:19.

[8] Wilhelm von Bode, with the assistance of Cornelis Hofstede de Groot, The Complete Work of Rembrandt, trans. Florence Simonds, 8 vols (Paris, 1897–1906), 8:22.

[9] Wilhelm von Bode, with the assistance of Cornelis Hofstede de Groot, The Complete Work of Rembrandt, trans. Florence Simonds, 8 vols (Paris, 1897–1906), 8:30.

[10] Valentiner’s publications on Rembrandt are numerous; for the most important see Wilhelm R. Valentiner, Rembrandt, des Meisters Gemälde, Klassiker der Kuinst in Gesamtausgaben 2, 3rd ed. (Stuttgart, 1908) and Wilhelm R. Valentiner, Rembrandt: wiedergefundene Gemälde (1910–1920), Klassiker der Kuinst in Gesamtausgaben 27, 2nd ed. (Berlin and Leipzig, 1923). For a full listing of his publications, see Masterpieces of Art: In Memory of Wilhelm R. Valentiner, 1880–1958 (Raleigh, NC, 1959), 297–319.

[11] Cornelis Hofstede de Groot, A Catalogue Raisonné of the Works of the Most Eminent Dutch Painters of the Seventeenth Century, trans. Edward G. Hawke, vol. 6 (London, 1916).

[12] See Catherine B. Scallen, Rembrandt, Reputation, and the Practice of Connoisseurship (Amsterdam 2004), 169–179.

[13] Abraham Bredius, Rembrandt, Schilderijen (Vienna, 1935); Jakob Rosenberg,Rembrandt, 2 vols (Cambridge, MA, 1948; rev. ed. 1964); Kurt Bauch, Rembrandt Gemälde (Berlin, 1966).

[14] Abraham Bredius, Rembrandt: The Complete Edition of the Paintings, rev. ed. by Horst Gerson (London, 1969). For an excellent assessment of late nineteenth- and early twentieth-century connoisseurship of Rembrandt’s work, see Catherine B. Scallen, Rembrandt, Reputation, and the Practice of Connoisseurship (Amsterdam, 2004).

[15] Initial funding for this project was provided by the Netherlands Organization for the Advancement of Pure Research (ZWO), later called the Netherlands Organization for Scientific Research (NWO). It also received support from the University of Amsterdam.

[16] See Ernst van de Wetering, ed., A Corpus of Rembrandt Paintings, vol. 4, The Self Portraits (Dordrecht, 2005), ix–xxii.

[17] A partial listing of the various reviews of the publications of the RRP is as follows: Editorial, Burlington Magazine 125 (1983): 661–663; J. R. J. van Asperen de Boer, "Review of vol. 1," Akt 9 (1985): 14–19; Peter Schatborn, "Review of vol. 1.," Oud-Holland 100 (1986): 55–63; Egbert Haverkamp-Begemann, “The State of Research in Northern Baroque Art,” Art Bulletin 69 (1987): 510–519, particularly 514–516; Walter Liedtke, “Reconstructing Rembrandt: Portraits from the Early Years in Amsterdam (1631–1634),” Apollo (May 1989): 323–331, 371–372; Sylvia Hochfield, “Rembrandt: The Unvarnished Truth?,” Art News (December 1987), 102–111; Mansfield Kirby Talley Jr., “Connoisseurship and the Methodology of the Rembrandt Research Project,”International Journal of Museum Management and Curatorship 8 (1989): 175–214; Leonard J. Slatkes, "Review of vol. 2," Art Bulletin 71 (1989): 139–144. For an excellent overview of the controversies surrounding the RRP, see Anthony Bailey, “The Art World ('The Polish Rider'),” New Yorker (5 March 1990), 45–77.

[18] To demonstrate how subjective interpretations of painting technique can be, it is necessary only to compare the results of the research of the RRP in the first three volumes of the Corpus with those of Claus Grimm in his Rembrandt Selbst: Eine Neubewertung Seiner Portraïtkunst (Stuttgart, 1991). Grimm, who also carefully analyzes artists’ painting techniques as a basis for attribution, has arrived at a very different list of paintings he believes were executed entirely by the master than that found in the Corpus. He is, in fact, more restrictive in the paintings he attributes to Rembrandt than the RRP.

[19] For an excellent assessment of Rembrandt’s workshop practice, examined from the vantage point of the RRP, see Josua Bruyn, "Rembrandt’s Workshop: Its Function and Production," in Rembrandt: The Master and His Workshop, ed. by Christopher Brown, Jan Kelch, and P. J. J. van Thiel (New Haven and London, 1991), 68–89, and in particular pages 83–85: “On the whole, one may say that with Rembrandt design and execution were closely bound up. Instead of making use of sophisticated workshop procedures which could in part replace the share of the master’s hand, he seems to have allowed invention and execution to be separated only in the early stages of the assistants’ activities. Later, they would be welcome to their own design and only rarely did they intervene with his own work.”

[20] See, for example, Josua Bruyn, "Rembrandt’s Workshop: Its Function and Production," in Rembrandt: The Master and His Workshop, ed. Christopher Brown, Jan Kelch, and P. J. J. van Thiel (New Haven and London, 1991), 85.

[21] In the Burlington Magazine 135 (1993), 279, J. Bruyn, B. Haak, S. H. Levie, and P. J. J. van Thiel wrote a letter saying that they were no longer involved with the RRP. They indicated that the project would in the future be headed by Ernst van de Wetering, who intended to eliminate the A, B, C categorization in future volumes of theCorpus.

