UvA-DARE (Digital Academic Repository)

Historical recipes for preparatory layers for oil paintings in manuals, manuscripts
and handbooks in North West Europe, 1550-1900: analysis and reconstructions

Stols-Witlox, M.J.N.

Publication date
2014

Link to publication

Citation for published version (APA):

Stols-Witlox, M. J. N. (2014). Historical recipes for preparatory layers for oil paintings in
manuals, manuscripts and handbooks in North West Europe, 1550-1900: analysis and
reconstructions. [Thesis, externally prepared, Universiteit van Amsterdam].

General rights
It is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s)
and/or copyright holder(s), other than for strictly personal, individual use, unless the work is under an open
content license (like Creative Commons).

Disclaimer/Complaints regulations
If you believe that digital publication of certain material infringes any of your rights or (privacy) interests, please
let the Library know, stating your reasons. In case of a legitimate complaint, the Library will make the material
inaccessible and/or remove it from the website. Please Ask the Library: https://uba.uva.nl/en/contact, or a letter
to: Library of the University of Amsterdam, Secretariat, Singel 425, 1012 WP Amsterdam, The Netherlands. You
will be contacted as soon as possible.

UvA-DARE is a service provided by the library of the University of Amsterdam (https://dare.uva.nl)

Download date:27 May 2024
General introduction

17
Unless otherwise stated, translations of historical recipe texts have
been made by the author

18
General introduction

There is nothing, perhaps, on which the durability of a picture so much depends

as on the goodness of the ground

Merrifield 1849 5

The influence of preparatory layers on the stability of oil paintings is a subject that has
occupied the minds of many authors of artist’s manuals, recipe books and related
historical sources. ‘Goodness of the ground’, however, comes in many guises. Depending
on their location in time and space, artists have had different views on the subject and
have selected different materials and techniques to prepare their support for painting.

In their position between the support and the paint layers, preparatory layers influence
the texture of a painting. Their colour has an impact on the tonality of the finished picture
and their absorbency influences both painting technique and paint chemistry. The
materials employed in preparatory layers co-determine the ageing and degradation of the
painting.
Because of these facts, the technique and materials of preparatory layers are of interest
both to paintings conservators, conservation scientists and (technical) art historians. For
painting conservators and conservation scientists, information on the layer structure of
preparatory layers, on their composition and on the degradation of their materials are
particularly important. However, a good understanding of the intention of the artist and
the original appearance of the painting is also necessary, since these influence
conservation decisions.

In the past, preparatory layers have been investigated by a number of methods. Both
scientific examination of paintings and paint samples as well as research of written
documents have been carried out. Previous recipe-based studies have investigated smaller
areas. The present research for the first time investigates a large group of recipes for
preparatory layers, dating over a long period. Investigating a longer period allows for the
identification of trends and provides a context for individual recipes. This results in a more
profound insight into the meaning of these texts and raises their potential to inform
conservators, scientists and art historians.6 By including reconstructions as a research tool,
a link can be made between written texts and the materials and condition of actual
paintings.

The following paragraphs take a look at the approach taken by earlier researchers who
investigated preparatory layers and give a short description of the results of those studies

5
Merrifield 1849 (reprint 1999): cclxxxi.
6
The importance of investigating a large collection of recipes and long period was discussed in Witlox and
Carlyle 2005

19
that are most relevant to this thesis, before the focus is directed to the present research,
its goals and methodology.

1.1 Earlier research on preparatory layers for oil painting

Initially, research into preparatory layers was executed from two distinct angles. Either
written texts on the subject were studied, in very few cases combined with
reconstructions, or the materials of historical paintings were examined by taking samples.
Publications tended to relate the results of either type of research. A more integrated
approach that combined both types of research became more common from the middle
of the twentieth century onwards.

1.1.1 Research into historic documents

Publications in the field of documentary research, dating from the second half of the
eighteenth century or from the early nineteenth century, are relatively unknown and not
widely available, unlike results of the research of Eastlake (1847) and Merrifield (1849),
whose transcriptions with translation of historical recipe sources are easily accessed
through a number of reprints and are still consulted. 7 Eastlake and Merrifield focused
mainly on Medieval and Renaissance recipes, with some reference to more recent
sources. They were motivated by a wish to understand developments in Italian and early
Flemish painting methods that led to the ‘invention’ of oil painting, which Vasari (1550)
had placed with the Van Eyck brothers.8 The treatises they transcribed and translated,
contained recipes for preparatory layers and they discussed the materials and function of
these layers in their introductions.
The efforts of Eastlake and Merrifield were followed by other publications of historical
sources. The Quellenschrifte für Kunstgeschichte und Kunsttechnik des Mittelalters und der
Neuzeit, a series published in Vienna within the last decades of the nineteenth century
and during the early twentieth century, was intended to make available historic sources to
art historians. The published transcripts included a number of historical treatises on
painting technique, amongst which was Filarete (Antonio Averlino), who describes
preparatory layers. 9 Simultaneously, German painter and art academy professor Berger
(1857-1919) was examining historical manuscripts and was making reconstructions.
Between 1893 and 1912 he published an impressive number of exerpts from historical
manuscripts in his books on historical painting methods, which were of great importance
for the dissemination of historical texts, in particular to German speaking countries. The
manuscripts he studied included important Baroque sources like the ‘De Mayerne

7
Merrifield 1849, 2 vols., was reprinted at least twice: New York: Dover 1967 in two volumes; Mineola, New
York: Dover publications 1999 in a single volume. Eastlake 1847 was reprinted as: New York: Dover 1960,
in two volumes. See for a discussion on earlier documentary research Nadolny et al. 2012
8
Although Merrifield in her introduction refers to many more recent publications and includes interviews
with contemporary artists, with whom she had conversations about the painting technique of sixteenth and
seventeenth painting. Merrifield 1849 (reprint 1999). Eastlake after describing Flemish painting techniques
devotes a detailed section to preparatory layers, discussing their alleged absorbency. Eastlake 1847 (reprint
1960), vol. 1: 369, etc. The fascination of Merrifield and Eastlake with the ‘discovery’ of oil paint is described
by Nadolny 2005: 1028-1030.
9
Founded in 1871 by Rudolf Eitelberger von Edelberg.
(http://www.dictionaryofarthistorians.org/eitelbergerr.htm, accesed 26-4-2012). 18 volumes were published
between 1871 and 1908. Oettingen 1890, Neue Folge, vol III is the volume that contains Filarete’s treatise.

20
manuscript’ 10 and expanded into the eighteenth century with sources such as Pernety,
Watin, Buonanni, Dossie and Cröker, 11 all sources that include recipes for preparatory
layers. Between 1915 and 1922, Bredius published partial transcripts of artists’
inventories.12 In the 1930s, British chemist A.P. Laurie, specialized in the analysis of
paintings, published a number of books on historical painting techniques. He based his
ideas regarding the methods of earlier painters on several historical sources and included
partial transcriptions. However, the transcriptions he provided were all based on earlier
work carried out by Merrifield and others. 13

Since then, many historical treatises or recipe books have become available in facsimile
editions and annotated (partial) transcriptions and/or translations. 14 Several extremely
useful bibliographies with titles and short descriptions of manuscripts and published
manuals on artists’ techniques, represent an important step towards developing more
broad-based research of historical recipes on painting technique.15 The most recent
development in the availability of historical recipes is the establishment of online
resources that offer access to scanned documents as well as print-on-demand facsimile
copies of historical sources. 16

