‘[B]efore we speak of the handling of colours in oil, it is no diversion from the subject if we say something about the priming, upon which, whether it is good or bad, depends the beauty & liveliness of the colours’, Theodore de Mayerne (1573–1655) wrote in the early seventeenth century.3 Priming, also called preparation or ground in other sources, is the term used for the layers that are applied to painting supports to prepare them for painting. Typically, the preparation of a traditional oil painting consists of the following layers: a size layer, often a glue, used to seal the support (usually wood, canvas or sometimes copper), followed by one or more layers of oil- or glue-bound pigments to fill irregularities and colour the support, and possibly a final thin isolating layer to regulate the absorbency of the image’s preparation or preparatory system.4

Artists used the ground colour to create a tonal harmony in their paintings. For example, the artist depicted by Vermeer in his Art of Painting (fig. 1), makes use of the grey ground. By leaving this grey uncovered in some areas and applying highlights and shadows, he uses the grey colour as a midtone. This tone would provide smooth transitions between light and dark passages and give a particular tonality to the painting. The ground’s colour co-determines the execution of a painting because an artist will employ different techniques depending on the colour of the ground. Had Vermeer’s artist painted on a white ground, this white could have formed the highlights in the painting, while the midtones and the shadows would have been applied in paint. It is difficult to imagine what a painting by Rembrandt would have looked like, had Rembrandt used a white ground for his highlights instead of the thickly sculpted application of lead white paint that is so characteristic of his style.

De Mayerne’s manuscript, Pictoria, Sculptoria et Quae Subalternarum Artium, is nowadays considered one of the most important original recipe sources on painting technique in the seventeenth century. De Mayerne noted down recipes that he obtained from artists at the British court, amongst others Netherlandish artists Peter Paul Rubens (1577–1640), Anthony van Dyck (1599–1641), Paul van Somer (c. 1577–1622) and Daniel Mytens (1590–1647).5 These recipes include a number of recipes for priming layers. They are practical instructions that describe the materials and their application, sometimes including notes about the desired characteristics of grounds and about the quality of the recipe.6 Such information adds a unique angle to research into artistic practices, and in this case, into the function and characteristics of the preparatory system. An example from the Dutch Huishoudelijk Woordenboek (1743) illustrates the value of recipes. Firstly, the recipe casts a light on workshop practices by describing the materials to be used and the steps the artist should take to prepare and apply a ground layer. Secondly, from the effort the author makes to describe a smooth and even surface for the artist to paint on, we can deduce the author’s ideas about the role of the ground layer. Thirdly, the recipe demonstrates that the author’s advice to artists was motivated by concerns about the longevity of the support, concerns that took into account both physical strength (resistance to tearing) and the influence of the ground on the long-term stability of colours that would be used in the painting:

While paintings have changed due to natural aging since they left the artist’s studio, and due to earlier restoration treatments, recipes like that in the Huishoudelijk Woordenboek highlight the qualities the author was seeking for in the freshly painted artwork. Recipes like these thus offer the possibility of bringing modern researchers significantly closer to understanding works of art as they were originally designed.

However, recipe study is not without problems and limitations, as the relation of written recipes to actual artistic practice is often difficult to establish. For example, how does the recipe that is published in a general dictionary like the Huishoudelijk Woordenboek, written for a general audience, relate to the daily practice of a specialized Dutch artist in his studio? In fact, the connection between this particular recipe and eighteenth-century Dutch practice is even more problematic. As will be discussed below, this eighteenth-century recipe is not an original, but has been taken and translated from a French source dated 1676.

This paper focuses on the relationship between historical recipes and artistic practice, between reading and painting. By examining the professional backgrounds of their authors, the authors’ stated motives for writing, their intended and actual audiences, this paper aims to assess the reliability of historical recipes as a source for learning more about historical artistic production. The paper builds upon earlier research performed by this author in which an extensive number of historic recipes on grounds in painting between 1550 and 1900 was collected, inventoried and stored in a custom designed a database.8 This database allows for a comprehensive view of the relationship between writing (author), reading (public) and painting (execution).The first section places the methodology followed in this paper in the context of recipe research within the fields of conservation, art history and the history of science, and introduces the corpus of recipes. In the second section, the professions of the authors and their motives for publishing these recipes are investigated. In subsequent sections, the contents of the recipes will be analyzed and compared, and the audiences of the ground recipes will be examined. Finally, a comparison between the grounds as described in recipes and grounds employed on actual paintings will be made, in order to evaluate the relation of ground recipes to painting practice.

