Depicting Human Flesh in Mummy Portraits from Roman Egypt
MA thesis
Conservation and Restoration of Cultural Heritage, Technical Art History
Author: Jan Matthijs van Daal Student number: 10645586
Contact: jan-van-daal@hotmail.com
Thesis supervisor: prof. dr. H.H.M. (Erma) Hermens Second reader: dr. H. (Herman) den Otter
Word count (pp. 8–98), excluding figure captions and tables: 17.835
University of Amsterdam
Corrected version submitted on 31-07-2019 Original version submitted on 21-06-2019
Ancient Incarnations:
Depicting Human Flesh in Mummy Portraits from
Roman Egypt
Cover: Mosaic of the four Allard Pierson mummy portraits examined in this thesis, imaged through four dif-ferent methods. Object accession numbers (from left to right): APM00724, APM08654, APM14232, APM14498. Imaging techniques (from top to bottom): visible light (VIS) photography, infrared (IR) photog-raphy, false-colour ultraviolet (UV) photogphotog-raphy, macro X-ray fluorescence spectroscopy (MA-XRF) maps of Pb-L emission lines. Credit: acquisition and processing of the MA-XRF maps by Erma Hermens, Annelies van Loon, Victor Gonzalez and Francesca Gabrieli; mosaic assemblage, photography and false-colour imaging by Jan van Daal.
Table of contents Samenvatting 8 Abstract 9 Map of Egypt 10 1 Introduction 11 1.1 Production sites 11 1.2 Dating 12 1.3 Conclusion 13
2 The mummy portrait as physical object 14
2.1 The importance of material-technical knowledge 14
2.1 Gaps in knowledge 15
2.2.1 Filling the gaps: Research question and strategy 16
2.3 Conclusion 17
3 The production of mummy portraits 18
3.1 Support 18
3.2 Preparatory layers 19
3.2.1 Size and ground 19
3.2.2 Compositional planning 20
3.3 Pigments 22
3.4 Binding media 26
3.5 Gilding 28
3.6 Toolmarks 29
3.7 Technique: Painting the flesh 29
4.3 Results 37 4.3.1 General observations and conservation history 38
4.3.2 Pigments 45
4.3.3 Paint 49
4.3.4 Incarnation-building 54
4.4 Conclusion 87
5 Painting technique: First steps towards sustainable terminology 88 5.1 Art-technological analogy versus authentic vocabulary 88 5.2 Flesh in mummy portraits in wider context 89 5.3 Incarnation-building: Preliminary catalogue of techniques 90
5.4 Conclusion 94
6 Discussion 95
6.1 Representability 95
6.2 Suggestions for future research 96
7 Concluding remarks 97
Acknowledgements 99
Bibliography 101
Appendices 107
Appendix I: Technical glossary 108
Appendix III: MA-XRF scans by Erma Hermens, Annelies van Loon,
Victor Gonzalez and Francesca Gabrieli 125
Appendix IV: VIL imaging of Egyptian blue by Luc Megens 133
Appendix V: XRD analysis by Luc Megens 145
schreven in het tweede semester van academisch jaar 2018/2019 voor de opleiding Conser-vering en Restauratie van Cultureel Erfgoed (specialisatie Technische Kunstgeschiedenis), in samenwerking met het Allard Pierson, de Rijksdienst voor het Cultureel Erfgoed en het Rijksmuseum Amsterdam.
Het onderzoeksproject focust op mummieportretten uit Romeins Egypte, geschil-derde portretten (op paneel of doek) van overleden individuen die werden begraven met de gemummificeerde lichamen. Mummieportretten zijn in groteren getale bekend sinds de late jaren ’80 van de negentiende eeuw. Kennis over de materialen en technieken waarmee zij zijn geproduceerd is echter tot op de dag van vandaag beperkt. Recente analyses hebben aangetoond dat het kan – en zou moeten – om in het bijzonder vooronderstellingen ten aan-zien van schildertechniek in twijfel te trekken. Voor mummieportretten is het nog altijd aca-demische conventie om visueel onderzoek uit te voeren en vervolgens de schildertechniek te definiëren als ofwel ‘encaustisch’ (d.w.z. geschilderd met een verf op basis van bijenwas) of ‘tempera’ (d.w.z. geschilderd met een verfmedium op waterbasis).
Deze scriptie stelt dat paradigma aan de kaak vanuit het idee dat het gebruik van terminologie zoals ‘encaustisch’ onhoudbaar is. Om een nieuwe wijze te presenteren van het kijken naar schildertechnieken in mummieportretten, wordt de techniek van het schilderen van menselijke huid bestudeerd. Het onderzoek bestaat uit drie hoofdonderdelen.
Eerst wordt een overzicht gepresenteerd van de materialen en technieken die werden gebruikt in de productie van mummieportretten. Dit was noodzakelijk om een algemene no-tie te ontwikkelen wat betreft huidtinten in mummieportretten. Er bestaat immers geen stan-daardwerk dat geraadpleegd kan worden voor dergelijke materiaaltechnische vraagstukken. Dit onderzoek, gebaseerd op literatuur, laat zekere tendensen zien. Het palet van pigmenten is overweldigend anorganisch, met loodwit en okers als hoofdcomponenten van incarnaties. Panelen van lindenhout en verven op basis van uitsluitend bijenwas lijken universeel popu-lair te zijn. Voorts lijkt er ook een zekere mate van standaardisering te zijn in de uitbeelding van huidtinten.
Een non-invasief, multimodaal technisch onderzoek van vier mummieportretten uit het Allard Pierson, het instituut voor erfgoedcollecties van de Universiteit van Amsterdam, vormt de hoeksteen van deze scriptie. Het onderzoek toont aan dat de incarnaties in de vier portretten zijn opgebouwd vanuit een complex samenspel van standaardtechnieken en schil-dertechnische eigenaardigheden. Het palet is eenvoudig. Ook in deze vier portretten vormen okers en loodwit de grondslag voor de huidtinten.
De laatste sectie van de scriptie adresseert duurzame terminologie. Door terminolo-gie te appropriëren uit verhandelingen zoals Gerard de Lairesse’s (1640–1711) Het groot schilderboek (1707) was het mogelijk om een catalogus op te stellen van schildertechnieken aangetroffen in de vier mummieportretten van het Allard Pierson. Deze methode biedt hand-vatten voor een beter begrip van het mummieportret als fysiek object en vermijdt tegelijker-tijd de valkuilen van de – niet-duurzame – conventionele terminologie.
Abstract
Ancient Incarnations: Depicting Human Flesh in Mummy Portraits from Roman Egypt is a MA thesis project at the University of Amsterdam by Jan van Daal under supervision of prof. dr. Erma Hermens. The thesis was written in the second semester of academic year 2018/2019 for the Conservation and Restoration of Cultural Heritage programme (speciali-sation in Technical Art History), in collaboration with the Allard Pierson, the Cultural Her-itage Agency of the Netherlands and the Rijksmuseum Amsterdam.
The research project focusses on mummy portraits from Roman Egypt, painted por-traits (on panel or canvas) of deceased individuals that accompanied the mummified bodies. Mummy portraits have been known in greater quantities since the late 1880s, but knowledge of the materials and the techniques with which they have been produced is limited to this date. More recent analyses have shown that specifically presuppositions regarding painting technique can – and should – be questioned. For mummy portraits it is still scholarly con-vention to conduct visual examination and subsequently define the painting technique as either ‘encaustic’ (i.e. painted with a beeswax-based paint) or ‘tempera’ (i.e. painted with an aqueous paint medium).
This thesis challenges that paradigm with the notion that using terminology like ‘en-caustic’ is unsustainable. To present a new manner of looking at painting technique in mummy portraits, the technique of depicting human flesh is studied. Three main components constitute the research.
