Too Hard to Handle. Cupping Paint on a Seventeenth-Century Glue-Lined Painting.

(1)

Too Hard to Handle

Cupping Paint on a Seventeenth-Century Glue-Lined Painting

Master’s Thesis

Sepha Carolina Wouda Student number: 10627685

Thesis supervisor: dr. Maartje Stols-Witlox Professors Conservation and Restoration: prof. dr. Maarten van Bommel

prof. dr. Ella Hendriks

University of Amsterdam 20 June 2017

(2)

Preface

This Master’s thesis is written within the programme of the master Conservation and Restoration of Cultural Heritage, Paintings Specialisation, at the University of Amsterdam. Its subject is the independent diagnostic research that has been carried out on a painting in the collection of the Dordrecht Museum. This research would not have been possible without the help of a large number of people to whom I would like to express my gratitude. First and foremost I would like to thank my thesis supervisor Maartje Stols-Witlox (University of Amsterdam) for her support, feedback and valuable advice. I also thank her for carrying out all SEM-EDX analyses. Special thanks go to Lidwien Speleers (Dordrecht Museum) for suggesting the thesis subject, as well as for sharing her knowledge about the painting, and for all the help and feedback she gave me during my research. I thank Sander Paarlberg (Dordrecht Museum) for sharing his expertise about Farrington and other painters from Dordrecht.

I owe much gratitude to Henk van Keulen (Cultural Heritage Agency) for performing the THM-Py-GC-MS analyses and answering difficult questions. I am also grateful to Bas van Velzen (University of Amsterdam) for analysing the canvas fibres. Sincere thanks go to Muriel Geldof (Cultural Heritage Agency) for placing the optical microscope at my disposal as well as for explaining its workings. I thank Ineke Joosten (Cultural Heritage Agency) for her advice on SEM-EDX. I thank Klaas-Jan van den Berg (Cultural Heritage Agency, University of Amsterdam) for his advice on Proteomics and Jitske Knip (UvA MA Student Science) for carrying out the Proteomics analyses.

At the Rijksmuseum, special thanks are due to conservator-restorers Susan Smelt and Gwen Tauber, who provided information about the paintings of Aelbert Cuyp, to Huub Baija for enlightening me on the painting’s frame and to Laura Raven for providing the FTIR report of the painting. I thank Carol Pottasch for receiving me at the Mauritshuis and giving me some useful information about Van Dyck. I also thank Abbie Vandivere (Mauritshuis, University of Amsterdam) for the tips and the information she gave me. I am much indebted to Kate Seymour (Limburg Conservation Institute SRAL, University of Amsterdam) for her explanation of lining techniques and answering my questions on that subject. Thanks are also due to Hans Bies (Bies Restoration) for his explanation of traditional lining methods.

I would like to thank my lecturers at the University of Amsterdam for their help and advice: Vera Blok, Emilie Froment, Saskia van Oudheusden and in particular Rene Peschar for his explanation of technical analyses and the correction of my texts and calculations. I thank René Lugtigheid for her clarification about the canvas. Furthermore I thank my professors Ella Hendriks and Maarten van Bommel (University of Amsterdam) for their insight, help and advice.

I thank my fellow students Rosa Boute, Marjolein Hupkes, Alma Jongstra, Noa Kollaard and Madeleine Vaudremer for their support and their company. To finish, I thank Jan and Esther for their support and help with the text; Joris for his technical support; and Camilla, Luca and Nora for their understanding and patience.

(5)

Abstract

This Master’s thesis describes the research that was carried out into serious degradation phenomena found in the paint layers of a seventeenth-century oil painting in the collection of the Dordrecht Museum. River Landscape with Hunters and a Harbour was painted by the English artist Richard Farrington who lived in Dordrecht in the mid-seventeenth century. The painting was lined in the past and shall undergo a restoration treatment on account of cupping paint and delamination issues. Consolidation, however, is complicated by the fact that the paint layers are extremely rigid and at the same time humidity sensitive. The purpose of the research, therefore, was to diagnose these phenomena, thereby laying the foundations for a suitable treatment plan for the painting.

Since very little is known about Farrington and his oeuvre, research was carried out into his life in Dordrecht and possible contacts with other painters. This investigation aimed at a better understanding of the historical context of the painting, including possible other works by the artist, as well as a dating of the painting. As Farrington can be considered a contemporary follower of Aelbert Cuyp, River Landscape was compared to paintings by this master. The similarity in grounds was examined so as to determine whether ready-primed canvases could have been used.

To assess the condition of the painting as well as the possible causes of the degradation, research was performed using several analytical techniques, such as optical microscopy and SEM-EDX. This gave insight into the material aspects and technique of the painting. It also yielded a global dating of the restorations and the format change of the painting that took place in the past.

Since the above-mentioned degradation phenomena are often related to the response of the different materials in a painting to fluctuating relative humidity, research was done into the current state of knowledge in this field. In addition, the adhesive of the lining as well as the binding media in the ground and paint layers were analysed using THM-Py-GC-MS. The lining adhesive consists mainly of animal glue, and bears the characteristics of a fish glue. This type of glue was also found in the chalk ground of the painting, in combination with linseed oil.

The research suggests that the binding media, in particular the lining adhesive, combined with exposure of the painting to high relative humidity, played a key role in the degradation. Further research is currently being conducted with Proteomics in order to ascertain which type of animal glue is present in the lining adhesive and the ground. This could provide an indication of the (differences in) sensitivity to humidity of the layers, information which is relevant for the pending restoration and conservation of the painting in the future.

(6)

Samenvatting

Deze masterscriptie beschrijft het onderzoek dat is uitgevoerd naar aanleiding van ernstige degradatieverschijnselen die zich hebben voorgedaan in de verflagen van een zeventiende-eeuws olieverf schilderij op doek, dat zich in de collectie van het Dordrechts Museum bevindt. Het schilderij Rivierlandschap met jagers en een haven werd geschilderd door de Engelse kunstenaar Richard Farrington die halverwege de zeventiende eeuw in Dordrecht woonde. Het schilderij is in het verleden gedoubleerd en zal binnenkort een restauratiebehandeling ondergaan vanwege opstaande verf en delaminatieproblemen. Consolidatie wordt echter bemoeilijkt doordat de verflagen extreem hard en tegelijkertijd vochtgevoelig zijn. Het doel van het onderzoek was dan ook om een diagnose te stellen met betrekking tot deze verschijnselen, zodat een passend behandelplan kan worden opgesteld voor het schilderij.

Omdat er nagenoeg niets bekend is over Farrington en zijn oeuvre is eerst onderzoek gedaan naar zijn leven in Dordrecht en mogelijke contacten met andere schilders. Dit had ten doel meer te weten te komen over de historische context en de datering van het schilderij alsmede over eventuele andere werken van de kunstenaar. Omdat Farrington wordt beschouwd als een eigentijdse navolger van Aelbert Cuyp is Rivierlandschap vergeleken met schilderijen van deze meester. Er is nagegaan of de schilderijen vergelijkbare gronderingen hebben die mogelijk zouden kunnen wijzen op het gebruik van voorgegrondeerde doeken. Een vergelijkbare dubbele grondering werd gevonden op een schilderij van Cuyp dat zich in het Rijksmuseum bevindt.

