In Search of the Origin of the Flood
Investigating a Painting with an Uncertain Attribution and Date
in the Collection
of the Rijkmuseum in Amsterdam
I L S E S T E E M A N
Student number - 10004518
Master Conservation and Restoration programme Technical Art History Faculty of Humanities
University of Amsterdam, Amsterdam
Thesis supervisors Dr. Machtelt Brüggen Israëls
Dr. Arjan de Koomen
Fig. 1. Catalogued as by an anonymous artist. The Flood. Ca. 1450-1500. Oil on Sundeala fiberboard, transferred from panel. 121.2 × 98.4 × 1.3 cm.
Table of contents
Acknowledgements 5 Abstract 6 Samenvatting 7 Introduction 8 Motivation 9State of research: attribution, date, provenance and function 9
State of research: previous technical examination 11
Aim, research question and methodology 12
1. Technical notes 16
1.1 Support 16
1.2 Ground layer 26
1.3 Underdrawing 36
1.4 Paint layers 47
2. Clothing and headdress 61
3. Iconography 73 4. Style 88 Conclusion 102 Further research 104 Bibliography 105 Archivalia 109 Photographic credits 110 Appendices 119
Appendix I - Supplementary technical information 119
Appendix II - Conservation report 1979-1980 120
Appendix IV - Braccia 142
Appendix V - XRF measurements 154
Appendix VI - Samples 169
Appendix VII - List of paintings associated with the
Acknowledgements
This thesis has been made possible with the help and support of many. Firstly, I would like to sincerely thank my supervisor Dr. Machtelt Brüggen Israëls of the Art History Department of the University of Amsterdam for her time, advice and her generous feedback on the drafts of the individual chapters, but most of all, for her trust in me and in my thesis. I would like to thank Dr. Arjan de Koomen of the Conservation and Restoration and Art History Department of the University of Amsterdam for taking the time to be its second reader.
My internship with the Paintings Conservation and Restoration Department at the
Rijksmuseum in Amsterdam was a great chance to examine the picture of the Flood that is the object of this thesis along with others in the collection of Italian painting. I especially wish to thank Dr. Duncan Bull, Dr. Petria Noble and Giulia de Vivo for granting me the internship, and supporting me throughout my research. I am particularly indebted to Giulia de Vivo for teaching me to do technical examinations, as well as for her suggestions, enthusiasm and patience. I thank Arie Pappot for his help in carrying out and interpreting XRF measurements, dr. Esther van Duijn for discussing the conservation history of the Flood and Dr. Sara van Dijk for discussing the clothing of the figures. I also wish to thank Prof. Dr. Arie Wallert for his stimulating interest in my research. I am indebted to Dr. Ineke Joosten of the Dutch Cultural Heritage Institute (RCE) who has performed the SEM-EDX analyses and has helped me to interpret the results. I would further like to thank Prof. Dr. Maarten van Bommel, Prof. Dr. Ella Hendriks, Dr. René Peschar, Dr. Maartje Stols-Witlox and Dr. Abbie Vandivere of the Conservation and Restoration Department of the University of Amsterdam for their advice and suggestions.
Last but certainly not least, I want to express my gratitude towards my family, friends and
Abstract
Central to this master thesis is a painting of uncertain attribution and date representing the Flood in the collection of the Rijksmuseum in Amsterdam. It depicts the Old Testament story of Noah in his ark who is watching how the sinful world is drowned by the Flood sent by God. In the past art historians, art dealers and collectors considered this striking picture a work by the Ferrarese artist Ercole de’ Roberti (ca. 1451–1496) or someone who worked close to him. However, more recently it has been argued that it originated in the circle of Hans Baldung Grien (1484/1485-1545), a painter who was active in the Danube region in Germany.
This research explores the Flood from different angles in order to get a better
understanding of its origin, including geographical provenance, dating, original context and possible authorship. Results of previous technical investigations are combined with new technical examination of the physical aspects of the painting, including its materials, techniques and
condition. In addition to these aspects, the clothing of the figures, the painting style as well as the iconography of the painting have been examined to gain a closer understanding of its origins.
The research yielded several indications that the Flood was most likely painted in northern
Italy in the late fifteenth- and early sixteenth centuries. The painting was presumably made in the context of predictions that a second Flood would occur in 1524. The predictions were
announced for the first time in 1499. The Flood can be attributed to the Paduan artist known as the Master of the San Nicolò Triptych, a painter who was active between 1490 and 1520. The Flood could be dated around 1499 and 1520.
Samenvatting
In deze master scriptie heb ik de Zondvloed, een schilderij met onzekere toeschrijving in de
collectie van het Rijksmuseum in Amsterdam onderzocht. Het verbeeldt het Oud-Testamentische verhaal waarbij Noach vanuit zijn ark ziet hoe de Zondvloed, gezonden door God, de zondige mensheid ten onder laat gaan. Het werk riep twijfels op wat betreft zowel de toeschrijving als de datering ervan. Kunsthistorici, kunsthandelaren en verzamelaars beschouwden dit opvallende schilderij als een werk van de Ferrarese schilder Ercole de’ Roberti (ca. 1451–1496) of iemand die in de buurt van hem werkzaam was. De meest recent aangedragen suggestie is dat het schilderij in de kring van Hans Baldung Grien (1484/1485-1545) ontstaan is, een schilder die werkzaam was in het gebied van de Donau in Duitsland.
Dit onderzoek belicht de Zondvloed vanuit verschillende invalshoeken om een beter begrip
te krijgen van de origine van het schilderij, waaronder de geografische herkomst, datering, originele context en mogelijke schilder. Vondsten uit eerder technisch onderzoek zijn
heroverwogen en gecombineerd met nieuw onderzoek naar de fysieke aspecten van het schilderij, waaronder de gebruikte materialen, technieken en conditie. Naast deze aspecten is ook de
kledingstijl van de figuren, de schilderstijl en de iconografie van het schilderij nader bestudeerd.
Uit dit onderzoek komen verscheidene aanwijzingen naar voren die erop wijzen dat de
Zondvloed geschilderd is in het noorden van Italië aan het einde van de vijftiende, begin zestiende eeuw. Het schilderij is vermoedelijk gemaakt in de context van de voorspellingen dat er een tweede Zondvloed zou komen in het jaar 1524. Deze voorspellingen werden voor het eerst aangekondigd in 1499. De Zondvloed zou toegeschreven kunnen worden aan een Paduaanse schilder met de noodnaam de Meester van het Sint Nicolaas Triptiek, die actief was tusssen 1490 en 1520. Het werk zou rond 1499 en 1520 gedateerd kunnen worden.
Introduction
According to the Old Testament, as told in Genesis, God saw that the earth was filled with violence and corruptness and that the wickedness of mankind was great. This made God repent that He had created humans and He wanted to destroy them. The judgment that God brought upon mankind was a worldwide Flood. There was only one righteous man who found grace in the eyes of God. His name was Noah. God urged him to build an ark as He was about to bring a Flood of waters upon the earth to destroy all flesh in which was the breath of life. God
commanded Noah to build the ark out of gopher wood and told him to establish a covenant with him. He was directed to enter the ark with his wife, sons and daughters in law and to bring every
sort of living being into the ark, a female and male of each.1 After seven days, after Noah did as
he had commanded, the waters of the Flood were upon the earth. The fountains from the deep broke up and the windows of heaven were opened. God let it rain for forty days and nights. The
waters increased and lifted the ark from the ground.2
This biblical story resembles one of the paintings in the Italian paintings collection of the
Rijksmuseum in Amsterdam (fig. 1). The painting depicts the Flood with in the middle the ark from which Noah is watching the rising water and the chaotic mass of sinners. Some of the damned are already drowned, while despairing sinners are still trying to survive by either trying to climb the ark, cling to a raft or flee into the mountains. The scene is set against darkening sky above a mountainous landscape with a city.