Firstly, an overview is presented of the materials and techniques employed in mummy portrait production. This was necessary to develop a general understanding about flesh tones in mummy portraits, since there is no standard work to resort to for such material-technical questions. This literature-based research shows certain trends. The pigment palette is overwhelmingly inorganic, with lead white and ochres predominating in the incarnations. Limewood panels and purely beeswax-based paints seem to be universally popular. Further-more, there seems to be a degree of standardisation in the depiction of flesh tones.
A non-invasive, multi-modal technical examination of four portraits from the Allard Pierson, the cultural heritage collection institute of the University of Amsterdam, forms the cornerstone of this thesis. The examination shows that the incarnations in the four portraits were built up in a complex interplay of staple techniques and painterly idiosyncrasies. The pigment palette is rather simple; also in these portraits ochres and lead white are the founda-tion of the flesh tones.
The final section of the thesis addresses sustainable terminology. By appropriating terminology from treatises like Gerard de Lairesse’s (1640–1711) Het groot schilderboek (1707) it was possible to draw up a catalogue of painting techniques observed in the four mummy portraits from the Allard Pierson. This method offers handgrips for a better under-standing of the mummy portrait as physical object, while simultaneously avoiding the pit-falls of the – unsustainable – conventional terminology.
Figure 1: Map of Egypt indicating the distribution of mummy portrait findspots. Credit: map by Jan van Daal; the information about mummy portrait findspots is derived from Walker and Bierbier 1997, 8.
1 Introduction
Mummy portraits, covering a mummy or inserted into its wrappings, are painted depictions of the deceased, a memento to their living image. These portraits constitute the sole surviving corpus of classical paintings on panel and canvas. When the Italian traveller Pietro della Valle (1586–1652) described the excavation of two portrait mummies from the Saqqara ne-cropolis near Cairo on 15 December 1615, he could not have been aware of the historical significance of his account. 1Della Valle would be the first to describe painted mummy
por-traits from Roman Egypt.
This story would nevertheless only be a prologue to the modern history of Romano-Egyptian painted mummy portraits. Della Valle purchased two mummies and had them shipped to Europe, where they changed hands several times after his death and ultimately became part of the Staatlichen Kunstsammlungen Dresden, where they reside to this date.2
These mummies never attracted much attention and only about two hundred years later more portrait mummies would come to light, during the early nineteenth-century excavations led by British and French consuls.3 The greatest number of mummy portraits however was
ex-cavated in the period between the late 1880s and the early 1910s, most notably as a result of the initiatives of the Austrian businessman Theodor Graf (1840–1903) and the excavations by the British archaeologist William Flinders Petrie (1853–1942).4
1.1 Production sites
The discoveries by Graf and Flinders Petrie have been formative for the study of Romano-Egyptian mummy portraits. Most of the portraits from the Graf collection are from the cem-eteries near er-Rubayat and Flinders Petrie mainly excavated the cemcem-eteries around Hawara, both areas in the Fayum oasis (Fig. 1). Whenever a provenance is known or can be plausibly assumed, mummy portraits are placed within the context of cemeteries, considering their apparent funerary function and the fact that no production site has yet been unearthed. Alt-hough physical evidence for Romano-Egyptian painting workshops remains elusive, the emergence of material-technical studies on groups of portraits does seem to point to mummy portraits having been produced in designated workshops.5
The vast majority of the portraits however has a provenance that is uncertain or points to the Fayum region. Klaus Parlasca’s four-volume catalogue, which lists virtually all known mummy portraits from Roman Egypt, attributes almost half of the thousand-odd known mummy portraits to either the er-Rubayat or the Hawara cemetries.6 This may suggest that
Romano-Egyptian painted mummy portraits were a typical tradition for the Fayum area; on the other hand, this image may very well be the result of the size of the operations by Graf and Flinders Petrie, on which the latter commented in 1911 that ‘… the portraits are so rare that they do not reward work on a small scale.’.7
As a result, this specific type of mummy portraits is commonly known as ‘Fayum
1 della Valle 1650, 372–91. 2 Borg 1998, 6.
3 Bierbier 1997, 23.
4 For a summary of the discovery and study of mummy portraits see Bierbier 1997, 23–4.
5 See Salvant et al. 2017 for a study of a group of portraits from a hypothesised workshop near Tebtunis.
6 Parlasca 1969, 13; Parlasca 1977, 15; Parlasca 1980, 13; Parlasca and Frenz 2003, 3.
the term refers to the Roman-era portraits painted on linen and panel and as such prevents potential confusion with the contemporary plaster portrait masks. These portrait masks are rather similar in style to the painted mummy portraits and fall within the same time period but are nevertheless a different class of artefacts.8 Since this thesis treats the practice of
painted mummy portraits as a Romano-Egyptian tradition and not with specific regard to the Fayum area, the term ‘Fayum portraits’ will be omitted in favour of the more neutral term ‘mummy portraits’. To avoid confusion of the painted mummy portraits with the sculptural plaster masks – which will not be treated in this thesis – it is important to note that the term ‘mummy portraits’ refers to the former, while the term ‘portrait masks’ is reserved for the latter.
1.2 Dating
As has been mentioned above, the tradition of mummy portraits is a feature of the Roman period in Egypt. This period is generally dated from 30 BCE, when the Ptolemaic kingdom was annexed by the Romans, until the Arab conquests and withdrawal of the Byzantine troops from Egypt in 642 CE. It is possible to assert a rather certain terminus ante quem for mummy portraits; emperor Theodosius I (346–395 CE) issued an edict in 392 CE, prohibit-ing the embalmprohibit-ing of bodies and consequently renderprohibit-ing the production of mummy portraits irrelevant.9 It has, on the other hand, not been possible to pinpoint a certain terminus post
quem for the tradition of mummy portrait production. Early in the history of the scholarship the portraits were believed to be Ptolemaic (305–30 BCE), but in the current scholarly tra-dition the earliest mummy portraits are generally assumed to be produced within the first quarter of the first century CE.10
The dating of individual portraits remains a thorny issue. Most of the portraits have been irrevocably separated from their original burial context without any record, resulting in the loss of vital information regarding provenance and chronology. It is thus more often than not the case that the only source of information about a mummy portrait is the object itself, which brings its complications.
The only absolute dating method that is currently available for mummy portraits is radiocarbon (or 14C-isotope) analysis of the organic support.11 Because of the destructive
nature of the technique and its potential error margin radiocarbon dating is only conducted incidentally, so a chronological framework based on absolute dating is non-existent.
The dating practice of mummy portraits has been debated since the late 1880s and its
8 Walker and Bierbier 1997 contains a multitude of portrait masks, but pp. 131–48 treat this object class more in-depth.
9 Thompson 1982, 10.
10 Corcoran 1995, 13.
11 Appendix I contains a technical glossary briefly explaining the analytical techniques mentioned throughout this thesis.
history has been discussed extensively by Barbara Borg.12 While traditionally the dating of
mummy portraits was more stylistically-oriented, the current system of dating is mainly based on Borg’s methodology in which fashion – most importantly, female hairstyles – takes a central role.13 The mummy portraits presented in this thesis have a dating based on this
method, which classifies them as products of the Roman era. Chronology is not the focus of this thesis and therefore the advantages and issues of Borg’s method will not be explicated. The latest official museum dating will accompany each portrait in its general description, even though they are subject to discussion.
