Decision Tree Branch for (2 Foolscap with Five...

Decision Tree Branch for Foolscap with Five-Pointed Collar

Fall 2015

Are roundels in a pyramid form (2 below, 1 on top) or pointing -

down (2 on top, 1 below)Second collar point &

ball fully extend beyondfront chain line — ► [la b , lac}-

on top (pyramid)

Front chain line *) bisects second collar

ball from front

Are the peaks divided (stripes in cap)?

Ea, Eb, F, Q, H, J . K, L M. N. Q, S

in------------- ---------------

Y ES n/a

A, B. C . D. 0 . U

FIG. 1 Decision branch diagram for Foolscap with Five-Pointed Collar watermark, developed by Louisa Smieska and Alison McCann, Fall 2015

Decision Trees and Fruitful CollaborationsThe Watermark Identification in Rembrandt’s Etchings (WIRE) Project at Cornell

Andrew C. Weislogel with

C. Richard Johnson, Jr., and

contributions from students1

Any additions to the file of

watermarks that has been

compiled will not only

improve our understanding of

Rembrandts progress on his

plates, but will enable us to

test and refine the conclusions

reached so far.2—Erik Hinterding

It is no exaggeration to say that Rembrandt’s etchings count among the most compelling examples of artistic production in the history of Western culture. The many catalogues of the artist's prints compiled since 1751- describing multiple states, ordering works by subject matter or by date, distinguishing his work from that of his pupils, and identifying copies and posthumous impressions —attest to the enduring fascination with Rembrandt’s technical experimentation and his versatility in interpreting the human psyche and capturing the natural world. Not surprisingly, the complexity and artistic creativity of these prints continue to inspire scholarly research, technical investigation, and multidisciplinary teaching approaches.As such, in addition to their acquisition by larger municipal museums, there is a history of collecting Rembrandt etchings among American college and university museums in the last century, a phenomenon explored further in Andaleeb Badiee Banta’s essay (pages 11-21).

Beginning with an initial collecting of prints during the 1940s,3 students’ encounters with Rembrandt as printmaker at Cornell University have been pursued through connoisseurship, exploration of subject matter and seventeenth-century Dutch culture, and an introduction to the artist’s process and materials. These efforts have been supported by a number of Rembrandt-related exhibitions at the Johnson Museum over the years,4 and were particularly catalyzed in the 1990s and 2000s thanks to the former Richard J. Schwartz Director Frank Robinson’s specialization in seventeenth-century Dutch art. Stephanie Wiles’s strong background in works on paper has supported further initiatives since she assumed the directorship in 2011; the present exhibition is the largest of these devoted to Rembrandt at Cornell.

W E I S L O G E L / J O H N S O N 33

Since 2015, a cross-disciplinary research project, Watermark Identification in Rembrandt’s Etchings (WIRE), has focused students’ attention and research efforts on the subject of Rembrandt’s watermarks. The project is meant to teach students about the importance of watermarks as a means to understand the artist’s printing practice, and to enable their contribution to an interactive online program designed to simplify the watermark identification process and further clarify aspects of Rembrandt’s printing process and chronology. From its inception, the project has proceeded as a series of semester-long courses, blending aspects of the art history seminar and the engineering research group and drawing students of various disciplines to participate, some for multiple semesters. This essay is intended to provide background for the current state of watermark study in Rembrandt’s prints, to elucidate the WIRE project’s dialogue between the disciplines of art historical inquiry and computational study, to chart its progress thus far, and to explore its implications for cross-disciplinary collaborations on the college and university level. It will make the case for a steady digital expansion of the taxonomy of watermark types compiled by Nancy Ash, Shelley Fletcher, Erik Hinterding,5 and others over the past three decades, and hint at useful possibilities its procedures have generated.

With each semester, new WIRE project students have been steeped in the lessons of previous watermark research. As such, it seems appropriate first to outline here the issues and history surrounding Rembrandt’s paper supports and watermarks, tapping the store of knowledge that has given rise to the present investigations. The following therefore might be considered a short primer for understanding Rembrandt’s papers and the usefulness of watermarks as a tool for reconstructing his printing practices.6

Interest in Rembrandt’s printing supports has a long history, dating almost as far back as the production of the prints themselves. At least as early as 1668, a year before the artist s death, collectors were admiring the artist’s use of Asian papers for the printing of certain impressions. Englishman Edward Browne, while traveling in Amsterdam, noted finding Rembrandt impressions on “Indian” paper.7 This expensive paper actually came from Japan, but it is sometimes confusingly referred to as “Indian” paper because the Dutch East India Company was the trade vehicle.8 Rembrandt seems to have first made prints on Japanese paper in 1647.? After this, he frequently used Japanese paper for early editions— print runs of perhaps twenty-five to fifty impressions—especially for prints with dark tonalities or heavy use of drypoint, because he understood the ability of these less-absorbent papers to show heavily-inked areas to more harmonious effect than European papers (cat. nos. 10,12, 29, 43, 48, 51, 52, 64, and

65).10 Rembrandt also printed on other supports, including coarse grayish cardoes papier (cartridge paper) meant for packaging gunpowder, and, also beginning around 1647, on vellum, or calfskin (cat. no. 44).11 Rembrandt’s use of different supports has been studied extensively and, as here, the inclusion of a range of these has become frequent in exhibitions of his prints.12

Of all of these support types, however, only laid papers of European manufacture can situate individual impressions in time, or group them with other prints from the same edition, through the characteristics imparted by the papermaking process. In the paper mills of Rembrandt’s Europe, this depended on a steady rotation of pairs of handheld paper molds or sieves (fig. 2), consisting of wooden frames with rows of very closely spaced horizontal wires called laid wires, fastened to the mold with rows of more widely spaced vertical wires called chain wires (fig. 3).13 A single overlay frame called a deckle fit both molds precisely, to confine the formed sheet to an even rectangle as each mold was dipped in turn into the vat of pulp.

Watermarks, designs made from bent wire,14 were attached to the grid of each paper mold to indicate the maker or geographical origin, or sometimes its intended market.15 Since two molds were used in the production process, these watermarks were made as identical as possible to each other so that the resulting batch of paper was uniformly marked. But because molds and watermarks were handmade,16 slight variations occurred between them, which is why the nearly identical watermarks from a pair of molds are referred to as “twinmarks”17. In addition, due to the increasing need for successful papermakers to set their papers apart from imitators, French and then German papermakers began to attach a second watermark, called a countermark (usually letters or words), to the mold during the sixteenth and seventeenth centuries.18 Typically, with European papers, the main watermark was found on the right side of the mold, while the countermark was found on the left. The tendency of the paper pulp to settle more thinly over the wires and watermark of the mold is what transfers these characteristics to the finished sheet of paper.

Especially in the case of larger sheets, watermarks and countermarks occupied relatively little of the area of the sheet, which means that an impression from a small copperplate will often be printed on a portion of the paper that does not include the watermark. The result is that only about one third of Rembrandt’s etchings printed on European papers display a watermark or part thereof.19 Even so, the presence of watermarks in this approximate third of Rembrandt's prints on European paper has enabled a significant increase in knowledge about Rembrandt’s production practices. In addition, certain geopolitical and

34 W E I S L O G E L / J O H N S O N

4

FIG. 2 Modern laid paper mold and deckle made by Timothy Moore. FIG. 3 Detail of Moore’s mold, showing horizontal laid wires, vertical chain wires, and wire watermark design.

mercantile circumstances can help explain the appearance or disappearance of certain papers in Rembrandt’s work.

Only a fraction of the paper used in the Netherlands during the first two thirds of the seventeenth century was produced there. This is largely because papermaking required fast-moving streams or rivers, rare in the Low Countries, to power paper mills.20 For example, many of Rembrandt’s earlier prints appear on Swiss paper marked with the canton of Basel’s heraldic sign, the Basel Crosier (see cat. nos. l and 59). With the gradual removal of wartime obstacles to seaborne trade,21 as well as the treaties of Munster and Osnabriick in 1648, which ended continental wars including that of the Dutch with Spain, paper from southwestern and central France began to appear in greater quantities on the Dutch market. This paper was often either made in French mills owned by Dutch entrepreneurs, or brokered by Dutch merchants.22 This explains both the relatively sudden advent of watermark types not previously seen in the Netherlands, such as the Phoenix and the Paschal Lamb, and perhaps, more importantly for Rembrandt studies, the disappearance of once common Swiss papers as the market shifted.23 This also resulted in many French papers marked specifically for the Dutch market, bearing, for example, the coat of arms of the city of Amsterdam, or referencing the seven provinces of the Dutch Republic.24

Scholars have been examining the watermarks found on artists’ papers at least since the late nineteenth century.25 But the study of watermarks in these early days relied largely on freehand copying or tracing of watermark designs, which, due to the difficulties of properly seeing watermarks on heavily inked sheets, complicated the process of assembling exact matches required to reunite

sheets from the same mold and confidently draw conclusions about printing practices.26 It has only been with the advent of photographic techniques such as beta- radiography, as pioneered since 1980 by Nancy Ash and colleagues in the United States, that it has been possible to systematically obtain clear radiographic images of watermarks.27 In the later 1980s, the members of the Dutch Rembrandt Paper Research Project developed a method for recording watermarks on Rembrandt prints in the collection of the Rijksmuseum using low-energy or grenz x-ray technology.28 These technology-intensive methods, and to a lesser extent, even simple transmitted and raking light photographs, have allowed researchers to overcome the inconsistencies of previous watermark tracing methods and enabled the exact matching of watermarks.29

The significance of watermarks to understanding Rembrandt as printmakerThe availability of images, combined with exhaustive research, has led to important conclusions about Rembrandt’s practice. The WIRE project at Cornell is based primarily on the pioneering scholarship contained in Erik Hinterding’s book Rembrandt as an Etcher: The Practice of Production and Distribution, which appeared in 2006 as an expansion of Nancy Ash and Shelley Fletcher’s 1998 publication Watermarks in Rembrandt’s Prints. Hinterding’s work has greatly enhanced the picture of Rembrandt as printmaker, offering conclusions about the types and quantities of paper the artist was accustomed to use.