This supposedly Italian picture is known as the Flood and was bequeathed to the
Rijksmuseum in the second year of the Second World War, together with a collection of other Italian paintings and sculptures, by the Dutch art collector J.W. Edwin vom Rath (1863-1940). Relatively little is known about the Flood. Although the attribution of the panel has generally been given to (the circle of) the Ferrarese Ercole de’ Roberti (ca. 1403–1482), painted somewhere between 1450-1500, both the attribution and date have been questioned more recently. Henk van Os (1994) suggested that the Flood would have been painted in the Danube region in the first half of the sixteenth century. Currently, the painting has been catalogued in the Rijksmuseum as a work by an anonymous artist. There are no clues regarding its commission or original placement and context. Its history before the 1920s seems to be undocumented. Also, it is unknown if this object was originally independent or whether it is a fragment of a larger original ensemble, such as a piece of furniture or an altarpiece.
1 The Holy Bible, Genesis 6:5-22. 2 The Holy Bible, Genesis 7:1-18.
Motivation
Currently, the Rijksmuseum’s Italian painting collection, including 215 pre-nineteenth-century works, is technically examined for the file catalogue of Italian paintings that is now in
preparation. Following my internship at Rijksmuseum’s Paintings Conservation and Restoration Department between February and June 2017, I selected the Flood as the subject of my master thesis. The Flood aroused my interests because of its crowded, colourful, and elaborate
composition; not even the smallest details were omitted. Most of all, I considered the debate on its attribution and date as an interesting and challenging call for further research.
State of research: attribution, date, provenance and function
The painting’s provenance extends back no further than the early 1920s. According to Ernst Buchner and Karl Feuchtmayr (1924) the work was owned by the art dealer Julius Böhler in Munich in 1923. At that time, Böhler considered the Flood to be a work by an artist from the
Ferrarese school.3 The Flood was put forward in relation to another painting of the same subject
in the Staatsgalerie in der Neuen Residenz, Bamberg. The work is dated 1516 and painted by the German Hans Baldung Grien (1484/1485-1545) (fig. 2). The authors were very much convinced that Baldung Grien derived elements in his painting from Italian prototypes and furthermore that
the Ferrarese Flood may have served as its source transferred either by a drawing or a print.4 At
some point in or after 1923, Dr. Glitza in Hamburg bought the painting from Böhler. The Flood
then passed through Paul Cassirer, who sold it to Edwin vom Rath in 1926.5
After the Flood entered the Rijksmuseum the painting was first mentioned in the
museum’s annual report of 1941. It is discussed as one of the most important objects that was part of the Vom Rath request, and it is stated that it derives from the Ferrarese school, painted
either by De’ Roberti himself or someone who worked close to him.6 It is furthermore described
as a peculiar Italian painting as it shows so many details without a striking main motif. In the exhibition catalogue of 1946 on the Rijksmuseum’s new acquisitions between 1940 and 1946, the
3 Ernst Buchner and Karl Feuchtmayr. Beiträge zur Geschichte der deutschen Kunst: Vol. 1: Oberdeutsche Kunst der Spätgotik
und Reformationzeit. Augsburg: Filser, 1924: p. 296, 294.
4 Buchner and Feuchtmayr 1924: p. 296. 5 Rijksmuseum files. SK-A-3418. Inventory card.
6 Verslagen Omtrent ‘s Rijks Verzamelingen van Geschiedenis en Kunst 1941. Deel LXIV. Den Haag: Algemn.
picture was still attributed to (the circle of) De’ Roberti.7 Enzo Carli (1958), also ascribed the
painting to the Ferrarese school as he indicated that the castle in the background is reminiscent of the Castello Estense in Ferrara. In 1963 Francesco Valcanover rejected that it could have been a
work by De’ Roberti.8 He even questioned whether the Flood could be considered as an Italian
painting or not. However, the attribution remained unchanged until the mid-1990s as the painting
was still designated as Ferrarese by Hanna Hohl (1967) and by Gert von der Osten (1981-82).9
Yet, in her brief visual analysis of the Flood, Hohl remarked that the stiffness of the figures reminded her of the Battle of Alexander (1529) in the Alte Pinakothek, Munich which is by the
hand of the Albrecht Altdorfer (fig. 3).10
Henk van Os devoted an article to the Flood in 1994.11 In his opinion, the artist of the
painting has a Northern touch. The pile of deliberate figures in such a crowded composition, which Van Os refers to as the indirect manner, seems to him characteristic of works by the
Danube school.12 Van Os repeats the suggestions by Buckner, Feuchtmayr and Von der Osten
that the Flood may have served as the source for Baldung Grien’s painting, but raises doubts about the way in which Baldung Grien could have come in contact with works from the circle of De’ Roberti. The only explanation, according to Van Os, is that the Rijksmuseum’s painting was not made in Ferrara but in the area where Baldung Grien was also active and thus that it must
have been painted in the circle of Altdorfer in the Danube region.13 Though the Flood depicts
compositional elements that are Italian. Van Os mentions the woman on the bull, who is similar to an Italian fifteenth-century depiction of Europa astride on a disguised Zeus and the naked figures who are clinging to the ark are similar to the nudes in Paolo Uccello’s fresco of the Flood
(1447-1448) in the Santa Maria Novella, Florence (fig. 4).14 The composition could also have been
borrowed from Francesco Rosselli’s print of the Flood (ca. 1470-1490) (fig. 5).15 Another reason
to assume that the painting was made in Germany, could be that the Flood was painted in a reaction to predictions by humanists around 1500. Astrologers, such as the German Johan
7 D.C. Roëll. Aanwinsten 1940-1946: Tentoonstelling van Schilderijen, Beeldhouwkunst, Meubelen, Kunstnijverheid, Teekeningen en
Prenten, Verworven in de Jaren 1940-1946. Amsterdam, 1946: p. 19.
8 Rijksmuseum files. SK-A-3418. Inventory card; Henk W. van Os. ‘Un Deluge á Amsterdam.’ In: Rosenberg, Pierre
ed. Hommage à Michel Laclotte: Études sur la Peinture du Moyen Âge et de la Renaissance. Paris: Electa Réunion des Musées Nationaux, 1994: p. 434.
9 Hanna Hohl. Die Darstellung der Sintflut und die Gestaltung des Elementaren. Diss. Eberhard-Karls-Universität, Tübingen.
Tübingen: 1967: p. 39;Gert von der Osten. Hans Baldung Grien: Gemälde und Dokumente. Berlin: Deutscher Verlag für Kunstwissenschaft. 1981-1982: p. 136.
10 Hohl 1967: p. 39.
11 Pieter J.J. van Thiel et al. All the Paintings of the Rijksmuseum in Amsterdam: a Completely Illustrated Catalogue.
Amsterdam: The Hague: Rijksmuseum SDU Uitgeverij Koninginnegracht, 1992: p. 644; Van Os 1994: p. 430-436.