Om een beeld te krijgen van de conditie van het schilderij en de mogelijke oorzaken van de degradatie is onderzoek verricht met verschillende analytische technieken zoals optische microscopie en SEM-EDX. Hierdoor kon inzicht worden verkregen in het materiaalgebruik en de techniek van de schilder. Ook kon een globale datering worden gegeven van een formaat wijziging van het schilderij en de restauraties die hebben plaatsgevonden in het verleden.

Omdat genoemde degradatieverschijnselen vaak verband houden met de reactie van de verschillende materialen in een schilderij op wisselende relatieve luchtvochtigheid, is onderzoek gedaan naar de stand van wetenschap op dit gebied. Dit gaf aanleiding tot het uitvoeren van bindmiddel-analyse met behulp van THM-Py-GC-MS. De analyses toonden aan dat de bedoekingsspecie voor het grootste deel uit dierlijke lijm bestaat, waarvan de geïdentificeerde componenten gelijkenis vertonen met die van een vislijm. Deze lijm was ook terug te vinden in de kalkgrondering van het schilderij, in een combinatie met lijnolie.

Het onderzoek geeft aanleiding te denken dat de bindmiddelen, met name de lijm van de doublering, een grote rol hebben gespeeld in de degradatie van het schilderij. Verder onderzoek wordt momenteel uitgevoerd met Proteomics om te bepalen welk type dierlijke lijm aanwezig is in de bedoekingsspecie en in de grondering. Dit zou een indicatie kunnen geven van de (verschillen in) vochtgevoeligheid van de lagen. Deze informatie is van belang voor de op handen zijnde restauratie en de conservering van het schilderij in de toekomst.

(7)

Introduction

In 2020, the Dordrecht Museum will organise a major exhibition about the Dordrecht artist Aelbert Cuyp (1620-1691) and the impact of his work on other landscape painters from Holland and abroad. Within that context, the canvas painting River Landscape with Hunters and a Harbour by the English artist Richard Farrington will be shown.1 Farrington, who was born in England between 1610-1630 and died in London between 1664-1681, lived and worked in Dordrecht from 1648 to 1664 and was greatly influenced by the work of Aelbert Cuyp. River Landscape has been in the collection of the Dordrecht Museum since 2011 when it was purchased from the auction house Dorotheum in Vienna. It came from a German collection but nothing more is known about the painting’s provenance. Other works of Farrington are described in seventeenth-century inventory lists of Dordrecht collections but these have been dispersed due to inheritance and sale. The

current location of these paintings is therefore unknown.2 Moreover, hardly anything is

known about the artist and his oeuvre.

The Dordrecht painting was glue-lined in the past and needs a restoration treatment because of severe cupping and delamination issues.3 There is no written evidence about the conservation and restoration history of the painting, but the object bears many traces of previous restorations. Besides the lining treatment and a possible format change, at least two retouching and filling phases are discernible. Several large fillings are present, which cover holes in the original canvas. The paint has cracked extensively and a network of largely random cracks has formed, which have opened and allowed the paint to cup (see cover image). Consolidation tests by Lidwien Speleers, Paintings Conservator of the Dordrecht Museum, showed that the cupped paint film is extremely rigid and difficult to relax with heat and pressure, as well as sensitive to treatment with moisture, making the upcoming restoration very complicated and time-consuming.

From research into the subject-related literature it is clear that many different causes may explain cupping and delamination of paint layers. Depending on the cause, different conservation methods would be chosen. Consequently, the present condition of the paint film and its behaviour during the consolidation tests cannot be understood well enough to proceed with the restoration treatment. The main question that will be researched in this master’s thesis is therefore:

How can the severe cupping, delamination and humidity-sensitivity of the paint film of the seventeenth-century glue-lined painting River Landscape with Hunters and a Harbour by Richard Farrington be explained?

1_{Personal communication of Sander Paarlberg, Curator of Old Master Paintings, Dordrecht Museum, 22}

February 2017.

2_{Personal communication of Sander Paarlberg, 4 December 2016.}

3_{Cupping is the phenomenon when islands of paint between cracks have lifted edges, which curled away}

(8)

By increasing understanding about the material composition and degradation that has occurred in the painting after its creation, better informed decisions can be made about the restoration treatment and measures can be taken to avoid further delamination or the recurrence of the cupping problem after restoration.

Research will be done into the history of River Landscape and other paintings by Farrington, to see whether the observed degradation can be related to his use of materials and painting technique or to the conservation history of the painting. Research into the historical background and date of the painting will also be relevant for the museum and the upcoming exhibition. Technical analysis will give insight into Farrington’s use of materials, which will be useful for further research into similarities with the works of contemporary Dordrecht artists. The information gained from both technical and comparative research will add to the knowledge about cupping paint on glue-lined paintings, a subject that has been given relatively little attention in the literature.4 It will also be valuable for conservator-restorers in the field who are confronted with comparable cases, particularly since the advanced analytical techniques that have been used for this research will seldom be available in practice.

The first chapter of this thesis will give a detailed description of the historical and artistic context of the painting and its creator, in order to find out more about the painting’s history, the artist’s oeuvre and possible contacts with contemporary painters who lived in Dordrecht during the 1650-60s. It will also discuss the influence of Cuyp on Farrington’s painting style, in order to come to a more accurate dating of the painting.

Different analytical techniques have been employed to determine the material composition, the painting technique and layer build-up of the painting. Insight into these aspects, described in Chapter two, is the first step towards understanding the observed deterioration. In addition, the coloured grounds used by Farrington and Cuyp will be compared, to see whether the former applied an unusual type of ground. The collected information will help with the dating of the lining treatment and other interventions.

Chapter three will discuss current scientific knowledge about the moisture response of oil paintings on canvas, since changes in (relative) humidity are often mentioned in literature as the main source of stress and stress-related degradation such as cracking, cupping and delamination of paint layers. This knowledge, combined with the results of binding medium analysis, will serve as a basis to explain the painting’s present condition and the observed degradation phenomena, the topic of Chapter four.

(9)

Chapter 1 Farrington, an English painter in the Netherlands

1.1. Records of his life in Dordrecht

The name Richard Farrington was first recorded by Abraham Bredius (1855-1946), a Dutch art historian, art collector and director of the Mauritshuis museum who studied Netherlandish seventeenth-century painting (Sorensen 2017, n. pag.). Bredius’ study in the Notarial Archive of Dordrecht delivered Een en ander over Dordtsche schilders, a lecture on the numerous painters that were born or active in Dordrecht during the seventeenth century. He stated that many of Rembrandt’s best apprentices came from there, such as Ferdinand Bol (1616-1680), Nicolaes Maes (1634-1693) and Samuel van Hoogstraten (1627-1678) (Bredius 1913, 4-5).

During this archival research, Bredius also encountered the name of Richard Farrington, an English painter living in Dordrecht, who was probably born in around 1625. In his Künstler-Inventare, Bredius collected numerous records of inventories of painters from the sixteenth to the eighteenth century. Some of the notarial deeds he recorded give account of financial transactions and legal acts Farrington was involved in between 1648 and 1664 (appendix IV.i) (Bredius 1917b, 1391-4):5

1648-07-17: “Richard Farington signs a contract together with the painter Pieter Fris.”6 1651-11-14: “Dirck Tegelberch, husband and guardian of Pieternella Mesian, Ritchard

Fariton, husband and guardian of Anna Mesian, and Maycken Mesian, unwed, are heirs of Maricken Claesdr (…).”

