RHA - Medieval Colours

Medieval Colours
between beauty and meaning

issn 218 2-3 294
N. I
SRIE W 2011
fac u l da d e d e c i n c i a s s o c i a i s e h u m a n a s u n l
Entrevista com Michel Pastoureau

conduzida por Adelaide Miranda e Ana Lemos
10
...............................................................................
Value and beauty: towards a double aesthetic of colours in early Romanesque book illumination
Anja Grebe
19
tude des rapports entre la couleur et le corps dans limage mdivale: lvangliaire dOtton III, fin x sicle
e
39
Marie Aschehoug-Clauteaux
Remarks on colours and pigments in the French court illumination of the 13th century
Xenia Muratova
50
...............................................................................
Byzantine recipes and book illumination
Doris Oltrogge
59
The Strasbourg family texts: originality and survival. A survey of illuminating techniques in medieval South Germany
Sylvie Neven
73
Les couleurs de lenluminure: recettes de Michelino da Besozzo et dAntoine de Compigne

Ins Villela-Petit
86
propos dune notice sur le vermillon

Horcio Peixeiro
103
Binding media in medieval manuscript illumination: a source research

Stefanos Kroustallis
112
...............................................................................
la recherche des pigments
Claude Coupry
126
Colours versus colorants in art history: evaluating lost manuscript yellows

Mark Clarke
138
ndice
Editorial
152
The colour of medieval Portuguese illumination: an interdisciplinary approach

Maria Joo Melo, Adelaide Miranda, Catarina Miguel, Rita Castro, Ana Lemos, V. Solange F. Muralha, Joo A. Lopes and Antnio e Pereira Gonalves
174
Illuminations: secrets, alchemy and conservation in three case studies

Marina Bicchieri, Michela Monti, Giovanna Piantanida e Armida Sodo
182
Early Netherlandish manuscript illumination: technical aspects of illuminations in the Rime Bible of Jacob van Maerlant
Arie Wallert
194
Analysis of the Anjou Bible

Marina van Bos e Lieve Watteeuw
205
That it seemeth to be the thing itself. The obsession of 16th century miniature painters to imitate the beauty of nature
Timea Tallian e Alan Derbyshire
...............................................................................
218
Non-invasive XRF and UV-Vis-NIR reflectance spectroscopic analysis of materials used by Beato Angelico in the manuscript Graduale n. 558
M. Picollo, A. Aldrovandi, A. Migliori, S. Giacomelli e M. Scudieri
229
A protocol for non-invasive analysis of miniature paintings

Maurizio Aceto, Angelo Agostino, Monica Gulmini, Eleanora Pellizzi e Valentina Bianco
242
The application of SR-XRF to the analysis of manuscript illumination a case study

Marco Battaglia, Laura Alidori Battaglia, Richard Celestre, Peter Denes, Dionisio Doering, Tae Sung Kim e Sarah Zalusly
252
Combining visible and infrared imaging spectroscopy with site specific, in-situ techniques for material identification and mapping
Paola Ricciardi e John K. Delaney
RECENSES CRTICAS
264
Michel Pastoureau. Noir. Histoire dune couleur. ditions du Seuil, 2008

Adelaide Miranda e Rita Carvalho
r e v i s ta d e h i s t r i a d a a r t e n. e s p e c i a l - 2 0 1 1
Mark Clarke. Mediaeval Painters Materials and Techniques: The Montpellier Liber diversarum arcium. Archetype, London, 2011
Stefanos Kroustallis
266
VARIA Des interactions entre scriptoria Portugais au XIIe sicle

Rmy Cordonnier
270
Highlighting manuscripts third dimension. Access, documentation and display of micrometric details
Ins Correia
282
volta de um vermelho. Apresentao de edio dO livro de como se fazem as cores, sob o olhar da cincia e tecnologia
Maria Joo Melo e Catarina Miguel
288
Interactive technology to explore medieval illuminations

Nuno Correia, Tarqunio Mota, Rita Carvalho e Andr Ricardo
292
NOTCIAS Colour in Medieval Illuminated Manuscripts: Between Beauty and Meaning

Maria Joo Melo
298
O projecto Imago
Ana Lemos
300
The awakening of the Manueline Foral charters

Ana Isabel Seruya e Maria Lusa Carvalho
301
Studies in medieval manuscript illuminations: Master and PhD thesis