Earlier research has formulated various theories about the role of recipes in artistic practice, and the relationship between recipes and artist-to-pupil instruction in a studio. Smith has pointed out that early modern recipe books were written for a variety of reasons, for example to collect information, as a ‘repository of memory’, but also to explain expert knowledge to laymen. In addition, writing may also have contributed to increasing an author’s social status. Smith stresses the significance of recipe books as places where the principles of artistic practices were discussed, principles that readers could use for their own improvisations.9 Smith believes that recipes may have played a role in ‘systematising and consolidating’ existing knowledge, but she considers much reprinted seventeenth and eighteenth century arts and crafts texts as being, ‘probably more about owning and connoisseurship than about making’.10 Prak agrees with Smith, but also considers the additional role that published recipes had as an active agent in increasing the knowledge of artisans with existing practical training.11 Clarke, in discussing the role of methods in the interpretation of historical sources, believes that ‘there is a greater correspondence between text and practice than is commonly appreciated’.12

Still, as Oltrogge writes, ‘it is extremely rare to prove the use of an art technological recipe collection by a practicing artist’.13 Some recipe books seem to offer good reason to doubt their influence on practicing artists, such as our recipe from the Huishoudelijk Woordenboek, a dictionary that was advertised as ‘a treasury of science to all houses and households, whatever their state or nature, serving to provide lessons that can be used to one’s advantage, or to obtain economical and innocent pleasure from it’.14 And not only does one wonder which artist would have used such a book, a second doubt concerns its author: what would an author – or compiler – of a general dictionary actually know about painting? Interestingly, such questions are not new. In fact, the publisher of the Huishoudelijk Woordenboek himself was sensitive to this issue. In his ‘message for readers’, translated from the introduction to the original 1709 French Dictionnaire Economique by Noël Chomel, on which the Huishoudelijk Woordenboek was based, the editor explained how its author, clergyman Noël Chomel (1633–1712),15 had become knowledgeable on a wide variety of subjects through varied professional occupations, which included managing an estate and a school, and working in a hospital.16

Researchers have suggested different means to overcome the problems sketched above, and methodologies have been described that include rigorous study of the varying contexts of individual recipes like marginalia and paratext.17 Reconstructions or recreations of historic recipes are increasingly employed for the interpretation of these historical recipes. When reconstructing a recipe and learning about its executability and effects, the credibility of a recipe can be evaluated and compared to what is known from examinations of historical paintings.18 Investigations by Carlyle and by Clarke have shown that the evaluation of individual recipes in the context of larger collections, as well as investigations of repetitions and re-workings of recipes, are all important aids in the evaluation of the relation between recipes and actual painting practice.19

While earlier studies have applied text-critical methods to individual recipe sources, no study has so far focused on the character of authorship, and the readers of a larger but well-defined and thus finite group of recipes from a particular time period. This is the methodology proposed and applied in the present paper. Focusing on recipes for preparatory layers dating from 1600 to 1800, a dataset has been assembled that contains all available recipes produced during this time.20 The recipes were taken from all types of recipe books, ranging from general encyclopedias and dictionaries to manuals written by professional artists and manuscripts intended for personal use. Studying a comprehensive collection of recipes on one topic allow us to uncover the general characteristics and specifics of recipe writing during a particular time period.

The dataset assembled for this purpose consists of 27 Dutch recipes for preparatory layers (table 1). The recipe in the Huishoudelijk Woordenboek quoted above demonstrates that Dutch recipes were influenced by international texts. This is not surprising, considering the availability of international literature in the Dutch Republic due to the extensive production and international trade in books in the Netherlands during the seventeenth and eighteenth centuries.21 The Dutch published recipes will therefore be compared with 66 contemporary foreign European recipe sources published between 1600 and 1800 containing 141 recipes (details on the individual recipes are provided in table 2). Most recipes describe preparatory layers for canvas supports. Panel grounds form the second largest group. Recipes for other supports (copper, stone, other metals, paper) are much less frequent, reflecting the use of each type of support during the seventeenth and eighteenth centuries. For comparisons between recipes and actual paintings, published data has been collected on 142 Dutch paintings from the period.22

Recipes for preparatory layers were produced by authors or compilers from varying backgrounds. The authors of Dutch seventeenth- and eighteenth-century recipes were professional artists, amateur artists and lexicographers (see table 3 A for author professions). Van Mander (1604), Beurs (1692), de Lairesse (1712), Horstok (c. 1800) and van Leen (1800) were professional painters. Simon Eikelenberg (1679–1704) was an amateur artist who wrote about painting technique due to a personal interest in the subject.23 Ground recipes also appeared in (arts) encyclopedias or dictionaries like the Huishoudelijk Woordenboek (1743) and the Nieuwen Verlichter (1777), written by authors whom we will call lexicographers. One Dutch recipe source, the Recepten-Boeck (c. 1650–1700), is a manuscript written by an anonymous author, who – judging by the contents of the manuscript– was a professional manufacturer of dyed fabrics or wool in Haarlem. Besides recipes for dying wool and other fabrics, the manuscript contains some personal notes by its author, amongst them a couple of recipes for preparatory layers obtained from friends or acquaintances.24

In a diagram visualizing author profession distribution over time, the continuous involvement of professional artists jumps off the page (fig. 2). Artists form the largest group of authors. This supports the idea that painter-authors had first-hand knowledge on the topic of grounds, which is not self-evident considering the availability of commercially-primed canvas in the seventeenth- and eighteenth century Netherlands.25