12 Van Os 1994: p. 430, “C’est ainsi que, de manière indirecte…”. 13 Van Os 1994: p. 434.
14 Van Os 1994: p. 430. 15 Van Os 1994: p. 430.
Stöffler (1452-1531), predicted that a new universal Flood would take place, as a consequence of
the conjunction of the planets Jupiter and Saturn.16
State of research: previous technical examination
As described in the Rijksmuseum’s annual report of 1980 the Flood has undergone a severe conservation treatment undertaken by conservators H.Ch. Coen and Teri Oikawa-Picante under
the charge of conservator Luitsen Kuiper in 1979-80 (app. II).17 The splitting and warping of the
panel caused cracks and delamination of the paint layers. After first treating the painting with a straightening method, the conservators decided to remove the original support. The ground and
paint layers were transferred to a fibreboard.18
The conservation report of this treatment records that technical visual examination was
carried out as part of the restoration campaign.19 The report describes the build-up of the
painting, as well as the phases of the conservation treatment with relevant photographs. Major conclusions were as follows. Regarding the old support, consisting of two vertical and one horizontal plank, it was assumed that the work had already been transferred in the past as the wood came off very easily during the removal of the panel. Additionally, the composition would
have been enlarged at some point by the addition of the horizontal plank.20 This assumption was
reinforced by the fact that the painting was done in egg tempera, though heavily overpainted with oil, except for the upper part, starting from just below the green hilly mountains up to the sky, which is done in oil, concluding that the upper part would possibly not have been part of the
original composition.21
16 Van Os 1994: p. 434.
17 Nederlandse Rijksmusea in 1980. Deel CII. Den Haag: Staatsuitgeverij, 1981: p. 18. 18 Idem.
19 Rijksmuseum files. SK-A-3418. Kuiper, Luitsen. Restauratie rapport: School Ferrarese 2e helft 15e eeuw. De
Zondvloed. Paneel. 122 × 98 × 1.6 cm. 1980: p. I-VI.
20 Rijksmuseum files. SK-A-3418. Kuiper 1980: p. II, VI. 21 Rijksmuseum files. SK-A-3418. Kuiper 1980: p. I, II.
Aim, research question and methodology
It is still unclear where the Flood was painted, with diverse suggestions as Italy and Germany. It is also unclear how the painting can be dated and what its original context and function may have been. The aim of the thesis is to reconsider and explore these aspects from different angles. It adds to the current state of knowledge by re-examining technical research that has been carried out in 1979-80 before the paint layers were transferred, combined with new technical research, visual analysis, iconography and historical research. The main research question is formulated as follows:
What are the geographical provenance, date, original context and possible authorship of the Flood in the Rijksmuseum in Amsterdam?
To answer this question, the thesis is divided in four chapters. Chapter one focuses on the technical aspects of the picture. The chapter gives answer to the question: what do the materials and techniques tell about the painting’s origin and date? The first part concentrates on the support. Although the painting has been transferred to a new support in the past, the
documented information of the old support during the transfer in combination with literature research on easel paintings has proved helpful in determining the wood type, woodworking and wood construction. A reconstruction has been made to visualize the original support. The second part examines the ground layer of the painting. The ground was examined using optical
microscopy, whereas the elemental composition was determined by X-ray fluorescence (XRF) and scanning electron microscopy with energy-dispersive X-ray spectroscopy (SEM-EDX). Part three focuses on the preparatory design of the Flood. It concentrates on the preparatory design of the painting, in which the results of infrared photography (IR) and infrared reflectography (IRR) are interpreted. This concerns how the underdrawing was executed and how the composition relates to similar existing imagery. The chapter also shows how IR(R) has proved helpful in determining the original format of the painting. The fourth part is concerned with the actual paint layers. The build-up, execution technique, colour palette, condition have been examined by visual examination, false-colour infrared photography (FC-IR) and X-radiography. Chapter two is concerned with the clothing and headgear worn by the figures in the painting. The question posed is: what does the clothing of the figures indicate about the painting’s origin and date? This
has been analysed by using visual analysis, literary sources and contemporary paintingsfor
painting has been examined using iconographical analysis. The questions being asked are: how does the painting relate to existing imagery of the Flood theme, which symbolic meaning can be suggested for the painting and what could have been the context of the Flood? The answers to these questions also provide insight into its origin and date. The last chapter is concerned with the style of the painting and gives answer to the questions: what is the style of the painting, how does it relate to a painting school or artist and how does that date the Flood? This has been
examined using visual analysis, contemporary paintingsfor comparison and literary sources.
Fig. 2. Hans Baldung Grien. The Flood. 1516. Mixed media on panel (linden). 82 × 65 cm. Staatsgalerie in der Neuen Residenz, Bayerische Staatsgemäldesammlungen, Bamberg
Fig. 3. Albrecht Altdorfer. The Battle of Alexander. 1529. Oil and tempera on wood (linden). 158.4 × 120.3 cm. Alte Pinakothek, Munich.
Fig. 4. Paolo Uccello. The Flood. 1447-1448. Fresco. 215 × 510 cm. Chiostro Verde, Santa Maria Novella, Florence
Fig. 5. Francesco Rosselli. The Flood (first state). Ca. 1470-1490. Engraving. 27 × 40.2 cm. British Museum, London
1. Technical notes
1.1 Support
This subchapter is concerned with the original support of the Flood. The type of wood, the construction of the panel, the dimensions of both panel and constituent planks, and the condition of the support are taken into account to get a better understanding of the place and time of origin of the painting.
The original support of the Flood is only documented in a photograph and conservation report. The original panel suffered from warping and continued to do so after it had been flattened in an intervention in 1979. In 1980 a second treatment followed, in which the wood support was completely removed by planing. The paint and ground layers were subsequently glued with
wax-resin onto a Sundeala fibreboard.22 The new support measures 121.2 by 98.4 by 1.3 cm.23 It was
backed by two linen cloths. At 120 by 96 cm, the painted surface is outmeasured by the new board along all edges (fig. 6).
Before the transfer, the painting was on a cradled hardwood panel measuring circa 122 by
98 cm with a thickness of 1.6 cm (fig. 7).24 It was composed of two vertically grained planks to
which a horizontally grained plank was attached at the top. Along all edges ran two strips of
wood. The left plank (as seen from the back of the painting) measured 102.4 by 46.3 and the right
plank measured 102.4 by 47 cm. According to Kuiper, they appear to have been cut from the same tree in adjacent positions since the right plank has an elongated large knot that is similar to the knot in the left plank (fig. 8). The planks were sawn tangentially and might be an indication that the panel was produced in Italy. The wood of most panel paintings from northern Europe has been cut in the radial direction, whereas the planks of Italian panel paintings were often
tangentially cut.25 The vertical planks were butt-joined and glued together. There were numerous
splits and cracks in the planks that still show in the ground and paint layers.26 The horizontal
22 Rijksmuseum files. SK-A-3418. Kuiper 1980: IV, V.
23 A Sundeala fiberboard is a compressed wood fibre board produced by the English firm Sundeala Ltd.
24 The cradling system consisted of seven vertical fixed members and eight sliding members that were not movable
any more.
25 Ciro Castelli. ‘Techniques of Construction of Wood Supports for Painting.’ In: Marco Ciatti ed. Panel Painting.
Technique and Conservation of Wood Supports. Florence: Edifir, 2006: p. 119.