Faculty of Social Sciences and Humanities and Faculty of Sciences and Technology, New University of Lisbon
302
Seminrio e Exposio. Os Livros de Horas do Palcio Nacional de Mafra e a cultura artstica do sculo XV
Ana Lemos
304
r e v i s ta d e h i s t r i a d a a r t e n. e s p e c i a l - 2 0 1 1
Editorial
Os cdices medievais iluminados so testemunhos da construo ideolgica e cultural Europeia, assumindose ainda como objectos artsticos. No caso de Portugal, nos scs. XII e XIII, constituem a nica arte da cor, j que se perderam as pinturas murais e sobre tbua. Este nmero especial da Revista de Histria da Arte apresenta o resultado cientfico do congresso Medieval colours: between beauty and meaning. An interdisciplinary conference on the study of colour in medieval manuscripts, que decorreu na FCT e na FCSH da Universidade Nova de Lisboa, de 10 a 11 de Setembro 2009, sendo organizado conjuntamente pelo Departamento de Conservao e Restauro e pelos Instituto de Histria da Arte e Instituto de Estudos Medievais. O estmulo para a organizao deste congresso nasceu da investigao efectuada no mbito de dois projectos financiados pela FCT-MCTES e coordenados por Adelaide Miranda e Maria Joo Melo. A equipa interdisciplinar pretende com estes projectos valorizar os manuscritos iluminados medievais (scs. XII/XIII), um patrimnio riqussimo, ligado formao da nacionalidade, que alia a importncia cultural dos textos valorizao artstica da iluminura. Pela primeira vez foi possvel reunir os mais notveis investigadores nesta rea, incluindo qumicos, historiadores da arte, cientistas da conservao e restauradores, entre outros. Os conferencistas convidados, Michel Pastoureau, Mark Clarke, Claude Coupry, Doris Oltrogge e Yarza Luaces apresentaram o estado da arte. As comunicaes dos participantes de diferentes origens e instituies (universidades, museus, laboratrios) possibilitaram uma discusso frutuosa das diversas abordagens de investigao. Concluiu-se que se encontravam reunidas as condies necessrias para a criao de uma rede de investigadores e de uma base de dados internacional, que teria como objectivo principal a construo de um Atlas dos pigmentos e corantes medievais. Este Atlas ser um passo importante no conhecimento e divulgao dos colorantes e significado da cor utilizada na iluminura medieval. Os artigos deste nmero encontram-se, tal como no congresso, organizados em torno de quatro eixos temticos: significado da cor, cor nas fontes escritas
Medieval illuminated manuscripts are among the most valuable objects of the European cultural heritage, and are testimonies to medieval ideas, religion and policy. For Portugal, in the 12th-13th centuries, they may be considered the only surviving artistic colour, because both mural and panel paintings are practically lost. This special issue of Revista de Histria da Arte is devoted to the conference Medieval colours: between beauty and meaning. An interdisciplinary conference on the study of colour in medieval manuscripts, organized by the Department of Conservation and Restoration (FCT) together with the Institute of Art History and Institute for Medieval Studies (FCSH), on 10th11th September 2009. The impetus for organizing this conference arose from research carried out under two projects funded by FCT-MCTES and coordinated by Adelaide Miranda and Maria Joo Melo. In these projects the interdisciplinary team aimed to evaluate medieval Portuguese illuminated manuscripts (12th-13th centuries), a rich heritage, linked to the formation of nationality, which combine the cultural importance of texts with the artistic appreciation of illumination. For the first time it was possible to bring together the most eminent researchers in this discipline: chemists, art historians, and conservation-restoration scientists. The invited speakers, Michel Pastoureau, Mark Clarke, Claude Coupry, Doris Oltrogge and Yarza Luaces, put into context the state-of-the-art and methods in this field. The communications presented by researchers of diverse geographic origins and institutions (universities, museums, chemistry laboratories) enabled a fruitful discussion of the different methods of research. This conference and this publication demonstrate conclusively that genuinely committed and open interdisciplinary collaboration is essential in colour research and manuscript studies, and indicates the value of Technical Art History across all fields of art history. A wide knowledge of cultural history helps us (for example) to avoid anachronistic interpretations of colour meanings. The careful observations of conservator-restorers, combined with chemical and physical analysis in the laboratory, helps us evaluate how colours have altered or been lost; these observations
r e v i s ta d e h i s t r i a d a a r t e s r i e w n. i - 2 0 1 1
medievais, os materiais da cor, novas perspectivas na anlise cientfica da iluminura medieval. Entrevista-se ainda Michel Pastoureau, historiador e pioneiro nos estudos da cor numa perspectiva social e simblica, que nos alerta para as armadilhas mais frequentes neste campo de estudos; mas que por outro lado, nos chama a ateno para a potencialidade e riqueza desta investigao desde que levada a cabo em contextos especficos e combinando resultados alargados. O estudo da histria das cores tambm atractivo e ldico, como nos conta Pastoureau (...) avec mes enqutes sur lhistoire des couleurs, je semblais par trop me faire plaisir. E esse prazer que gostaramos de poder partilhar com o leitor!
Maria Adelaide Miranda, FCSH-UNL