According to the system Hinterding adapted from Ash and Fletcher, within each type (Basilisk, Double-headed Eagle, Arms of Amsterdam, etc.) the differing variants are


FIGS. 4 , 5 Can you spot the differences? Radiographs of Doubleheaded Eagle twinmarks A.a.a. (top) on an impression of Naked Woman Seated on a Mound, ca. 16 3 1 , in the British Museum,

London, and A.a.b. (bottom), on an impression of Diana at the Bath, ca. 16 3 1 , in the Uffizi, Florence. (Images courtesy of Erik Hinterding)

assigned uppercase alphabetical designations—A, B, C,D, and so forth. Watermarks that share characteristics within these variants but are not identical to each other are termed subvariants, and given a further set of lowercase initials—A.a. and A.b., etc. Within subvariants, Hinterding distinguished twinmarks with a further set of lowercase letters. For example, papers designated A.a.a. and A.a.b. would be twinmarks from the same batch (figs. 4, 5).30

Each watermark subvariant stands for a batch of paper; all told, Rembrandt is believed to have used more than 250 such batches in his career. WIRE project discoveries have already begun to augment that number. Hinterding’s research also revealed Rembrandt’s practice of making plate changes for successive states in rapid succession, often within the availability of a single batch of paper. He also discovered that, of all the different plates printed on a given batch of paper, those with the most recent dates are almost invariably the freshest, darkest impressions, while impressions from earlier plates on the same paper often showed wear. This phenomenon revealed Rembrandt going back to his stock of earlier plates to reprint them along with new ones, and led to the rule of thumb that the latest in-plate date of a plate on a given paper often dates the group of all plates printed on it.31

Often, watermarks can be used to assist in the virtual reassembly of the remaining impressions from an edition. For example, research for the WIRE project has revealed that Harvard University Art Museums’ fourth-state impression of the Christ Crucified between the Two Thieves: ‘The Three Crosses’ (cat. no. 37) bears an example of the Strasbourg Bend D’.a. watermark. This discovery adds it to the only four other known impressions of this print in this edition of the fourth state, the paper datable to circa 1653.32 Another example is Christ Preaching ('La Petite Tombe')(cat. nos. 48-50): various impressions of this popular print are known on papers with the Q’.a.a. and Q’.a.b. twinmarks of the Foolscap with Five-Pointed Collar watermark, to which the WIRE project research has now added a new instance (cat. no. 49).”

As previously mentioned, watermarks can also be used to reassemble groups of plates all printed at roughly the same time. In the exhibition, impressions of Self-Portrait Leaning on a Stone Sill (cat. no. 3; fig. 6), The Hog (cat. no.23; fig- 7),34 and Ephraim Bonus, Jewish Physician (cat. no. 9) all appear on the same paper marked with an image of a Basilisk, the twinmarks A’.a.a. and A’.a.b. (figs. 8a,b).35 This shows Rembrandt going back to reprint earlier plates—in this case, the self-portrait is reprinted approximately eight years after its first creation. Twenty-five different plates appear on this large-format Basilisk A’.a.a. paper, indicating that Rembrandt must have bought a rather large batch of it


FIG. 6 Transmitted light detail of Self-Portrait Leaning on a Stone Sill (cat. no. 3), with Basilisk twinmark A ’.a.b., image reversed. (Image courtesy of Theresa Fairbanks-Harris)

FIGS. 8 a,b Radiographs of Basilisk twinmarks A ’.a.a. (left) and A ’.a.b. (right) both on impressions of Presentation in the Temple: Oblong Print, ca. 1640, in the Rijksmuseum, Amsterdam. (Images courtesy o f Erik Hinterding)


and continued to print on it for years.36 And while this is an unusual instance in Rembrandt’s practice, it is also a strong indication that Rembrandt had a printing setup in his home, since it is highly unlikely that a professional printer would have had access to the same batch of paper over such a long period of time.37 Further, black chalk additions to the self-portrait that widen the right side of the beret in Yale’s impression of Self-Portrait Leaning on a Stone Sill (cat. no. 3) could indicate that Rembrandt was still considering ways of perfecting the image, even at this later date.

The assigning of dates to prints and drawings via datable books or documents appearing on paper bearing the same watermarks has also long been practiced in watermark scholarship. This logic can be used to narrow down printing dates for Rembrandt etchings as well. In the past, when watermarks were normally traced by hand, these matches were of necessity suggestions rather than conclusions.With the advent of photographic methods for imaging watermarks, it became possible to be reasonably certain of a match. For example, the Morgan Library & Museum’s second-state impression of The Flute Player, etched in 1642 (cat. no. 59), is printed on paper with a Basel Crosier watermark, subvariant B’.a.a., that confirms its printing date—as well as the first and third states of the print—to 1642 also, because impressions of all three states appear on this same batch of paper.38 Yet, at least four impressions of the fourth state of this print are found on paper with Foolscap with Five-Pointed Collar variant H.b.39 While this paper is not precisely dated in Hinterding’s 2006 study, this watermark finds an exact visual match with a watermark on a document dated 1642 in the archives of The Hague.40 This demonstrates that Rembrandt could have made the changes leading to the fourth state, like the other three states, also in the year of that plate’s creation.

Hinterding’s watermark research also revealed an anomaly in Rembrandt’s print chronology. Through the first two decades or so of his print production, the artist’s earlier plates tended to appear on sheets from the same batch of paper with those just etched, but this practice ends somewhat abruptly in 1654, as if Rembrandt no longer had his older stock of plates on hand to print at that time.41 Various explanations for this have been advanced, most of them relating to Rembrandt’s financial difficulties of the 1650s.41 Some scholars believe the absence of Rembrandt’s copperplates in the 1656 inventory of his possessions drawn up during bankruptcy proceedings indicates that he had pawned or otherwise transferred them for safekeeping before the inventory was taken.43 Rembrandt did hold a sale of his possessions prior to this inventory, in December of 1655, but it seems unlikely to have included his plates, given the potential future income they would have represented.44 And he did demonstrably release or sell some of

his plates during his lifetime, notably portrait plates that remained with the sitter, presumably either as a gift or as part of the commissioning arrangement.

One such example is the plate for his famous portrait of Jan Six (cat. nos. 10 and 11), which remains with Six’s descendants in Amsterdam to this day. And, in at least one documented case, Rembrandt sold the plate of Abraham Casting Hagar and Ishmael, 1637 (NHD, no. 166), to the Portuguese painter Samuel ben Orta.45 Others obtained Rembrandt plates during the artist’s lifetime. For instance, the Amsterdam print publisher Dancker Danckerts owned the plate for Descent from the Cross: Second Plate (cat. no. 25) and an unidentified Saint Jerome subject, as documented in his inventory of 1667 found in the Amsterdam city archives.46 Other evidence in the form of a large shipment of 189 Rembrandt etchings to Italian collector Antonio Ruffo in Messina, sent just before Rembrandt’s death on October 4,1669, suggests that Rembrandt was still actively involved in distributing impressions of his prints from a variety of plates even at that late date.47

The Amsterdam print publisher and connoisseur Clement de Jonghe (ca. 1624-1677) appears to have been the recipient of the greatest number of Rembrandt’s copperplates—seventy-four of them—as shown by the 1679 inventory of his shop and home.48 Little is known about the relationship between Rembrandt and de Jonghe, but the two must have known each other at least since 1651 when Rembrandt etched his portrait (cat. nos. 12-15). No documentation is known showing exactly when de Jonghe obtained plates from the artist, although it is possible he received some during Rembrandt’s lifetime.49 Study of the watermarks found jointly on Rembrandt impressions and on other artists’ prints from Clement de Jonghe’s stock seems to argue for the latter’s obtaining plates from Rembrandt at least later than 1656.50 Increased access to watermark images and their comparative study proposed by the WIRE project may shed greater light on this topic in future.

The Cornell WIRE project:Goals, methods, and progressThe Watermark Identification in Rembrandt’s Etchings (WIRE) project at Cornell is a collaboration among the Johnson Museum and the School of Electrical and Computer Engineering in Ithaca, Cornell Tech in New York City, and key external museum and technical experts (see Appendix 1). Conceived on a model of participation by cross-disciplinary student teams with the oversight of Cornell faculty and Johnson Museum staff, WIRE is dedicated to ongoing research and interface development to significantly streamline access to and raise awareness about Rembrandt’s watermarks for printing practice and chronology purposes. The project aims to establish new


research tools available to a collaborative community of users—curators, students, collectors, art historians, archivists, librarians, and other humanities professionals. These tools are necessary because, given the hundreds of subvariants within the watermark types (with more being discovered), as well as the difficulties inherent in distinguishing between marks (even in high-quality photographic images), there remains significant risk of error in comparing an unidentified watermark with those illustrated in Hinterding’s publication.

The WIRE project originated from a related initiative, the Chain Line Pattern Matching Project (CLiP), launched in 2012 when paper conservator Margaret Holben Ellis encouraged Professor C. Richard Johnson, Jr. to consider applying his computational approach to thread counting for canvas weave matching in historic paintings, in which Professor Johnson had been successfully engaged for years, to mold screen pattern matching for laid papers. Johnson and Ellis set about matching chain line patterns in Rembrandt’s papers to determine the feasibility of finding moldmates (sheets of paper made from the same mold) without using watermarks. In 2014, the Dutch University Institute for Art History in Florence granted the project access to a large dataset of radiographs of Rembrandt’s papers, which featured many moldmates identified by matching watermarks. This spurred the creation of a semiautomated software tool for marking and matching chain line patterns, which prompts the user to click on three spots in the images and then automatically marks a user-selected number of spots along each chain line on the radiograph of a given sheet. The marked locations can then be used to compose straight line patterns drawn through the marked points.51 The similarity of chain line patterns in all pairings of images in the full dataset can be ranked.A top ranked set of four matching chain line patterns was thus discovered in the dataset from the Dutch University Institute for Art History. Three of the images contained only fragments of the top of a foolscap watermark, but their chain line pattern match helped confirm that all could be identified as Foolscap with Five-Pointed Collar, Hinterding subvariant K.a.a.52

The large size of the dataset, and the need to use watermarks for verification in assessing the accuracy of the CLiP similarity evaluation procedure, drove the need for a faster, computer-assisted means of classifying watermarks according to the Hinterding taxonomy. The shift from using watermark identification as a step in chain line pattern matching performance assessment to using chain line pattern matching in service to watermark identification, in particular with fragments and twins, gave rise to the first gathering of a cross-disciplinary student research group at Cornell in the fall semester of 2015. Dr. Hinterding’s direct

involvement in the project from its inception stemmed from a desire to preserve the watermark taxonomy he inherited and expanded. This connection was facilitated by Professor Johnson’s appointment as scientific researcher at the Rijksmuseum.

During the subsequent four academic semesters, the WIRE project has proceeded as a hybrid of an engineering research group and an art history seminar, composed of students with different academic backgrounds and skill sets (see Appendix 1). Guided by the present authors, they are responsible for foundational readings and discussion as well as individual progress on watermark decision tree branches by type, group feedback sessions, and presentations on related topics. The group is conceived as a learning laboratory where students contribute to the ongoing development of a cross-disciplinary learning model in the computational art history field, squarely aligned with the Johnson Museum’s teaching mission and emphasis on learning from original works of art.

The WIRE project’s objective is to build an interrogatory, computer-based interface that delivers a definitive watermark identification according to the fifty-four types of Rembrandt watermarks encompassed in Hinterding’s taxonomy. Its identification tool is based on the model of the decision tree, a branching graph that allows the researcher to visualize all possible outcomes of a series of decisions (fig. 1), a sort of visual algorithm or procedure for solving a recurrent problem. This particular approach was chosen in large part because it best approximated Hinterding’s winnowing process for separating subvariants when he assembled the 2006 publication.

In the WIRE project, students construct decision tree branches for watermark types that take the form of a sequence of comparison images and questions with binary answers— “yes/no,” “one/two,” “square/rounded,” etc.— about features of the watermark and its placement on the mold.53 These questions develop incrementally from close visual observation and comparison, and are vetted by the group for clarity. The questions guide the researcher quickly and consistently to the correct watermark subvariant, unlocking access to information about the dating of the print where known. Differentiation between watermarks frequently exploits the often crucial placement of watermark elements relative to chain lines in the laid paper mold as a unique distinguishing feature.54 All such identifying features are scale-invariant—they relate to relative position, number, and size of features within the image, but do not rely on matching the exact dimensions of a given watermark. This aspect allows the user to compare multiple image sizes and formats with the visual models provided in the decision tree, even images in different scales.