1653-04-08: “Ritschaert Pharinton, painter, stands surety for his brother-in-law Dirck Tegelberch, silversmith (…).”

1655-07-02: Ridchard Farington has a debt of fl. 4:10 st. for the priming of several canvases.

1657-07-20: “Somebody demands from Ritsi Fariton, agent of the English Court for Merchants, the payment of fl. 30,16 st. for a debt.”

1658-05-05: “The hornourable Ritsaert Farington, Citizen of Dordrecht, cancels the tenancy of a garden.”

1660-09-24: Richard Farrington, wine merchant, sells a consignment of Rhine wine to England and signs a warrant to receive the payment for the wine.

1662-05-20: “Richard Farrington rents a house in Dordrecht on the Cleyne vismarckt for fl. 220 per year.”

1662-06-20: “Richard Farrington, merchant, gives a mandate to Barent Verveer. He owes a person money and puts something in pledge.”

5_{These are my translations (between quotation marks) and summaries of a selection of data from Bredius}

(1917b, 1391-4).

Many different notations of names and words have been encountered, which was common practice in this period of history, as there was no standardised spelling yet. Communication of Sander Paarlberg, Curator of Old Master Paintings, Dordrecht Museum, email 2 May 2017.

6_{Pieter Fris (1628/9-1706), born in Amsterdam, was a painter of historical, religious and genre pieces. In}

1647 he signs as a witness in Dordrecht (Bredius 1919, 1986) and in 1648 he meets Farrington to sign a contract, but nothing is mentioned about the contents of this agreement (Bredius 1917b, 1391).

(10)

1663-11-13: “A baker demands from Ritchard Fariton the payment of fl. 60 for delivered bread.

1664-10-24: Richard Farrington lives in London and pays fl. 875 for delivered wine. 1681-07-03: “Anneken Barthouts Mesyan, widow of Richard Farrington in London,

inherits fl. 66 from a Dordrecht estate.”

These data reveal that Richard Farrington came to Dordrecht before the 17th of July 1648, left for London in 1663 or 1664 and died before July 1681. Other registered events in Farrington’s life are his marriage in 1650 to Anna Meschan and the birth of their four children, between 1651 and 1660 (appendix IV.ii) (Regional Archive of Dordrecht 2017).7 The marriage register states that Ritzart Farrington was a painter from Leicester who lived near the Vischmarkt (fish market) (figure 1.1).

Throughout the sixteenth and seventeenth centuries many painters from the Low Countries were employed by the English court and aristocracy and moved to Britain. Farrington was an exception to that unidirectional movement, as he left England to become active as a painter in the Netherlands (Brown 1996, 345-6). He was called a painter in 1650 (marriage register) and in 1653. Moreover, he had canvases primed in 1655 (see above). According to information of the Netherlands Institute for Art History (RKD), he did not join the Guild of St. Luke in Dordrecht (RKD 2017, n. pag.).8

Farrington’s stay in Dordrecht falls within the greatest period of Dutch painting, between 1640-1665. Both in terms of quantity and quality, the production of paintings in this period reached an amazingly high level. Due to the strong competition between artists, some of them could not make a living merely on the sale of their paintings and were compelled to take up an additional profession (Reiss 1975, 7-8). This may have been the case with

7_{According to Bredius (1917b, 1391), Farrington married Anna Mesyan in 1651, but the marriage register}

from the Dordrecht archive shows that they gave notice of the intended marriage on 29 May 1650. The actual marriage took place on 14 June 1650 in the Augustijnenkerk of the Nederlands Hervormde Gemeente.

8_{The information that Farrington did not join the guild could not be traced back in the mentioned sources on}

the RKD website.

Figure 1.1 Part of folio 182 from the marriage register (1642-1655) from the Regional Archives of Dordrecht records the marriage of Ritzart Farrington with Anna Meschan on 14 June 1650.

(11)

Farrington, who is mentioned as a merchant trading wine from Holland to England from 1657 onwards.9 By that time, he had to support a family with three children.

1.2. Contacts and chance encounters with other painters

Around 1650, Dordrecht was a wealthy city with a population of about 21,000 inhabitants (Nusteling et al. 2008, 78). It was a major trade port due to its advantageous geographic position between a number of waterways. It was well-known to English merchants and for a long time, the wine trade was an important source of income for the city (De Groot 1977, 60).10 Joan Blaeu, in his topographic atlas of Dutch cities of 1649, published a map of Dordrecht with a legend of its most important locations and streets (figure 1.2). In that period the entire city was situated within its medieval moat, the Spuy haven, and had the dimensions of the present historic centre. Given its compactness and the large number of artists who lived in the city in those days, it is very likely that Farrington knew at least some of his painter-colleagues. Moreover, Bredius’ archival finds suggest that the Englishman was embedded in diverse social and economic circles related to wine trade and art. It is interesting to investigate where contemporary painters lived in Dordrecht, to see whether Farrington lived in the same areas. The city map of figure 1.2 illustrates the streets where Farrington and other painters lived. The numbers on the map correspond with the numbers behind the street names mentioned in the text:

In 1650, Farrington lived on the Vischmarkt (1) and in 1662 he rented a house on the Cleyne Vismarckt (2) (Bredius 1917b, 1393). The Cleyne Vischmarkt was located on the current Palingstraat, a side street of the Wijnstraat, near ’t Groot Hooft (Lips 1974, 238-9). This was an area where other artists lived (7), among which Cornelis Bisschop (1630-1674), Abraham van Calraet (1642-1722) and Arent de Gelder (1645-1727).11

Aelbert Cuyp (1620-1691) and his father Jacob Gerritsz Cuyp (1594-1652) were the leading artists in Dordrecht over a period of some forty years from about 1630 onwards and during Farrington’s stay (Reiss 1973, n. pag.). Until his marriage in 1658, Aelbert lived in the house ‘Samson’ at the Nieuwbrug (3) on the side of the Voorstraat. In 1658, Cuyp moved to his wife’s house in the Hofstraat (4) and in 1663 they also moved to the Wijnstraat (5), near the Wijnkoperskapel, which was, at least until 1656 a meeting place for the wine merchant guild (Veerman 1977, 17-8). On Blaeu’s map, gardens and countryhouses are indicated outside the city walls, which were probably used by citizens to dwell in nature. Cuyp owned such a garden at the Matena’s pad (6) (Veerman 1977, 19) and Farrington obviously also had one, of which he cancelled the tenancy in 1658, according to an archival record.

9_{Holland was one of the seven provinces of the Dutch Republic, which existed from 1581 until 1795. On 15}

May 1648, the Peace of Münster made an end to the Eighty Years’ War of the Seven United Provinces against the Spanish Empire (Wiki 2017, n. pag.). Farrington’s presence in Dordrecht was registered in that same year, which might indicate that he actually arrived in 1648 after the war was over and not much earlier.

10_{“De bysondere gebouwen sijn op vele plaetsen groot en heerlijck,(…); oock vele op gewelfde kelders, om}

de Rhijnsche wijnen in te leggen, gebouwt” (Blaeu 1652, Dordrecht n. pag.).