If we are to believe their own words, the motives of these artist-authors centered around educating others. Van Mander intended to follow in the footsteps of illustrious earlier writers about the art of painting like Apelles, as he did not see any of his contemporaries expressing such a desire. He described himself as a person who knows painting from the inside out, and he intended to explain all the secrets (‘verborgenheden’) of painting in his book, in particular to instruct young practitioners.26 In the ‘dedication’, Willem Beurs expressed the hope that his Groote Waereld will ‘refresh the thoughts of the reader’, and that his explanations will be clear. The dedication to three of his female students emphasizes his educational goals.27 De Lairesse also had an educational purpose with his writings. This is evidenced not only from the description of his intended audience (see below), but also from the fact that he described his book as an instruction manual for painting.28 Horstok, who was, according to the title page of his book, a professional painter in Alkmaar, had a more limited purpose. His manual concerned the ‘invention’ of a new binder for painting, an emulsion of oil and egg. His purpose was to to lay a claim about his invention and to demonstrate the uniqueness of his method. The ground recipe he included describes a ground prepared using this new binder.29

The artist-authors of the manuscripts presented in the dataset only rarely state their motives directly. Eikelenberg and van Leen do not express their motives, Wiltschut and Dankers indicate their audience (‘young practitioners’) but do not provide a detailed explanation of their motives. The absence of authors’ motives are explained by the nature of such manuscripts. In contrast to publications, manuscripts were (still) in the hands of their owners or compilers, and even if authors envisioned their publication, as was true for example for Eikelenberg, they had not yet reached that stage.

A number of Dutch authors had international equivalents, i.e. authors with similar motives producing recipe books with similar characteristics (table 3A). Eikelenberg’s Aantekeningen shows some resemblance to De Mayerne’s Pictoria. Both authors demonstrate a keen interest in the practicalities of painting, assembled recipes and other information from artists and other sources with knowledge about painting, adding their own observations and performing experiments with these recipes. Of course there are differences, De Mayerne writing with a professional knowledge of chemistry and having direct access to important Golden Age painters and Eikelenberg living a more modest life in Alkmaar. However, they seem to have a similar interest in the practicalities of artistic methods, and like to perform their own experiments with them. De Lairesse’s treatise, described by the author himself as the materialisation of oral lessons he gave to students and painters, includes discussions about the role of ground colour in painting that seem very similar in character to French painter Jean-Baptiste Oudry’s lectures on painting, read to students of the Académie de Peinture in Paris in 1752. Like de Lairesse, Oudry focuses on the refinement of techniques, and addresses an audience of professional art students. For Horstok an international counterpart can be found in English author Dossie, who in his 1758 Handmaid to the Arts introduced innovations in the use of binders for grounds, out of a desire to improve the practice and longevity of painting.30 In the case of Horstok, his manual had the additional purpose of laying his claim on the method described.

Two recipe sources in the Dutch recipe corpus, the Huishoudelijk Woordenboek (1743) and the Nieuwen Verlichter (1777) were translated from earlier French recipes. Their authors, or compilers, fall within the group of lexicographers. The ground recipes in both sources are almost identical, and must have had an identical source (tables 1 and 2). As stated earlier, the main source of the Huishoudelijk Woordenboek was Chomel’s Dictionnaire Economique. However, Noël Chomel’s 1709 Dictionnaire did not contain any ground recipes, nor did any other French edition of Chomel that appeared prior to the 1743 Dutch edition. The recipe was introduced by the editors of the 1743 Dutch edition, and can be traced back to another French source, André Félibien’s Des Principes de l’Architecture, de la Sculpture, de la Peinture from 1676. Félibien (1619–1695) was historiographer to the French king, friend and biographer of Nicholas Poussin (1594–1665), and secretary of the Académie d’Architecture. He published art critiques, biographies of artists and wrote on art theoretical topics, amongst these his celebrated Entretiens sur les Vies et les Ouvrages des Plus Excellens Peintres Anciens et Modernes (Paris, 1666–1688). Four editions of Félibien’s Principes are known to have been published.31 The book had a considerable influence on Dutch authors; Félibien was mentioned by name in the ‘voorreden’ (introduction) to Wilhelmus Beurs’ Groote Waereld of 1692, in the Wiltschut manuscript, and by the anonymous author of the Nieuwen Verlichter.32 Judging from the repetitions of his ground recipe in later sources (table 4), Félibien not only influenced Dutch writing about grounds, but also in Denmark, Germany and England. As table 4 demonstrates, his ground recipe was repeated well into the nineteenth century.

Interestingly, the overview in table 4 demonstrates that authors did not always copy Félibien’s recipe word-by-word. Some made changes to the original text, possibly because they wished to introduce their own ideas about ground preparation. The recipe in the Huishoudelijk Woordenboek is not a direct copy of Félibien’s recipe. While the line ‘to make the colours which are applied on top keep better’ comes directly from Félibien, the description of the role of the glue and the sentence about the priming knife match the wording and sequence of a version of the same recipe in Bernard Dupuy du Grez’ Traité de la Peinture much more closely. Dupuy du Grez does not repeat Félibien’s comment about the role of the ground in ‘keeping’ the colours. Dupuy du Grez (1639–1720) was an amateur artist and lawyer from Toulouse. He founded the Toulouse art academy, later known as the Académie Royale de Peinture de Toulouse.33 Dupuy du Grez admired Félibien, hailing him in his preface as one of ‘our French writers’,34 and acknowledging him as a source for his own treatise.