26 - The splits and crack are positioned as follows: in the left plank there are two splits rising vertically from the
bottom, starting at 20.5 and 35 cm from the left edge of the new support, and rising along the whole height of the original left-hand vertical plank. To the left of these there is another split rising 17.6 cm from the left and curving round to the top of gable of the ark. In the right plank, a split starting 42.6 cm from the right and rising 34.5 cm vertically, another 39 cm from the right edge rising 13.5 cm vertically; and a third 23.5 cm from the right running the full height of the panel with an irregularity towards the bottom.
plank measured 15 by 93.8 cm and was attached with a lip joint and reinforced with three
butterfly keys on the back of the panel (fig. 9-10).27 The mortises of the butterfly keys were about
half of the planks’ thicknesses. Imprints of the three butterflies can still be seen in the paint surface (fig. 25). The butterflies measure 10 cm in length and 2.5 cm in width. The painted surface extended over the inner 1.2-cm-wide strips around all edges of the support, but not over
the outer 1.1 cm-wide ones.28 The outer strips may have been later additions to fit the painting in
a frame. Wood worm damage was visible on the back of the panel. All three planks show open tunnelling, suggesting they had been thinned (fig. 11). It is however hard to tell from the
photograph whether the panel had been thinned at some time. Its back was covered with a layer of shellac.
Two horizontal rows of five nail holes were visible on the back, positioned 12 cm below
the horizontal lip joint and 12 cm above the bottom. It was suggested they may have served to
attach battens.29 Fixed battens were used by both the original carpenters and by conservators to
prevent the panel from warping.30 Apparently, to Kuiper, there were no signs, such as nails or
other connection methods, indicating that the panel had been part of a larger ensemble.
The conservation report is not specific regarding the type of wood of the original
support. Unfortunately, the panel has not been preserved, nor have wood samples been taken. However, a photograph that was taken before the removal of the panel shows the back of the support and gives some clues as to the wood species (fig. 12). Both vertical planks had large conspicuous knots that were innate defects (fig. 8). The pattern of the grain is fairly coarse, and similar to the characteristics of poor-quality poplar. As of the third quarter of the thirteenth century the cheap, yet serviceable, white poplar (Populus Alba) was the most popular choice for
panel painting supports in Italy.31 The horizontal plank does not share the same characteristics,
on the contrary, it was straight grained with a fine, even texture (fig. 11) The wood species of the horizontal plank cannot be determined. The choice for a wood support does not affect the proposed date of the Flood, as although during the sixteenth century cloth supports became more
common, especially for larger formats, wood continued to be used.32 However, since to the north
27 It is unknown how the butterfly keys were attached to the panel. 28 Rijksmuseum files. SK-A-3418. Kuiper 1980: I.
29 Rijksmuseum files. SK-A-3418. Kuiper 1980: II.
30 Paul Ackroyd. ‘The Structural Conservation of Paintings on Wooden Panel Supports.’ In: Joyce Hill Stoner and
Rebecca Rushfield ed. Conservation of Easel Paintings. New York: Routledge, 2012: p. 460.
31 George Bisacca and Ciro Castelli. ‘History and Use of Panels or Other Rigid Supports for Easel Paintings: Italian
Panel Paintings.’ In: Joyce Hill Stoner and Rebecca Rushfield ed. Conservation of Easel Paintings. New York: Routledge, 2012: p. 74.
32 Luca Uzielli. ‘Historical Overview of Panel-Making Techniques in Central Italy.’ In: Kathleen Dardes and Andrea
Rothe ed. The Structural Conservation of Panel Paintings: Proceedings of a Symposium at the J. Paul Getty Museum, 24-28 April 1995. Los Angeles: Getty Conservation Institute, 1998: p. 114, Rutherford Gettens and George Stout. Painting Materials. A Short Encyclopaedia. New York: Dover, 1966: p. 270.
of the Alps poplar was only rarely used, Van Os his suggestion that the Flood would have been
produced in southern Germany in the circle of Albrecht Altdorfer comes into question.33 Linden
and coniferous wood are particularly prevalent in Renaissance paintings from southern
Germany.34 Altdorfer, for example, often painted on linden wood.35 If my identification of gattice
as the wood species used for the Flood its original support is correct, it is highly likely that the wood was obtained in Italy and that the painting was produced there. Italian artists used wood native to their region.36
Assuming that the panel came from Italy, the joints between the planks and the size of
the support can be evaluated in the light of knowledge of Italian panel constructions. It was speculated during the conservation treatment of 1979-80 that the painting had already been transferred in the past, and that the upper horizontal plank, was a later, not original, addition to
the lower vertical planks.37 This assumption was supported firstly by the fact that the upper part
of the painting, starting from just below the green hilly mountains up to the sky, was done in oil and not in tempera as in the lower part of the painting (fig. 13). The second argument was based on the wood of the vertical planks that came off very easily of the ground layer during the removal of the support. It would be an indication that at that time the panel resulted to have been applied against the ground layer, and not, in accordance with the build-up of layers such as the artist applied them, the other way around.
Before the panel was prepared with a ground layer, the lip joint between the horizontal
and vertical panel was covered by a 2.5-cm-wide piece of cloth.38 From the thirteenth to the
fifteenth century, it was often common practice, with few exceptions, to isolate and protect the whole surface of a Italian panel’s verso, even large panels, with a piece of linen before the
application of a ground layer as has been shown in many Italian paintings.39 Gradually during the
fifteenth century and sixteenth century the preparatory layers of the panel were simplified. An
33 Ackroyd 2012: p. 460: Bisacca and Castelli 2012: p. 74.
34 Ulrich Schiessl. ‘History of Structural Panel Painting Conservation in Austria, Germany and Switzerland.’ In:
Kathleen Dardes and Andrea Rothe ed. The Structural Conservation of Panel Paintings: Proceedings of a Symposium at the J. Paul Getty Museum, 24-28 April 1995. Los Angeles: Getty Conservation Institute, 1998: p. 209; Mark Leonard, Carole Namowicz and Anne Woollett. ‘Albrecht Altdorfer’s Crucifixion (Museum of Fine Arts, Budapest’. In: Marika Spring ed. Studying Old Master Paintings. Technology and Practice. The National Gallery Technical Bulletin 30th Anniversary Conference Postprints. London: Archetype Publications in association with the National Gallery, 2011: p. 97; Maryan Wynn Ainsworth et al. German Paintings in the Metropolitan Museum of Art, 1350-1600. New York: Metropolitan Museum of Art, New York, 2013: p. 277.
35 Leonard, Namowicz and Woollett 2011: p. 97.
36 Raffaella Bruzzone and Maria Clelia Galassi. ‘Wood Species in Italian Panel Paintings of the Fifteenth and
Sixteenth Centuries: Historical Investigation and Microscopical Wood Identification.’ In: Marika Spring ed. Studying Old Master Paintings. Technology and Practice. The National Gallery Technical Bulletin 30th Anniversary Conference Postprints. London: Archetype Publications in association with the National Gallery, 2011: p. 254.
37 No suggestions are given how the painting could have been transferred in the past. 38 Rijksmuseum files. SK-A-3418. Kuiper 1980: II.
entire cloth covering the panel’s verso made place for only some strips of cloth or fibres (linen or hemp). Mostly, they were only applied on sensitive areas such as knots, defects, and also joints as
was also the case with the lip joint of the support of the Flood.40 Cloth was applied over joints and
irregularities to prevent eventual movement of the wood from affecting the paint layers.41 The
very fact that there was a piece of cloth between the joints makes it very likely that the upper plank is an original (later) addition to the lower planks.