1.3 Conclusion
Although the history of their rediscovery begins in the early seventeenth century, the fasci-nation for Romano-Egyptian mummy portraits was born in the late 1880s. Especially the work of Theodor Graf and William Flinders Petrie effected the unearthing of many portraits. Since their initiatives focussed on excavation sites within the Fayum oasis, the painted mummy portraits from Roman Egypt have come to be known as ‘Fayum portraits’ for short. This suggests however that the tradition of painted mummy portraits was specific for the Fayum. As this is not the case, the more neutral term ‘mummy portraits’ will be used.
The matter of dating has been discussed continuously since the late nineteenth cen-tury. A chronological framework based on absolute dating is lacking to this day, so the cur-rent dating system is relative and strongly influenced by the methodology of Barbara Borg. Provenance-related terminology and dating remain worthwhile subjects for future research. Neither is the focus of this research and therefore the status quo for both topics will be main-tained throughout this thesis; mummy portraits are thus regarded as Romano-Egyptian arte-facts.
of provenance, dating, collection history and cultural context. Most of the oversight publi-cations from the second half of the twentieth century that are mentioned as the go-to author-ities on mummy portraits do contain a material-technical chapter.15 However, the content of
these chapters is often of the same, general nature, with material-technical aspects of mummy portraits being treated as background information rather than as research objective.
Publications dealing exclusively with the material-technical aspects of mummy por-traits start to appear in greater quantity in the 1970s, mainly emerging from the field of con-servation.16 Studies into the studio practice of Romano-Egyptian painters without
conserva-tion-oriented research questions also started to appear in limited number during the second half of the twentieth century. Perhaps the most notable example is the work by Caroline Cartwright on wood identification in mummy portraits on panel.17
Material-technical studies of mummy portraits in general have become more com-mon in the last decade. On the one hand, this surge in scholarship can be attributed to the increased availability and quality of analytical techniques. On the other hand, a significant impetus has been provided by the J. Paul Getty Museum’s Department of Antiquities Con-servation through their Ancient Panel Paintings: Examination, Analysis and Research (AP-PEAR) project. APPEAR, currently ongoing until 2021, was launched in 2013 with the aim of augmenting the knowledge on the material-technical aspects of mummy portraits through international scholarly collaboration.18
The material-technical studies of the past decade also show a certain trend. The stud-ies are relatively small-scale,19 material-oriented and focussed on specialised topics such as
red lake pigments,20 binding media analysis21 and imaging techniques.22 At the APPEAR
conference in May 2018 research questions on a more overarching material-technical level – such as the colour green in mummy portraits – were presented as well, but the conference papers are yet to be published.23
2.1 The importance of material-technical knowledge
Unavoidable is the question why there should be more technical studies of mummy portraits. As argued above, because of the undocumented removal from their burial context in most cases, the only sources of information about mummy portraits are the physical objects
14 E.g. Flinders Petrie 1911, 9–12.
15 E.g. Shore 1962, 20–4; Thompson 1982, 6–7; Doxiadis 1995, 93–104; Borg 1996, 5–18; Walker and Bierbier 1997, 21–2.
16 E.g. Buck and Feller 1972; Ramer 1979. 17 Cartwright 1997a; Cartwright 1997b.
18 < https://www.getty.edu/museum/research/ap-pear_project/ >
19 The sample groups rarely exceed a handful of mummy portraits and sometimes even focus on a
single portrait. Exceptions are Cartwright et al. 2011 and Salvant et al. 2017.
20 Miliani et al. 2010.
21 Granzotto and Arslanoglu 2017. 22 Delaney et al. 2017.
23 The conference papers are expected to be pub-lished by the winter of 2019/20. For abstracts of the papers see < https://www.getty.edu/mu- seum/research/appear_project/downloads/ap-pear_abstracts.pdf >
themselves. Questions about dating, provenance and cultural context have been asked and answered many times but often remain subject to debate. Studying the manufacture of mummy portraits allows for taking a step back. Through scientific analyses both invasively and non-invasively it is possible to assess specific materials and techniques used to create a mummy portrait.
This sketches the biography of a single object but combining data from other mummy portraits sheds light on overarching trends; standardised workshop practices and idiosyncra-sies become visible. This information can then be utilised as a framework of new data to reassess questions of provenance or dating of portraits currently shrouded in mystery. How-ever, as is the case with a radiocarbon analysis-based framework for absolute dating, there is currently not enough material-technical information on mummy portraits to do so. There-fore, to be able to truly understand these objects, it is imperative to keep expanding the da-taset of material-technical knowledge through new research questions.
2.2 Gaps in knowledge
Despite the growing number of publications there remain many unanswered questions re-garding mummy portrait manufacture. Most significant perhaps is the matter of technique. Through a combination of studies, it is possible to provide a general estimation of the mate-rials that were used by Romano-Egyptian painters, but there is less clarity as to how they utilised their materials. The lack of contemporary written sources and the fact that no Ro-mano-Egyptian painting workshop has been unearthed to this date are the main reasons for little information on painting technique of mummy portraits being available.
The limited understanding of Romano-Egyptian painting technique might have led to assumptions that are, retrospectively, too unnuanced or incorrect. This topic was ad-dressed at the APPEAR conference in the paper by Sabino et al., which discusses the prob-lem of finding the right nomenclature to characterise and categorise binding media in mummy portraits.24 The paper argues that mummy portraits are classified as being painted
either with an aqueous medium (tempera technique) or a wax-based medium (encaustic tech-nique), while modern binding media analyses reveal that the reality is often more nuanced or altogether different. The lack of an apt vocabulary to describe how Romano-Egyptian painters handled their paints in combination with presupposition-laden terminology such as ‘encaustic’ hampers the understanding of painting technique in mummy portraits.
Most of the authoritative literature indeed fits this paradigm. Mummy portraits are categorised as ‘encaustic’ or ‘tempera’ on visual characteristics, with Euphrosine Doxiadis extrapolating to the postulation that different wax-based binding medium preparations can be inferred from mummy portraits ‘Just by looking at them closely …’.25
Another feature of the existing literature is to default to Naturalis Historia because Pliny the Elder (23–79 CE) wrote a passage on the preparation of ‘Punic wax’, saponified beeswax that may be applied cold.26 This reference is problematic as ‘Punic wax’ has not
been identified undisputedly in any mummy portrait. Furthermore, Pliny himself does not seem to know Romano-Egyptian mummy portraits since they are unmentioned in Naturalis Historia; Pliny even laments that the tradition of portrait painting has disappeared altogether
24 See note 23 for the link to an abstract of this pa-per.
25 Doxiadis 1995, 95.
nately, there is little GC-MS data available for mummy portraits because the method requires the destruction of a microsample and a recurring theme in mummy portrait research is what Mazurek et al. call ‘… a reluctance to sample.’.29 This, combined with a reliance on an
ex-isting framework of visual observations, which in turn is indebted classical sources such as Naturalis Historia, has resulted in a oversimplified dichotomy between ‘encaustic’ and ‘tem-pera’ mummy portraits, which Sabino et al. desire to see reconsidered.
2.2.1 Filling the gaps: Research question and strategy
Answering a research question based on painting technique provides not only new infor-mation about mummy portrait manufacture; it also offers the possibility to develop a dis-course for discussing how the painters of mummy portraits handled their materials. Finding a new way to speak about mummy portraits allows for the circumvention of terminological pitfalls as discussed above. The call of Sabino et al. for more fitting nomenclature regarding paint in mummy portraits is thus in concordance with the idea of the portraits being their own main source of information, argued earlier in this thesis.
Mummy portraits are portraits in the most literal sense; they are depictions of people who lived, but there is surprisingly little information available on the technique of depicting those people. The incarnations or flesh tones in mummy portraits were established as a re-search focus only recently, at the 2016 conference ‘Inkarnat und Signifikanz’ in Munich.30
A general framework based on material-technical evidence through which this key aspect of mummy portraits can be understood, is however still lacking. The goal of this thesis is there-fore to propose such a framework, guided by the general research question ‘how did Ro-mano-Egyptian painters depict flesh tones to render individual humans in mummy por-traits?’.