The virtue of the WIRE project’s current focus on Rembrandt is the manageable and definable dataset going forward. The benefit of semiautomation rather than an automated, machine-learning approach to similarity assessment for this project is that the knowledge of the original watermarks experts is maintained and enhanced by the WIRE student researchers in the form of a verbal record of distinguishing features presented in the questions asked of the researcher. This subject expert terminology is already partially present in the headings for each watermark type in Ash and Fletcher’s 1998 catalogue and Hinterding’s 2006 study.55 But the WIRE project is systematically authoring a comprehensive description of the physical characteristics of every single Rembrandt watermark down to the subvariant level. As of this writing, WIRE student work groups have developed decision tree branches for fourteen watermark types, starting with the decision tree for the Foolscap with Five-Pointed Collar watermark.56 Students have presented their decision tree branches twice to curators and conservators at the Metropolitan Museum of Art. Additionally, students curated a study gallery installation on the WIRE project at the Johnson Museum.57

The WIRE decision tree software (figs. 9-10) was initially developed by Cornell Electrical and Computer Engineering graduate students Dong Ki Kim (M.Eng 2016) and Hao Lu (M.Eng 2016) in consultation with Dr. Louisa Smieska (PhD ’15) and Alison McCann (PhD candidate, History of Art and Visual Studies). Initially written in HTML, Cascading Style Sheets (CSS), and JavaScript, the program was designed to allow students with minimal programming experience to be able to code in individual branches of the decision tree. Each student is provided with a self-contained set of programs to download and instructions on how to proceed, and performs a sequence of fill-in-the-blank operations that insert the decision tree details into the code.

The resulting interface allows a user to upload a watermark image in any format in the image editor box at left, and select the matching watermark type from choices provided. The user then follows the short sequence of questions provided, with annotated images supplied for further guidance. The user may backtrack or start over at any time. Ultimately, the user will find an exact match with the subvariant letter designation found in Hinterding’s 2006 book. Students have been drafting reports for each individual branch they complete, culling out information from the 2006 publication about other collections holding prints with that watermark, and other plates that appear on that paper, to offer context for each subvariant “answer.” These reports are part of ongoing discussions with Dr. Hinterding for determining how, in the future, this information could appear online when the user arrives at the correct watermark subvariant.

The implications of new watermarks, and next stepsAlong with the systematic student development of the decision tree and the software allowing users to access its store of visual differentiations, the WIRE project is focusing on museums whose Rembrandt print holdings were not surveyed and imaged for previous watermark catalogues. This newly gathered data will drive the project’s second phase, that of developing an online database of Rembrandt watermarks in American collections. In addition, the present authors along with subsequent WIRE student cataloguing assistance are revisiting some collections whose Rembrandt watermarks were catalogued but not imaged in Hinterding’s 2006 publication. The purpose of this is to verify existing watermarks, catalogue those not previously noted, and collaborate with curators and conservators to establish imaging procedures and parameters for the proposed database. Many of these collections are among the lenders to this exhibition and have generously lent time and expertise toward these efforts (see Appendix 1). As hoped, this process has already yielded a crop of new instances of previously identified watermarks; it has also uncovered new watermarks not already present in the Rembrandt taxonomy or previously associated with Rembrandt’s prints.

For example, the plate for Rembrandt’s Self-Portrait (?) with Plumed Cap and Lowered Saber (see cat. no. 1), dated 1634, underwent significant changes during Rembrandt’s lifetime. In the first state, the composition began as a three-quarter length figure on a rectangular plate; in state two, the plate was cut down into a rough oval, keeping just the bust of the figure. In its third and final state, the oval plate was smoothed and regularized. Previous watermark study has shown that the initial cutting down of the plate happened relatively quickly, due to the existence of impressions of both first and second states on paper from the same batch bearing the Basilisk B.a. watermark.58 An impression of state two with an Arms of Wurttemberg watermark—also appearing on other Rembrandt prints dating to 1634 and earlier—further helps to temporally situate the printing of the second state.59 Finally, an impression of state three features an Eagle with Basel Crosier watermark also found on impressions of Abraham Casting out Hagar and Ishmael (NHD, no. 166) from 1637. This means the plate was smoothed to its final oval prior to that date.

Building on this background knowledge, research for the WIRE project and new photography by the Yale University Art Gallery’s paper conservation lab have revealed that the Yale impression of Self-Portrait (?) with Plumed Cap and Lowered Saber (fig. 11) also has an Arms of Wurttemberg mark (fig. 12) nominally of the A variant type. However, the mark exhibits previously unseen characteristics,60 indicating an apparently new subvariant, and thus a


F u ll Im a g e M a g n if ie d

Is there one central crown or two individual crowns?

- Select - *

One crown

Two crowns

Exam ple:

In the top im ages, the d o ub le-head ed e agle h a s o ne c

(m arked with red).

FIGS. 9,10 Screen shots showing two successive steps in the W IRE software for the identification of a test subject Double-headed Eagle watermark.


FIG. 11 Self-Portrait (?) with Plumed Cap and Lowered Saber, third state, 16 34 (Yale University Art Gallery, 1931.48). (Image courtesy of Theresa-

Fairbanks Harris) FIG. 12 Self-Portrait (?) with Plumed Cap and Lowered Saber, 16 34 (Yale University Art Gallery, 19 31.48), shown in transmitted

light, inverted and flipped horizontally. (Photograph by Andrew Weislogel) FIG. 13 Beta-radiograph of Self-Portrait (?) with Plumed Cap and Lowered Saber (cat. no. 1), inverted and flipped horizontally, showing Basel Crosier watermark with initials “M B” flanking the crosier inside the

shield. (Image courtesy of the Conservation Center, Institute of Fine Arts, New York University)

previously unidentified batch of paper used to print this plate. Correspondences with watermarks on archival documents further suggest that the Yale impression of Self-Portrait (?) with Plumed Cap and Lowered Saber was printed during the 1630s.61 Another third-state impression of this print has recently come to light, as a gift to the Johnson Museum from former director Frank Robinson in 2016 (cat. no. 1). This impression was found to show a Basel crosier watermark with initials “MB” flanking the crosier inside the shield (fig. 13). This watermark stood out because none of the variants published by Hinterding exhibits such initials. Indeed, in consultation with Dr. Hinterding it was deemed to be new to the Basel Crosier type.62 Since Swiss papers like this are rare in Rembrandt’s printing practice after the early 1640s,63 a printing date before 1642 might be advanced.64 Therefore, two new batches of paper can provisionally be added to the list of ten already associated with the printing of this plate between 1634 and 1651,65 both arguably early. As such, our picture of Rembrandt as printmaker continues to expand, and here also perhaps our sense of the popularity of the tronie type Rembrandt was creating by cutting down the plate.

A second example of a new watermark found on Cornell’s previously unpublished fifth-state impression of The Flight into Egypt: Altered from Segers (cat. no. 34) helps to reconstruct a batch of paper used for a later edition of the

plate. Upon first examination, the watermark on this sheet (fig. 14) appeared to be a countermark IHS subvariant A.b.66 Indeed, the only print previously found on this paper was another fifth-state impression of the same print, in the Rijksmuseum (fig. 15).67 So the Johnson Museum print and the Amsterdam print appeared to be printed on moldmate sheets. However, a renewed examination of the Cornell IHS countermark revealed that it is in fact not an exact match to the Hinterding A.b. subvariant, but instead nearly identical due to the variation in the shape of the S, at the top. Therefore, rather than being moldmates, it seems logical to assume these countermarks might instead be twinmarks from the same pair of molds.

Two other impressions of the same state of The Flight into Egypt: Altered from Segers appear on papers bearing the respective twinmarks of the Strasbourg Lily 4WR A.b. subvariant. Both this mark and the IHS countermark are found only on that same print—and only in the fifth state. Additionally, all four impressions discussed here are printed on half sheets.68 Thus, given the evidence of watermarks and paper size, the IHS countermarks on the Cornell and Amsterdam impressions could logically be the twinmarks for the two impressions showing Strasbourg Lily with initials 4WR A.b.a., and A.b.b., which are also half-sheets, and also twinmarks. Additionally, the fact that the three impressions classified as state V-a all appear with either


IHS or Strasbourg Lily marks makes further sense because IHS is the countermark for some Strasbourg Lily 4WR A variants.69 If this is the case, the new Cornell watermark helps to reconstruct a later edition of The Flight into Egypt: Altered from Segers, and also becomes a new subvariant to be added to the Rembrandt taxonomy.70 Finally, a new instance of this same Strasbourg Lily watermark has come to light on an impression of the eighth state of Adoration of the Shepherds: A Night Piece (NHD, no. 300);71 since the date of this print is ca. 1657, this watermark indicates a print date for all of the impressions of The Flight into Egypt: Altered from Segers of not earlier than 1657, which fits with the wear exhibited by the plate in these impressions.

Cases like these underscore the WIRE project’s dedication to incrementally filling in gaps in Rembrandt’s printmaking chronology. From the standpoint of the decision tree branch construction, the discovery of new watermarks necessarily forces a reconsideration, and in some instances a reordering, of the decision tree branch for that watermark. Because the existing taxonomy and its images—extensive as they are—are the sources for the construction of the branch, a new question has become apparent: how to integrate a newly discovered watermark into the decision tree without upsetting the existing structure? In the case of the aforementioned Basel Crosier watermark with the initials MB, this task fell to Riley Henderson, Cornell Class of 2018, who had already developed the decision tree branch for the Basel Crosier mark and its eleven preexisting subvariants. Based on similarities of various details other than the letters,

as well as the presence of a chain line running down the center of the shield, she proposed that the mark be added to the E variant group, where it has been given the subvariant designation “E’.b.” becoming the twelfth Basel Crosier subvariant, with Dr. Hinterding’s approval.

The discovery of new watermarks and the alteration of the decision branch for the Basel Crosier watermark made clear that the decision tree software must be adaptive. This requirement is now under consideration by Professor Johnson, Cornell Tech electrical and computer engineering professor Vikram Krishnamurthy, and his PhD student Sujay Bhatt, who are treating these questions as an advanced applied mathematics issue, developing an algorithm for automating the insertion of new parts into the decision tree as they arise. For this reason, in the spring of 2017, WIRE students not only developed individual branch diagrams for Rembrandt watermark types (see figs. 1 and 16) and coded them into the software, but were also directed to construct decision “tables” (fig. 17A). For every subvariant of a particular watermark type, a decision table asks and answers all questions used in identifying all subvariants within that particular type, not just the questions asked of a user following one specific path to a particular subvariant.This richer set of information enables computational algorithms to revise and upgrade the original branch.