10
The ‘De Mayerne manuscript’ was discussed earlier by Eastlake, but without many of the details and with
fewer transcriptions than Berger published.
11
Important in the context of this dissertation are: Berger 1901 (facsimile 1986), Berger 1904, Berger 1909,
Berger 1912 (reprint 2000). Kinseher (2012) provides background information on Berger.
12
Bredius 1915-22.
13
A.P. Laurie, published Materials of the painter’s craft in Europe and Egypt, from the earliest times to the
end of the XVIIIth century (1910), Pigments and mediums of the old masters (1914), Painter’s methods and
materials (1926).
14
See for instance the transcripts of historical sources by: Van de Graaf 1958; Maclehose 1960; Hawthorne
and Smith 1963; Borradaile 1966; Olszewski 1977; Miedema and Meijer 1979; Beal 1984; Veliz 1986;
Lehmann 2002 (Müller); Bartl, Krekel, Oltrogge 2005, Frezzato and Seccaroni 2010 and the English
translation of De Lairesse 1707 (De Vries 2011). Also partial transcripts or transcripts of single recipes should
be mentioned, for instance those by Thompson 1956, Van Schendel 1958; Thomson 1960, Rajnai 1993,
Hermens/Wallert 1998. Examples of reprints that have found practical use in the present research are: Boltz
von Ruffach 1549 (reprints München 1913, Vaduz 1988); Stalker and Parker 1688 (reprint London: 1960);
Dupuy du Grez 1699, (reprint Genève: Minkoff 1973), Watin 1774, reproduced by Kremer in Eichstetten, no
date [1987?], Le Pileur d’apligny: 1779 (reprint Genève Minkoff: 1973), Watin 1788 (reprint Paris: Laget:
1977), Dutens 1812 (reprint Genève: Minkoff 1972), Jay 1817 (reprint Genève: Minkoff 1973).
15
Ogden 1947; Bordini 1991; Schießl 1999; Clarke 2001; Bentchev 2004; Zindel 2010. More restricted in
scope but worth mentioning is Harley’s bibliography of manuscripts on painting technique in the British
Museum, which was published in 1969. Although much more than a bibliography of historical sources,
Harley’s Artists’ Pigments 1600-1835 provides in her first chapters a detailed overview of manuscripts that
fall within the period under study, with an emphasis on manuscripts and published books in the English
language. She provides interesting background details on most of the sources mentioned (Harley 1970).
Bregnhøi 2003 provided a list of instruction books for house painters presently available in Denmark. She
suggests that these same manuals were available in nineteenth century and early twentieth century Denmark;
Nadolny et al. 2012 provides appendices which contain an annotated overview of historical treatises dating
from the Middle Ages until the end of the nineteenth century as well as where background information and
examples of guild regulations and artists contracts can be located.
16
For instance: http://www.books.google.com; http://www.retrobibliothek.de; http://www.archive.org;
http://www.gallica.bnf.fr; http://www.zvdd.de; http://www.vd18-proto.bibliothek.uni-halle.de.
At the Fachhochschule in Cologne, Germany, Doris Oltrogge has established a database that contains a large
number of transcriptions of medieval and early modern art technological recipes from manuscripts produced
in the German speaking domain. Oltrogge, Doris, ‘Datenbank mittelaltlicher und frühneuzeitlicher
kunsttechnologischer Rezepte in handschriftlicher Überlieferung, (http//db.re.fh-
koeln.de/ICSFH/forschung/rezepte.aspx, accessed 29-6-2012. No other similar efforts are known to the
author.

21
While historical sources, in particular recipe books, have become more available, only few
publications have focused on recipes for preparatory layers. A general overview of
materials and techniques employed in Western painting from the Middle Ages until the
end of the nineteenth century, compiled by Rousseau and Van Sonnenburg (1963), made
reference to historical recipes for preparatory layers assembled from previously published
manuscripts and books (those published by Merrifield, Laurie, etc.). The 1968 paper ‘The
ground in pictures’ by Hendy and Lucas, incorporated large sections of the 1963 paper of
Rousseau and Von Sonnenburg, but was accompanied by an extensive overview of
photomicrographs of paint cross sections, prepared by conservation scientist Joyce
Plesters. Although not referred to directly in the text, the addition of analytical data to
accompany the overview could be taken as a desire to establish a link between historical
documents and actual paintings. 17

The last thirty years witnessed only few publications dedicated to recipes for preparatory
layers. Talley’s (1981) publication on portrait painting in England between the end of the
sixteenth century and 1700, investigated the techniques and materials described in
historical recipes from the period, including recipes for preparatory layers. Talley
discussed 22 historical sources from within this period and provided a description of each
source in which he included information on its author, intended audience and scope, thus
showing an awareness of the importance of context. 18
Bosshard’s (1989) paper on nineteenth century recipes for binding media in grounds and
paint layers used tables to organize the historical information in chronological order and
included a large number of sources.19 Massing (1998) gave a detailed description of
recipes for preparatory layers in seventeenth and early eighteenth century French recipe
books, summarizing the information and analyzing and comparing recipes from different
sources to form an overview of the techniques mentioned. Massing spoke about the
relation between recipes and actual paintings. 20
Carlyle’s dissertation (1991) and book (2001) on nineteenth century British artists’
handbooks, manuals and treatises provided a detailed overview and analysis of the
information contained in these sources and includes a comprehensive annotated
bibliography of the sources consulted.
In Callen’s (2000) publication on Impressionist painting technique, information from
French contemporary sources (and selected earlier sources or foreign sources) went hand
in hand with research into supplier’s archives. On a small number of occasions, reference
was made to the actual use of certain materials and techniques in paintings. Callen
discussed the concepts of ground absorbency and ground colour and placed these in a

17
Hendy and Lucas (and Plesters) 1968.
18
Talley 1981 included results of analysis.
19
Bosshard, Mühlethaler 1989. The sources discussed date from the second half of the eighteenth century
until the early years of the twentieth century and include the Technische Mitteilungen für Mahlerei, an
influential Munich periodical on painting technique, both historical and contemporary. See Kinseher 2012 for
background information on this periodical.
20
Massing 1998. It is unfortunate that for some general descriptions of contemporary practice only very few
recipes were cited. The application of a cold, gelled glue size layer to canvas was presented as the standard
method, while this was backed up by little documentary evidence. Chapter 11 of this thesis discusses glue size
layer options in more detail, and draws the conclusion that they were more diverse than suggested by
Massing. (Massing 1998: 349-50). Massing also contributed to Nadolny et al. 2012.

22
historical perspective. 21 Raft (2006) presented an overview of seventeenth and eighteenth
century recipes for multilayered primings. 22
Mayer and Myers (2011) wrote an overview of the techniques employed by American
painters from the colonial period up to 1860 that is based on recipe books, journals,
manuals and artists’ letters. Their book includes valuable information on the preparatory
layers employed by these artists.23 In addition, research into the materials and techniques
of Vincent Van Gogh made use of the artist’s correspondence. Results include everything
the artist said about his use of supports.24

Information from written sources is also found in archives or historical collections of paint
materials. Both are potentially interesting sources of information on materials employed
in preparatory layers. The archives that have been investigated to extract information
relating to painting materials vary from historical collections of paint materials, 25 historical
account books 26 to apothecary taxes. 27 Nineteenth century collections and archives have
received much attention recently, for instance in Haaf’s (1987) study on nineteenth
century pre-primed canvas, in the investigation on the archive of the nineteenth century
colourman Roberson (Woodcock 1995, 1997), in Clarke and Carlyle’s publications on the
Winsor & Newton archive (2005a, 2005b) and in Barrett’s (2009) publication on the
Belgian paint manufacturer Blockx’s archive. 28

Not exactly written sources, but related to this type of source, are depictions of artists’
studios that include unfinished paintings with the ground visible (see cover image) 29, and

21
Callen 2000. Although Callen discusses some results of scientific analysis, her publication is not described
in the paragraph that discussed studies with an integrated approach, since the emphasis lies much more on
documentary research.
22
Raft 2006. Raft concludes that although emulsions of flour and oil were mentioned in some recipes, they
were only used from the end of the nineteenth century on. He does not provide proof for this statement.
23
Mayer and Myers 2011.
24
A discussion on Van Gogh’s letters has been included in some of the publications on his painting technique
that were described earlier in this chapter. In the Netherlands, some artists’ archives are kept by the RKD
(Netherlands Institute for Art History, http://www.rkd.nl, accessed 21-2-‘12). Recently, the correspondence
by Vincent van Gogh has been published online in an annoted version by the Van Gogh Museum Amsterdam
(http://www.vangoghletters.org/vg/, accessed 21-2-‘12).
25
In this respect the research of Ineke Pey (1987) on the Hafkenscheid collection of paint materials (c. 1800-
1830s) and by Lisa Wagner (2007) on the collection of apothecary materials owned by Dr. Vigani in
Cambridge (c. 1805) are of particular interest.
26
Studies of historical accountbooks have also proven to be an interesting source on painting materials. Haack
Christensen (2011) published on the seventeenth century accounbooks of the Royal Danish court in the early
seventeeth century and is currently working on a PhD dissertation on that subject, Haller 2005 has published
on the sixteenth century accountbook of Wolfgang Pronner.
27
A currently running project at the Doerner Institute in Munich is studying German apothecary taxes
(http://www.doernerinstitut.de/en/projekte/taxenprojekt/index.html, accessed on 6-2-12), see also Krekel and
Burmester 2003
28
Naturally in this context one should not fail to mention Rosamund Harley’s earlier work on Winsor and
Newton preprimed canvases made in the early twentieth century and on her handbook on historical pigments.
Harley 1987, Harley 1970. Callen 2000 includes a number of illustrations of primed supports dating c. 1906,
Bourgeois factory.
29
Kleinert’s (2006) publication on seventeenth century genre paintings depicting artists’ studios investigated
whether studios depicted in paintings can be considered truthful depictions of artists’ practice. She concluded
that although such images show elements of actual practice, we have to consider the tendency to simplify and
only show what are generally considered archetypal images of painter practice. Kleinert 2006. Jonkman and
Geudeker’s (2010) book on nineteenth century depictions (photographs and paintings) of artists’ studios
includes photographs of unfinished paintings. The authors discuss the fact that especially later in the century

23
Van Hout’s study on unfinished paintings. 30 Depictions of unfinished paintings can be
helpful in investigations into ground colour.