The research strategy to answer this question consists of three steps. First, an inven-tory is presented with the current state of knowledge on the manufacture of mummy portraits. This overview does not aim to be exhaustive but should rather be taken as a baseline of what can be expected in mummy portrait manufacture, with a bias towards what is relevant for incarnations. Second, a case-study with four mummy portraits from the Allard Pierson, ex-amined specifically with regard to materials and techniques used to depict flesh tones, sheds light on the validity of the baseline. The third step is to outline a methodology – by reflecting back onto the case-study – for discussing technical aspects observed in mummy portrait in-carnations with more neutral yet fitting terms.
It should be emphasised that the aim of this thesis is methodological rather than de-scriptive. Based on an international corpus of approximately one thousand mummy portraits,
27 Nat. Hist. 35.4 in Rackham 1952, 262–3. 28 De mat. med. 2.105 in Wellmann 1907, 179–80. 29 Mazurek et al. 2014, 77.
30 This conference took place in the context of the ISIMAT project, which focussed on painted incar-nations from antiquity to the Middle Ages.
the case-study represents only a tiny fraction (c. 0.4%). The manner of depicting flesh tones in these portraits therefore cannot be regarded as normative. The goal of this thesis is to provide a foundation for further studies of incarnations in mummy portraits, while simulta-neously taking a step towards the use of more sustainable terminology.
2.3 Conclusion
Studying the mummy portrait as physical object has long been a background sketch rather than a main research focus. Material-technical studies became more common in the 1970s, emerging from the field of conservation. Through technological improvements and initia-tives such as APPEAR these studies increased in popularity, especially since the last decade. Material-technical information is crucial to better understand mummy portraits in general but key aspects of these artefacts are still poorly understood. Particularly problematic is the lack of understanding regarding the utilisation of paint, which Sabino et al. deem to be re-sponsible for faulty categorisations of mummy portraits.
Steps towards more suitable terminology for mummy portraits and better understand-ing of their technique can be provided through examinunderstand-ing how Romano-Egyptian painters depicted the most important element of each mummy portrait: the incarnation. An answer to this question shall be provided by combining data from existing studies with a case-study of four mummy portraits examined for this research project. Then, placing these results within a newly proposed framework of more neutral terminology serves the goal of this thesis, which is to provide methodology not only for the future study of incarnations in mummy portraits, but also for discussing mummy portraits within a fitting academic discourse.
about materials and techniques in mummy portraits, particularly that of relevance for flesh tones. The main physical components of mummy portraits (support, preparatory layers, pig-ments, binding media, gilding and toolmarks) are addressed and thereafter referenced passim in discussing the depiction of flesh tones. Without claiming to be exhaustive, this chapter should in the end function as a material-technical framework that aids in the contextualisa-tion of the data gathered from the examinacontextualisa-tion of the four Allard Pierson portraits.
3.1 Support
Based on primary supports alone, mummy portraits can be divided in two groups: those painted on panel and those on canvas. There are no absolute numbers readily available on the ratio of mummy portraits on canvas to mummy portraits on panel, but there seems to be a consensus that portraits on panel constitute the majority.31
Extensive work on the identification of wood species has been undertaken by Caro-line Cartwright. While the work on wood identification is ongoing to this date, lime/linden (Tilia species) wood has been identified as the material of choice for the majority (c. 72%) of the mummy portraits on panel.32 Not native to Egypt, limewood would have been
im-ported from Europe, where it was abundantly available.33
Before the final incorporation in the mummy wrappings, the portraits were placed over the face of the deceased and slightly curved to follow the shape of the head, with the longitudinally oriented wood grain ensuring minimal distortion in the face of the portrait.34
Limewood is ideal for this process as it is easy to cut exceedingly thin35, remains
flexible in doing so, with a low chance of warping and splitting as result of aging.36 Other
species of wood, such as oak or Lebanese cedar, have different properties but were used as well. No univocal answer is satisfying to explain the reason behind this diversity in utilised wood species; it seems to be a rather complicated phenomenon on the intersection of work-shop idiosyncrasies, economic factors and customer preferences.37
In contrast, mummy portraits on canvas have been studied sparsely. Some general remarks are still possible, however. Most canvas portraits are large, depicting entire busts or the portrayed from head to toe. These larger, single-piece portraits are also known as shrouds and would have been wrapped around or placed on top of the mummy, a practice dating back to pharaonic Egypt.38 Although based on conjecture, it seems plausible that Egyptian linen
would have been used, since most ancient Egyptian textiles are linen and flax has been cul-tivated in Egypt since prehistoric times.39
31 Thompson 1982, 10.
32 Cartwright et al. 2011, 56, Tab. 2. 33 Cartwright et al. 2011, 52, 57. 34 Spaabæk 2012, 67–8.
35 Panels as thin as 1 mm are not unheard of.
36 Spaabæk 2012, 67. 37 Cartwright et al. 2011, 57. 38 Vogelsang-Eastwood 2000, 295. 39 Vogelsang-Eastwood 2000, 269.
3.2 Preparatory layers
From a practical standpoint it seems plausible that instead of directly painting onto the pri-mary support, Romano-Egyptian painters first applied materials to facilitate the application of paint at a later stage. The problem with these ‘preparatory layers’ is that they usually lie hidden under the paint surface; cross-sections or coincidental visibility around existing dam-ages are often necessary for proper identification and analysis.
3.2.1 Size and ground
Through a compilation of existing literature, it has been possible to gain more insight into the usage of sizing layers and grounds in mummy portraits. These results are summarised in Table 1 and Figure 2. Even though the number of mummy portraits included in this overview is relatively low (representing c. 6.5% of the international corpus), a pattern does seem to emerge.
Firstly, size layers are rarely observed. They have been attested in two panels exam-ined by Ramer, a shroud examexam-ined by Rinuy and nine panels examexam-ined by Sand et al.40
Working with a size layer is beneficial for a painter since it reduces the absorbency of the support, which counteracts the leaching of binding medium from the ground and paint layers to a certain degree. It may be concluded that size layers were uncommon in mummy portrait production, but this seems unlikely when the practical advantages of sizing are considered. This point is also addressed by Sabino, who argues that although likely to have been used, sizing layers are particularly difficult to analyse due to their thinness, aging and sinking into the primary support over time.41
Ground layers, on the other hand, are more common. A ground layer would offer the painter a smoother, more uniform surface to paint on, as well as an extra layer of protection from an absorbent support. Ground layers were detected in almost two thirds of the portraits. So, they seem to be a common feature in mummy portraits, albeit not a necessary one. The grounds can be roughly categorised in three groups: calcium-based light grounds42, light
grounds based on lead white43, and dark grounds based on black pigment.44
Sample preparation for binding medium analysis of ground layers is notoriously dif-ficult and thus far less information on this topic is available. For the portraits analysed by Asensi Amorós et al., animal glue is mentioned as the binding medium.45 It should be noted
that these portraits all have the same provenance: Antinoopolis. Extrapolating these results should therefore be done cautiously, as this could be indicative of a regional workshop idi-osyncrasy rather than a general feature of mummy portrait grounds.46
40 Ramer 1979, 4–5; Rinuy 2001, 940; Sand et al. 2017a, 190–1.
41 Sabino 2016a, 4; Sabino 2016b, 4.
42 Asensi Amorós et al. 2001, 126, Tab. 2; Buck and Feller 1972, 804; Cartwright and Middleton 2008, 63, Tab. 4; Freccero 2012, 27–8; Hart 1980, 22; Hillyer 1984, 3; Miliani et al. 2010, 705; Oth-man et al. 2018; Ramer 1979, 4–5; Salvant et al. 2017, 818, Tab. 1a; Sand et al. 2017b, 198–9.