Further, careful reordering of the questions asked (fig. 17A) can lower the average number of steps required to receive a correct answer for any subvariant. For example, in

FIG. 14 Beta-radiograph of the IHS countermark on The Flight into Egypt: Altered from Segers (cat. no. 35). (Image courtesy of the Conservation Center, Institute of Fine Arts, New York University)

FIG. 15 Radiograph of the IHS countermark, Hinterding variant A.b. on The Flight into Egypt: Altered from Segers, ca. 16 53 (Rijksmuseum,

Amsterdam, O B .116). (Image courtesy of Erik Hinterding)


[ START |

FIG. 16 Decision tree branch for Fleur-de-lis watermark, developed by Phoebe Ross, Spring 20 17

fig- 17B, the average number of questions to a solution when starting with Question 1 is (2+2+2+4+4+3V6 = 17/6 = 2-833, which can be slightly reduced by starting with Question 4 producing an average of (2+3+3+3+3+2)76 =16/6 = 2.667• ln some instances, this efficiency is helpful.In other instances, the ability to generate different tree diagrams based on these tables might enhance the ability to better identify watermarks from fragments (when a watermark is cut off at the edge of a sheet). When a fragmentary watermark is encountered, it may prove useful to redraw the pertinent decision tree so that the user is asked questions relevant to what is visible in the fragment first, before other questions. This ability of the decision table and pertinent mathematics to bring to light all the different possible decision tree branches makes the decision table a powerful tool, which could ultimately provide the foundation for a better method of drawing up and expanding decision trees. This could have broad applications in a range of disciplines.

As the project progresses, certain modifications will allow for the growth of the decision tree, and for the enhancement of the user interface. For example, in the process of developing a consistent methodology for imaging and comparing watermarks during the 1980 and 1990s, Nancy Ash and other colleagues strongly recommended accounting

A: Decision TableVariant Q1:

Is themiddle petal divided by a central line?

Q2:Are the three bars at the center of the Lily the same width?

Q3:Are the petals of the Fleur-de-lis close together?

Q4:Are the ends of the three bars at the center square?

Q5:Is the outer tip of one of the smaller leaves (at bottom) close to the chain line?

A.a. N N N N YB.a. Y N Y Y NB.d. Y Y Y Y NC.a. N N Y N YC.b. N N Y N NC.c. N N Y Y N

B: Path TableVariant Starting with Q1: Starting with Q4:

A.a. Q1(N), Q3(N) Q4(N), Q3(N)B.a. Q1(Y), Q2(N) Q4(Y), Q1(Y), Q2(N)B.d. Q1 (V), Q2(Y) Q4(Y), Q1(Y), Q2(Y)

C.a. Q1(N), Q3(Y), Q4(N), Q5(Y) Q4(N), Q3(Y), Q5(Y)

C.b. Q1(N), Q3(Y), Q4(N), Q5(N) Q4(N), Q3(Y), Q5(N)

C.c. Q1(N), Q3(Y), Q4(Y) Q4(Y), Q1(N)

FIG. 17 Decision table for Fleur-de-lis watermark, developed by Phoebe Ross, Spring 2017

4 4 W E I S L O G E L / J O H N S O N

for the possibility of overlaying one watermark image over another in order to assure an exact match.72 Therefore, for future iterations of the WIRE interrogatory software, a digital registration or overlay component will be explored in order to compare and verify the watermark under examination with the correctly identified subvariant. Also, in cases where a single image or format cannot capture the whole watermark, a few different images of a given watermark may need to be uploaded to be compared with the illustrations in the decision tree; this is often the case when the ink of the printed image interferes with a clear reading of the watermark.

It has also been observed that, when watermarks are fragmentary, the chain line pattern marking and matching (CLiP) procedure referred to above is still very useful for identifying moldmates. This also suggests that chain line marking and matching could prove useful in distinguishing twinmarks, which are often extremely hard to tell apart. Thus, the key remaining task for the CLiP Project is the creation of a user-friendly software package for CLiP marking and matching, to be constructed in a freely available computational software package in order to encourage adoption by museums, independent conservators, and collectors. To address this issue and to continue to upgrade the user interface, a student software team was formed at Cornell in the 2017 spring semester. Amanda House (Computer Science ’17) and Katarina Martucci (Electrical and Computer Engineering ’20) are refining the WIRE user interface and consolidating the CLiP marking and matching procedure. They will continue working together throughout the fall 2017 semester and beyond.

In addition, to encourage standardized image gathering of watermarks and chain line patterns in Rembrandt’s etchings by museums with limited technical resources, a portable imaging system is under development in a collaboration with Paul Messier at the Yale University Lens Media Lab, John Delaney in the conservation lab at the National Gallery of Art, and Marc Walton and Oliver Cossairt in the NU-ACCESS program at Northwestern University. In this regard, Theresa Fairbanks-Harris, senior conservator of paper for the Yale University Art Gallery and the Yale Center for British Art, and Suzanne Boorsch, the Robert L. Solley Curator of Prints and Drawings at Yale University Art Gallery, have been especially helpful in reexamining Yale’s Rembrandt prints and in imaging watermarks to help consider questions of format and delivery of images, and have helped construct the WIRE project’s first nonradiograph image dataset (see Appendix 2).

The seeds for WIRE’S proposed database of Rembrandt watermarks in American collections were presciently sown by Nancy Ash and her colleagues in 1986.7! The database

proposed by the WIRE project will be informed by their original ideas, as well as by the International Society of Paper Historians cataloguing standards.74 The uses of such a database will be many, but among them will be the facility to find and compare multiple instances of a given watermark across collections, to compare and verify twinmarks across extant edition-mates of a given print, and to easily add new watermarks to the list of papers on which a given print is found. Perhaps most useful to researchers will be the ability to easily compare watermark and chain lines with the actual surface image of the print, with all of its visual cues about state, quality of impression, apparent plate wear, and inking and wiping details. These cues, as Erik Hinterding has demonstrated, can reveal so much in tandem with paper characteristics for situating an impression in Rembrandt’s chronology.

ConclusionDespite the important foundational work on Rembrandt’s watermarks accomplished over the past three and a half decades, it has become clear that the picture of Rembrandt as a printmaker can still benefit from refinement and clarification. Key paths to this refinement include seeking out the many watermarks still remaining to be discovered, expanding the taxonomy for Rembrandt’s watermarks, and making it broadly accessible to researchers at many levels. The procedures put in place by WIRE student teams promise to result in a useful and accessible decision tree structure that will in turn encourage the construction of a larger database of Rembrandt watermarks in United States collections.The recording of fifteen new examples of watermarks in the holdings of just the two organizing institutions of this exhibition suggests the likelihood of discovering further previously unseen marks. These additions also encourage engagement with many smaller college, university, and municipal museums with Rembrandt etchings in their collections. Appendix 2 details progress in cataloguing these previously uncounted collections, as well as the results of a renewed effort in the collection of the Yale University Art Gallery, which has revealed more watermarks than previously noted. Finally, investigations are expanding to include papers found in the drawings of Rembrandt and his students, as watermark evidence suggests the use of printing papers for some drawings supports in Rembrandt’s milieu.

The complex story of Rembrandt’s papers and their interpretation is humbling, but the practice of student involvement in research is a constant call to recalibrate and clarify expectations and goals, demanding that all participating researchers, whether students, faculty, or museum professionals, maintain vigilance as active learners. This research model offers various encouraging benefits, including an energetic and computer-savvy workforce who


have developed a range of skills, and a partial overlap of student cohorts from semester to semester that continually transforms learners into mentors as the project progresses. Further, the project offers students the opportunity to pursue research of value to the field without the necessity of physical access to large urban concentrations of museum collections.

The WIRE project has engendered dialogue among project staff and students who are open to learning about each other’s fields. As such, it can be viewed as a sort of crucible for the development of students interested in art who can speak the language of colleagues in technical and computational fields—colleagues who, in turn, can offer much toward the furthering of important questions in the study of art. The academic museum—with its object- based approach to works of art, and its pre-existing connections to related professions such as conservation and conservation science— provides a natural forum in which to assemble such a research group. And, in a broader sense, linking computational studies with traditional humanities content on the campus of a research university stands to inform other such relationships and projects in the growing field of digital humanities.

EN D N O TES

1 A full listing of student participants in the W IR E project is found in Appendix 1.

2 Amsterdam-London 2000-01, 34.

3 For further discussion of collecting at Cornell, see page 18.

4 Exhibitions featuring Rembrandt at Cornell include: Dutch and Flemish Prints from the Seventeenth Century (September 3 -2 9 ,19 6 8 ) ; Dutch Drawings o f the Seventeenth Century, from a Collection (November 6-December 2 3 ,19 7 9 ): Master Prints from Upstate New York Museums (October 26-December3 1. 1995); Ziet, de Dag Komt Aan: Dutch Landscape Prints and Drawings (June 13-A ugust 6, 1995); Things o f this World: Dutch Prints and Drawings o f the Seventeenth Century (December 2 1 , 1996-M arch 9, 1997); Rembrandt and the Art of Etching (January 10-M arch 15 ,19 9 8 ) ; Reflections to Astound: Seventeenth-Century Dutch Prints from a Private Collection (August 28-November 7,19 9 9 ); Fresh Woods and Pastures New: Seventeenth-Century Dutch Landscape Drawings from the Peck Collection (January 29-M arch 26, 2000); Etchings by Rembrandt from the S. William Pelletier Collection (January 17-A p ril 4, 2004); A Portrait o f the Artist, 15 2 5 -18 2 5 : Prints

from the collection of the Sarah Campbell Blaffer Foundation (September l-Decem ber 3 1 , 2005); Rembrandt Etchings Collected by Nancy and Nelson Schaenen, Jr. (April l-Ju n e 18, 2006); Rembrandt at 400 (April l-Ju n e 25, 2006); A Changing Landscape: Prints and Drawings 1570-1670 from the Arthur and Arlene Elkind Collection (February 3-A pril 8, 2007); The New and Unknown World: Art, Exploration, and Trade in the Dutch Golden Age (August 13-O ctober 2, 20 11). All but the 1968, 2006, and 2007 exhibitions were accompanied by a catalogue.

5 Erik Hinterding, Rembrandt as an Etcher: The Practice of Production and Distribution (Ouderkerk an den Ijssel: Sound and Vision, 2006), 3 vols. This publication, often referenced

in this essay, will be noted hereafter simply by means of the abbreviation EH 2006, followed by volume and page numbers.

6 For a glossary of terms and techniques, see pages 2 1 2 -2 1 4 .

7 Van Strien 1993, 255 as cited in Hinterding 1993-94, 264, note 62.

8 White 1999, 10. Archival documents from 1643 and 1644 respectively mention shipments of large quantities of Japanese paper to the Netherlands. See Stockholm 1968, 17 3, as cited in Van Breda 1997, not paginated. See also EH 2006,1, 182, note 336. It is also possible that some of Rembrandt’s Asian papers were Chinese, though the greater likelihood is that they are Chinese-style papers of Japanese craft.

9 EH 20 0 6,1, 112 .

10 White 19 9 9 ,10 .

1 1 Ibid., 262, note 33.

12 S. W. Reed, “Types of Paper Used by Rembrandt,” in Boston- New York 1969, 178-80 , was an early example. Sale catalogues and catalogues raisonnes have also often included discussions of Rembrandt’s papers; see Mayer Haunton, “Rembrandt’s Papers,” in Ritman 1995. Also, G. Biorklund, “Old Paper,” in Stockholm 1968, 16 5 -7 3 .