1.1.2 The evolution of source research: developments in source research methodology

The above cited publications appeared during a period that witnessed a growing interest
in research based on historical sources, in particular written documents. This growing
interest went along with the establishment of research groups and the development of a
more conscious approach to written documents containing information on painting
technique.

A decade ago, this growing interest in investigations of historical sources on art technique
led to the establishment of the Art Technological Source Research (ATSR) working group,
whose goals are: ‘to establish a forum for research on historical sources for artists'
materials and techniques; to systematize appropriate methodologies for this type of
research; and to provide an international platform for the dissemination of information
and research data’.31 Under the umbrella of the ATSR working group, now operating as an
ICOM-CC 32 working group, a number of symposia and workshops about historical source
research were organized. The electronic newsletter of this group circulates information
about the availability of information on historical sources.

The field of historical source research has evolved in the last decades, which is evident
from the number of papers that examine research methodologies. Some researchers have
raised the issue of the relevance of written recipes for actual painting practice or
discussed the possibilities and limitations in the interpretation of written sources.33 Van
de Graaf’s publication in 1963 was a forerunner in this respect. In this publication, Van de
Graaf demonstrated awareness of the importance of the context in which recipes were
written. He furthermore argued that it was important not to place too much emphasis on
single recipes but to study them in relation to earlier, contemporary and later sources and
to always attempt to relate recipes to actual painting practice.34 Carlyle published in the
late 1980s and 1990s on the subject of nineteenth century British artists’ handbooks and
manuals, and demonstrated how through careful analysis of the books, the context of the
source, the ideological basis on which authors wrote about painting practice and their
intended purpose could be incorporated in the interpretation of the texts they

more successful artists used two studios: a more formal studio or reception room in which they received
visitors and posed for photographs, and a second studio in which they executed their actual work. Jonkman,
Geudeker (eds.) 2010.
30
Van Hout and Huvenne 2012
31
The ATSR group was conceived by Ad Stijnman, then working at the ICN, Amsterdam (Netherlands
Institute for Cultural Heritage, now part of the RCE, Rijksdienst Cultureel Erfgoed), where it was founded in
2002. ATSR is now a working group under ICOM-CC. http://www.clericus.org, accessed 4-2-12. Quote from:
http://www.icom-cc.org/workinggroups/art-technological-source-research, accessed 4-2-12. Postprints of the
ATSR symposia were all published in London by Archetype.
32
ICOM-CC stands for International Council of Museums, Committee of Conservation.
33
Apart from difficulties with the interpretation of recipes, also the reason for transcribing and/or publishing
certain collections of historical recipes must be taken into account. Nadolny 2005 argues how both Eastlake
and Merrifield had a particular motive in publishing the recipes they transcribed: they wanted to find support
for their theory that during the Middle ages oil painting had been employed only for decorative painting and
outdoor painting, not as a standard binder for artistic paintings.
34
Van de Graaf 1962.

24
produced.35 Awareness of the importance of the social context within which sources were
written was also evident in publications by Massing (1994-2012). 36

The role and meaning of alchemical recipes, whose context necessitates a special
approach towards their interpretation, was the topic of Eamon’s (1994) publication on
‘books of secrets’. 37 Bucklow (1999-2011) published on the context and meaning of
recipes in alchemical treatises, focusing in particular on ‘impossible recipes’ that do not
lead to practical results but that make sense according to alchemic theory. 38

Discussions about the question of the practical value of written texts on painting
technique appeared in a number of publications, many of which focused on Medieval
treatises. Owen-Crocker (2009) wrote about issues encountered in the interpretation of
Anglo-Saxon manuscripts 39 and Clarke (2001-2011) published on the practical value of
Medieval manuscripts for contemporary artists’ practice. He argued that the practical
value can only be established by investigating authorship, intended audience, physical
appearance of the book/manuscript and by comparing the information contained in
different editions.40 In addition, a number of contemporary researchers included
discussions on their methods to establish the practical value of different editions and
copies of the same manuscript, again focusing on Medieval treatises. 41 Van Eikema
Hommes (2004) introduced a classification system in order to be able to group and
process historical information.42 Most recently, the overview of the state-of-the-art in
documentary source research, written by Nadolny et al. (2012), forms a very useful text
for those interested in research into documentary sources and provides some rules for
best practice.43

Nowadays, specialized researchers generally consider the context of a single textual

source, context meaning its author, the intended audience and the society that produced
the source. Researchers also recognize the problem of using historical texts in
transcription or translation and try to work with facsimiles, scans or originals. 44 These
factors will provide more depth to the interpretation of information contained in the
source itself. However, although specialized researchers employ this approach, it cannot
yet be considered standard practice. Two recent publications still mention
‘oversimplifications’, which are made because authors content themselves with a
relatively small number of easily accessible historical sources, as Nadolny et al. (2012)

35
See for example: Carlyle 1988, Carlyle 1990a, Carlyle 1990b, Carlyle 1991, Carlyle and Southall 1993,
Carlyle 1995, Carlyle 2001. In the introduction to her PhD thesis, Carlyle discusses the purpose and types of
sources published within this period. Carlyle 1991.
36
Massing 1995, 1998 Massing’s contribution to Nadolny et al. 2012.
37
Eamon 1994.
38
Amongst others in: Bucklow 1999, Bucklow 2000, Bucklow 2001, Bucklow 2009, Bucklow 2011.
39
Owen-Crocker 2009.
40
Clarke 2008; Clarke 2009; Clarke 2011a; Clarke 2011b: 19-24.
41
Villela-Petit 2006 ; Neven 2009; Neven 2011.
42
Van Eikema Hommes argued that classification of recipes according to their function can significantly
increase chances of their correct interpretation. In particular fragmentary information can thus be
complemented and compared with other recipes with a similar purpose. In this approach, also recipes from
different time periods and geographical areas may be included for comparison. Van Eikema Hommes 2004.
43
Nadolny et al. 2012.
44
Nadolny 2008: 7 provided the example of Thompson’s edition of Cennini. Thompson apparently was not
very consistent in his use of technical terms.

25
write. 45 According to Clarke (2008): ‘The same few sources are found to be used over and
over to answer all kinds of questions, even with reference to artefacts and practices far
distant in place and time from where and when those sources were compiled … If only a
few sources are known, no others will be cited, so no other will become well known. …
Researchers urgently need to make themselves aware of this range of material and to stop
using the same few sources, however good these sources may be’.46

1.1.3 Reconstructions for the investigation of preparatory layers

Historical recipes invite to be executed, and indeed a number of recipe-based studies have
extended their methodology to include reconstructions. The use of reconstructions to
understand the properties of paint materials and to understand the purpose of a layer
build-up is no recent development. Already Merrifield (1849) recorded her experiments or
reconstructions, carried out to help with the interpretation of recipes.47
Notwithstanding their potential, only few modern publications have employed
reconstructions to investigate the properties of preparatory layers. Brinkman (1993) made
reconstructions in an effort to shed light on the possible composition and function of the
isolation layer in Van Eyck’s paintings.48 He examined the working properties of a number
of binding media thought to have been employed, and based a hypothesis about the type
of oil that Van Eyck used to isolate his grounds on these reconstructions. Brinkman made
direct comparisons between his reconstructions and Van Eyck’s technique. His report on
the reconstructions does not mention awareness of possible differences between modern
and historic materials. Gypsum chemistry was investigated with practical experiments by
Federspiel (1995), Zillich (1998), Santos Gómez (2005) and Pombo Cardoso (2010). 49 These
authors all focused on the chemical reactions that take place when gypsum is burnt
and/or rehydrated for use in preparatory layers. They investigated burning temperatures
and the working properties of gypsum in different hydration states.
Van Laar (Keune 2011) based experiments with quartz grounds on Groen’s investigations
of the grounds employed by Rembrandt van Rijn. 50 Artist Van Laar actually employs similar
grounds for his own paintings; and also made a pictorial reconstruction of a Rembrandt
painting. Unfortunately, comparisons between his conclusions and historical practice are
difficult due to differences between the characteristics of the materials that he selected
and the materials available historically to Rembrandt.51