43 Cartwright and Middleton 2008, 63, Tab. 4; Freccero 2012, 28; Sabino 2016a, 4; Sabino 2016b, 4; Sand et al. 2017b, 198–9.
44 Asensi Amorós et al. 2001, 126, Tab. 2; Wrap-son 2006, 157.
45 Asensi Amorós et al. 2001, 125.
46 Three of these portraits exhibit a feature not yet encountered in other mummy portraits: a thin, pigmented isolation layer based on animal glue, covering the ground. See Asensi Amorós et al. 2001, 125.
subtle and lie hidden underneath opaque paint layers, only becoming visible through damage or specific imaging techniques. Difficulties in observing methods of compositional planning explain the paucity of available data; so only anecdotal evidence is available.
Seven of the twelve portraits examined by Salvant et al. were found to have white underdrawings, likely all calcium-based.47 The two portraits examined by Sabino also
con-tain preparatory drawings; both were applied in a liquid medium, one on basis of carbon black pigment and extending throughout most of the portrait48, the other in sepia following
the outlines and folds of the drapery, and grey on the proper right side of the neck.49 An
underdrawing similar to the one in Sabino 2016a was the only underdrawing observed in the portraits examined by den Hollander.50 Six out of thirteen portraits studied by Sand et al.
also contained underdrawings, all dark and applied liquid. Particularly interesting is 31161,52 from the Antikensammlung Berlin. The paint layers in the incarnation are virtually effaced, making it possible to study the detailed underdrawing underneath.51
Evidence of compositional planning is not always visible with the naked eye. One of the Tebtunis portraits analysed by Salvant et al. without observable underdrawing (6-21375) was found by Ganio et al. to have outlines of the face, eyes and nose painted with Egyptian blue pigment.52 There are no blue areas in the portrait, nor were these outlines visible with
the naked eye; only through visible-induced luminescence (VIL) of the Egyptian blue pig-ments it was possible to expose this feature. Another technique, visible-induced visible lu-minescence (VIVL) has been used successfully to image brushstrokes – now invisible – par-tially outlining figures in portraits from the Penn Museum and the Boston Museum of Fine Arts.53 radiography also proves to yield valuable information in some instances; the
X-radiograph of one of the portraits examined by Ramer and a female portrait from the Penn Museum clearly show that a paint containing heavy elements (probably lead) was used to sketch the outline of the portrayed.54
Further research is needed to shed more light on the topic, but from the examples above it becomes clear that compositional planning was used in the manufacture of mummy portraits and that Romano-Egyptian painters had a manifold of techniques at their disposal.
47 Salvant et al. 2017, 818, Tab. 1a. One of the portraits (6-21378a verso) is noted as having a black ink sketch as underdrawing. It is on the verso of a finished portrait and contains written instructions for the composition, so in this thesis it is not considered to be an independent portrait. This object is discussed in paragraph 3.7. 48 Sabino 2016a, 5–6.
49 Sabino 2016b, 5.
50 den Hollander 2011a, 53. 51 Sand et al. 2017a, 182–3.
52 Ganio et al. 2015, 817. However, Ganio et al. argue that the Egyptian blue in this instance func-tions as modulating pigment rather than as under-drawing material.
53 Mayberger 2016a; Svoboda 2018. 54 Ramer 1979, 2, Fig. 1; Mayberger 2016b.
Table 1: Overview of studies on mummy portraits where the presence or absence of a sizing and/or ground layer is mentioned. Credit: table by Jan van Daal.
Study Mummy
por-traits/fragments analysed
Size layers identi-fied Ground layers identified Ground layer types Asensi Amorós et al. 2001 11 0 11 Calcium-based (8); dark ground (3) Buck and Feller
1972 1 0 1 Calcium-based (1) Cartwright and Middleton 2008 4 0 4 Calcium-based (2); lead white-based (2) Freccero 2012 8 0 6 Calcium-based (5); lead white-based (1) Hart 1980 1 0 1 Calcium-based (1) Hillyer 1984 1 0 1 Calcium-based (1) den Hollander 2011a 5 0 1 Indeterminate (1)
Miliani et al. 2010 2 0 2 Calcium-based
(2)
Othman et al. 2018 2 0 2 Calcium-based
(2)
Ramer 1979 3 2 1 Calcium-based
(1)
Rinuy 2001 1 1 0 n/a
Sabino 2016a 1 0 1 Lead
white-based (1)
Sabino 2016b 1 0 1 Lead
white-based (1)
Salvant et al. 2017 12 0 5 Calcium-based
(2); Indetermi-nate (3)
Sand et al. 2017a/b 13 9 4 Calcium-based
(3); lead white-based (1)
Wrapson 2006 4 0 1 Dark ground (1)
Figure 2: Pie-chart visualising the use of grounds in mummy portraits, as presented in Tab. 1, above. Credit: figure by Jan van Daal.
3.3 Pigments
Pigments are a much-discussed matter in the scholarship on Fayum portraits. Ubiquitous are references to Pliny’s famous excursion on the pigments of the ancient world.55 Without
dis-crediting Pliny as an art technological source, it is important to note that a priori equating information on pigments in Naturalis Historia to the Romano-Egyptian palette would be problematic; many of Pliny’s statements, like the exploitation of iron earths, are yet to be backed by analytical evidence.56 It was therefore a deliberate choice for this section to base
the information on pigments in mummy portraits solely on results from technical examina-tions. Table 2 summarises the combined results of this literature study.57
From Table 2 it becomes clear that Romano-Egyptian painters had a wide variety of pigments at their disposal. While pigments like lead white and Egyptian blue would have been produced synthetically, most of the pigments originate from a natural source. The Ro-mano-Egyptian pigment palette is predominantly inorganic. The detection and analysis of inorganic pigments is often more difficult due to their deterioration with age but, on the other hand, the inclination towards inorganic pigments also shows the continuation of an art
55 Nat. Hist. 35.29–49 in Rackham 1952, 280–99. 56 Salvant et al. 2017, 824.
57 Tab. 2 is based on the following sources: Asensi Amorós et al. 2001, 125–6; Buck and Feller 1972, 804; Cartwright and Middleton 2008, 62–3; Corcoran and Svoboda 2010, 46; Delaney et al. 2017, 3–8; Dietemann et al. 2017, 203–4;
Freccero 2012, 28; Granzotto and Arslanoglu 2017, 172; Hart 1980, 21–3; Hillyer 1984, 8; den Hollander 2011a, 68–73, 80, Tab. 1; Miliani et al. 2010, 705–7; Othman et al. 2018; Ramer 1979, 5– 6; Rinuy 2001, 940; Sabino 2016a, 6–7; Sabino 2016b, 6; Salvant et al. 2017, 821–6; Wrapson 2006, 158.