(12)

Other well-known painters living in Dordrecht during Farrington’s stay were Nicolaes Maes (1634-1693), who bought a house in the Steechoversloot (8) in 1658 (Bredius 1913, 5) and Samuel van Hoogstraten (1627-1678), who had a studio on the Marktveld (9) and lived in the Steechoversloot (8) around 1667 (Regional Archive of Dordrecht 2017, n. pag.).

Adriaen Huybertsz Verveer (ca. 1620-1680) was a Dordrecht portrait and history painter and, like Farrington, also a wine merchant. He bought a house at the Nieuwe Haven (10) in 1664 and sold it in 1667 (Bredius 1917a, 905-6). Cornelis Tegelberg (c.1619-c.1667/87) was a local landscape painter influenced by Cuyp, who’s brother Dirck (see above) married Farrington’s sister-in-law. He was one of the founders of the Dordrecht Painters’ Guild in 1642 and was probably closely acquainted to Farrington (Paarlberg 2013, 278-9).

Figure 1.2 Joan Blaeu. Dordrecht. 1649. Copper engraving. 41 x 52 cm.

This map was published by Blaeu between 1649 and 1652 in Toonneel der steden van de vereenighde

Nederlanden, met hare beschrijvingen, part 1. It shows Dordrecht in the mid 1640s, around the time when

Farrington arrived. The numbers 1-2 indicate the streets where Farrington lived and the numbers 3-10 the streets where some other artists lived during his stay between 1648 and 1664.

1 3 2 10 4 5 6 7 8 9 8 7 7

(13)

1.3. Farrington’s paintings

The most important source of information about Farrington and his work in Dordrecht is Bredius’ archival research. In his Künstler-Inventare, several paintings by Farrington are mentioned. To place River Landscape in the context of Farrington’s oeuvre, descriptions of his paintings have been collected from Bredius’ publications and other sources. Information that gives insight into Farrington’s life as an artist is also described.

In the weekly magazine Kunstchronik of 16 May 1913, Bredius published important findings that contributed to the knowledge about Dordrecht painting schools. Richard Farrington is mentioned as an Englishman who painted in the style of Aelbert Cuyp (M.D.H 1913, 486-7). Farrington’s paintings, “large landscapes, one with a castle”, used to hang next to Cuyp’s work, in the house of the wealthy Dordrecht wine merchant Abraham Sam.12 Bredius argued that the paintings cannot have been of a too inferior quality if they were presented as pendants of Cuyp’s work. He suggested that Farrington might have been an apprentice of Aelbert Cuyp but he did not provide evidence for this (Bredius 1913, 3; 1917b, 1391).

Another painting by Richard Farrington is mentioned in the inventory of the Dordrecht painter Pieter van der Hulst (1651-1727) and his mother Maria van Rommerswael. Her inventory list of April 2nd 1674 recorded a rich collection of paintings, among which Een lantschap (landscape) by Richard Farenton and Koeyen (cows) by Aelbert van Kuijp (Bredius 1910, 13; 1917b, 1371-3).

Thieme and Becker published Bredius’ findings in their lexicon of artists and added information about three more paintings in the Dordrecht collections of Jacob Floryn (two landscapes in 1670) and Adriaen Huybertsz Verveer (one painting in 1660) (1915, 280).13

In Grant’s treatise about English landscape painters from the sixteenth to the nineteenth century, a painting is depicted by R. Farington, being in possession of a certain

F. Sabin of the Galleries in Cork Street (figure 1.3). Grant described the artist as an

“unknown but obviously talented member of the family”, referring to eighteenth-century painters such as George and Joseph Farington (Grant 1959, 274).14 Just the size of the painting is mentioned but not its title or dating. A few years later, the same painting appeared in a catalogue of English pictures and drawings of the auction house Christie, Manson & Woods and was sold at auction on 16 June 1961 in London. The title of this painting signed by R. Ffaringdon was A River Landscape with Monks and Anglers.15

12_{The inventory of Abraham Sam, Dordrecht 1692, mentions Een groot lantschap and Een lantschap kasteel}

by R. Faringhton (Bredius 1917b, 1394).

13_{This information obviously derived from personal communication with Bredius, as mentioned in their}

reference list, since it could not be found in the other sources.

14_{Bredius also remarked that later descendants of Farrington became well-known painters in England (1913,}

3; 1917b, 1391). Thieme & Becker mention the brothers Joseph and George Farington in the same context (1915, 280).

15_{It clearly concerns the same painting, as in the auction description a reference is given to the image in}

(14)

Through the ages, Farrington’s paintings in the above-mentioned old Dordrecht collections have been dispersed due to inheritance and auction.16 In his study about Netherlandish painting of the seventeenth century, Gerson argued that he could not discuss Richard Farrington’s art, as not a single painting had reappeared until then (Gerson 1983, 401).

On 29 January 2010, the painting River Landscape with Hunters and a Harbour by Richard Farrington was put up for auction at Sotheby’s New York (figure 1.4). It did not sell but almost two years later, on 12 December 2011, it was bought by the Dordrecht Museum from the auction house Dorotheum in Vienna. The auction catalogue mentioned that the painting derived from a German private collection but stated that nothing more was known about its provenance.17

16_{Personal communication of Sander Paarlberg, 4 December 2016.}

17_{An effort was made to trace back the painting’s history via its former owner. Unfortunately, due to privacy}

regulations, Dorotheum could not disclose their client but responded: “We were informed by the vendor, that the painting was given to him by his parents who appear to have bought it in a German Gallery.” (Personal communication of Camilla Tinnacher, Expert Assistant Old Master Paintings, email 9 March 2017)

Figure 1.3 Black and white picture of A River Landscape with Monks and Anglers by R. Farington in M.H. Grant’s treatise (1959, 274).

(15)

The signature on River Landscape clearly shows similarities with Farrington’s signatures on the various notarial records collected by Bredius (figure 1.5). Thieme and Becker

remarked that the artist, “strangely enough,” signed the documents with Ffarrington (1915,

280).18 However, in comparison with the painting’s signature it is perfectly clear that the

supposed double letter ‘f’ was Farrington’s specific way of writing a capital ‘F’. Several

differences can be noticed between the signatures from 1657-1658 and the one from 1662. In the latter, the ‘R’, has a long stroke, the ‘h’ has lost its stroke and the curl of the ‘d’ changed direction and is connected to the ‘F’. The 1662 signature is most similar to the one

on the painting, which may give an indication of the date of creation.19_{This would, in turn,}

imply that Farrington continued to paint during his work as a wine merchant in Dordrecht, and maybe until his return to England circa 1664.

18_{“Er unterschreibt sich merkwürdigerweise Ffarrington.”}

19_{The differences in style of writing between the signatures on the document and the one on the painting can}

obviously be explained by the differences between the used media, ink and paint.

Figure 1.4 Richard Farrington. River Landscape with Hunters and a Harbour. Ca. 1648-1664. Oil on canvas. 83.8 x 103.5 x 2.2 cm. Dordrecht. Dordrecht Museum.