mmac@fcsh.unl.pt
Maria Joo Melo, FCT-UNL

mjm@dq.fct.unl.pt
Mark Clarke, UvA

mark@clericus.org
and analyses, when further combined with the insights and knowledge from traditional art history, help us mentally reconstruct original appearances. Above all, interdisciplinary discussion was shown to be the royal road to formulating meaningful research questions for the curator in the gallery, the art historian in the study, or the chemist in the laboratory. The papers published here are organized thematically: the meaning of colour, colour in mediaeval written sources, the materials of colour, and new trends in the analysis of medieval manuscripts. An interview with Michel Pastoureau historian, anthropologist and pioneer of studies in the social and symbolic meanings of colour provides the perfect framework for the topic. As an historian who pioneered the social and symbolic studies in colour, he alerts us to the many traps that may be found in this field of study, highlighting at the same time the extraordinary impact and richness that the research on the social history of colour may bring. Colour: between beauty an meaning is also an extremely attractive field of research, or as Pastoureau stated: as a researcher we can even have fun with it! avec mes enqutes sur lhistoire des couleurs, je semblais par trop me faire plaisir. 1
1. with my researches in the history of colors, I seem to please myself greatly.
participantes do congresso medieval colours: between beauty and meaning na biblioteca da fct-unl, campus caparica, 10 setembro 2009
e n t r e v i s ta c o m m i c h e l pa s to u r e a u
michel pastoureau
ichel Pastoureau*, n le 17 juin 1947 Paris, est un historien mdiviste franais, spcialiste de la symbolique des couleurs, des emblmes, et de lhraldique. Michel Pastoureau est le petit-cousin de Claude Lvi-Strauss et le fils dHenri Pastoureau, proche des surralistes ; archiviste palographe, sa thse de lcole des chartes, soutenue en 1972, porte sur le bestiaire hraldique du Moyen ge. Il est historien, et directeur dtudes cole des hautes tudes en sciences sociales et lcole pratique des hautes tudes, o il occupe depuis 1983 la chaire dhistoire de la symbolique occidentale. Il a t lu le 28 avril 2006 correspondant franais de lAcadmie des
inscriptions et belles-lettres. Il est membre de lAcadmie internationale dhraldique et prsident de la Socit franaise dhraldique et de sigillographie. Il a publi une quarantaine douvrages, dont certains traduits dans plusieurs langues, consacrs lhistoire des couleurs, des animaux et des symboles. Notamment, Couleur, images, symboles, Bleu: lhistoire dune couleur, Une histoire symbolique du Moyen-Age occidental, Jsus chez le teinturier, Noir. Le 3 novembre 2010, il reoit le prix Mdicis essai pour son ouvrage Les Couleurs de nos souvenirs.
* Biografia adaptada da Wikipedia
10
com
conduzida por MARIA ADELAIDE MIRANDA ANA LEMOS
Michel Pastoureau, pouvez-vous nous raconter comment est ne votre attirance, et mme votre passion, pour les couleurs ?
Mon attirance pour les couleurs est ne trs tt, ds la petite enfance. Jai eu la chance de natre dans une famille o la peinture occupait une place importante. Trois des oncles de ma mre taient artistes peintres et, mme si je nai connu que lun dentre eux, du ct maternel tous les appartements de mes grand-tantes, de mes oncles et tantes, de mes cousins et mme de mon arrire-grand-mre, morte 96 ans, taient encombrs de tableaux, certains de grand format. Du ct de mon pre, les peintres nappartenaient pas la famille mais au cercle des amis proches. Ctaient tous des artistes sinscrivant de prs ou de loin dans la mouvance surraliste et essayant de vivre plus ou moins bien de leur peinture. Mon pre memmenait souvent dans leur atelier, terrain de jeux fascinant pour un petit garon. Parfois les peintres amis de mon pre me faisaient cadeau de tubes de peinture usags, destins la poubelle mais reprsentant pour moi le plus beau des prsents. Plaisir des yeux et du toucher plus que vritable bonheur cratif car ces tubes presque vides et desschs, souvrant mal ou pas du tout, ne mtaient gure utiles pour peindre une fois de retour la maison. Non seulement enlever le bouchon tait pratiquement impossible, mais je ne disposais ni du matriel ni des connaissances ncessaires pour peindre lhuile. Jtais cependant fier de dtenir un tel trsor, de classer les tubes par couleur, de caresser leur enveloppe de plomb, lourde et mate,
Entrevista
11
michel pastoureau
et surtout de les montrer firement mes camarades, qui ne possdait quune boite de crayons de couleur. Je dois galement mon pre une habitude qui ma accompagn tout au long de ma vie: la visite des muses et des expositions artistiques. Il my a emmen trs jeune, et ce got pour les muses et les expositions ne ma plus jamais quitt. Mais je dois aussi ma mre, passionne de botanique, une partie de mon attirance pour les couleurs. Elle tait pharmacienne. Sa pharmacie, situe Paris tout en haut de la butte Montmartre, tait galement pour le jeune enfant que jtais un magnifique terrain de jeux. Jaimais notamment observer comment les botes de mdicaments taient ranges et classes par couleur: bleus pour les calmants et les somnifres, orangs pour les produits toniques et reconstituants, bruns pour les laxatifs, verts pour les produits base de plantes. Le noir, signe de mort, ntait pas utilis en pharmacie, et le rouge servait attirer lattention sur les produits dangereux: Ne pas dpasser la dose prescrite. Je navais pas encore dix ans mais classer les couleurs tait dj chez moi un plaisir rcurrent. Que ce soit dans la pharmacie maternelle ou dans latelier des peintres, je mimaginais quil existait un ordre chromatique secret quil me fallait retrouver. En grandissant, jai dvelopp une sorte dhypersensibilit aux couleurs (ce qui ma parfois jou de vilains tours) et je me suis progressivement fabriqu un certain nombre de principes personnels. Des principes pas spcialement originaux, mais acquis de bonne heure et que par la suite je nai jamais eu motif ni envie de remettre en question, pas mme lorsque devenu historien jai compris quil ny avait pas de vrits chromatiques universelles mais quau contraire tout variait selon les poques et les socits. Ces quelques principes, forgs ds lenfance mais qui mont accompagn dans ma vie adulte, dans mon travail de chercheur et mme dans mes modestes activits de peintre du dimanche, peuvent se rsumer ainsi: 1. Le noir et le blanc sont des couleurs part entire. 2. Il nexiste que six couleurs de base: le noir, le blanc, le rouge, le bleu, le jaune et le vert. 3. Viennent ensuite cinq couleurs du second rang, que lon appelle tort des demi-couleurs: le gris, le brun, le rose, le violet et lorang. 4. Toutes les autres colorations ne sont que des nuances ou des nuances de nuances.
Comment sest pass le dbut de votre carrire ? Quel rle a pu jouer Georges Duby?
Jeune chercheur enthousiaste et naf, jai commenc mintresser lhistoire des couleurs tous problmes confondus au milieu des annes 1970. Lenvironnement familial, la frquentation des peintres, les visites rgulires faites aux muses, une attention maladive porte aux couleurs des vtements et une thse rcemment soutenue sur les armoiries mdivales: tout mavait prpar et conduit vers de telles recherches. Je les croyais faciles et estimables, bien acceptes dans le milieu des historiens et pratiques par de nombreux chercheurs. Je me trompais totalement.
12