The Dutch lexicographers of the 1743 edition of Chomel’s dictionary combined elements from the versions of Félibien (1676) and of Dupuy du Grez (1699).35 These editors were Jan Lodewyk Schuer, A.H. Westerhof, and a ‘certain amateur’. On the title page of the Woordenboek, the authors prominently announced that they had added ‘useful entries’ to Chomel’s dictionary while removing some entries that they considered of little relevance to the Dutch market.36 Apparently the entry on grounds was one of the new additions. Schuer and Westerhof were not painters. Schuer was a Dutch translator, editor and compiler of books, who had emigrated to the Netherlands from Hamburg.37 Westerhof, educated at the University of Leiden, was the principal of a Latin School, and author of several theological works. He also contributed to a Dutch-Latin dictionary, translated books and wrote poetry.38 Unfortunately the identity of the third contributor, ‘certain amateur’, remains unknown. The fact that Westerhof, Schuer and the ‘certain amateur’ turned to Félibien as a source of information was not surprising, considering Félibien’s reputation. Dupuy du Grez’ Traité on the other hand, was not widely known and only two editions of his work were published, the first in 1699 and a second in 1700.39 It is puzzling that lexicographers without documented first-hand knowledge about painting technique chose to change the text of a recipe on this subject, as was the case in the Huishoudelijk Woordenboek.

A second Dutch recipe in this group is a 1777 recipe for a ground on canvas in the anonymous Nieuwen Verlichter. Comparison of the details with other versions of the recipe (table 4) demonstrates that its most likely source was Traité de la Pratique de la Peinture published in 1730 by Philipe de la Hire (1640–1718).40

Studying the various sources authored in Dutch and the descriptions they encompass regarding ground recipes demonstrates that there are different levels and categories of authorship during this period, and by implication variation in the nature of authorship of ground recipes. Authors who expressed their motives, often had educational purposes, in particular the artists. Ground recipes found in encyclopedic works were primarily diffusing knowledge, and less concerned with a targeted audience such as students or artists.

Dutch authors operated in international circles and used foreign examples. fig. 3 provides an overview of German, English and French authors and their occupations. Comparison with fig. 2 demonstrates that in these countries, professional artists played a comparatively lesser role, while lexicographers played a larger part in the production of recipes for grounds than in Dutch sources.

Like Dutch authors, international artist-authors of ground recipes had various motives (fig. 3, and table 3B). Instructing young painters in the art of painting was mentioned by several.41 Hidalgo, for example, wrote his treatise for painting students, but he also believed that painting ‘is learned with greater perfection and brevity by those who can study under some great and skillful painter’.42 Painters did not only address students or other painters, however. Lebrun hoped that his treatise would help amateurs to speak on the subject of painting with ‘propriety’.43 Pacheco intended to accumulate information from earlier authors, information on which he commented and to which he added his own view.44 A special case was miniaturist Edward Norgate’s treatise on painting, which, according to Talley, was written at the request of De Mayerne, who hoped to learn Norgate’s methods for (miniature) painting.45

Table 3 demonstrates that many of the books produced by lexicographers were in fact French dictionaries and encyclopaedias (Furetière, Chomel, Dictionaire Universel des Arts, Pernety, Diderot and d’Alembert). An overview of different versions of the recipe for a canvas ground in the Huishoudelijk Woordenboek (table 4) shows that many of such dictionaries and encyclopaedias incorporated information about painting technique from Félibien, directly or indirectly, sometimes making small changes to Félibien’s recipe.46

Lexicographers typically addressed large audiences with little prior knowledge about painting technique. Furetière published to ‘conserve the complete language for posterity’,47 De la Marre, editor of the 1767 edition of Chomel’s Dictionnaire Oeconomique, hoped it would ‘contribute to the progress and solidity of our knowledge, multiply the number of genuine amateurs, able producers, intelligent economists’,48 Barrow’s focus on a general audience was evident in his assurance to his readers that ‘proficience [is] attainable much sooner than they might possibly expect’.49

These examples demonstrate how the intended audiences of several authors were readers without any direct connection to professional painting practice. While authors with a closer relation to painting practice wrote down recipes in manuscripts for personal use, as an aide-mémoire, or published to instruct other artists or amateurs, other authors published recipes out of a desire to spread knowledge. Possibly also the wish to raise the quality of artistic practice with experience from one’s own profession played a role, for instance for authors with backgrounds in chemistry or alchemy. However, this is not stated as such by these authors.

Authors provided different kinds of information, with varying levels of detail. As can be seen in table 1, most Dutch recipes for grounds included information about the pigments to be used and described the layer build-up, regardless of whether the recipe was written by artist-authors or by others. Table 2 shows that internationally, this is was also the trend. Reiteration of a recipe is evident only in the recipe in the Huishoudelijk Woordenboek.