Individual planks of a panel were usually placed with their longest side (coincided with the
grain direction) parallel to one another, as was also the case for the two vertical planks of the Flood. Less frequently, panels were formed by planks joined with their grain direction
perpendicular to one another.42 Lip-joints, however, have been found infrequently in original
Italian panel constructions. Planks were almost exclusively butt joined.43 It has been shown in
some cases that such a perpendicularly placed plank in an Italian painting was indeed a non-original addition, but the upper plank might also have been a modification by the carpenter in the
early stage of the preparation of the painting before the application of a ground layer.44 The artist
of the Flood might have used a panel that was already constructed by a carpenter independently from personal specifications, or at the request of the artist, the panel may have been enlarged,
because the artist adjusted his design.45 The butterfly keys on the back might pertain to the
original manufacture of the panel, although it is more likely that they are related to a later
restoration as appears to have been the case with numerous Italian panels.46
The original thickness of panels from central Italy between the thirteenth and fifteenth
century ranges between 3 and 4.5 cm and panels from the sixteenth to seventeenth centuries 4 to
4.5 cm.47. Between the fifteenth and sixteenth century the thickness of panel paintings in
northern Italy ranges between 0.5 and 3.1 cm (app. III). The thickness of the panel of the Flood suggests that the panel could be of northern Italian origin. This likelihood is bolstered further by the lengths and widths of the planks. The principal unit of length in Italian cities was usually based on the braccio (= the length from elbow to fingertip), which was subdivided according to
local custom and practical purposes and thus local and practical variations (app. IV).48 For
40 Uzielli 1998: p. 113; Castelli 2006: p. 116; Bisacca and Castelli 2012: p. 75. 41 Uzielli 1998: p. 113; Castelli 2006: p. 117.
42 Uzielli 1998: p. 119; Jørgen Wadum and Noëlle Streeton ed. ‘History and Use of Panels or Other Rigid Supports
for Easel Paintings.’ In: Joyce Hill Stoner and Rebecca Rushfield ed. Conservation of Easel Paintings. New York: Routledge, 2012: p. 96.
43 Bisacca and Castelli 2012: p. 76. 44 Uzielli 1998: p. 120.
45 Uzielli 1998: p. 120. 46 Uzielli 1998: p. 119. 47 Uzielli 1998: p. 118, 199.
48 Ronald Edward Zupko. Italian Weights and Measures. From the Middle Ages to the Nineteenth Century. Philadelphia:
instance one Paduan braccio da panno, that was used for measuring linen, equals 68.1 cm.
According to the braccio a muro (47 cm) that was used in several Emilia -Romagna cities, such as Varsi, Pelelgrino Parmense, Piacenza, Fiorenzuola d’Arda, the panel of the Flood would measure precisely 2.5 by 2 bracci.
The wood and construction of the support have proved to be good indicators of the date and place of origin of the Flood. The photographs and observations of the conservators who treated the painting before the destruction of the support, suggest it is was a poplar panel made in Italy. Its dimensions even indicate it is from the area Emilia, now known as Emilia -Romagna. The type of wood, joints, and preparation for painting moreover are characteristic for the fifteenth and early sixteenth centuries. Although perpendicularly planks have rarely been seen in original Italian panel constructions, it is likely that it surmounted the lower planks already in the early stages of the preparation of the panel.
Fig. 7. The back of the Flood in 1979.
Fig. 9 Reconstruction of the original support before 1980.
– — row of nail holes
Fig. 11. Reverse of the original support of the Flood.
The horizontal plank shows open worm tunnelling, indicating that it may have been thinned
Fig. 12. Verso of the original support (without the upper horizontal) after the cradling had been removed in 1979
Fig. 13. According to Kuiper the upper part of the painting was done in oil and not in tempera as in the lower part of the painting
1.2 Ground
This section is concerned with the ground of the Flood. Before the application of paint, a ground layer would first have been applied to the support. Since the composition of the ground might be an indication of the painting’s place of production, its characteristics have been examined. Gesso grounds (consisting of calcium sulphate and glue) are almost invariably found in Italian paintings, whereas chalk ground (consisting of calcium carbonate and glue) are mostly found in northern European paintings. Identification of the ground might therefore be helpful in determining
whether the painting is German or Italian.49
The conservation report (1979-80) records that the original panel had a reasonably irregular thick white preparation layer (fig. 14). During transfer, this ground was preserved, but was filled with a
mixture of chalk and colle totin, a rabbit glue, to level its thickness.50 The ground layer is now
visible at points where the paint layers are held in reserve (in for example the trees), thin, abraded, or have sustained (fig. 15-16). There is some evidence that the panel had moldings, as on the left there is a barb, suggesting that the painting was executed within a temporary or permanent frame molding that was applied to the panel.
Two samples, taken from the large mountain of the left vertical plank and one from the
water of the strip attached to it, were mounted as cross-sections (samples #02 and #09).
Examination of the cross-sections shows that the panel was prepared with a single coarse ground. Its colour is white beige to white yellow (fig. 17-18). The wax-resin that was used to make the ground and paint layers adhere to the new support partially infiltrated the ground. The infiltrated ground of sample #09 is therefore somewhat more yellow in colour. Sample #02 seems to be more homogenous, than sample #09. The latter has significant irregular, somewhat translucent large chunks, as well as some red and black tiny particles. The ground of both samples measure approximately 150 μm in thickness, but since the cross-sections do not include the entire ground, which moreover has been partially removed during transfer, it is invalid to compare these
measurements with those known of other Italian paintings.
There is no intermediate layer, or imprimatura, present between the ground and the paint
layers, although in the sixteenth-century, such a layer was often present in both northern
European and Italian panel paintings.51
49 Ainsworth et al 2013: p. 209.
50 Rijksmuseum files. SK-A-3418. Kuiper 1980: p. I, II.
51 Maartje Stols-Witlox. ‘Grounds, 1400-1900.’ In: Joyce Hill Stoner and Rebecca Rushfield ed. Conservation of Easel
XRF analysis indicates that the ground is probably composed of gypsum, which is usually associated with Italian late-Medieval and Renaissance paintings (fig. 19). Calcium sulphate
(sulphate in anhydrite or gypsum structure) was further confirmed by SEM-EDX analysis as the ground layer’s major constituent (fig. 20-21). The SEM micrographs of the gypsum show a compact lath-like or tabular microstructure, a feature that has also been noted in many Italian
gypsum grounds (fig. 22).52 XRF analysis furthermore showed considerable high strontium
contents throughout the painting. As strontium was detected in all colour regions of the Flood, it is unlikely not to be due to the presence of a strontium-containing pigment. A correlation between calcium and strontium was indicated by XRF analysis (fig. 23) and confirmed by SEM-EDX (fig. 24). Strontium is a common element found in calcium based grounds such as gypsum
and anhydrite and could be associated to celestite (SrSO4), an accessory mineral commonly found
in calcium sulphate deposits.53
The ground is now in a fair condition. There are losses to the ground along the filled and
retouched joins and splits of the planks of the original panel. There is a strong aging craquelure, especially in the upper part of the painting where the horizontal plank was originally positioned, but also around the butterfly reinforcements on both the horizontal and vertical planks, and around the areas where the large knots were positioned on the vertical planks (fig. 25).