A cette poque, chez les historiens, les archologues et les historiens de lart, personne, absolument personne ne sintressait aux couleurs. Mme dans des domaines o lon se serait attendu trouver des tudes sur un tel sujet, la couleur brillait par son absence. Lhistoire du vtement, par exemple, tait une histoire totalement achrome. Les documents existaient pourtant, et en grand nombre, mme pour les poques anciennes, mais les spcialistes du costume ne sy intressaient pas. Seules comptaient larchologie des formes et la nature des diffrentes pices composant le vtement au fil des ges. Lide dun systme vestimentaire, se situant au coeur de la vie en socit et au sein duquel les couleurs auraient pu jouer un rle essentiel, leur tait trangre. Comme leur tait tranger le nom de Roland Barthes Mais il y avait pire: lhistoire de la peinture. Dans une discipline o la couleur aurait d, par sa nature mme, occuper le premier rang, elle tait presque toujours passe sous silence. Des livres entiers, pais, savants, scrivaient sur luvre dun peintre ou sur tel ou tel mouvement pictural, sans que jamais leurs auteurs ne parlent des couleurs. Trois cents ou cinq cents pages sans formuler une seule ide, une seule remarque, un seul mot concernant les couleurs, pas mme des mots comme bleu, rouge, jaune: coup sr, ctait une performance! Elle tait courante, presque gnrale au milieu des annes soixante-dix. La couleur tait la grande absente de lhistoire de lart. Je tombais des nues. Mais je ntais pas au bout de mes dcouvertes. Aprs quelques mois denqutes et de rflexions sur lhistoire des couleurs, terrain quasi vierge, je constatais que mon travail tait mal reu. Soit un tel sujet semblait peu srieux, frivole mme (javais pourtant commenc par lhistoriographie), soit il semblait totalement vain, la fois inutile et impossible conduire. En outre, il paraissait plus ou moins immoral. Ctait lpoque o les historiens et dune manire plus gnrale les chercheurs en sciences humaines avaient des devoirs envers la socit mais fort peu de droits, et surtout pas le droit de se faire plaisir. Lide que le plaisir individuel du chercheur pouvait tre le moteur premier de sa recherche ntait pas dans lair du temps, ou bien tait condamne. Or, avec mes enqutes sur lhistoire des couleurs, je semblais par trop me faire plaisir. Ctait une attitude individualiste, presque indcente, sinon dangereuse, et en tout cas contraire lthique du chercheur. Heureusement, deux grands historiens mdivistes, qui ntaient pas historiens de lart, mont encourag dans mes recherches et en ont vu lintrt: Georges Duby et Jacques Le Goff. Je leur dois beaucoup. Duby, qui ses moments perdus peignait quelques tableaux pour se dtendre, dans un style semi-figuratif, a tout de suite vu comment la couleur pouvait tre un nouvel objet dhistoire. Il ma pouss conduire mes enqutes dans tous les domaines, du lexique aux symboles en passant par les emblmes et la cration artistique. Le Goff, avec qui jai fait un sminaire commun pendant un dizaine dannes, tait plus attir par lhistoire sociale des couleurs: vtements et codes vestimentaires, lois somptuaires, mtiers de la teinturerie, place des couleurs dans la vie quotidienne.
Votre approche de la couleur est tout fait novatrice. Elle se situe au carrefour de lhistoire sociale, artistique culturelle et symbolique. Quels
13
conseils donneriez-vous aux jeunes chercheurs qui sont de plus en plus nombreux sintresser ce domaine de recherche ?
Aux jeunes chercheurs qui veulent se lancer dans ltude des couleurs au Moyen Age (ou bien dans lAntiquit ou au dbut de lpoque moderne), je leur conseillerai de se mfier de tout anachronisme. Il est en effet impossible de projeter tels quels sur les images, les monuments, les oeuvres et les objets produits par les sicles passs nos dfinitions, nos conceptions et nos classements actuels de la couleur. Ce ntaient pas ceux des socits dautrefois (et ce ne seront peut-tre pas ceux des socits de demain...). Le danger de lanachronisme guette toujours lhistorien et lhistorien de lart peut-tre plus que tout autre chaque coin de document. Mais lorsquil sagit de la couleur, de ses dfinitions et de ses classements, ce danger semble plus grand encore. Rappelons par exemple que pendant des sicles et des sicles, le noir et le blanc ont t considrs comme des couleurs part entire (et mme comme des ples forts de tous les systmes de la couleur); que le spectre et lordre spectral des couleurs sont pratiquement inconnus avant le XVIIe sicle ; que larticulation entre couleurs primaires et couleurs complmentaires merge lentement au cours de ce mme sicle et ne simpose vraiment quau XIXe; que lopposition entre couleurs chaudes et couleurs froides est purement conventionnelle et fonctionne diffremment selon les poques et les socits. Au Moyen Age et la Renaissance, par exemple, le bleu est considr en Occident comme une couleur chaude, parfois mme comme la plus chaude de toutes les couleurs. Cest pourquoi lhistorien qui chercherait tudier dans un vitrail ou dans une miniature du Moyen Age ou bien dans un tableau de Raphal ou du Titien la proportion entre les couleurs chaudes et les couleurs froides et qui croirait navement que le bleu y est, comme aujourdhui, une couleur froide, se tromperait compltement et aboutirait des absurdits. Les notions de couleurs chaudes ou froides, de couleurs primaires ou complmentaires, les classements du spectre ou du cercle chromatique, les lois de la perception ou du contraste simultan ne sont pas des vrits ternelles mais seulement des tapes dans lhistoire mouvante des savoirs. Ne les manions pas inconsidrment, ne les appliquons pas, sans prcaution aucune, aux socits du pass. Prenons un exemple simple et attardons-nous sur le cas du spectre. Pour nous, depuis les expriences de Newton, la mise en valeur du spectre et la classification spectrale des couleurs, il est incontestable que le vert se situe quelque part entre le jaune et le bleu. De multiples habitudes sociales, des calculs scientifiques, des preuves naturelles (ainsi larc-en-ciel) et des pratiques quotidiennes de toutes sortes sont constamment l pour nous le rappeler ou pour nous le prouver. Or, pour lhomme de lAntiquit, du Moyen Age et encore de la Renaissance cela na gure de sens. Dans aucun systme antique ou mdival de la couleur, le vert ne se situe entre le jaune et le bleu. Ces deux dernires couleurs ne prennent pas place sur les mmes chelles ni sur les mmes axes; elles ne peuvent donc avoir un palier intermdiaire, un milieu qui serait le vert. Le vert entretient des rapports troits avec le bleu mais il nen a
14
aucun avec le jaune. Au reste, que ce soit en peinture ou en teinture, aucune recette ne nous apprend avant le XVe sicle que pour faire du vert il faille mlanger du jaune et du bleu. Peintres et teinturiers savent fabriquer la couleur verte, bien videmment, mais pour ce faire ils mlangent rarement ces deux couleurs. Pas plus quils ne mlangent du bleu et du rouge pour obtenir du violet. Pour ce faire, ils mlangent du bleu et du noir: le violet est un demi-noir, un sous-noir; il lest du reste encore dans la liturgie catholique et dans les pratiques vestimentaires du deuil. Lhistorien doit donc se mfier de tout raisonnement anachronique. Non seulement il ne doit pas projeter dans le pass ses propres connaissances de la physique ou de la chimie des couleurs, mais il ne doit pas prendre comme vrit absolue, immuable, lorganisation spectrale des couleurs et toutes les thories qui en dcoulent. Pour lui comme pour lethnologue, le spectre ne doit tre envisag que comme un systme parmi dautres pour classer les couleurs. Un systme aujourdhui connu et reconnu de tous, prouv par lexprience, dmont et dmontr scientifiquement, mais un systme qui peut-tre, dans deux, quatre ou dix sicles, fera sourire ou sera dfinitivement dpass. La notion de preuve scientifique est elle aussi troitement culturelle; elle a son histoire, ses raisons, ses enjeux idologiques et sociaux. Et sans mme solliciter la notion de preuve, que penser de lhomme mdival dont lappareil de vision nest aucunement diffrent du ntre qui ne peroit pas les contrastes de couleurs comme lhomme daujourdhui. Au Moyen Age, en effet, deux couleurs juxtaposes qui pour nous constituent un contraste fort peuvent trs bien former un contraste relativement faible; et inversement, deux couleurs qui pour notre oeil voisinent sans aucune violence peuvent hurler pour loeil mdival. Gardons lexemple du vert. Au Moyen Age, juxtaposer du rouge et du vert (la combinaison de couleurs la plus frquente dans le vtement entre lpoque de Charlemagne et le XIIe sicle) reprsente un contraste faible, presque un camaeu. Or pour nous il sagit dun contraste violent, opposant une couleur primaire et sa couleur complmentaire. Inversement, associer du jaune et du vert, deux couleurs voisines dans le spectre, est pour nous un contraste relativement peu marqu. Or cest au Moyen le contraste le plus dur que lon puisse mettre en scne: on sen sert pour vtir les fous et pour souligner tout comportement dangereux, transgressif ou diabolique!