The nature of the preparation specified in recipes depended on the characteristics of the support. For panels, most authors advised a first layer of chalk and glue followed by an oil-bound, pigmented layer. A number of authors dispensed with the first layer and advised only oil-bound layers for panels. For canvas supports, authors described the application of a size layer consisting of animal glue or flour paste, followed by one or more layers of oil-bound pigments.50 Recipes for grounds on copper supports are few, and described the use of a single layer of oil-bound pigments. The materials mentioned in Dutch ground recipes show many similarities with those described in neighbouring countries. For example, the use of glue size layers, of ground layers containing pigments like chalk, lead white, ochres, umbers, blacks, the number of layers applied to the supports, these are always mentioned. Internationally, more variation in the use of materials is incorporated into recipes, and a considerable number of these materials are not mentioned in Dutch recipes from the same period. This is not surprising considering the fact that the total number of international recipes is roughly seven times higher; more recipes lead to more room for variation. The fact that comparison between Dutch and foreign recipes for the production and cleaning of lead white pigments and for animal glue as an ingredient in paintings, also shows such similarities, demonstrates that this is a more general trend.51

Most recipes include details that reveal something about the reasoning behind the instructions (table 1 and 2). Detailed procedures described for ground application show how much care was taken in smoothing the surface of the ground, smoothness that could be obtained by scraping the ground with a sharp tool or by smoothing with tools like pumice stone. In the case of copper supports, measures were taken to create a surface texture in the ground in order to increase the adhesion of subsequent paint layers to the ground, illustrating that with copper supports adhesion of the paint layers could be problematic.

Comments about the choice in materials and the layering of the ground appear in the recipes, regardless whether their authors are painters or not. They demonstrate authors’ beliefs that grounds influence a painting’s longevity and in their pronounced effect on the characteristics of the paint layers applied on top. In particular a ground’s resistance to flaking (fig. 4a and fig. 4b), the sinking in of paint layers and the impact of the ground’s colour receive much attention. The term ‘sinking in’ is used to describe how matt spots or areas appear in a painting when the oil binder is absorbed in unequal measure by a lower layer. It is an undesired effect, because the sinking in of colours changes not only their gloss, but also results in lighter and less saturated tones. While sinking-in may be remedied afterwards by applying oil or varnish to re-saturate matt areas, it prevents an artist from judging the effect of his work during the painting process. Therefore sinking in is a topic that would have been of particular interest to those practically involved in painting. In the Dutch recipes it is mentioned only by Eikelenberg in a recipe for a ground on canvas (fig. 5):

Dutch sources Beurs, Wiltschut and Danckers, de Lairesse, the Huishoudelijk Woordenboek and the Nieuwen Verlichter all wrote about ground colour, discussing how ground colour should be suited to the subject of the painting, or pointing at its effect on the long-term stability of the colours of the painting, by ‘keeping the colours beautiful’.54

The recipes provided by van Mander, Wiltschut and Dankers and by de Lairesse stands out within the group, because these authors did not include detailed information about the pigment composition or the layer build-up of the ground but only provided detailed information on specific aspects of grounds.

Van Mander wrote in detail about the grounds used by his ‘moderne voorders’ (‘recent predecessors’), who would prepare their panel for painting by applying a chalk and glue ground, covered with a thin pinkish second layer which was semi-transparent.55 His description of methods employed by his contemporaries on the other hand only mentions the primed support in passing, in a paragraph explaining how nowadays some painters combine elements from preparatory drawings or sketches and set up their composition on the primed support with a liquid paint.56

Wiltschut and Dankers pointed out how the ground can assist an artist in creating beautiful colours, and provided instructions on ground pigmentation. Possibly this focus is explained by the fact that discussions on the tone of the ground and its influence on the colours of a painting appear in a section on the painting of flowers, which demands a very precise use of colour.57

De Lairesse focused on the relation between ground colour and subject. According to de Vries, de Lairesse wrote for an audience of ‘young professionals, either apprentices in the last phases of their training, or young masters who had not yet fully developed their own style and repertoire of subjects’.58 The fact that de Lairesse did not explain the basics of ground preparation, but instead gave instructions that helped artists refine their use of ground colour fits this theory.

While the focus of these authors, all professional painters, might suggest that painters were less interested in explaining the technicalities of ground preparation and application, other artist-authors did include basic information on ground preparation. For instance the recipes provided by Wilhelmus Beurs in his Groote Waereld of 1692 described all the steps taken in ground preparation in detail, as did Willem van Leen in his Over Teeken- en Schilderkunst. Possibly a diverging focus is not necessarily related to an author’s distance from practice, but should be explained by the audiences the authors intended to serve with their recipes.

Intended audiences are relatively easy to establish, as they were often described in the preface, and sometimes the title or subtitle betrays the audience the author or compiler was addressing. For example, the title page of the 1777 Nieuwen Verlichter states that the book was written for an audience consisting of ‘artists, varnishers, gilders and marblers, and all lovers of these praiseworthy arts’.59 Is this typical for the audiences of Dutch recipe books with ground recipes? Who read Dutch ground recipes and is there any evidence that artists were amongst the readers?

The character of the text and the level of information in a recipe can help establish the type of audience the author hoped to reach. One might assume that those authors who provide detailed, step-by-step instructions had a practical use in mind for their recipes, as only those who actually paint would need to know how to prepare supports for painting. While this may be the case, we have to consider the possibility that even when authors included detailed and complete instructions, their audience did not consist of (aspiring) artists. This is evidenced by Félibien. While he provides step-by-step instructions for ground preparation and application, Félibien’s preface described his intended audience as people who needed to discuss the making of art with those who made it, and of individuals who wished to acquire a general knowledge about art.60 He was not explicitly writing for the artists themselves.

fig. 6 gives an overview of intended audiences of Dutch recipes.61 It demonstrates that artists were often included in the descriptions of target audiences. Van Mander proclaimed to write for ‘painters, art lovers and poets’, Wilhelmus Beurs described his audience as ‘lovers and students of the honorable art of painting’ and de Lairesse as ‘art loving readers’ and ‘young students’.62 The fact that Eikelenberg expected his audience to make practical use of the recipes, shows from the disclaimer he included in the first pages of his manuscript:

Internationally (fig. 7),64 similar groups of readers were mentioned as those addressed by the Dutch: authors, artists, amateur painters and ‘art lovers’ forming the largest audiences. The only exception is the mention of manufacturers as an audience in the German Praktisches Handbuch für Mahler und Lakirer of 1795.65 No Dutch author expressed writing for such readers.