In conclusion, there is a gesso ground present throughout. The ground consists of calcium sulphate components, which is more typical for paintings that were produced south of the Alps.
52 Nicholas Penny, Ashok Roy and Marika Spring. ‘Veronese’s Paintings in the National Gallery Technique and
Materials: Part II.’ In: National Gallery Technical Bulletin. Vol. 17 (1996): p. 40; Berry et al. 2011: p. 132.
53 Carole Namowicz, Marc Walton and Karel Trentelman. ‘Determination of Strontium Content in Calcium-based
Ground with XRF and ICP-MS.’ Presented at the 9th International Conference on NDT of Art, Jerusalem Israel, 25-30 May 2008: p. 8; Antunes et al. 2014: p. 1031.
Fig. 14. Edge of the right vertical plank in 1979.
The original support had a reasonably irregular thick white ground layer
Fig. 15. Detail of the Flood and photomicrograph (magnification 60x). Ground visible through paint loss in the lower right corner of the painting
Fig. 16. Detail of the Flood.
The artist intentionally left the ground visible between the leaves of the tree
Fig. 17b. Photo SK-A-3418_02 2017 in ultraviolet fluorescence - magnification 400x. Ground
Fig. 18b. Photo SK-A-3418_09 2017 in ultraviolet fluorescence - magnification 400x. 1) Ground 2) Underdrawing 3) Varnish
Fig. 20. EDX spectrum of the ground of SK-A-3418_02 2017. The sulphur, calcium and oxygen contents are high
Fig. 21. EDX spectrum of the ground of SK-A-3418_09 2017. The sulphur, calcium and oxygen contents are high
Fig. 22. Cross-section of the gesso ground as a backscattered electron image in the SEM, showing a coarse ground layer mainly consisting of lath-like or tabular particles
Fig. 23. Correlation curve between calcium and strontium
1550 2050 2550 3050 3550 4050 4550 5050 0 500 1000 1500 2000 2500 3000 3500 4000 Sr -K Ca-K
40 kv
Fig. 24. EDX spectrum of the ground of SK-A-3418_02 2017. The strontium, oxygen and sulphur contents are high
1.3 Underdrawing
The underdrawing of the Flood is the next step in the build-up of the painting. The underdrawing provides information about the compositional strategies and techniques of the artist. It also allows for an identification of changes that have been made either during underpainting or painting.
The painting has been examined with IRR and IR. These techniques are able to reveal the carbon based materials of layers, such as an underdrawing, below a painting’s surface. The composition was very well planned as both imaging techniques revealed an extremely clear, extensive and detailed underdrawing. However, there is no visible evidence that a preparatory design was transferred mechanically by tracing, pouncing or squaring. The underdrawing is visible
throughout and even extends to the areas painted over the inner additional strips (fig. 26). The large mountain, architecture and the heads of the winds at the top, except for the one at the far right, in the area where the horizontal plank was originally situated are underdrawn as well (fig.
27-28).54 Moreover, the underdrawing in the upper parts is identical in style to that found in the
areas corresponding to where the lower vertical planks were. For example, the head of the wind at the far left and the face of the man who is floating on a tree trunk are drawn in the same way. They share the same facial features, especially the eyes (fig. 29). The presence of the
underdrawing confirms that both the inner strips and horizontal plank were part of the original composition. The assumption that the upper plank would have been a later addition seems therefore improbable. It is likely that the upper plank was an early modification to the construction of the woodwork, made before the draughtsman and painter set to work.
A fair amount of underdrawing can also be seen with the naked eye, but it does not seem
that it has been left visible on purpose, but rather that it is the result of pigments that have faded or that the paint has become more transparent through aging (fig. 30-31). However, an area at the right of the large mountain, where a cross-section was taken from, only consists of ground, underdrawing and varnish (fig. 32). This can either mean that the sample area corresponds to a part that was left unfinished, incidentally not touched by the brush, or deliberately done so held in reserve.
The design was directly executed onto the gesso ground and was established in one
campaign probably by one hand. The underdrawing appears to be executed with a liquid black
54 The head at the far right was probably either added later in the painting stage or at some other time as part of a
medium. The lines seem to be drawn with a brush as the lines vary considerably in width and often have a tapering end. Evidence from IR(R) and a SEM-EDX show that it contains a carbon-based black pigment, such as lampblack or charcoal (fig. 33-34).
The underdrawing defines the contours of large and small figures, animals, trees and
architecture. Even details, such as folds of clothes, hair and facial features, are carefully underdrawn. The artist used hatching with parallel lines to suggest shadow in for instance the architecture on the left side of the painting and the side, body and thighs of the woman with the sagging breasts (fig. 35-36).
There are some compositional changes between the underdrawing and the painted
composition. The largest modifications occur in the ark. The latch in the roof of the ark has been repositioned a few times, and was originally planned somewhat larger. It may have been shifted to the right as otherwise there would be too much overlap with the figure of Noah who peeks out (fig. 37). The hand of the desperate man with the black hair who is clinging to the ark was corrected. It seems that the hand did not fit perfectly over the beam and was therefore
repositioned (fig. 38). There are other minor changes in other figures and buildings. The woman behind the naked man holding a paddle, for example, is watching the viewer, whereas in the underdrawing she was looking down (fig. 39). A window of a building was replaced at the painting stage by a cross at the rooftop of a church in front of it, and the man with the pink overgown (cioppa) has a circlet around his head in the underdrawing that is absent in painting (fig. 40, 66).
In summary, the composition of the Flood was extensively underdrawn as IR(R) shows an elaborate underdrawing. The underdrawing is in some areas visible to the naked eye, probably due to pigments that have faded, but also because the artist left the area at the right of the large mountain unfinished or did deliberately so. The underdrawing was worked out in great detail in a carbon-based medium. IR(R) has proven to be very effective to confirm that the upper
horizontal plank and inner strips were original to the painting, since they have underdrawing identical to that of the rest of the painting. Although there is no direct evidence, the carefully executed underdrawing suggests that the artist transferred the design mechanically onto the support, even though he subsequently made minor changes, both in the drawing and eventually in painting.
Fig. 26. Detail of the Flood in visual and infrared light (wavelength band 900-1700 nm). Underdrawing also present on the origina inner strips that were attached to the panel
Fig. 27. Detail of the Flood and photomicrograph (magnification 60x). The large mountain at the right of the painting was underdrawn
Fig. 28. Details of the Flood in visual and infrared light (wavelength band 900-1700 nm). The heads at the top, except for the one at the far right, are underdrawn
Fig. 29. Details of the Flood in infrared light (wavelength band 900-1700 nm).
Fig. 30. Detail of the Flood and photomicrograph (magnification 60x). Underdrawing visible through paint layer and through paint loss
Fig. 31. Detail of the Flood and photomicrograph (magnification 60x). Underdrawing visible through paint layer and abrasion
Fig. 32. Detail of the Flood and photomicrograph (magnification 60x).
Area of the large mountain at the right of the painting that only consists of ground, underdrawing and varnish
Fig. 33. Photo SK-A-3418_09 2017 in ultraviolet fluorescence - magnification 500x. 1) Ground 2) Underdrawing 3) Varnish
Fig. 34. EDX spectrum of the ground of SK-A-3418_09 2017. The carbon content is the highest. The drawing medium is a carbon-based black pigment, such as lampblack or charcoal
Fig. 35. Detail of the Flood in infrared light (wavelength band 900-1100 nm). Underdrawing of the architecture at the right of the painting. Hatching suggests shadow
Fig. 36. Detail of the Flood in infrared light (wavelength band 900-1100 nm).