Quelles sont les autres difficults que rencontre lhistorien des couleurs?
Elles sont au moins de deux sortes. Sur les monuments, les uvres dart, les objets et les images que les sicles passs nous ont transmis, nous voyons les couleurs non pas dans leur tat dorigine mais telles que le temps les a faites. Lcart est parfois considrable. Or ce travail du temps, quil soit d lvolution chimique des matires colorantes ou bien laction des hommes qui, au fil des sicles, ont peint et repeint, modifi, nettoy, vernis ou supprim telle ou telle couche de couleur pose par les gnrations prcdentes, est en lui-mme un document dhistoire. Ds lors, que faire? Faut-il, avec des moyens techniques aujourdhui trs sophistiqus, restaurer les couleurs, tenter de
15
les remettre dans leur tat dorigine? Il y a l un positivisme qui me parat la fois dangereux et contraire aux missions de lhistorien: le travail du temps fait partie intgrante de sa recherche; pourquoi le renier, leffacer, le dtruire? En outre, les grands peintres savent trs bien que leurs pigments vont voluer, que leurs couleurs vont se transformer: ils agissent en consquence, et ce quils souhaitent pour la postrit ce nest pas ltat premier du tableau ou du panneau mais un tat ultrieur, queux-mmes ne verront pas. Ainsi, lorsque nous essayons de remettre tel tableau ou tel uvre dart dans son tat premier nous allons parfois lencontre de la volont de lartiste. La ralit historique nest pas seulement ce quelle a t dans son tat dorigine, cest aussi ce que le temps en a fait. Mais jusquo peut-on laisser le temps faire son oeuvre? Autre difficult documentaire: nous voyons aujourdhui les images et les couleurs du pass dans des conditions dclairage trs diffrentes de celles quont connues les socits antrieures au XXe sicle. La torche, la lampe huile, la chandelle, le cierge, la bougie produisent une lumire qui nest pas celle du courant lectrique. Cest une vidence. Et pourtant quel historien, quel visiteur de muse ou dexposition, quel amateur dart ancien en tient compte? Aucun. Or loublier conduit parfois des absurdits. Prenons pour exemple la rcente restauration des votes de la chapelle Sixtine et les efforts considrables tant techniques que mdiatiques pour retrouver la fracheur et la puret originelle des couleurs poses par Michel Ange. Un tel exercice stimule certes la curiosit, mme sil agace un peu, mais il devient parfaitement vain et anachronique si lon claire la lumire lectrique les couches de couleurs ainsi dgages. Que voit-on rellement des couleurs de Michel Ange et de ses lves avec nos clairages modernes? La trahison nest-elle pas plus grande que celle qua lentement opre le travail du temps depuis le XVIe sicle? Plus inquitante aussi, quand on songe lexemple de Lascaux ou celui dautres sites prhistoriques, dtruits ou endommags par la rencontre funeste des tmoignages du pass et des curiosits daujourdhui. Mais, dun autre ct, il est impossible, absurde mme, de remettre dans la chapelle Sixtine des chandelles ou des lampes huile. Que faire? Les clairages du pass sont tous produits par des flammes. Celles-ci font bouger les formes et les couleurs des images et des tableaux, elles les animent, les font vibrer, les rendent mme cintiques (pensons un document comme la broderie de Bayeux regard la lumire de torches ou de chandelles). Nos clairages lectriques, au contraire, sont relativement statiques, ils ne font bouger ni les formes ni les couleurs. Do un cart de sensibilit considrable entre notre regard et celui de nos anctres. Quon le veuille ou non, nous ne percevrons jamais comme eux un objet, un document, une uvre dart. Pour un il antique, mdival ou moderne, les couleurs sont toujours en mouvement Aristote souligne dj combien toute couleur est mouvement. Pour lil daujourdhui les couleurs ne bougent pas, ou gure, elles semblent immobiles: la diffrence de perception est immense. De mme, nous navons aucune difficult pour clairer de manire uniforme une grande surface. Dans un muse contemporain, un tableau de trois mtres de haut sur cinq de long ne prsente aucune zone moins claire que les autres. Grce aux
16
spots et des lumires artificielles de plus en plus perfectionnes, le tableau est parfaitement clair dans son entier sans que la couche picturale en souffre. Un tel exercice est impossible pour les socits du pass. Quelles que soient leurs natures et leurs performances, les clairages base de flammes ne peuvent clairer de faon gale une surface un peu importante. Il y a toujours des zones bien claires et dautres qui restent dans lombre. Do ces jeux de clairs-obscurs auxquels sont si sensibles les artistes et les publics dautrefois. Larrive de llectricit a totalement modifi le rapport du spectateur lobjet, luvre dart, limage et, peut-tre plus encore, la couleur.
Quelles informations peuvent fournir, selon vous, les analyses matrielles et scientifiques de la couleur dans les documents et les uvres dart du Moyen Age? Vont-elles apporter lhistorien un regard nouveau sur ce champ dtude?
Les analyses de pigments et de colorants peuvent apporter beaucoup. Je souhaite quelles se multiplient et quelles sappliquent tous les supports de la couleur. Mais cela deux conditions. Dune part, il faut tenir compte des dangers de lanachronisme dont je viens de parler. Le savoir mdival nest pas le savoir daujourdhui, et mme sil est lgitime dutiliser nos savoirs daujourdhui pour essayer de mieux comprendre le pass, il faut le faire avec une certaines prudence. Dautres part, les analyses en laboratoire nont de raison dtre que si elles sont relies une problmatique. Que cherche-t-on exactement? Faire des analyses pour faire des analyses ne sert rien. Or jai limpression quassez souvent cest ce qui se passe aujourdhui: on fait des analyses sans but trs prcis, sans problmatique. Ou bien on oublie en cours danalyse ce que lon cherchait au dpart. Il faut une collaboration plus troite entre les scientifiques qui effectuent les analyses en laboratoire, et les historiens qui indiquent dans quelles direction les conduire. Il faut srier les problmes, les buts, les enjeux. Il faut constamment se demander: Des analyses, pour quoi faire?. Dans certains domaines, ltude en laboratoire des pigments apporte des informations importantes. Prenons un exemple simple : un panneau peint italien du XVe sicle reprsentant une Vierge au manteau. A loeil nu, le bleu du manteau de la Vierge, place au centre du panneau, nous semblera identique aux autres bleus que lon peut voir ailleurs sur le panneau pour peindre dautres figures. Mais en laboratoire, on pourra constater que le manteau de la Vierge est peint avec du lapis-lazuli un pigment trs cher au XVe tandis que les autres figures sont peintes avec des pigments plus ordinaires, de lazurite par exemple, ou bien de lindigo. Loeil ne voit pas la diffrente mais elle existe et est importante, notamment pour le commanditaire et pour lartiste. Dans ce choix de diffrents pigments bleus pour peindre les diffrentes figures du panneau, il y a des enjeux conomiques, cultuels, symboliques. La Vierge est mise en valeur, honore, vnre mme par le choix dune matire colorante de trs grand prix, et pourtant le spectateur ne sen rend pas compte. Cela nous en dit long sur les pratiques mdivales de la couleur et de la matire.