13 For details on the construction of paper molds, see Loeber 1982, 13 -2 6 ; De Lalande 17 6 1-8 9 . For more detail on the papermaking process itself, see for example Hunter 1947, 17 7 -7 9 .

14 Loeber 1982, 31.

15 EH 20 0 6,1, 45.

16 Loeber 1982, 31.

17 Stevenson 19 5 1 -5 2 , 65.

18 Loeber 1982, 50.

19 As Erik Hinterding relates about his experience of the ratio of watermarked to nonwatermarked prints, “A total of 2,369 complete or fragmentary watermarks were found in the more than 7,000 etchings by Rembrandt examined, and 1,998 of them could be identified.” EH 2006,1, 42.

20 Chapel Hill-Ithaca 1999, 30.

21 Laurentius 2008, vi.

22 Chapel Hill-Ithaca 1999, 31.

23 Laurentius 2008, vi.

24 Voorn i9 6 0 ,10 2.

25 See, for example, Eugene Dutuit, Voeuvre complet de Rembrandt, Paris, 18 8 3 -8 5 . As cited in Amsterdam-London 2000-01, 23, and Boston-New York 19 69-70, 18 1-8 3 .

26 As Theo and Frans Laurentius note, during the eighteenth century paper molds and were made with thicker wires, to speed up the drainage of water from the mold, increasing efficiency and profits. This results in more clearly legible chain lines and watermarks. Seventeenth-century paper characteristics therefore can be harder to read than eighteenth-century ones, which explains some of the difficulties researchers of Rembrandt watermarks have traditionally encountered. See Laurentius 2008, v.

27 Beta-radiography uses a polymethacrylate sheet labeled with radioactive carbon-14 to produce an image on a radiographic film, which reveals the watermark and other mold characteristics without obstruction by the ink o f the design. This method produces high quality and high contrast images of watermarks and other texture characteristics of paper. For a fuller description of these techniques, see Tydeman 1967; Ash 1986, 64; and Chapel Hill-Ithaca 1 9 9 9 ,11 7 -2 4 .

28 Laurentius et al. 1992, 35 3 -8 4 . See also Erik Hinterding’s interview, pages 59-6 3.

29 Amsterdam-London 2000-01, 23.

30 EH 2006, II, 10.

3 1 EH 20 0 6,1, 25.


32 Ibid., 18 5-8 6 .

33 Ibid., 135 . Since Christ Preaching (‘La Petite Tombe’) is a single-state print, only watermarks and, to a less precise extent, wear evident to the plate can suggest the chronological placement of a given impression. One wonders in this case if microphotography of impressions designed to examine plate wear and wiping scratches to effect a sequence for the printing of certain Rembrandt plates, as Angela Campbell has initiated for the Meisterstiche of Albrecht Dtirer, would be warranted for further establishment of a chronology of such impressions. See Campbell and Raftery 2 0 1 2 ,1 5 - 1 6 .

34 Thanks to Elizabeth Martinson, Cornell Class of 2017, of the W IRE project for identifying this watermark on the Johnson Museums impression of The Hog. Martinson also developed the decision tree branch for the Basilisk watermark.

35 EH 2006, II, 65-66.

36 EH 20 0 6,1, 4 1, and II, 369-70.

37 Ibid., I, 41.

38 EH 2006, II, 62.

39 Ibid., 124.

40 Laurentius 2007,483, 2 1 1 . The archival documents consulted for this study are all expressly of the most common size imported at that time, approximately 32 x 4 1 cm, which corresponds to the half-sheet format of The Flute Player. See Laurentius 2008, xviii.

4 1 EH 20 0 6,1, 13 9 -4 1 .

42 For these, see Crenshaw 2006, especially 56-68.

43 Hinterding 19 9 3-9 4 , 258.

44 Van Eeghen 1969, 168.

45 De Groot 1976, 7 1 -7 7 .

46 Amsterdam 2 0 1 1 , 1 1 1 - 1 5 .

47 Ricci 19 18 , 30, as cited in Hinterding 19 9 3-9 4, 258.

48 Laurentius 2 0 1 0 ,1 2 7 -5 7 , especially 14 7-4 9 .

49 Hinterding 1993- 94. 260-62. De Hoop Scheffer and Boon 1971. 4-

50 EH 20 0 6,1, 143.

5 1 Xi et al. 2016, 135.

52 Ibid, 138.

53 The images of watermarks to be used in the decision tree that will be posted for public use are from Dr. Hinterding’s personal photograph collection that includes images of nearly all of the subvariants in his study.

54 Stevenson 19 5 1 -5 2 , 66; Ash, Nicholson in Milwaukee 1982.

55 See, for example, EH 2006, II, 104: “EAG LE, SIN G LEH EAD ED: A. Crowned, single-headed eagle with a crosier (not the Basel variety) on its breast. Double-layered wing feathers point downwards.”

56 The list of types is: Basel Crosier, Basilisk, Countermark Letter I, Single-headed Eagle, Single-headed Eagle with Basel Crosier, Double-headed Eagle, Fleur-de-lis, Foolscap with Five-Pointed Collar, Grapes, Countermark IHS, Paschal Lamb, Seven Provinces, Strasbourg Bend, Strasbourg Lily.

57 W IRE 2016.

58 Laurentius et al. 1992, 36 6-6 7; Guillaud and Guillaud 19 86,6 75.

59 EH 2006, II, 5 3 -54 ; Laurentius et al. 1992, 367.

60 It cannot be a match for published variants because in addition to some variations in the quartered arms within the shield, the “4B” pendant below the shield is reversed with respect to the coat o f arms. The watermark for this print was previously catalogued under “Unrecognizable ZZ.zz” (EH 2006. II, 241). The print is numbered there as 1931.46.

6 1 Similar Arms of Wiirttemberg marks on papers from Montbeliard, in eastern France near the Swiss border, including one with a similarly reversed numeral 4, appear on documents in the city archives of Middelburg, Rosendaal, The Hague, and Flushing, all with dates between 16 32 and 1639. Laurentius 2007, # 2 11 , 106.

62 Authors’ meeting with Erik Hinterding, New York, November 2, 2016.

63 Dates of Rembrandt plates found on paper with other Basel Crosier variants start in 16 3 1, but end in 1642. See EH 2006,II, 60-64. Only two batches of paper Rembrandt used after 1648 are likely of Swiss manufacture. These are Double-headed Eagle variants E.a. (1654) and D.a. (1656). See EH 2006, II, 4 1 5 - 1 6 . One paper batch with Basilisk A ’.a., appears until 1647, but this seems to have been a large-format paper also perhaps purchased in a large quantity around 16 4 1 and therefore took longer to use up. EH 2006, II, 36 7-6 8.

64 The quality of the Johnson Museum impression compares favorably with other early third-state impressions as confirmed by watermarks, for instance in the richness of the web of lines in the cloak below the gorget, and the hatching lines around the upper half o f the oval that frame the head.

65 In much the same way, the number of papers associated with Self-Portrait in a Velvet Cap with Plume of 16 38 has just been expanded by an instance o f the Horse and Rider A.b. variant found on an impression of state ii. See Boerner 2016, 62-63, no. 32.

66 EH 2006, II, 154.

67 Ibid. It is listed as White and Boon state six of seven, which corresponds to state five in New Hollstein. See NHD, no. 271.

68 Ibid., II, 258. One is in the Rijksmuseum (O B.116), and the other two have a Strasbourg Lily with Initials 4W R watermark, subvariants A.b.a. (Boston, Museum of Fine Arts, P.444) and A.b.b. (Musee Jenesch, Vevey, Switzerland, P.55). The rest of the twelve are all printed on smaller uncut sheets, most of them with the Arms of London watermark A.a.

69 Ibid., II, 15 4 and 199.

70 This hypothesis can be further explored by an examination of the quality of the impressions found on these papers. The Cornell print, according to the latest catalogue of Rembrandt’s etchings, is an intermediary state “V-a”, described as “ The marks in the sky not clear anymore. This may have been caused by either wear or burnishing.” NHD, no. 2 7 1 . An examination of the Amsterdam impression with the IHS A.b. watermark, also noted in N H D as state V-a, shows it to be o f comparable quality to the Cornell impression; both show traces of the vertical line at center and smudgy marks in the sky— the Cornell impression perhaps more faintly than the Rijksmuseum impression, but they are nonetheless present. The impression on paper bearing the Strasbourg Lily A.b.b. with initials 4W R, in the Musee Jenesch, Vevey Switzerland (M JV P. 55) is also listed as a state V-a impression. Finally, the Museum of Fine Arts, Boston impression, on the Strasbourg Lily A.b.a. twin paper to the Swiss impression (EH 2006, II, 197) is catalogued in NH D as a bona fide state V print; it does appear earlier from the plate than the Cornell or Amsterdam impressions, but the marks in the sky are nearly worn away here, as well. Thus, given their similarities in quality, it is feasible to imagine the printing of all four of these impressions within the span of a single batch of paper.

7 1 SUArt Galleries, Syracuse University 1963.932. For further information, see Appendix 2, page 52.

72 Ash 1986, 65; Nicholson in Milwaukee 1982, 2; Skalka 1995, 4 0 -4 1.

73 Ash 1986, 67.

74 International Association o f Paper Historians International Standard fo r the Registration of Papers with or without Watermarks, version 2 .1 .1 (2013). Accessed May 17 , 2017. http://www.paperhistory.org/standards/.


http://www.paperhistory.org/standards/

A P P E N D IX 1

WIRE project participants, 2015-present

Project leadersA n d re w C . W eislogel, the Seym o u r R. A sk in, Jr. ’47 Curator, Earlier European and A m erican A rt,

H erbert F. Johnson M useu m o f A rt, C o rn ell U niversity

C . Richard Johnson, Jr„ Jeffrey S. M . H ed rick Senior

Professor o f Electrical and C om p uter Engineering,

C orn ell U niversity / Jacobs Fellow in Com putational

A rts and H um anities, C orn ell Tech / Scientific

Researcher, Rijksm useum , A m sterd am / C om putational

A r t H isto ry Advisor, R K D — N etherlands Institute for

A r t H istory (The H ague)

Project advisory groupM argaret H olben Ellis, Eugene Th aw Professor o f

Con servation , C on servation C en ter o f the Institute

o f Fine A rts, N e w York U niversity

E rik H interding, C urator o f Prints, R ijksm useum , A m sterdam

N adine O renstein, D rue H einz C urator in Charge,

D epartm ent o f Prints and D raw ings, M etropolitan

M useu m o f A rt

O ya Y ild irim Rieger, A ssociate U niversity Librarian,

Scholarly Resources and Preservation Services,

C orn ell U niversity L ib rary

Stephanie L. W iles, the Richard J. Schw artz Director,

H erbert F. Johnson M useu m o f A rt, C orn ell U niversity

Cornell University: WIRE project students (2015 to present)Sujay Bhatt, P h D candidate, Electrical and C om p uter