Around 1992-1993, Carlyle introduced the concept of ‘historical accuracy’ in

reconstructions, drawing attention to the importance of selecting historically appropriate
materials for reconstructions that are used to understand the properties and ageing
characteristics of historical materials.52 By investigating the historical production methods

45
Nadolny et al. 2012: 10
46
Clarke 2008: 20
47
Merrifield 1849 (1999): liv-lv described reconstructions of enamel colours; Ibid.: ccxxxv reconstructions of
recipes to purify and prepare oil for painting;
48
Brinkman 1993.
49
Federspiel 1995, Zillich 1998, Santos Gómez 2005; Pombo Cardoso 2010.
50
Keune 2011; Groen 2011b.
51
The paper describes the use of white and yellow pottery clay, river clay, fine sand, silversand or aquarium
sand. Keune 2011: 25.
52
This concept was introduced while executing varnish reconstructions at the Canadian Conservation Institute
during 1992-3, and developed further in subsequent reconstruction projects. The term ‘historically accurate

26
and material properties of materials employed for reconstructions, and subsequently
choosing materials that resemble such materials as closely as possible, Carlyle seeks to
produce reconstructions which are more comparable to the original paintings. She
recognizes that it is impossible to exactly recreate the circumstances in which paintings
were created in former centuries, compromises are unavoidable. However, attempts to
work with ‘historically accurate’ materials instead of the highly refined and purified
materials produced today will, according to Carlyle, result in reconstructions that may be
considered more similar to historical objects.
Carlyle herself executed ‘historically accurate’ reconstructions in a number of projects
with materials that were selected after a careful analysis of historical preparation
methods, amongst which were studies of nineteenth century flour paste grounds and the
grounds used by Van Gogh. 53 The reconstructions of Van Gogh’s grounds resulted in
practical observations about the handling and application properties of the materials and
about the colour and absorbency of ground types similar to those employed by Van Gogh.
Carlyle furthermore demonstrated how well-documented reconstructions may serve a
purpose as reference sets for instrumental analysis. 54

The influence of Carlyle’s approach is felt in reconstruction-based studies executed by

Vandivere (2011, 2013). Vandivere published on the influence of imprimatura isolation
layers on the visibility of underdrawings55 and worked as much as possible with materials
prepared according to historical methods, such as stack-process lead white. Although she
had to compromise in a number of instances, an awareness of the effect of these
compromises has been included in her conclusions.

Recently, Bucklow (2012) suggested replacing the term ‘historically accurate’ with
‘historically informed’ as, he rightly states, historical accuracy is impossible to achieve.56
The term ‘historically informed’ is nowadays recommended for the performance of period
music with contemporary instruments and following our present-day knowledge of
historic aesthetic tastes. 57 Carlyle suggests the term ‘historically appropriate’. This is the
term adopted in this thesis. 58

Reconstructions for didactic purposes, to inform the general public of painting methods or
to replace missing elements or decorations, were an important topic in the Studies in
honour of Renate Woudhuysen-Keller (2012). 59 For such reconstructions, attention to
historical accuracy is not as important as it is for those that are compared directly with the
results of scientific investigation into historical materials.

reconstruction’ was first used by Carlyle in a publication in 2001. Email from Carlyle dated 21-2-‘12. Carlyle
explains and applies the principle in a number of publications: Carlyle 2006; Carlyle et al. 2008; Carlyle
2012a, Carlyle 2012b.
53
Carlyle 2006; Carlyle et al. 2008a; Carlyle et al. 2008b
54
Carlyle, Boon, Haswell and Stols-Witlox 2008.
55
Vandivere 2011; Vandivere 2013.
56
Bucklow in Wrapson et al. (eds.) 2012: 26
57
See Butt 2002.
58
Carlyle, oral communication, June 2013.
59
Wrapson et al. (eds.) 2012

27
The use of reconstructions and reconstruction methodology is still in development. The
requirements that should be placed on reconstructions, reconstruction documentation
and reconstruction storage are also being explored. Carlyle (2012a, 2012b) describes
current views on the use of ‘historically accurate’ reconstructions.

1.1.4 The start of scientific investigations of preparatory layers

Nadolny (2003) places the beginning of the scientific examination of paintings in the late
eighteenth century. During the nineteenth century, a number of leading chemists of the
day, like Faraday and Chaptal, investigated paintings and analyzed their composition. The
question when oil was introduced as a binder for artistic painting, one of the main issues
that occupied the minds of nineteenth century art historians and chemists, was
investigated by swiping tests on paintings, 60 but also by analysis of paint samples.
Scientists examined properties such as solubility, melting point, smell, taste and reaction
to heat. 61 Nadolny’s 2003 publication suggests that the first analyses focused more on
paint layers than on preparatory layers, however her study shows that some scientific
analysis of the composition of preparatory layers was indeed published during the
nineteenth century. 62 Merrifield (1849) mentioned ‘experiments on the grounds of the old
Venetian pictures’ executed by ‘a Venetian professor’ who as a result of his experiments
considered himself able to conclude that these grounds were all made of gesso,
sometimes with a little black added, and covered with a glue isolation layer. 63 In Nadolny
(2005) we find how Eastlake had a paint sample analyzed by a chemist because he wanted
to know if pumice powder was present in the preparatory layers. 64

Wülfert (1999) dates the introduction of microscopy and microchemical testing to analyze
works of art in Germany in the second half of the nineteenth century. 65 Certainly, by 1929
when Dutch paintings conservator and chemist A.M. de Wild published his PhD research
on pigment microscopy and micro-chemical analysis of pigments in works of art, methods
were considered to have reached an academic standard. 66

Although some of the conclusions of nineteenth century investigations may still be valid, 67
early studies are, today, not usually referenced as scientific evidence. The earliest
publications on the analysis of preparatory layers for oil paintings that are still quoted as
scientific support in more recent publications date from the 1950s. That was the time
when Gettens and Mrose as well as Bones devoted papers to the analysis of calcium
sulphate grounds in Italian pictures and when Coremans, Gettens and Thissen published
their investigations of the painting techniques of early Flemish painting. 68

60
Swiping tests involve testing the solubility of the layer with different solvents. As solubility depends on the
nature of the material, such tests can result in a rough determination of the binder of the layer.
61
Nadolny 2003: 42.
62
Nadolny 2003: 44-47.
63
Merrifield 1849 (1999): cclxxxviii.
64
Nadolny 2005: 1031.
65
Wülfert 1999: 3-4.
66
De Wild 1928.
67
For instance some of the pigment identifications of Humphry and Haslam (Rees-Jones 1990).
68
Gettens and Mrose 1954. Gettens published results of chemical analysis of ground layers earlier, in a 1935
paper on an Italian painting, but this publication was not solely focused on grounds (Gettens 1935: 165-73);
Bones 1954; Coremans, Gettens and Thissen 1952.

28
The 1950s and 1960s saw a number of publications discussing the scientific analysis of
high-profile paintings.69 These publications were followed by several papers that
communicated the results of scientific analysis of paintings that fell within specific periods
or that covered certain geographical areas, like Rioux’s (1973) publication on French
seventeenth and eighteenth century red grounds.70

In contrast to the more limited scope of these publications, Plesters (1968) published an
overview consisting of photomicrographs of paintings dating from the Middle Ages until
the nineteenth century. Her synopsis, mentioned earlier in relation to the paper by Hendy
and Lucas on ‘grounds in paintings’, was the first scientific publication on the subject of
preparatory layers with such a wide scope. 71

Methods for instrumental analysis of pictures were evolving quickly and ‘new’ methods
were applied for instance to the study of lead white and calcium carbonate, important
materials in preparatory layers. 72 In 1976 the field was ready for a survey of methods
employed between 1961 and 1972 for the scientific investigation of painting materials.
This overview was written by Plesters, who included developments in documentary source
research and focused on an integrated approach that above all should not ‘lose sight,
behind the serried ranks of grey boxes of electronics, of the work of art’. 73

1.1.5 From the 1970s onwards: more combinations of recipe research and instrumental
analysis

Plesters was not alone in feeling the need for an integrated approach. A desire to combine
scientific analyses with the study of historical documents in order to answer questions
relating to painting technique, was evident in Kühn’s 1967 publications on lead white 74
and in Von Sonnenburg’s writings on Rubens’s painting technique (1979). 75 An interesting
study that compared information from a written source, Thomas Bardwell’s treatise The
practice of painting (1756), with the painting techniques used by this artist, was published