[22] These changes were, in fact, made. Please see Ernst van de Wetering, ed., A Corpus of Rembrandt Paintings, vol. 4, The Self Portraits (Dordrecht, 2005), xviii.

[23] For discussions of the workshops of Rubens and Anthony van Dyck, see Arthur K. Wheelock Jr. in Cynthia P. Schneider, Rembrandt’s Landscapes: Drawings and Prints, exh. cat. (Boston, 1990), 11–16, and Barnes in Cynthia P. Schneider,Rembrandt’s Landscapes: Drawings and Prints, exh. cat. (Boston, 1990), 17–25. Josua Bruyn ("Rembrandt’s Workshop: Its Function and Production," in Rembrandt: The Master and His Workshop, ed. Christopher Brown, Jan Kelch, and P. J. J. van Thiel [New Haven and London, 1991], 83), who does not believe that Rembrandt had assistants help him in the manner of Rubens, expressly contrasts the types of commissions Rembrandt and Rubens received and notes that no evidence of workshop participation has been found in the few large-scale commissions Rembrandt did receive, among them The Blinding of Sampson (Städelsches Kunstinstitut) and The Night Watch (Rijksmuseum, Amsterdam). Bruyn, however, is not consistent in his belief; he argues that the horse in Rembrandt’s Portrait of Frederick Rihel on Horseback, 1663(?) (National Gallery, London), was executed by an assistant.

[24] See, in particular, Ernst van de Wetering, “Problems of Apprenticeship and Studio Collaboration,” in A Corpus of Rembrandt Paintings, ed. Josua Bruyn et al. (The Hague, Boston, and London, 1982), 2:45–90; Josua Bruyn, "Rembrandt's Workshop: Its Function and Production," in Rembrandt: The Master and His Workshop, ed. Christopher Brown, Jan Kelch, and P. J. J. van Thiel (New Haven and London, 1991), 68–89; and R. de Jager, "Meester, leerjongen, leertijd. Een analyse van zeventiende-eeuwse Nord-Nederlandse leerlingcontracten van kunstschilders, goud-en zilversmeden," Oud-Holland104 (1990): 69–111.

[25] Ernst van de Wetering, ed., A Corpus of Rembrandt Paintings, vol. 4, The Self Portraits (Dordrecht, 2005), xvi.

[26] For a discussion of these works, see Peter Schatborn, Jan Lievens, 1607–1674: Prenten & tekeningen, exh. cat. (Amsterdam, 1988).

[27] Walter L. Strauss and Marjon van der Meulen, The Rembrandt Documents (New York, 1979), doc. 1656/12. Inventory numbers 25, 27, 28, 33, 120, and 123 record paintings retouched by Rembrandt, while inventory number 79 mentions a painting “designed by” Rembrandt. None of these works have been identified.

[28] One large painting that I am certain was executed with the help of assistants is Belshazzar’s Feast, c. 1635 (National Gallery, London). Although the attribution of this painting solely to Rembrandt’s hand has never been questioned in print, the peripheral figures in this composition are executed in a range of styles that are inconsistent with Rembrandt’s own manner of painting.

[29] It has been recognized that an assistant painted the hands in one of Rembrandt’s most prestigious commissions, his Portrait of Johannes Uyttettbogaert, 1633 (Rijksmuseum, Amsterdam). See Wouter Th. Koek and Guido M. C. Jansen, Rembrandt in a New Light: Presentation of Seven Restored Paintings by Rembrandt (Amsterdam, 1993).

[30] The large number of unfinished (“onopgemaeckt”) paintings listed in the inventory of Rembrandt’s estate taken on October 5, 1669 indicates that the artist did not immediately complete paintings that he had commenced. Perhaps he intended some of these to be worked up by assistants. For the inventory, see Walter L. Strauss and Marjon van der Meulen, The Rembrandt Documents (New York, 1979), 586–589, doc. 1169/5, nos. 8 and 11.

[31] Josua Bruyn, "Rembrandt’s Workshop: Its Function and Production," inRembrandt: The Master and His Workshop, ed. Christopher Brown, Jan Kelch, and P. J. J. van Thiel (New Haven and London, 1991), 70–71.

[32] The one exception is Old Woman Plucking a Fowl, which was given by Dr. and Mrs. Walter Timme in 1956.

[33] Wilhelm von Bode, “Alte Kunstwerke in den Sammlungen der Vereinigten Staaten,” Zeitschrift für Bildende Kunst 6 (1895): 13–19, 70–76, does not explicitly mention visiting Widener, but he does write that he went to Philadelphia. Since he was actively visiting museums and collectors, however, it hardly seems possible that he and Widener did not meet. Widener had been collecting Dutch paintings prior to 1893, but only in 1894 did he begin to acquire major examples (Rembrandt’s Saskia, for example, from Sedelmeyer in Paris, the dealer who would publish Bode’s corpus on Rembrandt). In that same year he also bought a number of other important Dutch pictures, including Meindert Hobbema’s The Travelers, Pieter de Hooch’s The Bedroom, and Aelbert Cuyp’s Lady and Gentleman on Horseback. At about this time he also began to sell minor works from his earlier collection.

[34] We have also benefited enormously from discussions with colleagues from other institutions, in particular Ashok Roy, David Bomford, Ernst van de Wetering, and Karen Groen.