Engin eering, C orn ell Tech

M argaret C an field ’2 0 , H istory o f A rt/C h e m istry

K atrina Ferreira ’2 0 , H unter R. Raw lings III

C orn ell Presidential Research Scholar, Industrial and

L ab o r Relations

V irgin ia G irard T 7 , H isto ry o f A rt

R iley H enderson T 8 , H isto ry o f A r t

A m an d a H ouse T 7 , C o m p uter Science

D o n g K i K im , M .E n g T 6 , Electrical and C om p uter

Engin eering

So Jeong Lim ’20

H ao Lu, M .E n g T 6 , Electrical and Com p uter

Engin eering

A n ju m M alik T 6 , A n th ro p o lo gy

Elizabeth M artin son T 7 , H istory

K atarina M artu cci ’2 0 , Electrical and C om p uter

Engineering

A liso n M cC a n n , P h D candidate, H istory o f A rt

K ira N icolai T 6 , H isto ry o f Art/Statistics

M argaret O rbon ’ 17 , H isto ry o f A rt

O scar Rieveling T 6 , H isto ry o f A rt/F ren ch

Phoebe Ross T 7 , Industrial and L ab o r Relations

Jason Setter, M .E n g T 6 , Electrical and C om p uter

Engin eering

Sam antha Siegler T 8 , H istory o f A rt

N in a Sim pkins T 9 , H istory o f A rt/In form ation Science

Xuelei X i, M .E n g Y 7 , Electrical and C om p uter

Engin eering

Cornell University: Faculty and staffM ichele Brow n , b ook conservator, O lin Lib rary

Paper Con servation Lab

M ichele H am ill, paper and photograph conservator,

O lin Lib rary Paper Con servation Lab

Jill Iacchei, O lin L ib rary Paper Con servation Lab

Professor V ik ram Krishnam urthy, Electrical and

C om p uter Engin eering, C orn ell Tech

Professor Steve M arschner, C o m p uter Science

Caitlin M oore, O lin L ib rary Paper C on servation Lab

Brittany Rubin, print room curatorial assistant,

H erbert F. Johnson M useu m o f A rt

Jennifer Scheuer, printm akin g specialist,

D epartm ent o f A rt

B ruce W alter, Sen ior Research Associate,

P rogram o f C om p uter Graphics

48 A P P E N D I X 1

Outside Cornell

OBERLIN COLLEGE

A n daleeb Badiee Banta, C urator o f European and

A m erican A rt, A llen M em orial A r t M useu m

UNIVERSITY OF IOWA

T im o th y Barrett, Director, C en ter for the B ook

FRANCES LEHMAN LOEB ART CENTER,VASSAR COLLEGEPatricia Phagan, the Philip and L yn n Straus C urator

o f Prints and D raw in gs

METROPOLITAN MUSEUM OF ARTLouisa Sm ieska (Cornell P h D T 5 ), A n d re w W. M ellon

Foundation C on servation Fellow in the D epartm ent o f

Scientific Research

MORGAN LIBRARY & MUSEUMReba Fishm an Snyder, paper conservator

NATIONAL GALLERY OF ARTJohn Delaney, Sen ior Im aging Scientist,

Scientific Research D epartm ent

NEW YORK UNIVERSITYLydia A iken h ead, conservation student,

Con servation C en ter o f the Institute o f Fine A rts

Saira H aqqi, C on servation C en ter o f the Institute

o f Fine A rts

NORTHWESTERN UNIVERSITY

Professor O liver C ossairt, Electrical Engin eering and

C o m p uter Science, N orthw estern U n iversity/A rt

Institute o f C h icago C en ter for Scientific Studies in the

A rts ( N U -A C C E S S )

Professor M arc W alton, M aterials Science and

Engin eering, M c C o rm ic k School o f En gin eerin g

and A pp lied Science, N U -A C C E S S

STATE UNIVERSITY OF NEW YORK AT CORTLAND

Professor C harles Heasley, Printm aking and N e w M edia

SYRACUSE UNIVERSITYD om en ic J. Iacono, D irector, Syracuse U niversity

A rt Galleries

UNIVERSITY OF WISCONSIN, MADISOND evin Conathan , D epartm ent o f Electrical and

C o m p uter Engin eering

Professor W illiam A . Sethares, Electrical and

C om p uter Engin eering

YALE UNIVERSITY

Suzanne B oorsch, the Robert L. Solley C urator o f

Prints and D raw in gs, Yale U niversity A r t G allery

Theresa Fairb ank s-H arris, Sen ior C on servator o f

Paper, Yale U niversity A r t G allery and Yale Center

for British A rt

Paul M essier, Lens M ed ia Lab, Institute o f the

Preservation o f C ultural Heritage

INDEPENDENT SCHOLARSG eo rg D ietz, independent paper expert

T im o th y M oore, paper m oldm aker, C o n co rd , M ichigan

We also acknowledge the assistance provided by the Dutch University Institute for Art History, Florence.

A P P E N D I X 1 49

A P P E N D IX 2

Annotated list of previously unpublished watermarks catalogued by the WIRE project, 2015-presentBy collection, with selected illustrations

I. Herbert F. Johnson Museum of Art,Cornell UniversityCompiled by Elizabeth Martinson, Cornell ’17, and Brittany Rubin

The Small Lion Hunt (with Two Lions), ca. 16 2 9

Etching

Bartsch/W hite and Boon 1 1 5 only state; N H D 28, ii/ii

Gift o f Paul Ehrenfest, Class o f 19 3 2 , 84 .0 29 .0 13

w atermark : Foolscap with Five-Pointed Collar, Hinterding subvariant K.i. (for reference: E H 2006,II, 1 3 3 ; illus., Ill, 2 3 3 )

The Rat Catcher, 1 6 3 2

CAT. NO. 18

Etching

Bartsch/W hite and Boon 1 2 1 iii/iii; N H D 1 1 1 iii/iii

G ift o f W ellington R. Burt, Class o f 19 2 6 , 6 5.0 23

w atermark : Counterm ark, Hinterding subvariant P D B ’.a.a. (for reference: EH 2006, II, 89; illus. Ill, 14 6)

The Angel Appearing to the Shepherds, 1 6 3 4

Etching, engraving, and drypoint

Bartsch/W hite and Boon 4 4 iii/iii; N H D 1 2 5 iii/vi

Bequest o f Louis J. Heizmann, 74.057.009

w atermark : M iscellaneous, Late, similar to Hinterding subvariant B.a. “O C T U IE R 17 4 9 ” (for reference: E H 2006, II, 238 , illus. Ill, 503)

new proposed tw in m a r k : M iscellaneous, Late B.a.b.

Self-Portrait (?) with Plumed Cap and Lowered Saber, 1 6 3 4

CAT. NO. 1

Etching and engraving

Bartsch/W hite and Boon 23 iii/iii; N H D 1 3 5 iii/iii

G ift o f M argaret and Frank Robinson, in honor o f Susan Lynch, 2 0 16 .0 39 .0 0 1

w atermark : Basel Crosier, similar to Hinterding subvariant E ’.a. with the addition o f letters M and B to either side o f the shield bearing the crosier (for reference: E H 2006, II, 64; illus., Ill, 108)

new proposed subvariant : Basel Crosier E ’.b.

(For further discussion and watermark illustration, see page 42.)

Three Oriental Figures (Jacob and Laban?), 1 6 4 1

Etching and drypoint

Bartsch/W hite and Boon 1 1 8 ii/ii; N H D 19 0 ii/ii

Gift o f the Estate o f Louis H. and Katherine W. Poliak,T R 9705.005

w atermark : Partial A rm s o f Am sterdam , Hinterding subvariant D.d. (for reference: E H 2006, II, 29; illus. Ill, 32 )

The Hog, 16 4 3

CAT. NO. 23


Bartsch/W hite and Boon 1 5 7 i/ii; N H D 2 1 5 ii/ii

Acquired through the generosity o f H elen-M ae and Seym our R. Askin, Jr., Class o f 19 4 7 , N a n cy and Nelson Schaenen, Jr., Class o f 19 50 , and through the M argaret and Charles Treman, Class o f 19 30 , Fund, supplemented by M argaret and Frank Robinson, 2003.086

watermark: Partial Basilisk, Hinterding subvariant A ’.a.a. (for reference: E H 2006, II, 65, illus. Ill, 1 1 1 )

(For further discussion and watermark illustration, see pages 36-38.)

The Flight into Egypt: Altered from Segers, ca. 1 6 5 2CAT. NO. 35


Bartsch/W hite and Boon 56 vi(?)/vii; N H D 2 7 1 v-a/vi

M em bership Purchase Fund, 8 4.0 10

watermark: Counterm ark IH S, closest to Hinterdingsubvariant A.b. (for reference: E H 2006, II, 15 4 ;illus., Ill, 300)

new proposed tw in m a r k : Counterm ark IH S A.b.a.


2. Allen Memorial Art Museum,Oberlin CollegeCompiled by Katrina Ferreira, Cornell '20; Phoebe Ross, Cornell ’17; with Andaleeb Badiee Banta and Andrew C. Weislogel.

Man in Cloak and Fur Cap, Leaning againsta Bank, ca. 1630Etching

Bartsch/W hite and Boon 1 5 1 i/iii; N H D 48 i/iii

Special Gifts Fund, 19 6 6 .2 1

w atermark : Partial foolscap, variant indeterminate

The White Negress, ca. 16 3 0

Etching

Bartsch/W hite and Boon 3 5 7 ii/ii; N H D 56 iii/iii

Gift o f the fam ily o f Ralph F. H irschm ann (O C 19 4 3 and D. Sc. 19 6 9 ), R C 2 0 17 .4

w atermark : Partial Arches, not in Hinterding


The Angel Appearing to the Shepherds, 1 6 3 4


Bartsch/W hite and Boon 4 4 iii/iii; N H D 12 5 iii/vi

R. T. M iller Jr. Fund, 19 5 7 .2 2

watermark: Basel Crosier, Hinterding subvariant A.a.b. (for reference: E H 2006, II, 6 1 ; illus. Ill, 10 2)

Christ Driving the Money Changersfrom the Temple, 1 6 3 5

Etching with touches o f drypoint

Barsch/W hite and Boon 69 i/ii; N H D 13 9 , i/iv

Gift o f the M ax Kade Foundation, 19 6 8 .12 8

watermark: Foolscap with Five-Pointed Collar,Hinterding subvariant K.c.b. (for reference: E H 2006,II, 1 3 1 ; illus. Ill, 228)

Jan Uytenbogaert, Preacher o f the Remonstrants, 1 6 3 5

Etching

Bartsch/W hite and Boon 279 iv/vi; N H D 1 5 3 iv/ix


w atermark : Partial Double-headed Eagle, Hinterding subvariant C.a.a., ca. 1 6 3 7 (for reference: E H 2006, II, 66; illus. Ill, 17 9 )

Young Man in a Velvet Cap (Petrus Sylvius?), 1637CAT. NO. 6

Etching

Bartsch/W hite and Boon 268 ii/ii; N H D 16 4 ii/ii

Gift o f the M ax Kade Foundation, 1 9 6 8 .1 3 1 A

w atermark : Fragm ent, three roundels together in pyram idal form , possibly from a foolscap

Abraham and Isaac, 16 4 5

Etching

Bartsch/W hite and Boon 3 4 only state; N H D 2 2 4 i/ii

G ift o f M rs. John A . Hadden, 19 6 8 .2 3

w atermark: Partial Strasbourg Lily, Hinterding subvariant B.b.a. (for reference: E H 2006, II, 200; illus. Ill, 4 19 )

Ephraim Bonus, Jewish Physician, 1 6 4 7CAT. NO. 9


Bartsch/W hite and Boon 2 7 8 ii/ii; N H D 2 3 7 ii/ii


w atermark : Basilisk, Hinterding subvariant A ’.a.a.(for reference: E H 2006, II, 65; illus. Ill, 10 1)

(For further discussion see pages 36-38.)