17
v a l u e a n d b e a u t y : t o w a r d s a d o u b l e a e s t h e t i c o f c o l o u r s i n e a r ly r o m a n e s q u e b o o k i l l u m i n at i o n
Resumo
A investigao sobre manuscritos medievais iluminados ainda muito dominada por uma anlise baseada no estilo e na iconografia. Alm disso, a maioria dos estudos sobre texto-e-imagem concentram-se no contedo e no significado simblico da iluminura, enquanto questes ligadas materialidade e paginao, que se situem para alm da anlise codicolgica, so muitas vezes ignoradas. O mesmo acontece com a anlise das cores: a investigao tem-se focado no seu significado religioso, dinstico ou ritual, considerando-as categorias abstractas sem uma materialidade especfica. Mesmo um especialista em pigmentos medievais como Heinz RoosenRunge (Roosen-Runge 1967), seguiu a tradio iconogrfica da histria da arte e interpretou as cores com base na sua funo, por exemplo, como meio de representao transcendental de ideias e ideais. Pelo contrrio, historiadores como Michel Pastoureau (Pastoureau, 1990) trabalhando para uma histria social das cores, sublinharam a importncia dos pigmentos e do seu simbolismo especfico, e no de conceitos abstractos, para um melhor conhecimento da percepo medieval da cor. A tese de Pastoureau baseou-se principalmente na sua investigao sobre o papel da cor no vesturio e txteis medievais. Com esta contribuio, desejo demonstrar que se tomaram atitudes semelhantes sobre a materialidade da cor em manuscritos iluminados. A recente investigao interdisciplinar sobre o Codex Aureus Epternacensis, produzido em torno a 1045 no mosteiro beneditino de Echternach, conservado no Germanisches Nationalmuseum em Nuremberg, lanou uma nova luz sobre a conscincia que os iluminadores e seus patronos tinham da materialidade da cor e do seu valor. As minhas reflexes basearam-se na anlise cientfica do Codex Aureus levada a cabo por Doris Oltrogge e Robert Fuchs (Fuchs and Oltragge 2009) e, nas minhas prprias pesquisas sobre o uso e esttica das cores na Idade Mdia (Grebe 2007). O precioso Evangelirio contem mais de 40 miniaturas a pgina inteira, pintadas por vrios iluminadores de oficina. Os iluminadores usaram uma paleta muito variada de cores e tons, por exemplo, o vermelho ou prpura no so apenas representados por um, mas por dois ou mais pigmentos. Entre eles encontramos os muito raros e dispendiosos pigmentos exticos, como o lpis-lazli ou o quermes, que foram abundantemente usados neste manuscrito quando comparado com outros manuscritos romnicos. Esta observao permite colocar como hiptese, que os iluminadores de Echternach e os seus patronos quiseram de facto usar o maior nmero possvel de cores e as mais dispendiosas, para incluir a mxima variedade de cores no Evangelirio Dourado. Outro nvel de conscincia cromtica revelou-se pelo facto de muitas miniaturas do ciclo com a Vida de Cristo e a miniatura em duplo folio da Maiestas Domini (fol. 2v-3r) terem sido repintadas durante a produo do manuscrito. O retoque da maioria das carnaes e fundos pode ser interpretado, numa primeira anlise, como um restauro esttico. Surpreendentemente, os iluminadores no alteraram as composies na sua globalidade ou as formas das cabeas ou mos, em vez disso, mudaram apenas o tom da pele e os fundos. O motivo foi, provavelmente, o de evitar contrastes violentos e unificar cromaticamente os duplo flios executados por diferentes iluminadores. O Codex Aureus Epternacensis , assim, a prova da elevada sensibilidade do perodo romnico aos valores cromticos e virtudes materiais de cores e pigmentos, que sero discutidos neste artigo.
palavras-chave
evangelirio manuscrito iluminado esttica das cores otoniano lpis-lazli
19
Abstract
Research on medieval manuscript illumination is still dominated by the analysis of style and iconography. Also, most word-and-image studies concentrate on the content and the symbolical meaning of a miniature, whereas questions of materiality and layout beyond codicology are often neglected. The same is true for the analysis of colours: research so far focuses on the religious, dynastical or ritual symbolism of colours that are considered as abstract categories without a specific materiality. Even a specialist in medieval pigments like Heinz Roosen-Runge (Roosen-Runge 1967) followed the iconographical tradition of art history and interpreted colours with regard to their function, e.g. as media to represent transcendental ideas and ideals. On the contrary, historians like Michel Pastoureau (Pastoureau 1990), working towards a social history of colours, have underlined the importance of pigments and their specific symbolism instead of abstract colour terms for a better understanding of the medieval notion of colours. Pastoureaus thesis is based primarily on his investigations into the role of colours in medieval clothing and textiles. In my contribution, I am trying to show that similar attitudes towards the materiality of colours are manifested in medieval manuscripts. Recent interdisciplinary research on the Codex Aureus Epternacensis, made around 1045 in the Benedictine monastery of Echternach, now in the Germanisches Nationalmuseum in Nuremberg has shed new light on the consciousness illuminators and patrons had of the materiality of colours and their respective value. My reflections are based on the technological examination of the Codex Aureus carried out by Doris Oltrogge and Robert Fuchs (Fuchs and Oltrogge 2009) and my own research on the use and aesthetics of colours in the Middle Ages (Grebe 2007). The precious Gospel Book contains more than 40 full-page miniatures painted by several illuminators of the Ottonian Echternach workshop. The illuminators used a very varied palette of different colours and tones, e.g. the colours red or purple are not only represented by one, but by two or more pigments. Among them we find many rare and expensive exotic pigments like lapis lazuli or kermes, which have been abundantly used in this manuscript compared to other Romanesque manuscripts. This observation leads to the hypothesis that the Echternach illuminators and their patrons explicitly wished to use as many and costly colours as possible to include the maximal variety of colours in the Golden Gospel Book. Another level of chromatic awareness is revealed by the fact that many miniatures in the cycle with the Life of Christ and the double page miniature of the Majestas Domini (fol. 2v-3r) have been painted over during the making of the manuscript. The reworking of most of the fleshy parts and backgrounds can tentatively be interpreted as aesthetical restoration. Surprisingly, the illuminators did not alter the whole composition or touch the forms of the heads and hands, instead, they only changed the tone of the complexion and backgrounds. The reason was probably to avoid violent colour contrasts and to chromatically unify double-pages executed by different illuminators. The Codex Aureus Epternacensis thus proves the high sensibility of the Romanesque period towards chromatic values and material virtues of Agradecimentos por ajuda na configurao do texto e sugestes article.Catarina Sousa, Pedro Fialho colours and pigments, which will be discussed in the de Ana
de Sousa, Justino Maciel, Felix Teichner e Heidi.
key-words
Gospel Book Illuminated Manuscript Aesthetic of Colours Ottonian Lapis lazuli
value and beauty: towards a double aesthetic of colours in early romanesque book illumination
a n ja g r e b e
Department of Medieval Art History University of Bamberg, Bamberg (Germany)
1. From Material to Medium