(16)

1.4. The influence of Aelbert Cuyp

From the descriptions in literature and the painting in the collection of the Dordrecht Museum, it appears that Farrington was a painter of landscapes. During the seventeenth century, landscape painting prospered as an independent genre in the Dutch Republic. Since the art market encouraged painters to specialize, most of them were inclined to stick to a certain subject and repeat their most favoured designs with only slight changes (Brown et al. 1987, 60). Paarlberg argued that Farrington’s landscape is highly influenced by the late work of Aelbert Cuyp, the great interpreter of the Dutch landscape in the Italianate manner (Paarlberg 2013, 280). Cuyp, in turn, was influenced by contemporary landscape painters active in Utrecht, particularly Jan Both, and started to exploit their southern lighting effects in his landscapes from the late 1650s onwards (Spring and Keith 2009, 71).

24 April 1657

20 May 1662

Figure 1.5 The signature on the painting exhibits similar characteristics as the signatures on some notarial records. 5 May 1658

Infrared photograph (830-900 nm) of the signature on the painting. Photograph: Lidwien Speleers (Nikon D3300 camera converted to full spectrum)

(17)

This supports a late dating of Farrington’s landscape, as suggested earlier in connection with the comparison of his signatures.

Besides creating light effects of the setting sun, Farrington tried, like Cuyp, to capture the harmony of humanity and nature in his landscape. The diagonal composition of the hilly scenery along the waterside with a tall tree on the right and bushes in the foreground, shepherds with their livestock and the figure in red dress, are well-known motifs in Cuyp’s later work (figures 1.6-7) (Paarlberg 2013, 280).

Cuyp spent his entire artistic career in Dordrecht and was almost completely unknown outside this city in the seventeenth century. His landscapes and equestrian portraits were essentially created for a local market of prosperous citizens. The paintings were also popular among wealthy patricians and merchants such as Abraham Sam (see above), who owned at least four large landscapes of Cuyp (Chong 2001, 40-42). It is therefore possible that Farrington knew Cuyp’s work from the collections of his fellow wine merchants.

Aelbert Cuyp’s style was imitated by other painters during his lifetime but it is not known whether he had an active workshop with apprentices or assistants. One of the brothers Van Calraet, probably Abraham, is thought to have been an apprentice of Cuyp (Paarlberg 2013, 276). Farrington was clearly very familiar with Cuyp’s painting, but given certain differences in style, he must have been a contemporary follower, rather than an assistant in Cuyp’s (supposed) workshop, as stated by Paarlberg (2013, 280-2).

It is not known where Farrington and Cuyp bought their artists’ materials, but they probably obtained them from an apothecary or grocer. In the second quarter of the seventeenth century, some groceries in the major cities of the Dutch Republic started to specialise in the trade and sale of artists’ materials. In Dordrecht, Cornelis Bolenbeek was the local dealer who offered almost all painting materials, including ready-made paint and primed canvases (Henny 1994, 48-50). Stols-Witlox quoted a number of authors who stated that by the seventeenth century, the priming of canvas was mostly carried out by

Figure 1.6 Richard Farrington. River Landscape

with Hunters and a Harbour. ca.1648-1664.

Oil on canvas. 83.8 x 103.5 x 2.2 cm. Dordrecht, Dordrecht museum.

Figure 1.7 Aelbert Cuyp. The Flight into Egypt.

ca.1655. Oil on canvas. 67.9 x 90.8 cm. Los Angeles,

Los Angeles County Museum of art. Source: www.lacma.org

(18)

professional primers (Stols-Witlox 2012, 170). An archival record from 1655 indicates that Farrington had to pay a debt for the priming of several canvases (Bredius 1917b, 1392). It is therefore well possible that he used similar ready-primed canvases like other artists such as Aelbert Cuyp. To investigate possible similarities, a comparison will be made between the grounds used by Farrington and Cuyp, in paragraph 2.5.

1.5. Discussion

On the basis of archival records, the researched literature and River Landscape, an image can be formed of the unknown Richard Farrington and his life in Dordrecht. The marriage certificate and the early archival deeds provide a strong indication that he came to Holland as a painter. Almost ten years later, his activities seem to have focussed on wine trade, possibly out of necessity to support his family.20Since Farrington can be considered as a

contemporary follower of Cuyp, whose Italianate style developed from the late 1650s

onwards, an earlier creation date of the painting is very unlikely. Also based on the described similarities between signatures, the painting possibly dates from the 1660s.

Bredius’ archival finds showed that Farrington lived in the city for about sixteen years and that he was engaged in diverse social and economic circles related to wine trade and art. Given the size of Dordrecht and the fact that the many artists who worked there halfway the seventeenth century lived relatively close to each other, it is highly likely that Farrington had contact with other painters. Although it is not certain whether he knew Aelbert Cuyp in person, he was certainly familiar with Cuyp’s work.

Farrington probably did not leave a major oeuvre, but the future display of his painting in the Dordrecht Museum may lead to more recognition of the minor master and eventually to the rediscovery of more of his paintings.

Research into the historical context did not bring clarity about the cause of the observed degradation in River Landscape, as its history could not be traced down and no other paintings by Farrington have been found. Since the investigation showed that most traces of the past can be found on the painting itself, these will be thoroughly examined in the next chapter.

20_{This contrasts with the current entry on the RKD website: “Farrington was a businessman (a}

(19)

Chapter 2 Examination of the painting

In order to find explanations for the observed degradation phenomena, it is essential to know which materials are present in the painting. To determine the material composition, the painting technique and layer build-up of the painting, various analytical techniques have been applied. Starting point of the research was close examination of the painting with the naked eye, a head loupe, a Dino-Lite microscope and a stereomicroscope. The painting was examined in visible light, raking light and ultraviolet light (appendix I). To identify materials, canvas fibre analysis with polarized light microscopy and analysis of paint cross-sections with optical microscopy and Scanning Electron Microscopy combined with Energy Dispersive X-ray Spectroscopy (SEM-EDX) were carried out (appendices VII-IX). Insight into the material aspects of the painting will also help with the dating of the lining treatment and other interventions, which took place in the past.

2.1. Support

2.1.1. Canvas: material and weave properties

The painting was executed on a canvas support of plain-weave linen fabric. At an unknown moment in the past, it underwent a lining treatment. The lining canvas also consists of plain-woven linen.21 In the seventeenth century, artists’ canvases were traditionally made from linen, which was manufactured from the bast fibres of flax plants. As a result of its superior qualities of durability, linen has been used for painting supports for centuries and was also commonly used for lining canvases (Villers 2003, 157, 163).

Seventeenth-century canvases were hand-woven and generally have a coarser and more irregular structure than machine-woven canvases from post-industrial periods (Henny 1994, 44). Although the original tacking margins of the painting were cut off during the lining treatment, the abraded right edge reveals parts of the original canvas, which is clearly coarser than the lining canvas. Since no X-radiograph of the painting was available, a sound thread count, which is a measure of the weave density of a fabric, could not be executed. On two locations along the edge of the original canvas, 8 yarns per cm were counted in horizontal direction (figure 2.1). These yarns are about 0.5 mm thick and probably indicate the warp direction of the fabric since in vertical direction the threads are about twice as thick, which is more usual for filling (or weft) threads.22

On eight different locations on the lining canvas, a manual thread count was carried out (appendix II.ii). The average number of yarns per centimetre in horizontal direction is 15. In vertical direction, the average number of yarns per centimetre is 12 (figure 2.2). Since the number of yarns and the thread thicknesses in both directions is similar, the weave can be considered balanced (Collier, Bide and Tortora 2009, 274). The weave

21_{Fibre analysis was carried out with Bas van Velzen, Lecturer MA Book & Paper Conservation at the UvA,}

on 6 March 2017. Fibres of both canvases were examined with the Herzog test and could be identified as flax fibres (see appendix VII).