Saint Jerome beside a Pollard Willow, 16 4 8CAT. NO. 30

Etching with drypoint

Bartsch/W hite and Boon 10 3 ii/ii; N H D 24 4 iv/iv


watermark: Foolscap with Five-Pointed Collar,Hinterding subvariant K.a.b. (for reference: E H 2006,II, 12 8 ; illus. Ill, 224)

Landscape with Three Gabled Cottages beside a Road, 16 5 0CAT. NO. 56


Bartsch/W hite and Boon 2 1 7 iii/iii; N H D 248 iii/iii

G ift o f the M ax Kade Foundation, 19 6 7 .4 4

w atermark : Strasbourg Lily, Hinterding subvariant E ’.a.a. or E ’.a.b. (for reference: E H 2006, II, 2 1 1 , 2 1 3 ; illus. Ill, 454 , 4 5 5 )

Presentation in the Temple in the Dark Manner, ca. 1 6 5 4


Bartsch/W hite and Boon 50 only state; N H D 295 only state

G ift o f the M ax Kade Foundation, 19 6 7.4 6

w atermark : Counterm ark PB, Hinterding subvariant A.a.b. or PB A.b. (for reference: E H 2006, II, 87; illus.HI, 14 2 )

Christ Preaching ( ‘La Petite Tombe’), ca. 1 6 5 7CAT. NO. 50


Bartsch/W hite and Boon 6 7 only state; N H D 298 i/ii

Gift o f M rs. John A . Hadden, 19 6 8 .2 1

w atermark : Strasbourg Lily, Hinterding variant A 4 W R or B (subvariant not in Hinterding)

(fig- 1)

FIG. 1


3. Syracuse University Art GalleriesCompiled by Elizabeth Martinson, Cornell ’17; Brittany Rubin; and Andrew C. Weislogel; with DomenicJ. lacono.

Three Heads o f Women, One Asleep, 1 6 3 7

Etching

Bartsch/W hite and Boon 36 8 i/i; N H D 1 6 1 i/iii

Gift o f Mr. C loud Wampler, 19 6 3.9 3 6

w atermark : Counterm ark N B , Hinterding subvariant A.a.(for reference: E H 2006, II, 85; illus. Ill, 140)

n o te : This print has not been previously docum entedwith this watermark. The m ark appears on only one otherknown print: N ude M an Seated on the G round with OneLeg Extended, 16 4 6 (N H D 2 3 4 ii/ii), in the collectiono f Teylers M useum , Haarlem (K G .3 7 6 1). The date o fthat print suggests 16 4 6 as the earliest print date for thisimpression.

Landscape with Cottage and a Large Tree, 1 6 4 1

Etching

Bartsch/W hite and Boon 226 i/i; N H D 19 8 only state

Gift o f Mr. Cloud Wampler, 6 3 .9 23

watermark: Strasbourg Lily, Hinterding variant D.b.(for reference: E H 2006, II, 208; illus. Ill, 444)

n o t e : This waterm ark also appears on an impression o f Landscape with Cottages and a Hay Barn: Oblong (N H D 19 9 only), in the collection o f the National G allery o f A rt, W ashington, D C (19 4 3 .3 .7 10 8 ).

(For more information on this print, see entry for cat. no. 54.)

Self-Portrait Etching at a Window, 16 4 8


Bartsch/W hite and Boon 2 2 iv/v; N H D 240 iv/ix

G ift o f Mr. Cloud Wampler, 19 6 3 .1 0 5 3

watermark: M ost likely Strasbourg Lily, Hinterding subvariant G.c.a. or G.c.b. (for reference: E H 2006, II,2 1 8 ; illus., Ill, 4 6 3 -6 4 ) . These are catalogued as twin watermarks, and m ay even be the same waterm ark due to their similarity.

n o te : D ue to in-plate dates o f other prints appearing on this paper, this impression is likely datable to 1 6 5 1 or after.

A Blind Hurdy-Gurdy Player and Family Receiving Alms, 16 4 8


Bartsch/W hite and Boon 17 6 iii/iii; N H D 24 3 iii/v

Gift o f Harold M . (’ 1 7 ) and Alice Jalonack, 6 5 .1 1 5

w atermark : Partial, three roundels arranged in a triangular formation, possibly from a Foolscap with Seven-Pointed Collar or Single-headed Eagle

A Scholar in his Study ( ‘Faust’), ca. 1 6 5 2


Bartsch/W hite and Boon 270 ; N H D 2 70 iv(?)/vii

G ift o f Mr. Cloud Wampler, 19 6 3 .10 4 5

watermark: Partial, with a distinct “4W .” M ost likely a Strasbourg Lily, possibly Hinterding subvariants 4 W R A.b.a, 4 W R A.h., or 4 W R B.i. (for reference: E H 2006, II, 19 7 , 19 9 , 202; illus. Ill, 4 0 3 , 4 1 3 , 4 2 7), or a Strasbourg Bend, possibly Hinterding variant F.c. (E H 2006, II, 18 8 ; illus. Ill, 3 8 1) .

Adoration of the Shepherds: A Night Piece, ca. 1 6 5 7


Bartsch/W hite and Boon 46; N H D 300 viii/xi

G ift o f M r. C loud Wampler, 19 6 3 .9 3 2

watermark: Strasbourg Lily with initials 4 W R , Hinterding subvariant A.b.a. (for reference: E H 2006, II, 19 7 ; illus. Ill, 403)

n o t e : This print has not been previously docum ented with this watermark. The same waterm ark appears on an impression o f The Flight into Egypt: Altered fro m Segers in the Boston M useum o f Fine A rts (H D P.444, N H D 2 7 1 v/vi). Although Rembrandt’s alteration o f Segers’s plate dates from circa 16 5 3 , the fourth N H D state o f the plate must date after circa 1 6 5 7 as corroborated by the date o f Adoration o f the Shepherds: A Night Piece.

4. Yale University Art GalleryCompiled by Theresa Fairbanks-Harris; Virginia Girard,Cornell ’17; Sarah Schlick, Yale ’18; and Andrew C. Weislogel; with Suzanne Boorsch and C. R. Johnson, Jr.

Selected previously published waterm arks are included where new inform ation has arisen or a waterm ark image is now available.

The Rat Catcher, 1 6 3 2

Etching

Bartsch/W hite and Boon 1 2 1 iii/iii; N H D 1 1 1 iii/iii

Gift o f M rs. Vincenzo Ardenghi, 19 3 1 .4 9

w atermark : Basilisk, not found in Hinterding (for reference: E H 2006, II, 7 1 ; illus., Ill, 1 1 9 )

new proposed tw in m a r k : BasiliskE.a.b.

n o t e : The waterm ark on this print is a Basilisk E variant, w hich is different from all other Basilisk marks because the chain lines run horizontally across the design instead o f vertically. The loops o f the tail do not match the variantE.a. in width or curve pattern.

The Raising o f Lazarus: The Larger Plate, ca. 1 6 3 2

Etching and engraving

Bartsch/W hite and Boon 7 3 viii/x; N H D 1 1 3 v/ix

Fritz Achelis M em orial Collection, G ift o f Frederic George Achelis, B .A . 19 0 7 ,1 9 2 5 .9 0

w atermark : Basilisk, Hinterding subvariant B.c.(for reference: E H 2006, II, 69; illus III, 1 1 5 )


The Good Samaritan, 1 6 3 3


Bartsch/W hite and Boon 90 i/iv; N H D 1 1 6 i/iv

Fritz Achelis M em orial Collection, Gift o f Frederic George Achelis, B .A . 1 9 0 7 ,1 9 2 5 .9 1

watermark: A rm s o f Burgundy and Austria,Flinterding subvariant A.a.a. (for reference: E H 2006,II, 4 2; illus., I ll, 63)

n o te : The presence o f this waterm ark on a number o f impressions o f the first state o f this print in the same state indicates an edition printed together, with a likely date o f circa 16 3 2 , before Rembrandt’s added signature and date o f 1 6 3 3 in the fourth state. Ten other impressions are docum ented on this paper and its twinm ark, A.a.b.

The Good Samaritan, 1 6 3 3


Bartsch/W hite and Boon 90 iv/iv; N H D 1 1 6 iv/iv

Gift o f Lydia Evans Tunnard, 19 8 0 .4 3 .4 7

watermark: Star, Hinterding subvariant A.a.(published E H 2006, II, 1 8 1 , not illus.)

(fig- 2)

Self-Portrait with Raised Saber, 1 6 3 4

Etching, with touches o f burin

Bartsch/W hite and Boon 18 ii/ii; N H D 134/ii/ii)

Fritz Achelis M em orial Collection, Gift o f Frederic George Achelis, B .A. 19 0 7 ,1 9 2 5 .9 3

w atermark : Basel Crosier (crosier only, without surrounding shield and crown), not in Hinterding, cf. H eawood 1 1 8 4

new proposed subvariant : Basel Crosier H.a.

(fig- 3)

Self-Portrait (?) with Plumed Cap and Lowered Saber, 1634 Etching and engraving

Bartsch/W hite and Boon 2 3 iii/iii; N H D 1 3 5 iii/iii

G ift o f M rs. Vincenzo Ardenghi, 19 3 1 .0 4 8

w atermark : A rm s o f W iirttem berg, not in Hinterding (for reference: E H 2006, II, 5 2 - 5 3 ; illus. cf., Ill, 82)

new proposed subvariant: A rm s o f W iirttem berg A.b.

no te : This waterm ark appears to be a new subvariant o f the A rm s o f W iirttem berg, variant A . In addition to some variations in the quartered arms w ithin the shield, the “4B ” pendant below the shield is reversed with respect to the coat o f arms. The waterm ark for this print was previously catalogued under “ Unrecognizable Z Z .z z ”(E H 2006, II, 2 4 1 , there catalogued as 19 3 1.4 6 ) .


Jan Uytenbogaert, Preacher o f the Remonstrants, 1 6 3 5

Etching

Bartsch/W hite and Boon 2 7 9 vi/vi; N H D 1 5 3 viii/ix

Gift o f M rs. H ow ard C. M orse, 19 5 7 .4 5 .3 5

watermark: (Partial) Foolscap with Seven-Pointed Collar, Hinterding subvariant L.a. (for reference: E H 2006, II, 1 4 5 - 4 6 ; illus. Ill, 276)

n o t e : This print in this state is not catalogued with this waterm ark in Hinterding or in N H D . This paper also features the counterm ark PvL.a. This waterm ark does appear on state v.

(fig- 4 )

FIG. 2 FIG. 3


FIG. 4

Old Man with a Divided Fur Cap, 16 4 0


Bartsch/W hite and Boon 26 5 i/ii; N H D 1 8 2 i/ii

Fritz Achelis M em orial Collection, Gift o f Frederic George Achelis, B .A . 1 9 0 7 ,1 9 2 5 . 1 0 9

w atermark.- Partial Foolscap, Hinterding subvariant Z Z .z z (published E H 2006, II, 3 1 2 , not illus.)