historical tradition that has its roots in early Dynastic (i.e. pre-Roman) Egypt.58
When compared on a large scale – like in Parlasca’s overview catalogue – mummy portraits clearly show certain preferences regarding colour use. Black, red, white and yel-low/brown are the key hues in each portrait. Blue is seldomly observed and pure green only in the context of certain items of jewellery.59 Although the term undercuts the tonal varieties
and subtleties in each mummy portrait, this palette of four basic colours is called ‘tetrachro-mic’; the tetrachromic palette is typically associated with the ancient Greek manner of paint-ing that took form in the fourth century BCE under the influence of famous painters such as Apelles.60
The main organic pigments in mummy portraits, madder and indigo, are both ex-tracted from plants as dyestuffs. Not only were these substances already known for a long time in Dynastic Egypt, but indigo and madder were also the most commonly used organic dyestuffs for blue and red textiles, respectively.61 Madder and indigo were used – both
sep-arate and mixed – in mummy portraits to create a range of blue, pink, purple and red hues. Interestingly, various technical examinations of mummy portraits have shown that these or-ganic pigments were used mainly to colour the garments of the portrayed.62
Apart from being attested through analyses of mummy portraits, there are also art technological sources that corroborate the use of several of the colourants mentioned in Ta-ble 2 in Roman Egypt. Six pottery saucers from the British Museum dating to the first cen-tury CE containing powdered pigments were excavated in 1888 by Flinders Petrie at the cemetery of Hawara (Fig. 3). While it cannot be proven that these saucers once belonged to a mummy portrait painter, analysis of the pigments shows that these were exactly the pig-ments also found in mummy portraits.63 The label ‘Romano-Egyptian’ is particularly fitting
for this miniature palette; the blue pigment is traditional Egyptian blue and the orange-red pigment is confirmed as minium, which only came into use in Egypt from the late Ptolemaic or the early Roman period.64 Additionally, the use of lead white and madder as pigments is
also unknown in Egypt before the Roman age.65
The famous Egyptian (al)chemical treatise known as the ‘Stockholm Papyrus’, was written around 300 CE and as such is contemporary with mummy portrait production.66 The
Stockholm Papyrus contains several recipes for dyeing textiles. Madder is mentioned in some recipes for dyeing in red and purple;67 indigo dye extracted from woad (Isatis tinctoria)
is used to dye blue;68 and even the combination of madder and indigo is mentioned as a
method of dyeing purple.69 Sources like the recipes in the Stockholm Papyrus and the saucers
from the British Museum provide insight into Romano-Egyptian studio practice. Features of mummy portrait production like the colourants used in garments do not exist in a vacuum;
58 Lee and Quirke 2000, 104–5.
59 E.g. the green gemstone necklace discussed in Delaney et al. 2017, 6–7.
60 Nat. Hist. 35.50 in Rackham 1952, 298–9; Dox-iadis 1995, 98–9.
61 Vogelsang-Eastwood 2000, 278–9.
62 Delaney et al. 2017, 7–8; Miliani et al. 2010, 706–8; Salvant et al. 2017, 824–6.
63 Cartwright and Middleton 2008, 63, Tab. 5. 64 Lee and Quirke 2000, 114.
65 Cartwright and Middleton 2008, 64. 66 The Stockholm Papyrus is often mentioned alongside its contemporary counterpart the ‘Ley-den Papyrus X’. The Ley‘Ley-den Papyrus X mainly deals with metalworking and contains no relevant recipes for manufacturing (colourants in) mummy portraits.
67 Caley and Jensen 2008, 74–5, 79. 68 Caley and Jensen 2008, 73–4, 77. 69 Caley and Jensen 2008, 84–5.
Figure 3: Six pottery saucers containing pigments, excavated by Flinders Petrie at Hawara in 1888. British Museum, London. Contents, measurements and accession numbers (from left to right): blue (Egyptian blue), H. 6.35 x Ø 10.16 cm, 1888,0920.23; yellow (jarosite), H. 7.62 x Ø 10.16 cm, 1888,0920.26; pink (madder + gypsum), H. 5.08 x Ø 10.16 cm, 1888,0920.27; orange (minium), H. 5.08 x Ø 10.16 cm, 1888,0920.28; white (gypsum), H. 6.35 x Ø 8.89 cm, 1888,0920.25; dark red (red ochre + haematite), H. 7.62 x Ø 10.16 cm, 1888,0920.24. Credit: photography by the British Museum (© The Trustees of the British Museum). Use of image permitted under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0) license.
Table 2: Overview of pigments used in mummy portraits. Credit: table by Jan van Daal.
Colour Common name Scientific name Chemical formula
Black Carbon blacks (bone
black, lamp black, vine black, etc.)
n/a Varying
Blue Egyptian blue Calcium copper silicate /
Cuprorivaite
(Ca,Cu)Si4O10
Indigo
2,2'-Bis(2,3-dihydro-3-oxoin-dolyliden) / Indigotin
C16H10N2O2
Brown Brown ochre (goethite) Iron(III) oxide-hydroxide α-FeOOH
Umber Iron oxide + manganese oxide +
aluminium oxide
Fe2O3MnO2
nH2OSiAlO3
Green Copper greens Copper acetates /
Copper carboxylates
Varying
Egyptian blue + orpiment n/a n/a
Egyptian green Cuprowollastonite (Ca,Cu)3Si3O9
Green earth Aluminosilicate K[(Al,Fe+3),(Fe+2,Mg]
(AlSi3,Si4)O10(OH)2
Malachite Copper carbonate hydroxite Cu2CO3(OH)2
Grey Black + white n/a n/a
Pink Haematite + white n/a n/a
Madder + white n/a n/a
Red ochre + white n/a n/a
Purple Madder 1,2-dihydroxyanthraquinone /
Alizarin
C14H8O4
Madder + Egyptian blue n/a n/a
Madder + indigo n/a n/a
Red Cinnabar / vermillion Mercury(II) sulfide α-HgS
Haematite Iron(III) oxide Fe2O3
Madder 1,2-dihydroxyanthraquinone /
Alizarin
C14H8O4
Minium / red lead Lead(II,IV) oxide Pb3O4
Realgar Arsenic sulfide α-As4S4
Red ochre (goethite) + haematite
Iron(III) oxide-hydroxide +
iron(III) oxide (α-FeOOH)Fe
2O3
White Alunite Aluminium potassium sulfate KAl3(SO4)2(OH)6
Calcite / chalk Calcium carbonate CaCO3
Gypsum Calcium sulfate dihydrate CaSO42H2O
Huntite Magnesium calcium carbonate Mg3Ca(CO3)4
Lead white Lead(II) carbonate hydroxide 2PbCO3Pb(OH)2
Yellow Yellow ochre (goethite) Iron(III) oxide-hydroxide α-FeOOH
Yellow ochre (jarosite) Potassium-iron(III) sulfate KFe3+3(SO4)2(OH)6
Yellow ochre (natrojaro-site)
Sodium-iron(III) sulfate Na+Fe+3
3 (SO4)2(OH)6
Orpiment Arsenic sulfide As2S3
visual examination and limitations in data interpretation.
As argued earlier, currently the best method to assess binding medium composition in mummy portraits is through GC-MS analysis. GC-MS requires the destruction of a small paint sample, which is why many institutions housing mummy portraits are reluctant towards binding medium analysis.70
It is therefore still common practice to rely on visual observation in binding medium interpretation. Impasto is associated with ‘encaustic’, flat surfaces and hatching are associ-ated with ‘tempera’. This method has proven to be imprecise; beeswax-based paint was at-tested in analysis of one of the portraits examined by Sabino, as well as others, categorised as ‘tempera’ on account of their visual characteristics.71
The third complicating factor is the limit of what binding medium analysis is cur-rently available to tell. ‘Punic wax’ is likely to have been used in mummy portraits because of its utility as paint. With GC-MS it is possible to identify saponified beeswax. In the past this has been done and subsequently interpreted as proof for Punic wax.72 However, to
in-terpret GC-MS data in this manner would be incorrect. Table 2 shows that metals make up a significant portion of the Romano-Egyptian palette. To date, binding medium analysis is not able to distinguish between beeswax treated with an alkali beforehand and saponification as result of long-term interaction between a lipid medium and metal-containing pigments.73
So, in sketching an overview of binding medium usage in mummy portraits it is im-portant to stay close to analytical data, without extrapolating unnecessarily. Compiling the information available in literature was again deemed the best method for sketching an over-view. Scarcity of published data from binding medium analysis for mummy portraits proved to be a significant hinderance.