To what extent intended audiences represent actual audiences is difficult to say. However, some indications can be given about the audiences reached by the authors of these recipes. While fellow authors were not mentioned as target audiences, repetitions and re-workings of recipes in later sources demonstrate that they did read recipe books. Some mentioned the authors they admired or that inspired them by name. The anonymous author-compiler of the Nieuwen Verlichter mentioned, amongst others, van Mander, Félibien, du Fresnoy, Dupuy du Grez and Van Gool. The anonymous Swiss author(s) of Mss.Hist.Helv.xvii, 233 (A and B) and Mss.Hist.Helv.xvii.234 referred to the treatise by Wilhelmus Beurs (in the German translation of 1693), but also to a French Traité de Peinture, to de Piles’ Élemens de la Peinture Pratique (1684) and to Leonardo da Vinci’s treatise (fifteenth century).66 This case shows a surprisingly wide distribution of sources: a Swiss author in the Bern region is quoting from books published at least three quarters of a century earlier in Amsterdam and Paris, and even refers to a treatise that originated in Florence some three hundred years ago.67

Unfortunately, there is little information about artists’ reading habits, making it difficult to check whether the authors of recipes reached their intention and educated artists. There is, however, some evidence to suggest that artists were interested in recipe books. Some seventeenth- and eighteenth-century book auction catalogues concerned artists’ book collections. Such catalogues reveal that the collection of painter Pieter Saenredam (1597–1665, auction in 1667) contained the ‘book of secrets’ of Carel Baten, a book that provided recipes for paint manufacture. The collection of painter Robert Duval (1649–1732) and ‘an amateur’ (auction in 1732) included d’Emery’s Noveau Recueil des Secrets and the treatise by Félibien. The catalogue of the collection of painter Frans van Mieris ii (1689–1763, auction in 1764) mentioned Leonardo da Vinci’s Traité de la Peinture, the anonymous Italian Abecedario Pittorico and du Fresnoy’s Van de Schilderkonst; the catalogue of the 1785 auction of the book collection of artist Jan Stolker (1724–1785), of writer Abraham van Alleplas and unknown ‘K.B.’, included van Mander, Chomel’s Dictionnaire Oeconomique, de Lairesse’s Cours de Peinture and an anonymous Traité de Vernis.68 A second source of information on artist book ownership is probate inventories. Bredius’ Künstlerinventäre demonstrates that painter Cornelis Dusart (1660–1704) owned works by van Mander and de Lairesse, as well as ’hundreds more books’; Bartholomeus van der Helst (1613–1670) was the owner of Baten’s Secreetboeck. Unfortunately, most inventories only provided specifications for the most expensive items, and books are not always described in full.69

Although evidence on book ownership by artists as presented in this section constitutes only a very small sample, it does demonstrate that seventeenth- and eighteenth-century artists owned recipe books. It furthermore shows that books written for general audiences were amongst these books. This implies that the division between specialized information for professionals and books for amateurs was less clear-cut than is sometimes suggested, and opens up the possibility that artists derived information from the second category of sources.

It could be argued that reading recipes does not automatically imply that the reader actually used them in his painting. Perhaps readers were interested in recipes for reasons that mirrored the various reasons for writing, as discussed by Smith. Perhaps artists read treatises as an intellectual entertainment, or out of mere curiosity, and not necessarily as practical instructions.70 One thing is certain, artist-authors themselves did not always practice what they preached.71 De Lairesse even warned his readers that in his book they might find advice that would contradict what they had seen de Lairesse himself do in his studio. He acknowledged that he did indeed sometimes paint differently than described in his book. The reason was that reflection had produced new insights since his painting days.72

Obtaining more clarity about the relevance of recipes for painting practice requires systematic comparisons between preparatory systems advised in recipes and grounds that were actually employed in paintings from the same time period. For, only if recipes match contemporary painting practice, it is possible that recipes served as sources of information about actual painting practice.

Using scientific data on 142 Dutch paintings from the period, the main types of grounds and the colours of these grounds can be compared in a systematic manner to seventeenth- and eighteenth-century recipes, in order to investigate whether materials and techniques described in recipes match those employed in contemporary paintings.73 As earlier researchers have focused their attention mainly on those periods and areas that are considered of larger art historical relevance, information from painting analyses is not divided evenly across the period covered by the present study. Much detail is available about grounds in seventeenth-century Dutch paintings, but the eighteenth century is not very well represented.74 However, taking a step back and looking at the seventeenth and eighteenth centuries as a whole, some patterns in the layer build-up and use of ground colour can be compared.