Underdrawing of the woman with the pendulous breasts at the right of the painting. Hatching suggests shadow
Fig. 37. Detail of the Flood in visual and infrared light (wavelength band 900-1100 nm). The skylight of the ark has been repositioned a few times, and was originally planned somewhat larger
Fig. 38. Detail of the Flood in visual and infrared light (wavelength band 900-1100 nm).
The hand of the sinner corrected, probably the hand did not fit perfectly over the beam and was therefore repositioned
Fig. 39. Detail of the Flood in visual and infrared light (wavelength band 900-1700 nm).
The woman behind the naked man holding a paddle, for example, is watching the viewer, but in the underdrawing she is looking down
Fig. 40. Detail of the Flood in visual and infrared light (wavelength band 900-1700 nm). A window of a church was replaced by a cross in the painting stage
1.4 Paint layers
Close visual examination of the paint layers provides insight into the pigments and techniques used by the artist and is useful in understanding the appearance of the painting, along with its provenance and date. Pigments used in more recent interventions will also be identified.
The artist applied the paint after he had rigorously established the design. The artist closely followed the underdrawing during the paint stage, making only some small changes. Contours of the underdrawing, such as the eyes of the figures, were often refined during painting (fig. 41).
The work has been heavily overpainted in the past. It is hard to establish the exact extent
of the different restoration campaigns. Major retouches, which are clearly visible to the naked eye and in UV light, are largely confined to the areas of the inner strips, the splits in the old panels and the joins of the individual planks. Notwithstanding the transfer and the mechanical stress related to the old support, the paint layers of the Flood are in a reasonable condition.
Kuiper, Coen and Oikawa-Picante who examined the painting in 1979-80 thought, that the medium of the lower part of the Flood was egg tempera, whereas the upper part including the
sky and the green mountains was oil.55 These conclusions were presumably drawn because of the
striking contrast between the lower and upper part, as well as the painting’s supposed date of 1450-1500. However, recent close visual re-examination of the appearance of the application of the paint layers suggests that the medium is oil. The blended tones, creamy, long and fluid brushstrokes are more typical of an oil medium and are not usually seen with tempera painting (fig. 42). This observation may date the painting towards the early sixteenth century since oil as
the principal medium came largely into use during that period.56 The painting might also be in
tempera grassa (egg tempera to which oil has been added), a medium that is hard tell apart from oil. Tempera grassa might be an indication of a late fifteenth-century date.57 Further analysis was not
carried out to confirm the binding medium.
The composition was built up in very few layers of paint and mostly painted from light to
dark. This might be another indication that the painting is in oil, a technique that often appears in
oil paintings.58 The skin tones for example were usually built up in two or three pinkish layers (fig.
43). The artist held many areas in reserve, including the trees, figures, buildings and most of their
55 Rijksmuseum files. SK-A-3418. Kuiper 1980: p. I, II.
56 Paul Craddock. Scientific Investigation of Copies, Fakes and Forgeries. Oxford: Butterworth-Heinemann, 2009: p. 297. 57 Jill Dunkerton. ‘Modification to Traditional Egg Tempera Techniques in Fifteenth-Century Italy.’ In: Tonnie
Bakkenist et al. Early Italian Paintings: Techniques and Analysis. Maastricht: Limburg Conservation Institute 1997: p. 19-34.
58 Dawson William Carr. Andrea Mantegna. The Adoration of the Magi. Los Angeles: The J. Paul Getty Museum, 1997: p.
chimneys (fig. 44). The paint handling of the upper part is unlike that of the lower part. The upper part of the painting including the green hilly landscape and sky with dark clouds is done in a rough way and worked in long broad and open brush strokes. The artist used his fingertip to blot the wet surface paint in the green bluish mountains. The lower part including figures and architecture was executed in a fine technique. Here, the paint layers have been carefully applied with short smooth brushstrokes, with some impasto in white areas, such as beards or the manes of a horse (fig. 45). Another notable technique is the grid of a window that has been incised with a stylus (fig. 46).
As IRR has already confirmed, the underdrawing of the upper part is consistent with the
underdrawing of the lower part and thus original to the overall, original composition of the Flood. Although the sky, and in particular the white area, is largely overpainted, it does not seem that there are paint layers present underneath the upper non-original layers that would be more consistent with the composition in the lower part of the painting. X-radiography that is able to reveal the presence of compositional alterations from an earlier stage, did not show changes in the paint layers (fig. 47). It is therefore very probable that the upper part, although different in execution technique, is part of the original composition.
There are ruled black lines, approximately 0.2 cm in thickness, along the left, bottom and right edges of the composition, that are visible in IR(R) and only in one area to the naked eye (fig. 48). The lines in IR(R) are often interrupted as they are obscured by overpaint and therefore partially visible. Since there is no line visible at the top edge of the dark sky, this area is probably completely overpainted. Cross-sections have not been taken to confirm this argument.
Analysis with XRF has been carried out to distinguish between original pigments and materials used in conservation treatments (app. V). FC-IR also facilitated the identification of pigments that are present at the surface. In a false colour infrared image the colour of a normal visual image is produced differently (by combining the image with an infrared photo and altering colour channels) in order to discern the response of pigments in the near-infrared spectrum. Many pigments in the Flood correspond to those used in fifteenth- and sixteenth-century Italian painting.
As is also clear in X-radiography, all white areas and mixtures with white were painted in
lead white (2PbCO3.Pb(OH)2) (fig. 47). The light parts of the sky were done in lead white and
some lead-tin yellow (Pb(Sn,Si)O3). High concentrations of zinc white (ZnO), a pigment that has
49).59 Lead-tin yellow has also been found in the clothing of figures. As indicated by high
concentrations of arsenic evidenced in XRF most yellow areas are in orpiment (As2S3). The
leopard consists of lead-tin yellow and yellow ochre (Fe2O3.H2O), while the yellow dress of
woman who is holding a child only consists of yellow ochre.
Black pigment is used sparingly and is presumably also carbon-based and similar to the
one that has been used for the ink of the underdrawing.
The water parts in the painting are now rather uniform and appear to have a dark colour.
However, IR(R) shows that the artist defined the waves with fluid brushstrokes during painting (fig. 50). XRF indicates the presence of copper in the water parts, probably indicating azurite
(2CuCO3.Cu(OH)2) (fig. 51). Azurite has been known to degrade over time. When bound in oil, it
discolours and becomes dark.60 This seems to be the case in the Flood and might be an extra
indication that the artist used oil as a medium. A cross-section that was taken from the water parts confirms that the binder has discoloured (fig. 52). The water has largely been overpainted
with a mixture of zinc white and probably Prussian blue (Fe4[Fe(CN)6]3 · xH2O) as in XRF these
areas are rich in iron. Prussian blue came into use during the eighteenth century.61 The artist may
have executed the blue clothing with ultramarine blue (Na7Al6Si6O24S3) as the clothing appears to
have a characteristic bright red in FC-IR (fig. 53), and, as is clear in XRF, is rich in silicium, which is one of the components in ultramarine.
The greens, such as the hills, grass, leaves of the trees and clouds, consist of verdigris and
probably green earth. The green hilly mountains may have increased in transparency with age as the lower lead white layer of the sky is shining through.