1. Letter to Jakob Heller dated November 4, 1508, ed. by Rupprich 1956, 67-68. Cf. the letters to Heller dated March 21, 1509: the colours I used are worth more than 24 guilders (Rupprich 1956, 69), and July 24, 1509: I have been working on it [= the altarpiece] for more than a year using ultramarin for approximately 25 guilders. (Rupprich 1956, 71).
In 1508, Albrecht Drer started to work on a large altarpiece depicting the Assumption of the Virgin (Anzelewsky 1991, 221-228; Kutschbach 1995, 71-80; Decker 1996; Grebe 2006, 82-86; Pfaff 1971). The execution of the painting, which took Drer more than one and a half years to complete, was accompanied by an extensive correspondence with the patron, the Frankfurt merchant Jakob Heller. Heller repeatedly complained about the delayed completion of the altarpiece and threatened to withdraw his commission. Drer countered the reproach by pointing to the quality of the execution and his use of the best and most beautiful colours available. These included vltermarin, or lapis lazuli, worth more than 20 ducats, which he claimed to have applied in 5 to 6 layers.1 And instead of apologizing to his patron, he doubled the price of the altarpiece. His arguments were: a painting executed with the maximum amount of materia and ingenium needed more time to finish, and it had its price which Heller actually agreed to pay when he finally saw the painting. The Heller correspondence underlines the importance which was attached to certain colours and pigments at the beginning of the 16th century. Lapis lazuli, for instance, did not only provide a blue tone, but also represented quality and prestige (Pastoureau 2001; Fuchs and Oltrogge 1990; Bender 1990). The Latin name, ultramarin, points to the far-off, exotic, and even mystical origin of the colour prepared on the basis of Afghan lapis lazuli, which was the most expensive pigment apart from gold and was therefore used as distinguishing feature in painting throughout the Middle-
21
colour and meaning
Ages. Drer probably used lapis lazuli for the blue cape of the Virgin thus enhancing its inherent symbolical value by the use of the precious and prestigious material.2 If it had been his aim to communicate only the chromatic symbolism of blue as celestial colour he could have used a less expensive pigment like azurite instead. In his letters to Heller, however, Drer does not mention the religious symbolism of colours, but insists on the value of the colour in its literal, material sense. One should think that the primacy of materia would have been outdated in the age after Alberti. In his treatise De Pittura (1435/36), which Drer knew from a copy in possession of a Nuremberg humanist (Fara 2002, 171-347), Leon Battista Alberti clearly put the referential value of colour, even of gold, above its material value (Alberti 2000, 290-291). According to Monika Wagner, Alberti marks the beginning of modern colour aesthetics: Since Leon Battista Alberti [] colour was no longer judged by its material value. Now, its task was to simulate all kinds of other materials on the surface of the image. (Wagner 2001, 17-18).
2. Unfortunately, the original panel was destroyed by a fire in the Ducal Palace in Munich in 1729. Before the painting was sold to Archduke Maximilian I of Bavaria in 1614 the Nuremberg painter Jobst Harrich made a faithful copy of the original, now in Frankfurt/Main, Historisches Museum, Inv. Nr. B 265, which, however, does not allow any judgement about the paints or pigments used. 3. Cf. Welchman 1990, esp. 8: From the time of Plotinus (A.D. 205-70) until the High Renaissance, writers touching upon the arts and mentioning color were forced to wrestle with the troublesome and dangerous question of sensuality. Although colors were perceived through sight the highest of senses they were still a part of the non-celestial substance of which humankind was made. The only way to elevate the colorful, and hence intrinsically sensual, human sense of sight was to celebrate the divine nature of light.
2. Colours in medieval book illumination: Colour systems and attempts at interpretation