The APPEAR project sparked the generation of a new dataset in the form of numer-ous GC-MS analyses by Joy Mazurek from the Science Department of the Getty Conserva-tion Institute. The results of those analyses are incorporated into the APPEAR database, which is currently accessible only for project members. Joy Mazurek intends to publish her results later in 2019 but was kind enough to share an overview with the author, for which he is truly grateful. Only through combining references in the literature with her unpublished data it was possible to approach binding media in mummy portraits quantitatively. The re-sults of this assessment are summarised in Table 3 and Figure 4.74
70 GC-MS is not the sole available method for binding medium analysis. There are possibilities for non-invasive assessment of the binding me-dium or methods where the sample does not need to be destroyed. The results of those methods however are less precise than GC-MS. There are analytical methods offering results even more pre-cise than those of GC-MS analysis. See note 76 for examples.
71 Sabino 2016a, 6, 16. 72 Ramer 1979, 3, 6–7.
73 Stacey et al. 2018, 472, 486.
74 Tab. 3 is based on the following sources: Asensi Amorós et al. 2001, 124; Cuní et al. 2012, 666; Delaney et al. 2017, 4; Dietemann et al. 2017, 209–10; Freccero 2012, 27; Granzotto and Arslanoglu 2017, 173–4; Hillyer 1984, 8; Ma-zurek 2019, Tab. 1; Ramer 1979, 6–7; Rinuy 2001, 939; Sabino 2016a, 6; Sabino 2016b, 6; Sal-vant et al. 2017, 818, Tab. 1a, 820-1; Wrapson 2006, 159.
This overview, currently the most extensive of its kind, represents approximately 7.5% of the known mummy portraits. Suggested is that beeswax-based paint was used for most of the mummy portraits. In less than 25% percent of the sample group only a paint on basis of an aqueous medium (i.e. ‘tempera’) could be attested. Considering this dataset, it seems unfitting to simply adhere the label ‘encaustic’ or ‘tempera’ to a mummy portrait, especially because there are also instances of ‘mixed techniques’ like beeswax with glue.
Beeswax-based paint has surprising versatility and can be used to create a multitude of surface techniques. This is evident from portraits examined by Freccero and Sabino that were believed to be painted with a tempera medium, which turned out to be beeswax.75 To
achieve these surface techniques, there must have been various ways to modify the beeswax medium, which analytical techniques are not yet capable of detecting unambiguously. Re-cent studies have shown that it is possible to identify the sources of animal glues and plant gums in ancient paintings, but this type of analysis is still far from being common practice.76
This information underlines the importance for developing a new way to interpret the paint-erly elements in mummy portraits, without relying on conjecture about their binding media.
Table 3: Overview of studies on binding media in mummy portraits. Credit: table by Jan van Daal.
Study Mummy portraits/fragments
analysed for binding media
Binding media identified
Asensi Amorós et al. 2001 1 Beeswax (1)
Cuní et al. 2012 1 Beeswax (1)
Delaney et al. 2017 1 Beeswax (1)
Dietemann et al. 2017 4 Beeswax (2); Beeswax + animal
fat (1); beeswax + egg (1)
Freccero 2012 8 Beeswax (8)
Granzotto and Arslanoglu 2017 1 Gum (1)
Hillyer 1984 1 Glue + gum (1)
Mazurek 2019 45 Beeswax (30); beeswax + glue
(2); glue (11); glue + gum (1); unidentified ‘tempera’ (1)
Ramer 1979 177 Beeswax (1)
Rinuy 2001 1 Gum (1)
Salvant et al. 2017 878 Beeswax (8)
Wrapson 2006 3 Beeswax (3)
Total 75 See Fig. 4
75 Freccero 2012, 27; Sabino 2016a, 6, 16. 76 Mazurek et al. 2014 were able to prove that the glue binding medium of a Romano-Egyptian trip-tych was made from cow hide. Granzotto and Arslanoglu 2017 identified gum from the Acacia
seyal Del. (gum arabic) together with locust bean
gum as binding medium of a fragmentary mummy portrait on canvas.
77 Two of the portraits analysed by Ramer were also analysed by Mazurek. To avoid redundancy, only the portrait not analysed by Mazurek is in-cluded in the table.
78 Two of the portraits analysed by Salvant et al. were also analysed by Mazurek. To avoid redun-dancy, only the portraits not analysed by Mazurek are included in the table.
Figure 4: Pie-chart visualising the use of binding media in mummy portraits, as presented in Tab. 3, above. Credit: figure by Jan van Daal.
3.5 Gilding
Some mummy portraits received additional embellishment in the form of gilding. Gilding mummy portraits was done selectively. Most common are gilded funerary wreaths, back-grounds, jewellery or ornamental details.79 A rare example of gold leaf on flesh tones is
11.2892 from the Boston Museum of Fine Arts, which has gilded lips.80 This suggests that
gold leaf sometimes had symbolic as well as decorative value. Even though gilded wreaths are encountered most often, surveying Parlasca’s catalogue suggests that less than 10% of the mummy portraits has a gilded wreath.81 Gilding in mummy portraits is thus relatively
uncommon.
Little is known about the practice of gilding mummy portraits, but gold leaf on top of painted surfaces would have to be applied with a mordant. 82 The late-third or early-fourth
century CE Leyden Papyrus X contains five recipes for writing with gold leaf.83 The gilding
practice as outlined in these recipes could, theoretically, be analogous to the gilding of mummy portraits. Further study in the form of historical reconstructions would be desirable to shed more light on this matter.
79 Doxiadis 1995, 99.
80 < https://www.mfa.org/collections/object/por-trait-mummy-of-a-youth-45976 >
81 Cf. Parlasca 1969; Parlasca 1977; Parlasca 1980; Parlasca and Frenz 2003.
82 Billinge et al. 1997, 31.
3.6 Toolmarks
The main source on tools used by Romano-Egyptian painters is again Pliny the Elder. Fol-lowing Mayhoff’s emendation of Nat. Hist. 35.149, three tools would have been key, at least to the encaustic painter: the brush, the cauterium, a palette knife-like tool that might have been heated, and the cestrum, a small engraving tool.84 There is no archaeological evidence
confirming the use of these tools by mummy portrait painters, but it is nonetheless plausible that tools of this nature would have been available to painters in Roman Egypt. Through close observation of mummy portraits it is possible to identify brushstrokes, surfaces seem-ingly blended with a palette knife and subtle engravings into the impasto.85 So,
notwithstand-ing individual painters’ preferences regardnotwithstand-ing tools, there is no reason to assume that brushes, cauteria and cestra were not part of the standard equipment in mummy portrait painting.
3.7 Technique: Painting the flesh
With portraits but no manuals on portrait painting surviving from antiquity, gauging how Romano-Egyptian painters strived to depict human flesh remains a tantalising effort. A sin-gle source with instructions on painting a mummy portrait has survived, possibly perchance, as it is the backside of a portrait. This object (6-21378a verso) is housed by the Phoebe A. Hearst Museum of Anthropology, Berkley and is dated to the mid-second century CE. On the bare wood there is a portrait sketch with Greek instructions in black ink (Fig. 5). These instructions are directed towards a painter and provide general guidelines for the portrait. Two instructions pertain to the incarnation: σιτόχρωμος and ἡδυτέρου[ς] ὀφθαλμούς, mean-ing ‘fair-coloured [skin]’ and ‘[paint] the eyes softer’, respectively.86 What fair skin and soft
eyes entailed for the Romano-Egyptian painter remains beyond the reach of the knowable. This sketch does however prove that qualitative factors were attributed to painted flesh tones in Roman Egypt, suggesting that incarnations in mummy portraits were understood and de-picted within a shared conceptual framework.