(20)

direction cannot be determined with certainty, as the lining canvas does not have a selvedge. However, particular weave faults occur in the vertical direction, as occasionally a pair of linked yarns run together over some distance or throughout the entire height of the canvas (figure 2.3). This phenomenon was described by Van de Wetering as a criterion to distinguish the warp direction of linen weaves in paintings (2009, 267). Moreover, little knots are visible, which suggest that broken threads have been reconnected. According to textiles conservator René Lugtigheid, the observed weave faults may indeed be caused by the breaking of warp threads during the weaving process. This suggests that the warp of the lining canvas is positioned in vertical direction relative to the painting. Although during a lining treatment, the weave directions of the original and the lining canvas should be aligned, as stated by Ackroyd (1995, 85), the warp of the original canvas is probably in opposite direction.

Figure 2.1 Thread count of the yarns in horizontal direction of the original canvas, which is visible on the right edge of the painting. The horizontal yarns (vertical in the picture) are about half the thickness of the filling yarns. (detail, see overview appendix II.i)

Figure 2.2 Tread count of the yarns in vertical direction of the lining canvas on the reverse of the painting in the right lower quadrant.

Figure 2.3 Weave fault in vertical direction of the lining canvas. The irregularity occurs when a warp thread breaks and two adjacent warp threads are woven together over some distance. (detail, see overview appendix II.ii)

(21)

2.1.2. Stretcher: materials and construction

The lining canvas is attached to a wooden stretcher with a margin of the fabric folded over to the side edges. The tacking margins are attached with 168 ferrous, machine-cut nails of two different types. The heads of 118 nails that have been attached during the lining treatment are irregular in thickness, size and shape. These nails closely resemble a type that was manufactured around 1825-1830. Machine cut nails produced after ca. 1830 cannot be easily distinguished from those made nowadays, since they were all cut and headed uniformly (Goist 2004, n. pag.). Since there are clear differences in form and surface finish between the 118 nails in the painting and comparable modern ones, the first mentioned may be manufactured before 1830 (figure 2.4). The fact that the majority of the nails are rusty indicates that they have been exposed to moisture or high relative humidity in the past.

The stretcher is not original to the painting as in the seventeenth century only strainers were used as supports for paintings on canvas. Probably the painting was mounted onto the stretcher during the lining treatment. Based on her experience and knowledge about historical linings, Kate Seymour, Head of Education at Stichting Restauratie Atelier Limburg (SRAL), dated the treatment between 1800 and 1880.23

The stretcher is composed of four bars of stained coniferous wood with adjustable corner joints and two crossbars, which offer additional strength and stability (figure 2.5). The stretcher bars are connected to each other with closed bridle joints with mitred corners and keys, which is a type of mortise-and-tenon joint common to Italy (figure 2.6).

23_{Personal communication of Kate Seymour, SRAL (Art Conservation and Research, Maastricht), 12}

January 2017.

Figure 2.4 Right: The nails used to attach the lining canvas and the modern nails in the foreground both have irregular shapes, but the modern nails have rounder heads and a smoother surface finish. Left: illustration of a machine cut and headed nail as manufactured around 1825-1830, resembling the ones in the painting. Source: Table by Kim Cullen Cobb in Goist 2004.

(22)

The so-called Buck Type 2e stretcher has also been reported to have occurred in Russia (Buckley 2012, 158; Bockrath and Buckley 2007, n. pag.). In Hampel’s treatise about the restoration of paintings, the glue-paste lining of paintings is described and a similar stretcher type is depicted (figure 2.7). It was regarded as the best stretcher then available (Weimar, halfway the nineteenth century), both for new and lined paintings (Hampel 1846, 40-6).

The stretcher bars are bevelled in profile, which prevents the canvas being marked by the wood. The bevelling of stretcher bars did not occur prior to the nineteenth century. In addition, keyed stretchers were not widely used and machine-cut nails were not produced prior to the late eighteenth century (Buckley 2012, 150-3). Consequently, the stretcher and therefore also the lining treatment cannot be dated earlier than the nineteenth century.

The stretcher is in a good and stable condition. It bears four labels from the auction houses Dorotheum and Sotheby’s and one old paper label with a number (figure 2.6). The limited number of labels suggests that the painting stayed in one place for a long period of time.

Figure 2.6 Left: right upper corner of the stretcher with closed bridle joint with mitred corner and keys. Right: Buck Type 2e, which is a simple type of mortise-and-tenon joint as encountered in the stretcher of the painting. Source: Bockrath and Buckley 2007.

Figure 2.5 The stretcher is composed of coniferous wood with adjustable corner joints and two cross bars.

(23)

2.1.3. Format change

The vertical crossbar is currently placed asymmetrically in the stretcher, which indicates that a format change may have taken place. Close examination reveals that the joints at the left side have been adjusted and do not fit as well as the joints at the right side. Moreover, pencil marks to indicate the new position of the left vertical stretcher bar are visible on the horizontal bars (figure 2.8). Assuming that the crossbar was originally placed in the centre of the stretcher, it can be deduced that 12.2 cm was cut off and that the stretcher’s original width must have been 115.7 cm (appendix III.i).

A strip of the original paint surface of about 1.5 cm wide is folded over the adjusted edge of the stretcher and attached with nails.24 At this side of the painting, the tacking margin of the lining canvas is absent (figure 2.9) (appendix I.iii). This indicates that the format of the painting was changed at a certain moment after the lining treatment was carried out and that circa 10.7 cm of the painted surface may have been cut off. On this side of the painting, ten old nail holes are visible on the upper edge, and eight on the lower edge of the tacking margins. Probably the lining canvas was partly removed from the stretcher during the format change. After the adjustments, the canvas was reattached with 50 new nails, which are less irregular and have larger and thinner heads than the old nails described above (figure 2.10).

When the painting was purchased at the auction house, it was mounted in a frame. According to frame conservator Hubert Baija, it is a late nineteenth-century frame with Louis XIV elements and ornaments, which are inconsistent copies of the original style

24_{When the painting was purchased, a wooden slat covered the right side edge of the painting, which was}

attached with 7 nails. The slat was removed in the conservation studio of the Dordrecht Museum on 6 February 2017 during examination of the painting.

Figure 2.7 Illustration from Hampel’s treatise depicting a similar stretcher as used for the painting after the lining treatment. Source: Hampel 1846, 43.

(24)

(appendix I.vii).25 The reverse shows a non-solid, and therefore wood saving construction, which is characteristic for frames from about the 1870s onwards. The open construction was adjusted to the painting’s size by removing the left batten and moving the top batten upwards in order to enlarge the rebate of the frame (appendix I.viii).

In summary, at some point in history, but not earlier than at the end of the nineteenth century and after the painting had been lined and attached to its current stretcher, the format of the painting was adjusted to fit the mentioned frame.