Descriptions of the Dutch and European recipes earlier in this paper demonstrated that three main types of grounds were advised relatively frequently: single or double lead white based grounds; two-layer (double) grounds made up of a first layer of earth or clay in oil, covered by a second layer containing lead white mixed with some pigments, also in oil; and grounds consisting of one or more layers of earth pigments in oil. Figs. 8-11 give the division over time of these groups of grounds, both in the recipes and in painting examinations. These figures allow cross-comparisons between Dutch recipes and paintings, and comparison with recipes and paintings from other European countries.

Focusing first on recipes and occurrences in paintings of single and double lead white/oil grounds, there seems to be a reasonable correspondence between recipes and practice (figs. 8 and 9). Recipes for this type of ground are present throughout the period, and lead white-based grounds are identified in paintings from both the seventeenth and eighteenth centuries, in the Netherlands and abroad.

The occurrence of double grounds consisting of a first layer of cheaper earth or clay pigments, covered by a second layer of lead white is very different in recipes and paintings. Within the Netherlands, such grounds are mentioned in eighteenth-century recipes, while so far they have not been identified in eighteenth-century Dutch paintings. And although these grounds were not mentioned in Dutch seventeenth-century recipes, double grounds of this type were in fact commonly employed by Dutch and Flemish seventeenth-century painters, like Antony Van Dyck, Rembrandt van Rhijn (1606–1669), Aelbert Cuyp (1620–1691), Caesar van Everdingen (1616–1678), Bartholomeus van der Helst (1613–1670) and Caspar Netscher (1635–1684).75 While this diversion between theory (recipes) and practice (paintings) might be taken as a sign that recipe authors looked back to earlier Dutch practices, another explanation seems more likely: all these recipes are based on Félibien’s recipe. Therefore we may conclude that these eighteenth-century recipes provide no evidence that eighteenth-century authors were describing earlier Dutch methods. It is more likely that the motivation of the authors to include recipes for this type of ground can be linked to a desire to connect to a famous seventeenth century French academic authority. These authors looked to the literature for inspiration, not to earlier artists.

Internationally, the situation is different. Already in the seventeenth century this type of ground was described by de Mayerne and in Manuscript Sloane 1990 (fig. 10). These recipes have direct links to contemporary practice. The first came from a professional primer (‘from the Wallonian primer, now living in London’), the second from Flemish portrait painter Paul van Somer i (Antwerp c. 1577 - London 1622). These sources confirm the authenticity of the recipes.76

As in the Netherlands, internationally the eighteenth-century recipes for this type of ground were influenced by Félibien. All but one recipe for this type of ground that date from the second half of the eighteenth century are versions of Félibien’s, demonstrating once again its enormous impact.77

Both in the Netherlands and internationally, many of these later versions of Félibien’s recipe appeared in encyclopedic works aimed at general audiences (table 4). Analyses of paintings from the eighteenth century show this type to have been used very infrequently during the first half of the eighteenth century; during the second it appears to have been completely absent (fig. 12). Both facts combined lead to the tentative conclusion that the relation between this particular recipe and actual painting practice was by that time very weak. However, as the number of painting analyses from the second half of the century is low (sixteen paintings in total), more painting examinations would have to be performed to confirm this hypothesis.

Single earth- or clay-based grounds, as they were used by a number of seventeenth- and eighteenth-century artists, both Dutch and international, are all but absent from the Dutch recipes. Only Simon Eikelenberg provided a recipe for a clay ground, consisting of a single-layer of potters’ earth and linseed oil.78 Internationally, earth pigment or clay-based grounds were mentioned more frequently. While their use seems to have gradually diminished in favour of other types of grounds, their mention in historical sources did not decrease in frequency. An important cause for this is again Félibien’s recipe, which includes the option to only apply the first, earth-based ground layer, omitting the lead white-based second layer (one Dutch recipe and thirteen recipes in other European countries). Here, the same argument applies as for the two layer ground consisting of a layer of earth pigments covered by a lead white based layer: the connection between this particular recipe and actual painting practice seems to weaken over time.

Besides layer build-up, the colour of the ground was a topic that strongly interested the authors of ground recipes. Ground colour was considered an important issue, and authors wrote about actively engaging the ground colour in the design and execution of paintings. In his Groot Schilderboek of 1712, Gerard de Lairesse for example recommended differently coloured grounds for different subjects: a pearl coloured ground for landscapes, umber toned grounds [greyish brown] for indoor scenes, and grounds tinted with Cologne earth or umber and black for night scenes.79 De Lairesse explained that he wished the ground to have ‘something in common with the nature of the subject: the first with the blue of the sky; the second with the reflections; and the third with the shadows’.80 Other authors during the period were not as elaborate as de Lairesse, but nonetheless some of their recipes provide evidence that they considered the topic of a painting in their choice of ground colour.81 So far, examples in painting investigations of the practice of connecting ground colour to the topic of a painting are rare. Noble discusses two examples, both concerning pendant portraits where the male sitter is executed on a ground with a darker tonality than the ground used for the female sitter. One series is painted by Caspar Netscher (1635–1684), the other by George van der Mijn (1723–1763).82

Unfortunately, not all recipes contain information about the colour of the ground. Twelve colour descriptions are available (fig. 12). The ground colours of eighteenth-century Dutch paintings have not been frequently analysed (fig. 13). The grounds on eighteenth-century canvas paintings that have been investigated, have single-layer grounds of a varying colour (whitish, ochre, brown and grey).83 The three eighteenth-century recipe sources mention white, reddish brown and grey ground colours. This limits possibilities for comparing recipes and paintings from in particular the first half of the eighteenth century.