69
Coremans and Thissen’s publications on Van Eyck’s Ghent altarpiece included the results of the scientific
examination on the preparatory layers Coremans 1953, Coremans and Thissen 1962. See also Rees–Jones,
S.G. 1990. Also their 1962/63 investigation or Rubens’s Descent from the cross discussed preparatory layers.
Coremans and Thissen 1962, See also Phillipot and Phillipot 1963. In 1965 Kühn presented the results of
scientific analysis of ground materials used by Rembrandt. Kühn 1965. His investigation of Rembrandt
grounds was initiated to provide context for the quartz containing ground found in the Stuttgart self-portrait,
which authenticity was contested at the time. Paint samples were not usually mounted as cross sections during
this investigation, which limited the conclusions possible. Van de Wetering 1986: 18. In 1967 Kühn wrote
about Vermeer’s pigments and grounds. Kühn 1967. Kühn simultaneously continued publishing on
Rembrandt’s grounds (Kühn 1976, 1977). In these later investigations paint samples mounted as cross
sections allowed for a more precise determination of ground materials and layer build-up. Groen 2005b: 319.
70
Rioux 1973.
71
Hendy, Lucas (and Plesters) 1968.
72
See for instance Houtman and Turkstra 1964; Kühn 1966 and Kühn 1967; Coremans and Thissen 1959;
Keisch 1970; Keisch and Callahan 1973; Johnson and Packard 1971; Winter 1975. In 1986, Mairinger and
Schreiner described investigation methods for the investigation of preparatory layers (Mairinger and Schreiner
1986: chapter IX). The development of methods for paint investigation and the evolution of the profession of
conservation scientist in the Netherlands was described in Groen 2011a.
73
Plesters 1976: 5.
74
Kühn 1967.
75
Von Sonnenburg 1979 referred to the De Mayerne manuscript, De Piles and the manuscripts published by
Berger as historical sources on preparatory layers.

29
by Talley and Groen in 1975. The paper investigated the link between one artist’s writings
and his actual practice. The authors were able to conclude that although Bardwell did
follow the methods he described, his actual paintings tended to be less complicated than
the advice issued in his book. 76
In 1979, Miedema and Meijer published a study that incorporated visual observation,
documentary evidence from historical sources and previously published results of
scientific analyses, to investigate the transition from whitish to coloured grounds in
sixteenth century Netherlandish art. The authors counted instances of white versus
coloured grounds on panel and on canvas in a number of exhibition catalogues. By
complementing these results with the outcomes of earlier investigations on ground
colour, they were able to draw tentative conclusions about the transition from panel to
canvas in the Netherlands and to relate this transition to the introduction of coloured
grounds in the sixteenth century Netherlands. 77 Notwithstanding the limited data they
had available, many of their conclusions still stand.

The above and other publications were quoted by Straub and Koller in their thorough and
well–referenced reviews of the state-of-the-art of research on historical painting
techniques from the Middle Ages until the twentieth century, that were published in the
three volume Reclam’s Handbuch (1984). In their essays, both authors presented
chronological outlines of painting techniques, consistently combining the findings of
earlier scientific research with historical documentary evidence. 78 The authors listed and
in some cases briefly summarized individual historical sources. There was no room for
complete content of the recipe sources nor for the context in which they appeared.

Between 1984 and the present, more information has been published about the
preparatory layers of single artists and smaller groups of artists. 79 Some of these studies
are of particular importance to our topic. Papers on respectively Italian painting
techniques and artists’ materials between 1400 and 1550 and on seventeenth century
painter’s practices by Billinge (1997) and Kirby (1999), helped to build a context for the

76
Talley and Groen 1975.
77
Miedema and Meijer 1979.
78
Reclam’s Handbuch, 3 volumes, 1984, written by different authors amongst which Manfred Koller and Rolf
E. Straub. Volume 1 deals with historical painting techniques.
79
The following publications are used elsewhere in this thesis and not described in the main text of this
paragraph: Goldberg et al (1998) on Dürer’s techniques and materials; Pérez (1992) publised on Velázquez;
Martin et al. (2003) described the grounds of a number of painting produced in Rome and Bologna in the
1620s. Many publications exist on Rubens’s painting technique. The overview by Sonnenburg 1979 gives a
detailed description of Rubens’s procedures and includes descriptions of preparatory layers on both panel and
canvas; the techniques of Rubens’s paintings, including preparatory layers, are also described in Van Hout
1998; Van Hout (unpublished) 2005, Van Hout 2007, Van Hout 2008, Martin et al. 2005, Martin 2008. A
number of publications describe Van Dyck’s use of preparatory layers, amongst others: Christensen et al
1991; Roy 1999a and White 1999. Van Eikema Hommes (2012) describes the preparatory layers employed by
Ferdinand Bol; Van Eikema Hommes and Speleers (2011) the materials employed by painters who decorated
the Oranjezaal in Paleis Huis Ten Bosch; the grounds employed by Flinck, Jordaens, Lievens, De Groot, Bol
and Ovens for the the Royal Palace Amsterdam are described in Van Eikema Hommes and Froment (2011);
Costaras 1998, Levy-van Halm 1998 and Wadum (1998) discuss different aspects of Vermeer’s technique;
Duval (1992) wrote an overview of the preparatory layers employed by Poussin, Boucher, Coypel and French
contemporaries; Duval (1994) also examined the grounds of 26 paintings by Poussin; Hackney, Jones and
Townsend (eds.) 1999 share the results of the technical examination and instrumental analysis of a selection
of British paintings dating from the end of the sixteenth to the twentieth century, Townsend 2004 describes
the techniques of the Pre-Raphaelites.

30
recipes discussed in this thesis. Kirby’s paper includes a wealth of information on
documentary sources for this period. It discusses the different types of sources available
(artists’ handbooks, legal and commercial records, pharmacopoeias) and provides
background information on a number of these sources. The section describing
seventeenth century grounds, uses brief quotes from documentary sources (recipe books,
inventories, painter’s contracts) combined with the results of visual and scientific
examination of paintings. Documentary evidence on the preparation of panels for painting
focuses mainly on the practice in Antwerp, but Kirby’s description of the preparation of
canvases also discusses English and Italian methods.
Also Dunkerton and co-authors’ two volumes on Italian painting technique (1991, 1999)
are very relevant to this thesis. They share the results of the analysis of a number of the
preparatory layers of paintings from the collection of the National Gallery London and
place them in the context of historical and technical developments. 80 Historical recipes
play a smaller role in these overviews than the results of scientific investigations.

Besides publications that focus on a larger area and longer period, this PhD research also
benefits from prior research into the techniques and materials of single artists or within
smaller geographical areas. Preparatory layers employed by French seventeenth and
eighteenth century painters have been the subject of studies by Duval, Bergeon and
Martin. Their publications present a mixture of scientific data and historical recipes, with
an overall emphasis on the analysis. 81 A comparable study of ca. 100 Italian baroque
paintings has been carried out by Hamsík (1993), although the resulting paper
unfortunately is very brief. 82 Noble’s (2004) overview of the materials used in the grounds
of portrait paintings in the Mauritshuis provides an interesting review of the materials and
colours employed by a number of seventeenth and eighteenth century artists. 83
An important contribution to the field has been made by the Rembrandt Research
Project.84 A concise overview of Groen’s research on preparatory layers employed by
Rembrandt and contemporary Amsterdam and Haarlem painters was published in the
fourth volume of the Corpus of Rembrandt paintings (2005). 85 Publications by Hendriks on
the painting technique of Frans Hals and on grounds used by Vincent van Gogh in his
Antwerp and Paris periods are valuable sources of information within the context of the
present research. 86 Hendriks (2005) also wrote about the results of the scientific
examination of preparatory layers used by Haarlem painters active between c. 1550 and

80
Dunkerton et al. 1991; Dunkerton et al. 1999.
81
Bergeon and Martin 1994 provide an overview of French recipe books from the seventeenth and eighteenth
century and attempt to link these to results of scientific investigation of paintings; Martin 2008 builds on the
extensive analysis executed by Duval on coloured preparatory layers in French paintings and attempts to
group preparatory layers used on canvas between 1600 and 1640 based on the material composition of the
lowest layer only.
82
Hamsík 1993.
83
Noble 2004.
84
Van de Wetering 1997 describes preparatory layers on panel and canvas as used by Rembrandt.
85
Groen 2005. The investigations focused on the elemental composition of the preparatory layers, binding
medium analysis was not included as a standard method and was referred to once in the text. Groen 2005:
325.
86
Hendriks 1990, Hendriks and Geldof 2005; Hendriks 2006a and 2006b. Also the recent monograph on Van
Gogh’s studio practice (2013) contains information on the preparatory layers employed by this artist, in
particular on the preprimed canvas that he ordered from the Paris colourmerchant Tasset et l’Hôte and on the
quantative analysis by SEM-EDX of lead white and extenders in his ground layers. Vellekoop et al. 2013.