Since the sketch at Berkley is unique, further source material on the depiction of flesh tones in mummy portraits must come from examination of the objects themselves. Inorganic materials are the main components; in the few instances that technical examination of the flesh tones was undertaken, white pigments (especially lead white) and ochres are richly employed. For the general flesh tones, the yellow ochres goethite and jarosite frequently recur. Natrojarosite was identified in the flesh tone of the portrait examined by Delaney et al.87 Ramer, in his examination of portraits from the Petrie Museum mentions minium being
84 Borg 1996, 5–8. The tools of painters not work-ing in wax-based paint are not mentioned in an-cient literature.
85 Doxiadis 1995, 96–7.
86 Fournet 2004, 97–9. Fournet’s interpretation and translation of the Greek text is generally ac-cepted.
idence. An organic red pigment was detected by Lalli et al. in a cross-section from the cheek of a portrait.90 Additionally, there is a
portrait from the British Museum where madder may be present in the flesh tones.91 Reddish-brown
umber was detected through the presence of manganese in all por-traits examined by Salvant et al., specifically in the areas outlining details (e.g. nose, eyes) in the flesh tones.92
Egyptian blue seems to fulfil a special function in certain portraits. Ganio et al. observed Egyptian blue in the flesh tone contours of one portrait.93 It could
be interpreted as a compositional planning feature, but since more obvious materials for underdraw-ings were available it is plausible that the Egyptian blue was instead used to modulate certain areas of the flesh tones, especially since there is no visible blue in the por-trait.94
The most extensive study on incarnations in mummy por-traits has been conducted by Sand et al., who examined thirteen mummy portraits from German museum collections as part of the ISIMAT project. Sand et al. cate-gorise the portraits in three groups, relating to the visual
88 Ramer 1979, 5.
89 Salvant et al. 2017, 825. 90 Lalli et al. 2012, 38–9.
91 Cartwright and Middleton 2008, 63, Tab. 4.
92 Salvant et al. 2017, 827–8. 93 Ganio et al. 2015, 814, Fig. 1. 94 Ganio et al. 2015, 817.
Figure 5: Infrared reflectogram showing the verso of a nearly effaced mummy portrait with a sketch in black ink, accompa-nied by instructions (most likely for the painter) in ancient Greek. Anonymous, Mummy portrait with sketch on reverse, Phoebe A. Hearst Museum of Anthropology, Berkeley CA (museum number: 6-21378a), 96–192 CE, Tebtunis (Tell Umm e-Bara-gat), Cemetery 7 or 8, encaustic on wood (recto), black ink on wood (verso), L. 35.2 x W. 24.3 x D. 1.2 cm. Credit: courtesy of the Phoebe A. Hearst Museum of Anthro-pology, photography by the J. Paul Getty Museum (Cat. no. 6-21378a). Use of image permitted under an Attribution-Non-Commercial-ShareAlike 4.0 (BY-NC-SA 4.0) Creative Com-mons License.
characteristics of the (presumed) binding media: tempera, hot encaustic and cold encaustic. They present their findings accordingly.95 This thesis strives to move away from the
para-digm of classifying mummy portraits on basis of hypothetical binding media; so, in summa-rising their results here, the trichotomy of Sand et al. will not be followed.
A general feature of the mummy portraits examined by Sand et al. is that the back-ground is painted first, sparing out the figure. Painting the flesh tones is the next step, fol-lowing – and covering – the markers of the earlier compositional planning. Painting the in-carnations begins with the application of a rosy and ochreous ground tone. The eyes and mouth, and sometimes the eyebrows and nasal contours, are spared out in this process. De-pending on the direction of the light, the face is partitioned in lighter and darker areas. Paint-ing in the eyes and mouth is a multi-step process; Sand et al. reconstructed this process for the eyes of one portrait in fourteen steps.96 Lips are red, painted in varying degrees of detail
but always with a darker upper lip.
Filling in the larger fleshy portions is done in distinct areas of warmer, rosy tones and cooler whitish-yellow tones. Even though these areas are blended later, they remain visibly distinct.97 Shadows also appear to follow a system: brownish-red hues in the form
shadows, more greenish-brown in the cast shadows.98 The contours are depicted as form
shadows in brownish-red, albeit darker and more distinct. Doxiadis observed that red and black tones predominate in male portraits, and lighter yellowish and white tones in the fe-male ones, with red being more localised as blush on the cheeks and lips.99
Transitional areas are blended differently, possibly in relation to the handling prop-erties of the paint. Particularly thin paint surfaces are characterised by sharper distinctions between light and dark, while seamless blending characterises the slightly thicker flat sur-faces.100 Pastose paint is blended more roughly, especially around the cheeks. This blending
can be done with a (stiff) brush, but alternative methods are also observed. For example, in one portrait the pastose paint was blended with a rhombic tool – interpreted as a heated cauterium – of which the impressions are still clearly visible.101 For this frequently occurring
surface treatment Doxiadis coined the term ‘wounding’, which is a fitting description re-gardless of tool-specific interpretations.102
The final stages of depicting the incarnation revolved around details. Contours would be emphasised if necessary and smaller areas of light and dark would be painted in loosely with a brush. The final stage of the incarnation consists of applying bright white highlights. These highlights are located mainly in the neck, above the lips, on the nose and in the eyes. For one mummy portrait studied within the ISIMAT project Thieme et al. studied this com-plex process of incarnation-building in-depth, resulting in a map of features in the flesh tones as well as a partial reconstruction of the primary stages of modelling.103
It should be kept in mind that there would always have been an interplay between protocols for painting flesh tones and artistic liberties. There are not enough studies on mummy portrait incarnations available to assess how the ISIMAT portraits relate to the in-ternational corpus, but the usage of pigments does suggest that these portraits are
95 For the full discussion on the painting tech-nique of (flesh tones in) these mummy portraits see Sand et al. 2017a, 178–92.
96 Sand et al. 2017a, 174, Fig. 1. 97 Sand et al. 2017a, 188, Fig. 23. 98 Sand et al. 2017a, 188.
99 Doxiadis 1995, 100.
100 Sand et al. 2017a, 181, 192. 101 Sand et al. 2017a, 184–6. 102 Doxiadis 1995, 97. 103 Thieme et al. 2017, 167–9.
3.8 Conclusion
Studying mummy portrait production yields information about a manifold of Romano-Egyp-tian studio practices. Certain materials such as limewood and beeswax-based paints were evidently preferred. The pigment palette seems to be a unifying factor as well; mummy por-trait painters utilised a similar range of mostly inorganic pigments. On the other hand, there is an abundance of material-technical diversity. Ground layers appear in different types, if at all. Furthermore, binding media were not always beeswax-based and the existence of ‘mixed techniques’ shows that experimenting was possible.
This image also reflects in the depiction of flesh tones. The same pigments predom-inate in incarnations: lead white, minium, ochres and (possibly) red lakes. In some mummy portraits there is evidence for the potential use of Egyptian blue as modulation pigment in flesh tones.
The examination of thirteen mummy portraits in German collections provided valu-able insights about flesh tone depiction. The incarnations were built up in a multi-phase layering process and adjacent areas were blended according to the properties of the paint. Similarities in the build-up of the flesh tones suggests a protocol for painting incarnations in mummy portraits. Further research into this topic remains nevertheless desirable if these findings are to be assessed for their representativeness.
104 Sand et al. 2017b, 198–9. Red lake pigments are also mentioned in various instances, but this
could not be corroborated (non-invasively) within the scope of the project.