Both in recipes and in paintings, greyish ground colours are frequent. They form roughly half of the colours described throughout the period. There is a difference between the relatively high frequency of reddish grounds in recipes and their much more limited actual use in Dutch paintings. The first mention of a red ground, in the anonymous Recepten-boek om Allerlei Kleuren te Verwen, dating from the second half of the seventeenth century, is rather puzzling. The recipe advised to prepare a cloth or silk for painting with a mixture of warm glue and red bole, but such a mixture would cause flaking, as the glue was likely to become too stiff for a flexible canvas support. Its content is unique and no other recipe proposes a similar preparation for canvas.84 The two other recipes for red grounds, in the Huishoudelijk Woordenboek and in the Nieuwen Verlichter, belong to the series of recipes derived from Félibien (table 4). In these recipes, red was one of the options. Artists could apply another (greyish) layer on top.

Yellowish grounds are featured both in recipes and in paintings of the period. In recipes, their colour is the result of the use of earth pigments (ochres), or in van Leen’s recipe of lead white heated to a temperature where it changes colour to a light yellow.85 Painting investigations describe the use of ochre and other earth pigments, sometimes mixed with lead white and chalk.

Flesh coloured grounds are absent from the Dutch recipes. Only Karel van Mander mentioned them when he discussed the historical practices of ‘our modern ancestors’.86 Hendriks’ investigations of the ground colours of paintings dating from the sixteenth to the eighteenth century produced in Haarlem demonstrates that flesh colour or pink grounds were employed by seventeenth-century Haarlem artists.87 However, in other artistic centres their use seems to have been more limited, and the absence of Haarlem recipe books from the period may explain this particular difference between recipes and paintings.88

Both comparisons between the layer build-up and the colour of grounds in recipes and paintings demonstrate similarities. There is a reasonable correspondence between recipes and contemporary paintings. Yet, differences have also been identified. In particular repetitions of Félibien’s recipe over the course of the eighteenth century seem to result in a decrease in correspondence between recipes and paintings later in that century. Many of the pigments described by the Dutch authors are also found when performing pigment analysis on paintings. As these pigments are part of the standard palettes of painters during the seventeenth and eighteenth centuries, this may not seem very significant. However, with these pigments, a wide range of colours would be possible, and the fact that ground colours (mentioned in recipes) do match those in paintings to a large degree, demonstrates the relevance of the correspondence between recipes and paintings.

Tracing the road ‘from reading to practice’ is a challenge. This paper has highlighted how combining different types of evidence, and investigating single recipes in the wider context of total recipe production, leads to more insight into the reliability of individual recipes. The recipes analysed in this paper demonstrate the possible gain of this approach.

The extensive re-use of Félibien’s recipe serves as a warning about the complexities of evaluating the relevance of recipes for painting practice, and thus about the use of historical recipes in studies into paint materials. It also underscores the need for broad-based and interdisciplinary study. Examining the many derivatives of Félibien’s recipe and the numerous reproductions of them in an international context demonstrated that the adaptation of recipes from earlier sources was not a uniquely Dutch phenomenon; Félibien’s recipe was translated and published in English, German and Danish sources as well. Félibien’s recipe demonstrates that such copies of recipes, even if they appeared in encyclopaedias aimed at general audiences, were nonetheless changed and updated by the compilers of such sources. These updates, which modify or further explain procedures, might even suggest that recipes still had a practical relevance in settings that at first sight appear remote from actual painting practices. The fact that recipes were still ‘alive’ and changing implies that such recipes have a higher relevance than was previously assumed, a thought that finds support in the fact that the book collections of some artists contained publications aimed at a general audiences. Yet, comparisons between recipes and actual paintings also leads to opposite conclusions, as they demonstrated that Félibien’s ground was found less and less in actual paintings of the later eighteenth century. This obviously means that the role of texts like Félibien’s recipe changed. While at the time of its first publication, the recipe corresponded to contemporary painting practice, later re-workings and replications of the recipe must have served other purposes, probably reproduced within a literary rather than a practice-oriented tradition.

This paper has been able to show that a number of recipes correspond closely to artistic practice of the period; in particular ground colours described in recipes show a good resemblance to those used in paintings that are contemporary to their writing or publication. From comments in these recipes we furthermore learn about artists’ motives, for example about the decision of some artists to match their grounds to the painting they wished to create, deliberately adapting and using the ground layer’s characteristics like colour and texture. Knowledge of this kind proves especially important in the interpretation of painting investigations, as it explains its findings. For instance, recipes help explain why differently coloured grounds were observed in two pendant portraits by Caspar Netscher and another series by George van der Mijn – to match the tonality of the sitter. Such information makes recipes rich sources of information. They can serve an important purpose, both in painting examinations as well as in studies into the development of the society that produced them. While the anonymous author of the Nieuwen Verlichter is right in saying that his great example, Apelles, ‘did not use his tongue to produce such beautiful paintings’, it is clear that the tongue– or the pen or printing press in our case – had an importance of its own.89