31
1700 in an essay on Haarlem paint materials.87 The nineteenth century use of India rubber
or caoutchouc in preparatory layers has been discussed by Carlyle 88 amongst others, and
was the subject of a detailed study by Labreuche (2011). 89
Similar studies into South European ground materials seem less frequent. The monograph
by Bruquetas-Galán (2002) about painting materials and techniques in Golden Age Spain is
an interesting and valuable exception. It is based on a wide variety of documentary
sources and links these to the results of paint analyses. 90

Interest in the analysis of preparatory layers of eighteenth century paintings has been
limited. The few publications that include information on eighteenth century ground
materials tend to focus on the first half of the century, in examinations that have an
interest in their relation to seventeenth century painting techniques. 91

The preparation of less frequently employed supports has been discussed in a number of
publications. First Van de Graaf (1976) and later Horovitz (1986, 2012) wrote on
preparatory layers for copper supports, based on painting investigations and documentary
evidence, 92 Reifsnyder (1999, 2012) published on painting on lapidary surfaces in fifteenth
and sixteenth century Italy. 93 Fötzsch (2008) wrote about the use of laminated sheets of
paper, cardboard, fibre boards in late nineteenth century studies by a German painter, 94 a
support that was also discussed by Cove (2004) in relation to Constable’s oil sketches.95 A
number of studies have focused in particular on the materials and techniques used for
preparatory layers in North America. 96

For investigations of preparatory layers also publications that investigate individual

materials used in these layers are important. Studies in this area that are interesting in the
context of this thesis include both general overviews such as the Pigment Compendium or

87
Hendriks and Geldof 2005.
88
Carlyle 1991, Carlyle 2001.
89
Labreuche 2011 discussed the materials and condition of paintings on India rubber ground, patents and
other documentary evidence.
90
Bruquetas-Galán 2002. Veliz 1998 also provides important documentary evidence and discussed results
from the investigation of Spanish panel paintings in an article that describes their material properties from the
Middle Ages until the seventeenth century.
91
One of the few publications with a specific focus on eighteenth century painting technique is Groen and De
Keijzer’s paper on the materials used in nine Dutch paintings dating from the first half of the eighteenth
century (Groen and De Keijser, 1996). Wallert 1999 discusses some eighteenth century flower paintings.
Reports on Eighteenth century Italian paintings have on occasion been published and usually contain a
description of their preparatory layers. (For instance: Bomford and Roy 1993 on paintings by Canaletto; Keith
1994 on a painting by Tiepolo). Duval and colleagues included French eighteenth century paintings in some
of their investigations (e.g. Duval 1992).
92
Van de Graaf 1976; Horovitz 1986; Horovitz 2012.
93
Reifsnyder 1999; Horovitz and Reifsnyder 2012.
94
Fötzsch 2008.
95
Cove 2004.
96
Although outside of the scope of the present study, they are mentioned briefly here: Fulton et al. 2005
published on the materials of Meade, Quimby 1974 provides an overview of techniques employed in
American paintings before 1776; Quandt 1971 and Quandt 1972 discusses grounds in 18th century American
pictures, Zucker 1999 in 19th century American paintings, Goldberg 1993 and Currie 1995 describe the use of
scored wooden panels and grounds in 19th century American paintings; Muller 1992 describes an early
plywood panel (1880) used by an American painter; Katlan 1992: 41-3 discusses the introduction of ready-
made plywood panel supports for paintings; Katlan 1999 provides an overview of 19th century technical
innovations for painters in America.

32
the Glossaire des Matériaux and investigations focusing on particular pigments, such as
the Artists’ pigments series. 97 Red and yellow earth pigments were singled out for
investigation in Grygar et al.’s study on the use of such earth pigments in the grounds of
Baroque paintings. 98
Recently, a number of new online resources has become available that share results of the
technical analyses of paintings, including information on preparatory layers: 99 the Cranach
digital archive, 100 the website established by the KIK/IRPA in Brussels with the reports of
technical investigations of the Ghent altarpiece,101 the Rembrandt Database 102 and the
online publication of the research project ‘Painting techniques of Impressionism and Post-
impressionism’. 103

1.1.6 The achievements and limitations of previous research

The above synopsis of previous research demonstrates that preparatory layers received
interest from the nineteenth century onward. Their importance was recognized by many
authors, who investigated their composition and function both through documentary
research and through the scientific examination of paintings and paint samples. It is these
investigations that have built our present state of knowledge on the materials and
techniques employed for preparatory layers.

Previous researchers had different motives to investigate preparatory layers. Throughout

the approximate 150 years of research on preparatory layers, the desire to understand
paint technological developments was an important motivation. It is evident in many
publications, all the way from Eastlake (1847) to the present day, and it seems to have
been the driving force behind the overviews of Plesters (1968) and Koller & Straub (1984),
and of monographs that deal with smaller geographical areas and shorter periods.
Reasons for wishing to understand technological developments vary, but in general we
can say that with knowledge of painting technique, a more multifaceted understanding of
the history of paintings was developed, and a context was created for the history and
material identity of specific objects.
The materiality of preparatory layers has also been studied in order to inform questions of
authenticity and chronology. This was one of the motives of, for instance, the Rembrandt
Research Project for analyzing Rembrandt’s preparatory layers and the grounds of his
contemporaries (see Groen 2005).

Has previous research provided enough information about preparatory layers to support
conservators, conservation scientists and (technical) art historians in their work? And also,

97
Artists’ Pigments, four volumes: Feller (ed.), 1986, Roy (ed.) 1993, West Fitzhugh (ed.) 1997, Berrie (ed.)
2007; Eastaugh et al. 2004; Guineau 2005.
98
Grygar et al. 2003.
99
In addition to these project, a research project at CITAR in Portugal, on painters from the North of Portugal,
will also result in an online resource for information on this topic . Calvo and Matos Lopes Pinto Leão Aguiar
2011.
100
http://www.lucascranach.org/relatedprojects_en.html, accessed 22-3-2013.
101
http://www.closertovaneyck.kikirpa.be/#home, accessed 5-6-2013.
102
http://www.rembrandtdatabase.org/Rembrandt, accessed 22-3-2013.
103
http://www.museenkoeln.de/ausstellungen/wrm_0802_impressionismus/info_e.asp?lang=en, accessed 22-
3-2013.

33
have the data that have become available through previous research been interpreted in
accordance with present-day research standards?
It is evident that our current view of preparatory layers for oil paintings depends more on
instrumental analysis than on the investigation of recipe books and other written
documents or on reconstructions. Some of the analytical results are of a relatively early
date and in particular information on binders and media may need to be re-evaluated.104
Only a relatively small group of recipes for preparatory layers has received general
attention, and their interpretation is not always abreast with current standards for recipe
study. In few cases, reconstructions have been applied to understand the practicalities of
preparatory layers, and only a small number of researchers have done so. Most
publications involving reconstructions of preparatory layers are of a relatively recent date.
While early reconstruction based studies do not discuss the material properties of
historical materials, more recent investigations, in particular those by Carlyle, make an
effort to employ materials that bear as close a resemblance to historical materials as
possible. The potential of such reconstructions to both enrich investigations into painting
technique, and as a means to study degradation mechanisms, has been demonstrated in
these studies.