25_{Personal communication of Hubert Baija, Senior Conservator of Frames and Gilding at the Rijksmuseum}

Amsterdam, 21 January 2017.

Figure 2.8 Pencil marks on the horizontal stretcher bar indicate the new position of the left vertical bar. The new joint does not fit as well as the joint depicted in figure 2.6. A label of the auction house Dorotheum is visible on the top left side.

Figure 2.9 Top: Left upper corner of the painting showing the tacking margins of the lining canvas folded over the edges. Bottom: Right upper corner of the painting showing a strip of the original paint surface folded over the edge of the stretcher and used as a tacking margin.

Figure 2.10 Two old nail holes and two new nails are visible on the left side, indicating that the canvas was partially detached during format change. Three old nails, which date from the time that the painting was lined, are visible on the right side. Note the differences in appearance between the two types of nails.

(25)

2.2. Preparatory layers

2.2.1. Sizing layer

In the seventeenth century, the binding medium for grounds on canvas was frequently oil. To protect the canvas from the corrosive action of drying oil, an isolating size layer was usually advised (Stols-Witlox 2012, 173). In the lacunas, a shiny layer is visible on top of the canvas. This may be a size layer (figure 2.11). Size layers were also applied to lay down canvas fibres, fill the interstices and reduce the absorbency of the support and therewith the penetration of ground layers. In historic documentary sources, starch, oil and different types of animal glues are recommended as sizing materials, although oil is also described as damaging to the canvas (Stols-Witlox 2008, 148). Moreover, thick layers of glue and the use of too strong glues were discouraged, as these were said to result in cracks and paint delamination, particularly in humid environments (Stols-Witlox 2014, 235).

Since the colour of the original canvas is remarkably dark, it may have been sized with a drying oil, which darkens upon ageing. The shiny layer might also be the lining adhesive, which penetrated the canvas, or a mixture of lining and sizing materials. In almost all lacunas the canvas is visible, which indicates that delamination occurred at the interface between the canvas and the ground. Some paint samples, which delaminated from the canvas, were mount into cross-sections (appendix VI, VIII). At the bottom side of the paint cross-sections, a similar dark layer is visible, which can be seen to fluoresce in ultraviolet illumination (figure 2.12).

Figure 2.11 In the lacunas, the dark canvas is visible with a shiny layer on top, which may be a size layer. Left: lacuna in the sky area. Right: paint delamination on the right upper corner of the painting. Delamination obviously occurred at the interface between the canvas and the first ground. Dino-Lite photomicrographs in visible light. (details, see overview appendix II.i.)

(26)

2.2.2. Grounds

With a microscope, the preparatory layers can be seen in some areas where paint losses have occurred. The priming of the canvas consists of a beige first ground and a grey second ground on top (figure 2.13). In the seventeenth century, greyish single and sometimes double grounds were commonly used for canvas preparation in the Netherlands (Stols-Witlox 2012, 174). Paint cross-sections provide more information about the painting’s layer structure and confirm the presence of a double ground (figure 2.14). The same preparatory layers are visible in all paint samples (appendix VIII).

Figure 2.12 Microscopic image of paint cross-section DM-011-977-10a from a green-brown paint layer, in visible light (left) and ultraviolet light (right) (see appendix VIII.iv). A dark layer is visible on the underside of the paint cross-section, which shows fluorescence in UV illumination. The layer is similar in appearance as the dark layer on top of the canvas in figure 2.11.

Figure 2.13 Lacuna in the paint layer, revealing the dark brown canvas, a beige first ground and a grey second ground. Dino-Lite photomicrograph in visible light. (detail, see overview appendix II.i)

Figure 2.14 Visible light microscopic image of paint cross-section DM-011-977-3a from the sky area. The ground consists of a thick beige layer with a thinner grey layer on top (see appendix VIII.i).

(27)

In the first ground, oval-shaped particles are present, which have been identified as calcareous microfossils (figure 2.15).26 These fossilised creatures are present in chalk of natural origin (Spring 2009, 80). SEM-EDX analysis established that the first, beige ground contains a large amount of calcium (figures 2.16-17).27 These are indications for the presence of a chalk-containing ground. At the top and bottom of this ground, thin brown layers or discolorations are visible (figure 2.14 and 2.18). The BSE images in figure 2.16 reveal that the brown layers are part of the beige ground and the EDX spectra show that the same elements are present and about the same amount of calcium (see appendix IX.iii). This indicates that the brown layers are part of the chalk-containing, beige ground.

26_{Personal communication of Ineke Joosten, Specialist Conservation and Restoration at RCE, email 31 May}

2017.

27_{SEM-EDX analysis was carried out with Maartje Stols-Witlox, Lecturer and Coordinator MA Paintings}

Conservation at the UvA, on 21 March and 29 May 2017.

Figure 2.16 Backscattered electron image and detail of paint cross-section DM-011-977-3a from the sky area. EDX analysis was carried out in area 1 of the first ground and area 2 of the second ground, as well as on two particles in the second ground, pt3 and pt4 (see appendix IX.i).

Figure 2.15 Left: microscope detail in ultraviolet light of a cross-section from paint sample DM-011-977-4a. In the first ground, calcareous microfossils (small shells) are visible (see appendix VIII.ii). Right: SEM-BSE image of the same sample showing the oval-shaped particles in the first, calcium-containing ground.

(28)

The grey ground layer contains large white pigment particles, a black pigment and small orange-red pigment particles. The black pigment has splintery particles, which leads to the assumption that it is a plant black (figure 2.18). SEM-EDX analysis showed that the grey ground contains, among others, lead, calcium and carbon (figures 2.19-20). The black particle (pt3 in figure 2.16) contains a large amount of carbon and the white particle (pt4) contains mostly lead (appendix IX.i). Probably the grey ground consists of a mixture of lead white (basic lead(II)-carbonate), carbon (charcoal) black, chalk (calcium carbonate) and some red earth pigment.

Figure 2.17 EDX spectrum of area 1 of the first ground layer (pt1 in figure 2.16). The peaks in the interval 3.5-4.0 KeV demonstrate that, compared to the other elements, a large amount of calcium is present in this area. Carbon and oxygen are also present and the sulphur/ lead peak is low. This, in combination with the presence of calcareous microfossils (figure 2.15), indicates that the ground contains calcium carbonate (chalk) instead of calcium sulphate (see appendix IX.i for more details).

Figure 2.18 Visible light microscopic image of paint cross-section DM-011-977-3a, showing the first beige ground (1), the second grey ground (3) and the brown layer (2), which is part of the first ground. From the photograph it is clear that the grey ground consists of large white, splintery black and very few orange-red particles (see appendix VIII.i).

1

2

3

(29)

Figure 2.20 EDX spectrum between 6 and 16 KeV of area 2 of the second ground layer (pt2 in figure 2.16). Although the amount of counts is low, the peaks at 10.55 and 12.61 KeV indicate the presence of lead, which was also present at 2.34 KeV in the spectrum of figure 2.19.