The reds in the headdresses, clothing and roofs are predominantly vermillion (HgS) as
XRF revealed high concentrations of mercury (fig. 54). The same measurements also indicates concentrations of arsenic, possibly related to orpiment as an impurity in early modern European
vermillion.62 The pink clothes are done in vermillion mixed with lead white.
XRF analysis also indicated frequently concentrations for barium, which may be barium
sulphate. Barium sulphate is found as an accessory mineral with azurite, and also became one of the modern nineteenth- and twentieth-century white pigments as an alternative to lead white.
59 Herman Kühn. ‘Zinc White.’ In: Robert L. Feller et al. Artist’s Pigments: A Handbook of Their History and
Characteristics. Vol. 1. Washington and London: National Gallery of Art, Washington and Archetype Publications, London, 1986: p. 169.
60 Annelies van Loon, Petria Noble and Aviva Burnstock. ‘Ageing and Deterioration of Traditional Oil and Tempera
Paints.’ In: Joyce Hill Stoner and Rebecca Rushfield ed. Conservation of Easel Paintings. New York: Routledge, 2012: p. 214-241: p. 234.
61 Gettens and Stout 1966: p. 150.
62 Nicholas Eastaugh et al. Pigment Compendium: A Dictionary and Optical Microscopy of Historical Pigments. Oxford:
However, barium neither seem to be related to azurite nor to a white pigment, such as zinc or titanium white. It is more likely that barium sulphate was used as an extender (in order to increase the paint’s volume and film thickness) in the paints that has been used in later restorations. It has
been known as an extender in manufactured paints of the nineteenth century.63
The Flood is lightly to moderately abraded, and probably as a result the skin tones often
show a stripy appearance. Small losses are scattered across the paint surface. In addition to the crackle in the ground, the paint layers show a very fine irregular aging craquelure mostly confined to the areas of light green passages, the water and the dark green of the trees (fig. 55). The varnish and wax-resin are discoloured and uneven in saturation.
Close examination with the naked eye, XRF, IR(R), FC-IR has largely revealed the composition of the original paint layers that are now much overpainted. The artist used pigments that are of fifteenth- and sixteenth century Italian painting, such as lead white, azurite, verdigris, vermillion, carbon black, lead-tin yellow and orpiment. XRF cannot detect and verify all pigments, and therefore the colour palette may be more varied. Although in the past the principal binding medium of the Flood has been reported as egg tempera, there are several indications that it is actually oil. The discoloration of the blue also suggests that the binding medium is oil. The binding medium and the pigments indicate that the Flood was made in Italy around 1500.
Fig. 41. Detail of the Flood and photomicrograph (magnification 60x). After the underdrawing had been established, the lines of the eyes were refined
63 Giorgia Mancini et al. The Sixteenth Century Italian Paintings. Volume III. Bologna and Ferrara. London: National Gallery
Fig. 42. Detail of the Flood and photomicrograph (magnification 60x).
The pink overgown (cioppa) of the man is painted wet-in-wet, which is typical for oil paint
Fig. 43. Detail of the Flood and photomicrograph (magnification 60x). Skin tones of the man who is holding a paddle
Fig. 44. Detail of the Flood and photomicrograph (magnification 60x). Chimneys were often left in reserve
Fig. 45. Detail of the Flood and photomicrograph (magnification 60x). Impasto white paint, which is typical for oil paint
Fig. 46. Detail of the Flood and photomicrograph (magnification 70x). Grid of a window that has been incised in the paint layer
Fig. 47. The X-radiograph of the Flood shows the presence and distribution of lead white. It does not reveal changes in composition
Fig. 48. Detail of the Flood in visual and infrared light (wavelength band 900-1700 nm). Vertical ruled black line at the right edge of the composition
Fig. 49. XRF spectrum of the overpainted white yellowish sky. The spectrum shows high peaks for lead, and a small peak for tin, indicating lead-tin yellow. The small peak for zinc indicates that the sky was overpainted with zinc white.
Fig. 50. Detail of the Flood in visual and infrared light (wavelength band 900-1700 nm). The blue paint of the water has discoloured. The infrared reflectograph shows that the water was refined with waves
Fig. 51. XRF spectrum of the darkened water. The spectrum shows a high peak for copper, indicating azurite and a small peak for iron, indicating Prussian blue
Fig. 52. Photo SK-A-3418_01 2017 in dark field - magnification 500x. Sample taken from the painted water shows that the binder has discoloured.
Fig. 54. XRF spectrum of the red bonnet of a man in the water. The spectrum shows high peaks for mercury, indicating vermillion
2. Clothing and headdress
The Flood depicts numerous figures, dressed in a variation of clothes. There are strong indications, as has been shown in the previous chapter, that the Flood was made in Italy. This chapter is concerned with the chronological and topographical evidence of clothing and
headgear. Clothing and headdresses depicted in the Flood will be compared to other fifteenth- and early sixteenth-century paintings, as well as to contemporary woodcuts and descriptions in the fashion treatise Habiti Antichi et Moderni di Diversi Parti del Mondo (1590) by the Venetian Cesare Vecellio. This publication deals in great detail with Italian clothing and calls attention to the variety of styles, starting with ancient Rome and then continuing with Medieval and Renaissance
Italy, and in particular Venice.64
Styles in fashion had always spread fact across the various countries of Western Europe, but during the late fifteenth century and early sixteenth century the foreign influence was particularly
strongly felt in Italy.65 This had to do with an increase in the number of visits to Italy by scholars
and artists from abroad. For example, Milan was well acquainted with the fashions of Burgundy
and France and the south German fashion was of great influence in the Italian Veneto region.66
Nonetheless, countries also remained true to their own fashion styles.
In the Flood, the woman who sits astride a bull wears a dress that can be dated with some precision. One of the indications is its very low neckline, which would be unusual in Italy much
before 1500 (fig. 56).67 The detachable sleeves, also date it to the end of the fifteenth century. The
sleeve was separate from the bodice of the dress to which it was attached with eyelets and laces.68
Similar low necklines and sleeves can be seen in the dresses depicted in Vittore Carpaccio’s Two Venetian Ladies (1520) in the Museo Correr, Venice and in Ridolfo Ghirlandaio’s Portrait of a Woman (1506-1510) in the Uffizi, Florence (fig. 57-58).
The man with the pointy beard in the lower right corner of the painting wears a red
capperone, a hat worn wrapped around the head with frontward folds (fig. 59). Such a capperone was worn throughout Europe, but only during a limited period of time. Similar headgear appears in Saint Lucy before the Judge, a fresco that Altichiero da Zevio painted between 1378 and 1384 (fig. 60). The capperone became one of the predominant headdresses between 1400 to 1429 and
64 Cesare Vecellio. The Clothing of the Renaissance World: Europe, Asia, Africa, the Americas: Cesare Vecellio's Habiti Antichi et
Moderni. Trans. Margaret F. Rosenthal and Ann Rosalind Jones. London: Thames and Hudson, 2008: p. 8.
65 Jacqueline Herald. Renaissance Dress in Italy 1400-1500. London: Bell & Hyman Limited, 1981: p. 193. 66 Herald 1981: p. 195.
67 Nicholas Penny. The Sixteenth-century Italian Paintings. Volume I. Paintings From Bergamo, Brescia and Cremona. [mus.cat.]
London: National Gallery Company, 2004: p. 23.