(fig- 5)

Cottages and Farm Buildings with aMan Sketching, ca. 1 6 4 1

Etching

Bartsch/W hite and Boon 2 1 9 only state;N H D 2 0 1 only state

Fritz Achelis M em orial Collection, Gift o f Frederic George Achelis, B A 19 0 7, 1 9 2 5 . 1 2 1

watermark: Counterm ark LB, Hinterding subvariant LB.a.b. (for reference: E H 2006, II, 83; illus. Ill, 13 6 )

no te : This is the counterm ark to Foolscap with Five- Pointed Collar, Hinterding subvariant K.a., the paper datable to circa 16 50 .

The Flute Player ( ‘Flet Uilespiegeltje’), 16 4 2


Bartsch/W hite and Boon 18 8 iv/iv; N H D 2 1 1 iv/iv

The W alter R. Callender, B A 18 9 4 , M em orial Collection, G ift o f Ivy Lee Callender, 19 6 2 .4 5 .4 18

watermark: Partial foolscap, three roundels, two on top, one on bottom (for reference: E H 2006, II, 12 6 ; illus.,Ill, 2 1 9 -2 0 )

n o t e : Other impressions o f this print show Foolscap with Five-Pointed Collar, Hinterding subvariant I.zz., partial (just the roundels), but sim ilarly two on top, one on bottom. This is likely an I variant (see E H 2006, II, 292 cf. Rijksprentenkabinet impression). The likely printing date is ca. 16 48.

The Three Trees, 16 4 3


Bartsch/W hite and Boon 2 1 2 only; N H D 2 1 4 only

Fritz Achelis M em orial Collection, Gift o f Frederic George Achelis, B A 1 9 0 7 , 1 9 2 5 . 1 1 9

watermark: Strasbourg Lily, Hinterding subvariant E ’.a.b. (for reference: E H 2006, II, 2 1 2 , illus. Ill, 4 5 5 )

no te : D ue to in-plate dates o f other prints appearing on this paper, this impression is likely datable to circa 16 5 2 .

Abraham and Isaac, 16 4 5

Etching

Bartsch/W hite and Boon 3 4 i/ii; N H D 2 3 4 ii/ii

Fritz Achelis M em orial Collection, Gift o f Frederic George Achelis, B .A . 1 9 0 7 , 1 9 2 5 . 1 2 2

watermark: M iscellaneous, Late, not in Hinterding, last letter o f the waterm ark T or L (?)

(fig- 6)

FIG. 5


FIG. 6

FIG. 7

Jan Asselijn, Painter ( ‘Krabbetje’), ca. 1 6 4 7


Bartsch/W hite and Boon 2 7 7 iii/iii; N H D 2 36 iv/vii

Fritz Achelis M em orial Collection, Gift o f Frederic George Achelis, B A 1 9 0 7 ,1 9 2 5 .1 2 5

watermark: A rm s, unidentified, similar to Hinterding subvariant B ’.a.b. (published E H 2006,II, 56; for reference: illus. Ill, 89)

( % • 7 )

Saint Jerome beside a Pollard Willow, 16 4 8

Bartsch/W hite and Boon 10 3 ii/ii; N H D 244 iv/iv

G ift o f Mr. and M rs. Christopher Tunnard, 19 7 3 .1 5 6 .6

watermark: Partial Foolscap with Seven-Pointed Collar, possibly Hinterding subvariant C.a.a. (for reference:E H 2006, II, 14 0 ; illus., Ill, 2 5 7 )

Panorama near Bloemendael Showing the Saxenburg Estate (‘The Goldweigher’s Field’), 1 6 5 1


Bartsch/W hite and Boon 2 3 4 only state;N H D 2 5 7 only state

Gift o f Edw ard B. Greene, 1900, 1 9 3 1 . 1 0

watermark: Paschal Lamb, Hinterding subvariant A.c. (E H 2006, II, 16 4 ; illus. cf. Ill, 324 )

n o t e : This waterm ark is likely to be an unidentified twinm ark o f the Paschal Lamb, variant A .c. Hinterding notes that it is likely the entire variant A consists o f various twinm arks in different stages o f deformation, and that one batch o f paper m ay encompass the whole group o f watermarks.


The Circumcision in the Stable, 1 6 5 4

Etching

Bartsch/W hite and Boon 4 7 i/ii; N H D 280 iii/v

Everett V. M eeks, B A 19 0 1 , Fund, 20 0 7.16 .2

watermark: Counterm ark G I, Hinterding subvariant Gl.a.b. (for reference: E H 2006, II, 76; illus., I ll, 12 6 )

n o te : This counterm arks letter I is not exactly parallel with the adjacent chain line; rather it bends away from it; this m atching detail confirm s the subvariant.

Christ Presented to the People: Oblong Plate, 16 5 5

CAT. NO. 46

Drypoint

Bartsch/W hite and Boon 76 viii/viii; N H D 290 viii/viii

Fritz Achelis M em orial Collection, Gift o f Frederic George Achelis, B A 1 9 0 7 , 1 9 2 5 . 1 3 7

watermark: Grapes, Hinterding subvariant A.zz. (published EH 2006, II, 14 8)

(fig- 8)

Christ Preaching ( ‘La Petite Tombe’), ca. 1 6 5 7


Bartsch/W hite and Boon 6 7 only state; N H D 298 i/ii

Fritz Achelis M em orial Collection, G ift o f Frederic George Achelis, B A 1 9 0 7 ,1 9 2 5 . 1 3 4

watermark: Foolscap with Five-Pointed Collar, Hinterding subvariant Q ’.a. (published E H 2006, II,1 3 5 , illus. Ill, 2 4 1)

no te : Closer inspection confirms the Yale waterm ark as a subvariant Q ’.a.b. (see cat. no. 49, fig. 3 1 )

(fig- 9 )

FIG. 9


A P P E N D IX 3

Further sources for Rembrandt watermark study

Books and articlesC . M . Briquet, Les filigranes. D ictionnaire historique des m arques du p a p ier des leur apparition vers 12 8 2 ju sq u e n 1600, 4 vols., Paris, 19 0 7 .

W illiam A lgern o n C hurchill, W atermarks in P aper in H olland, England, France, etc. in the X V I I a n d X V III Centuries a n d their Interconnection, A m sterdam , 1 9 3 5 .

Joh n B archam G reen , P aperm aking by H and, M aidstone, 19 6 7 .

E rik H interding, “W aterm ark Research as a Tool for

the Study o f R em bran dts Etchings,” in R em bran dt the P rintm aker (A m sterd am -Lon do n 2 0 0 0 -0 1) , 2 3 - 3 5 .

C . R ichard Johnson, Jr., W illiam A . Sethares, M argaret

H olben Ellis, and Saira H aqqi, “H unting for Paper M oldm ates A m o n g R em bran dts Prints,” IE E E Signal Processing M agazine (Special Issue on Signal Processing f o r A rt Investigation) 3 2 (Ju ly 2 0 1 5 ) , 2 8 - 3 7 .

D an Kushel, “R adiographic M ethods U sed in the

R ecordin g o f Structure and W aterm arks in H istoric Papers,” in Fresh Woods an d Pastures N ew (C h ap el H ill—

Ithaca 19 9 9 ), 9 - 5 3 .

E m ile-Josep h Labarre, A D ictionary o f P aper a n d Paper- M aking Terms with E quivalents in French, G erm an,D utch a n d Italian, A m sterd am , 19 3 7 .

Em ile-Josep h Labarre, “ The Sizes o f Paper, their N am es, O rigin & H isto ry” in Buch u n d P ap ier buchkundliche un d papiergeschichtliche A rbeiten ; H. Bockw itz zum 65. Geburtstage dargebracht, Leipzig, 19 4 9 , 3 5 - 5 4 .

Johann Lindt, The P aper-M ills o f B ern e a n d their W atermarks 14 6 5 - 18 5 9 , H ilversum , N etherlands, 19 6 4 .

N ad in e O renstein et al., “Print Publishers in the N etherlands 1 5 8 0 - 1 6 2 0 ,” in The D aw n o f the G olden Age: N orthern N etherlandish A rt 15 8 0 - 16 2 0 , edited by G e r Luijten, et al., N e w H aven, 1 9 9 4 ,1 6 7 - 2 0 0 .

M ark van Staalduinen, “C ontent-based Paper Retrieval

Tow ards Reconstruction o f A r t History,” P h D thesis, D elft U niversity o f Technology, 2 0 10 .

M ark van Staalduinen, Jan C . A . van der Lubbe,E ric Backer, and Pavel Paclik, “ Paper Retrieval Based on Specific Paper Features: C h ain and Laid Lines,” in M ultim edia Content Representation, Classification a n d Security, Lecture Notes in C om puter Science 4 1 0 5 , 2006,

346 - 53 -

A llan H . Stevenson, “ N e w Uses o f W aterm arks as

Bibliographical Evidence,” in Studies in B ibliography 1 , 1 9 4 8 - 4 9 , 1 4 9 - 8 2 .

A llan H . Stevenson, “ Chain-Inden tation s in Paper as Evidence,” in Studies in B ibliography 6, 1 9 5 3 - 5 4 , 1 8 1 - 9 5 .

A d Stijnm an, E n gravin g an d Etching 14 0 0 -20 0 0 :A H istory o f the D evelopm ent o f M an u al Intaglio Printm aking Processes, London , 2 0 1 2 .

W alter Friedrich Tschudin and Em ile-Josep h Labarre,

The A n cien t Paper-M ills o f Basle a n d their M arks, H ilversum , N etherlands, 19 5 8 .

D avid L. V ander M eulen, “ The Identification o f Paper w ithout W aterm arks: The Exam ple o f Popes ‘D unciad,’ ”

in Studies in B ibliography 3 7 , 1 9 8 4 , 5 8 - 8 1 .

H en k V oorn , “Som e C urious E arly Experim ents in D utch Paperm aking,” in The P a p er M a ker 3 2 , no. 1, 1 9 6 3 , 1 1 - 1 9 .

Selected online resourcesA n g lia T V , “Paperm akin g b y hand at H ayle M ill,

En glan d in 19 76 ,” video, 1 5 : 1 8 , http://w w w .youtube .co m /w atch ?v =X s3 Pfw OItto.

“Bernstein: The M em o ry o f Paper,” last m odified

June 26, 2 0 1 7 , http://w w w .m em oryofpaper.eu /BernsteinPortal/appl_start.disp.

D utch U niversity Institute for A r t H istory, Florence, “ W aterm arks D atabase o f the D utch U niversity Institute for A r t H isto ry” last m odified Jan u ary 24 , 2 0 1 4 , http://w w w .w m -portal.net/niki/index.php.

U niversity o f Iow a C en ter for the Book, “C h an ce ry

Paperm aking,” video, 10 :00, 2 0 1 3 , http://now.m ow a .ed u/20i3/o5cen ter-b ook -releases-sh ort-film -ab out -paperm aking.


http://www.youtube

http://www.memoryofpaper.eu

http://www.wm-portal.net/niki/index.php

http://now.mowa

Decision Tree Branch for (2 Foolscap with Five...

Documents

Transcript of Decision Tree Branch for (2 Foolscap with Five...