Figure 2.19 EDX spectrum between 0 and 4.5 KeV of area 2 of the second ground layer (pt2 in figure 2.16). Among others, the presence of lead/ sulphur, carbon and calcium could be established. The peaks of Pb at 2.34 KeV and S 2.31/2.46 KeV coincide and both elements are present in this spectrum (see appendix IX.i).

s

(30)

2.2.3. Comparison with Aelbert Cuyp’s grounds

Comparison of the ground in Farrington’s landscape with the grounds Aelbert Cuyp used in his paintings is interesting because the availability of ready-primed canvases in the seventeenth century could result in both artists using the same commercial ground type (see paragraph 1.4). In that case, similar forms of degradation could occur in the paintings by Cuyp. A study by Sheldon et al. of eleven paintings by Aelbert Cuyp revealed changes in his use of pigments in paints and grounds throughout his career. It was found that the artist used single or double grounds, which were rather variable in their colour and pigment content (Sheldon et al. 2016, 118-20; Spring 2001, 65-73).28 The same was encountered in paint cross-sections from three canvas paintings by Cuyp in the collection of the Rijksmuseum Amsterdam. The painting Landscape with Cows and a Shepherd Boy (1650-1670) has a double greyish ground, which is similar to the ground on Farrington’s painting (figures 2.21-22).29 Noteworthy is the presence of brown layers at the top and bottom of the chalk ground in both samples. Of seven paintings on canvas by Cuyp, examined by Sheldon et al., three have double grounds of grey over red but no double ground has similar colours and pigments as the priming on Farrington’s painting and the Rijksmuseum Cuyp. Although Cuyp and Farrington may have purchased canvases from the same professional primer, no firm conclusions can be drawn in that respect.

Both the Rijksmuseum landscape by Cuyp and by Farrington received a glue lining treatment at an unknown moment in the past. The condition report of Cuyp’s landscape mentions tears in the canvas, fine craquelure and lifted paint layers, being accurately fixed by the lining.30

28_{Different grounds on seven canvas paintings were encountered: one double and three single beige grounds,}

and three double grey grounds. The grey grounds consist of a first red earth layer (red earth, chalk, lead white, one with umber, one with red lead) and a pale grey second layer (lead white, charcoal black, chalk and one with umber), which were common grounds in seventeenth-century Europe. The grey grounds all date from the late 1650s (Spring 2001, 67, 72-3; Sheldon et al. 2016, 118-20).

29_{The examination and pigment identification was carried out with Susan Smelt, Paintings Conservator of}

the Rijksmuseum, on 13 April 2017. Two paintings (SK-A-78 and SK-A-4118) have single, reddish-brown earth grounds and one (SK-A-3754) has a double, grey ground.

30_{The condition report of SK-A-3754 was drawn up on 26-10-1998 by Willem de Ridder and Michel van de}

Laar at the Rijksmuseum and mentions: “scheuren goed gerepareerd d.m.v. doublure”, “conditie en hechting verflaag matig/ fijn craquelé, gekrompen en opstaande verf: dit is gefixeerd met doubl.”

Figure 2.21 Visible light microscopic image of paint cross-section DM-011-977-5a (Farrington). The ground consists of a grey layer (lead white, charcoal black, chalk, an earth pigment) over a beige layer (chalk, possibly an earth pigment) (see appendix VIII.iii).

Figure 2.22 Visible light microscopic image of paint cross-section SK-A-3754-100-16 (Cuyp). The ground consists of a warm grey layer (lead white, charcoal black, chalk, an earth pigment) over a beige layer (chalk, an earth pigment) (see appendix VIII.v).

(31)

2.3. Paint layers

2.3.1. Materials

The paint layers of both light-toned paint samples of the sky contain large, angular-shaped particles surrounded by a white pigment (figure 2.23). SEM-EDX established the presence of, among other elements, lead, silicon, potassium and possibly cobalt (figure 2.24-25) (see appendix IX.i). This may indicate the presence of the pigment lead white mixed with smalt (potassium glass containing cobalt).

Figure 2.23 Photomicrograph in visible light of paint cross-section DM-011-977-3a from the sky area. The paint consists of angular shaped greyish particles in a matrix of white pigments (see appendix VIII.i).

Figure 2.24 BSE image of paint cross-section DM-011-977-3a from the sky area. EDX analysis was carried out in area 4 of the paint layer and two pigment particles, pt1-2 and pt3 (see appendix IX.i).

Figure 2.25 EDX spectrum between 0 and 4.5 KeV of area 4 of the light-toned paint layer in figure 2.23. Among others, the presence of lead/ sulphur, silicon and potassium could be established. (see appendix IX.i for the other spectra of this analysis)

(32)

From the cross-section taken of the tree, it is clear that the brown paint was applied on top of the light-coloured paint of the sky (fig 2.26). In the brown paint layers, greenish blue angular-shaped pigment particles are visible. SEM-EDX established the presence of copper, which probably indicates that a copper-containing pigment such as azurite or copper green is present (figure 2.27). This may have been used in combination with a yellow lake on a calcium substrate to obtain green, since calcium was also identified in the layer. Among others, smalt, lead-tin yellow and earth pigments are suspected to be present in the brown paint layers (see appendices VIII-IX for all analyses).

Figure 2.26 Microscopic image of paint cross-section DM-011-977-10a from a green-brown paint layer, in visible light (left) and ultraviolet light (right) (see appendix VIII.iv). The paint consists of three layers on top of the ground, which can be distinguished in UV light. Note the relatively large fossil in the first, beige ground.

1 2 3 pt1

Figure 2.27 EDX spectrum between 0 and 4.5 KeV of pt1 (see figure 2.26) of a blue particle in a green-brown paint layer. Despite the low amount of counts, copper peaks could be established around 0.9 KeV and 8.03-8.90 KeV, which indicate that a copper-containing pigment is present in the green-brown paint layer (see appendix IX.iv for the other spectra of this analysis).

(33)

2.3.2. Previous restorations

Besides the lining treatment, at least two restoration campaigns can be distinguished by the different colours of fillings and retouching paints used in the same area (figure 2.28). An intrinsic risk of lining is the flattening of impasto and canvas weave emphasis in the paint surface (see paragraph 3.5). In this painting however, no indication was found of any such distortions.

Many retouches are visible in the sky and on paint samples that were collected from the folded-over edge of the painting. These retouches must have been applied when the painting still had its original format, considering their position on the folded edge. From the cross-sections it can be deduced that zinc white (zinc(II)-oxide) was used for the retouches, as the white pigment can be seen to fluoresce in ultraviolet light (figure 2.29 and 2.28) (Eastaugh et al. 2008, 865). The presence of zinc could also be established with EDX (appendix IX.i-ii). The large-scale production of zinc white for oil painting started about 1850 (Kühn 1986, 171). Therefore, the retouching may have taken place from the second half of the nineteenth century, before the size of the painting was adjusted to the frame.

Figure 2.28 Microscopic detail of the left edge of the painting showing the lining canvas and the trimmed edge of the painting. Photographed with a Dino-Lite microscope in visible light (left) and ultraviolet light (right). Varnish residues and retouches can be seen to fluoresce in UV light (detail, see overview appendix II.i.)

Figure 2.29 Microscopic image of paint croos-section DM-011-977-3a from the sky area in visible light (left) and ultraviolet light (right). On top of the sample two layers of retouches can be distinguished. The upper paint layer can be seen to fluoresce in ultraviolet radiation, revealing the presence of zinc white (see appendix VIII.i).

Too Hard to Handle. Cupping Paint on a Seventeenth-Century Glue-Lined Painting.