Since Straub and Koller’s essays in Reclam’s Handbuch, some 23 years ago, no
comprehensive overviews of developments in the use of ground materials over longer
periods and larger geographical areas have been published. Straub and Koller’s overviews
are supported mainly by the results of scientific analysis, to which some information from
a small number of historical sources has been added. Although the reviews of these
authors are excellent, they are not up to date with current knowledge.
Studies from the past twenty years have tended to focus on local developments in artistic
centers or on the techniques of individual artists, sometimes revisiting, refining or
complimenting earlier research. As a result, small ‘islands’ exist of up to date and in-depth
knowledge on the preparatory layers employed in particular artistic centers or artists,
while limited attention has been paid to other areas. This is unfortunate, since often the
connection between these islands can only be made at a superficial level, which results in
a fragmented image.
The overview of current knowledge about the materials of preparatory layers that was
published as Stols-Witlox (2012), can be considered a recent effort in the spirit of Straub
and Koller. This publication is based on investigations of preparatory layers by previous
authors, with reference to a selection of the more easily available historical recipe
sources. 105 It demonstrates that sufficient information is available from previous technical
investigations for a more updated general overview of developments of the materials and
techniques employed in preparatory layers for oil painting in post-Medieval Europe 106.
However, while such a synopsis may provide valuable information to support research and

104
See Spring and Higgitt 2006, Higgitt and White 2005.
105
The chapter on preparatory layers appeared in a book for painting conservation students and professionals
that is presented as a ‘ … comprehensive text on the history, philosophy, and methods of treatment of easel
painting that combines theory with practice’. Hill Stoner and Rushfield (eds.) 2012: back flap. The aim of the
chapter was to provide a general overview of what the field at present believes to be the main developments in
the materials and layering of preparatory layers in European oil painting, and to guide the readers towards
more focused publications for more detailed information on certain aspects. Stols-Witlox 2012
106
Although not too much emphasis must be placed on earlier binding medium analyses, as recent research
has demonstrated its limitations. See for instance Spring and Higgitt 2006, Higgitt and White 2005.

34
conservation, it does not tell the whole story: Why did artists choose certain binders,
pigmentations and layer build-up? How were the ingredients that they used prepared?
Were minor additions included? And finally, are the materials we now detect in
preparatory layers the result of the artist’s intent or of earlier conservation inventions? All
these questions are highly relevant to the work of conservation scientists, practicing
conservators or (technical) art historians and can not be answered by instrumental
analysis.

The role of preparatory layers in determining the state of conservation of paintings has
rarely been the focus of earlier publications. If addressed at all, this topic was usually
mentioned only in passing. 107 Exceptions are found in a very small number of
investigations, like Labreuche’s study (2011) about paintings executed on India Rubber
grounds or Carlyle, Young and Alpine’s study (2008b) on the mechanical responses of
different types of grounds. Carlyle’s (1991, 2001) study of British painting handbooks
includes a section about artists’ views on the degradation of materials, with comments
about the role of preparatory layers.108

Outside of the period and area investigated by Carlyle, we do not know if artists were
aware of the long-term consequences of their use of materials in preparatory layers and if
they took these into account when choosing or preparing their preparatory system.

1.2 The central aims of this thesis

The development of conservation science has now reached a point where investigation
methods are sensitive enough to often detect trace components. Thus, the influence of
paint additives and minor components on paint chemistry, in this particular case on the
chemistry of preparatory layers, can be more effectively studied.109 As preparatory layers
influence the visual characteristics of a painting as well as its ageing behaviour, their
degradation and their influence on other layers in the painting are an important topic for
investigation.

Without knowledge of the original composition of the paint, the internal conditions of
degraded paint layers are difficult to understand. Detailed information about materials
that we know to have been used historically, helps with the interpretation of data
obtained from instrumental analysis. 110 Study of historical documents provides such

107
For instance in Vandivere (2011), the increased translucency of oil and lead white based imprimatura
layers only plays a small role in her chapter on the use of translucent flesh-coloured imprimatura layers. In her
section on the visibility of the underdrawing through the primuersel, Vandivere mentions the increase in
refractive index of an aged linseed oil but fails to mention the increased transparency of lead white containing
layers through saponification. Vandivere 2011: 83.
108
Carlyle 1991, vol. 1: 347-370; Carlyle 2001: 257-268.
109
The chemistry of aged oil paintings has recently been studied in research projects focusing on reaction
mechanisms that take place drying and degrading oil paints. Inside the Netherlands, I am referring to research
executed under the umbrella of the the MOLArt Project (1995-2002) and the De Mayerne Programme (2002-
2006). This work is continued in the PAinT project, part of the Science4Arts programme (2012-2016), all
multi-disciplinary programs sponsored by NWO. See http://www.nwo.nl for more information on these
programmes.
110
A further complicating factor regarding the interpretation of scientific data, is identified by Keune and
Carlyle (2005): paint components within the layer structure of a painting can move through the layer system.

35
information. It tells us which materials we should consider because of their availability
and/or the fact that they were advised for use in similar layers.
If scientific investigations are preceded or accompanied by in-depth research into
historical written sources, more accurate interpretations of the results of instrumental
analysis are within reach. Instrumental analysis needs to be complemented with research
that explores the origin of ground materials, that investigates preparation and application
methods and that focuses on the role of possible additives. This will lead to a more full
and multi-faceted understanding of these layers, which is vital for the conservation field.
For only if we know more about the starting point of the ageing process of paintings, can
we come closer to a full understanding of the paths of degradation.
Documentary research clearly has its limitations too, in particular regarding the availability
of sources and their interpretation, which is not always straightforward. Research based
on documentary sources does not lead to absolute certainties. But if both instrumental
analysis and documentary source research are carried out, the results can complement
each other. This relation between documentary and scientific research is of a symbiotic
nature.

Paragrah 1.1 has indicated that information on painting materials and preparation
methods has previously been sought and found in several types of historical documents,
such as bills for materials or lists of materials purchased, artists’ correspondence,
historical collections of (paint) materials, paint manufacturers catalogues, guild
regulations, historical recipes and artists’ inventories.

Of all these categories of historical documents, recipes are the only type of document that
not only lists materials, but provides systematical instructions on how to use these
materials in the creation of preparatory layers. Recipes form a separate category of
written records, and they are the category that this thesis will focus on, because recipes
provide the most complete answer to the question: which materials can be expected in
prepratory layers, how were such materials employed and why were they selected for
use?

The potential value of research of historical recipes as a means to create clarity about the
intended purpose of materials and as a means to inform the interpretation of analytical
results, has been proven in earlier studies. In particular in combination with
reconstructions, recipe-based research can significantly increase our understanding of the
use and degradation of painting materials.
Reconstructions play an important role in establishing a link between recipes and actual
painting technique. They lead to insight into the actual effects of materials and
procedures, and can be used to study the causes of degradation phenomena observed in
paintings. In some cases, chemical characterization by Instrumental analysis is executed to
support these investigations.

In this thesis, for the first time historical recipes for preparatory layers from a long time
period and large geographical area will be thoroughly and systematically investigated.

Keune analyzed the location of paint binder components in reconstructions executed by Carlyle. Binder
components were proven to move to other layers in the build-up, even in the presence of intermediary
isolation layers. Keune 2005: 74.

36
This investigation focuses both on the material characteristics, chronology of use and on
the reasons why they were advised for use in the preparatory system. The recipes are
investigated by textual analysis and by reconstruction.
The wide-angle approach taken, allows the researcher to take a step back from individual
recipes in order to identify trends. A second result of this approach is the fact that the
sometimes limited, unclear or incomplete information given in one recipe can be clarified
through this context. By combining information from different recipes, missing details can
be filled in and a more sound basis for conclusions is possible.

For this dissertation, the choice was initially made to gather recipes from a long period,
starting from around the time when the Van Eyck brothers were active (early fifteenth
century) to the late nineteenth century, and from a large geographical area, North West
Europe. However, as will be discussed in Chapter 2, which describes the size and scope of
the recipe collection, insufficient recipes were available from before 1550 to allow for a
recipe-based approach to the subject of preparatory layers. Therefore the decision was
made to take 1550 as a start date.

The value of recipe-based research into preparatory layers for conservation-related

research has been outlined above. But the outcomes of this investigation are also relevant
for practicing conservators. As knowledge about the composition and function of
preparatory layers increases our understanding of the chemical and physical state of a
painting, it influences and co-determines decisions for both active and passive
conservation.
Art technological investigations will benefit from information in the recipes about the
motives of artists to include certain materials in their preparatory layers. Recipe texts tell
us why artists chose particular ground colours or how they dealt with the working
properties of the different supports and grounds. Discussions on the advantages and
disadvantages of certain types of preparation shed an interesting light on what drove
artists, and how they reflected on their own practice.

1.3 Research methodology

As introduced above, this PhD dissertation reports on research that integrates the results
of three different approaches or research methods: the textual analysis of a large
collection of recipes, reconstructions of historical recipes and instrumental analysis.
Recipes form the backbone of the research and are the starting point of all investigations.

As a first step, a recipe collection is established that includes European recipes for
preparatory layers from c. 1400 to 1900. Simultaneously, research into secondary
literature on the topic of preparatory layer materials and on methodologies for recipe
study is executed. 111 As stated above, the boundaries of the period and area investigated
are determined by the scope of the recipe collection. Paragraph 2.1 describes the size and
scope of the recipe collection and explains why the decision was made to focus on North
West European recipes dated between c. 1550-1900.

111
The search strategy employed is explained in Paragraph 2.1.

37
As a second step, the contents of recipes are described in detail. Chronology and the use
of materials are investigated. Recipe texts are analyzed by a ‘thick description method’, 112
which investigates the recipe texts in detail while taking into account the context of the
recipe (source, author, location, period, relation to contemporary, earlier and later
recipes). Information about the professions of the authors of recipe books and the
audience they wished to address are included in cases where they provide important
information about the role of particular recipes.

These descriptions act as a reference guide for future investigations into preparatory
layers and form the basis of later chapters. From the recipe descriptions, a number of
themes that attract attention throughout the time period are selected for further study.
Topics are chosen for focused attention when the sources provide information that is vital
for our level of knowledge about workshop practice or when their investigation leads to
more insight into the degradation of preparatory layers. Some themes or questions that
are selected, focus on the characteristics of fresh grounds, while other questions concern
the aging of preparatory layers. Questions that fall within the first category are: the
working properties and the aesthetic effects of grounds, both in colour, texture and
saturation and the role of commercial ground preparation. Questions that focus on the
aged grounds, deal with the influence of specific pigments on the degradation of a
painting, the role of the ground binder or binding medium and the influence of ground
layering and thickness on the degradation of paintings.

Further study of these themes leads to more knowledge about the motivation of artists,
explains certain characteristics observed in preparatory layers and results in a more
thorough understanding of the material aspects of ground preparation, application and
degradation in general.

Throughout the thesis, bar graphs are employed to represent chronology in recipe
content. These bar graphs, such as for instance Figure 5.1, are a tool to analyze the recipe
groups, and clarify general developments regarding materials and layer build-up. In order
to produce recipe groups that are large enough to be meaningful, some generalization is
unavoidable. For instance slight ground colour variations cannot be accounted for, and the
the exact nature of an earth pigment is not visible in the bar graphs. While this is
unfortunate, grouping cannot be avoided, as without it, the meaning of the graphs is lost
in the multitude of variables. However, when reading these graphs, readers have to be
aware of this fact and are advised to refer to the transcripts of individual recipes which are
provided on CD for more detail.

112
‘Thick description’ was introduced as a qualitative research tool in anthropological research by Gilbert
Ryle in 1971. Ponterotto 2006 discusses the meaning of the term and provides a working definition, focused
on anthropological research but applicable to recipe research if instead of ‘social actions’ we think about
recipes (writing and/or publishing a recipe can be interpreted as a social action). Ponterotto’s definition is:
‘Thick description refers to the researcher’s task of both describing and interpreting observed social action (or
behavior) within its particular context.’… ‘Thick description accurately describes observed social actions and
assigns purpose and intentionality to these actions, by way of the researcher’s understanding and clear
description of the context under which the social actions took place. Thick description captures the thoughts
and feelings of participants as well as the often complex web of relationships among them’. Ponterotto 2006:
543.

38
Through reconstruction-based studies, links are established with actual paintings.
Reconstructions are executed in order to investigate the properties of different groups of
materials that are mentioned in the recipes. Reconstruction-based chapters investigate
the handling properties of the materials, their effect on the painting process and their
long-term effect on the ageing of the paintings. Such issues cannot be investigated solely
through instrumental analysis of aged and/or restored paintings or through the textual
analysis of the recipes only.
As in these cases a direct comparison between original objects and reconstructions is
intended, historical accuracy is of crucial importance and is pursued at the highest
possible level. Much emphasis is therefore placed on sourcing suitable materials.
If the highest level of historical accuracy cannot be achieved without personal danger to
the practitioner, or within a feasible time frame, or if insufficient data are available about
the original materials, compromises are unavoidable. An example of such a compromise is
the decision to use distilled or demineralized water. This decision has been made because,
even though water with this degree purity was not available in former centuries, it
ensures that the water used does not contain salts or acids that would not have been
present historically. 113 The choice for demineralized or distilled water furthermore raises
the level of reproducibility of the reconstructions, because it does not introduce materials
of unknown and varying composition. The pH of the water and the amount of dissolved
salts will however influence the chemical reactions that take place and possibly also the
characteristics of the final product, for instance during glue preparation 114 and during
washing of lead white pigment.115

It is important to realise and acknowledge that even if materials are located that can be
directly compared to the materials available in former centuries, we are operating in a
field with a large number of variables and unknowns. Therefore a reconstruction will at its
best approach the process of a particular artist in a particular situation, it can never claim
to be a perfect replica. In addition, as Carlyle writes, direct comparisons must always be
made with care because of another important difference: the difference between a
practiced workman in the setting of a studio or factory and the modern paint researcher
preparing small samples inside his chemical lab. 116

Instrumental analysis is employed to link the results of reconstructions to data from actual
paintings which has been obtained from the literature. Analytical methods employed are:
light microscopy, polarised light microscopy, Scanning Electron Microscopy (SEM-EDX) and

113
For instance in the Netherlands, rain water nowadays may contain ammonia (from fertilizers), NO3, SO4
(emission of SO2 from the industry), heavy metals (Fluor, Nickel, Zinc, Cadmium, Lead). See: Swaluw et al.
2010. Tapwater in Amsterdam, according to the report on water quality in the first three months of 2012,
contains trace amounts of aluminium, arsenic, fluoride, ammonia, nitrate, and larger amounts of chlorine,
magnesium, sodium, sulphates.
http://www.waternet.nl/media/407425/waterkwaliteit%20kwartaal%201%202012.pdf
, accessed 14-5-2012. Naturally water in former centuries would also have contained ionized salts, etc.
However the exact amount and composition would have depended on the period (think of the influence of
industrial pollution, etc.) and on the geographical area. By choosing demineralised or distilled water, a more
constant water quality is available, which raises the possitiliby of reproducible reconstructions.
114
For which a low salt content is advised, see Chapter 11
115
See Chapters 13 and 14
116
Carlyle discussed problems of downscaling briefly in Carlyle 2012b: 111.

39
X-ray diffraction (XRD). Microscopy has been performed by the author, SEM-EDX and XRD
analysis have been performed by or in cooperation with conservation scientists.

1.4 Structure of the dissertation

This dissertation consists of two main parts, followed by general conclusions. Part I of this
thesis concerns the textual analysis of the recipes for preparatory layers that have been
gathered. Part II assembles studies that employ reconstructions in order to understand
phenomena observed in actual paintings.

The first chapter of Part I, Chapter 2, describes the recipe collection, its size and scope. In
this chapter, the choice to focus on the time period 1550-1900 is made and explained. This
chapter furthermore contains an analysis of the character of the sources and their authors
and it discusses the possible impact of historical recipe books. Chapter 3 describes the
terminology employed for the description of the different preparatory layers in this thesis
and places terminology in a historical context. Chapter 4 provides a synopsis of
developments regarding ground layers during the time period preceding the main period
under investigation. It acts as an introduction to the main period. Chapter 5 consists of a
detailed description of the recipe content, ordered thematically and chronologically.
Chapter 6 discusses the materials that are mentioned in the recipes. It provides
information on the origin and preparation of these materials and describes their
chronology. Chapter 7 focuses on the application methods and handling properties of
preparatory layers. This chapter addresses the questions how preparatory layers were
applied and smoothed, why certain application methods were advised and how
application methods relate to texture of preparatory layers. Together, Chapters 5 to 7
function as a reference section on recipes for preparatory layers. Chapter 8, on the subject
of ground colour, investigates the role of ground colour in painting. Was ground colour
considered important, which ground colours were advised and for what reason? How did
discussions on ground colour change throughout the period? Was ground colour subject-
related? Did ground colours change with commercial manufacture? Chapter 9 focuses on
the role of commercial ground preparation. It traces the role of commercial suppliers of
prepared supports throughout the period, and investigates the relation between
commercially primed and artist-primed supports. Chapter 10 investigates comments on
the ageing and degradation of preparatory layers as they appear in the sources. It
discusses the influence of individual pigments, the role of ground colour and ground
absorbency, the influence of layer thickness on the degradation, and analyses advice in
historical recipe books to prevent degradation phenomena.

The recipe-based reconstructions discussed in Part II examine preparatory layer

characteristics and the influence of preparatory layers on the finished painting and on its
degradation. The subjects that are investigated through reconstructions are: the effect of
animal glue size layers on the properties of the preparatory system (Chapter 11), starch
and flour paste as size and ground layers (Chapter 12), lead white quality and the effects
of lead white washing on pigment composition (Chapters 13 and 14) and binders used for
streaky imprimaturas (Chapter 15).

40
Part II is followed by general conclusions, which discuss the main findings and identify
areas that require further study.

41