In his Theory of Colours (Farbenlehre) of 1810, Johann Wolfgang von Goethe gave priority to the metaphysics as opposed to the materiality of colours (Schmidt 1965). He described colours as acts of light, referring to medieval colour theories, where colours are described as reflections of the immaterial divine light which Saint Augustine had designated as queen of all colours. 3 Referring to Fritz Haeberleins Essential Features of a Post-antique Iconography of Colours (Grundzge einer nachantiken Farbenikonographie), Lorenz Dittmann characterised the colours in medieval art as intended to represent the divine (Vorstellungsfarben). Released from all referential tasks colours could now function as vocabulary in the system of colour symbolism (Dittmann 1987, 2-3). Although Heinz Roosen-Runge, in his fundamental study on the Colours and Techniques of Early Medieval Book Illumination, puts his main focus on the material nature of colours, he nevertheless interprets them as a means of representation for transcendental truth (Roosen-Runge 1967, vol. 1, 26-29). Today, research on medieval manuscript illumination continues to be dominated by the analysis of style and iconography. Also, most word-and-image studies concentrate on the content and the symbolical meaning of a miniature, whereas questions of materiality and layout, which go beyond codicological issues, are often neglected. The same is true for the analysis of colours: research focuses mainly on the religious, dynastical or ritual symbolism of colours which are considered as abstract categories without a specific materiality. In view of the role of colours as bearers of meaning (Bedeutungstrger) in medieval art and culture one has to ask for the specific use of colours as well as the value(s) attached to them in the literal and figurative sense to fully understand a work of art.
22
4. Cf. Pastoureau 1990, 38: Avant tout codage extra-pictural, venant de lextrieur, la couleur est dabord code de lintrieur, par et pour un document donn. [...] Elle est la fois surface, matire, lumire, clat, fluide, rythme, construction, syntaxe, marque, emblme, symbole, ide, mode, atmosphre, style, art, technique, prix, beaut tous termes certes vagues et discutables mais dont la liste [...] souligne parfaitement ce caractre polysmique et plurifonctionnel de la couleur.
Until now, such questions have mainly been asked by historians. Michel Pastoureau, above all, has been working towards a social history of colours (histoire sociale des couleurs), that takes into consideration the percipient against the backdrop of his culture, ideas, experiences and sensitivity (Pastoureau 1990, 22). According to Pastoureau, research into the use of colours in the Middle Ages should not be based on abstract theory but on the objects and their specific colouring in which the polysemantic and multifunctional character colours is manifested.4 This leads to the hypothesis, that people in the Middle Ages in many cases did not perceive colours as abstract chromatic values, but in their material form as pigments. The case of the red pigment called sanguinis draconis, for instance, clearly shows that it is necessary to develop a kind of pigment symbolism instead of abstract colour terms to help with our understanding of the medieval notion of colours: Quand au XVe sicle un peintre utilise pour tel lment de son tableau le clbre pigment sang-dragon plutt que tel ou tel autre pigment rouge, cest sans doute pour des raisons la fois conomiques, techniques et artistiques, lies au prix de ce pigment, son pouvoir couvrant, aux effets quil permet dobtenir, mais cest aussi et surtout pour des raisons symboliques, lies au nom mme de ce pigment et aux rves que ce nom suscite chez lartiste et son public. (Pastoureau 1989, 39-40)
3. The values of colour between consciousness of material (Materialbewusstsein), symbolic function and aesthetics
Michel Pastoureaus thesis is based primarily on his investigations into the role of colours in medieval clothing and textiles. In the following, I would like to show that medieval manuscripts manifest similar attitudes towards the materiality of colours. The investigations of Heinz Roosen-Runge and Johan Jakob Tikkanen (Tikkanen 1933, 352-443) have shown that medieval illuminators possessed a very detailed knowledge of colours and pigments, which was transmitted orally, by written treatises such as the Mappae Clavicula, the Heraclius or the Theophilus, and by numerous individual collections of recipes and guidelines (Oltrogge 2006; Gullick 1995; Roosen-Runge 1967; Bartl et. al. 2005, 17-34). These texts consist mainly of practical instructions in the preparation and handling of colours. Sometimes however, the introductory remarks contain theoretical reflections on the meaning of colours as well as on the conjunction of materia and ingenium. A 12th-century copy of the Mappae Clavicula provides an example: The arts are learned little by little, step by step. The art of painting is preceded by the preparation of colours. Thus, you should first pay attention to the mixing of each colour. Only then should you start on the actual work, but you
23
colour and meaning
should always strive to let your painting appear decorative as well as natural. This book shows how artistry in combination with ingenuity will enhance your work. (cf. Halleux 1990, 179) 5 In the Early Romanesque period, it is not only treatises, but the works of art themselves which testify to the awareness of illuminators and patrons towards the materiality of colours and their respective value. Up to now, nearly all attempts to draw up a kind of general chromatic grammar or system of colour symbolism based on the use of colours in illuminated manuscripts have failed. It has become clear that the application and distribution of colours varies from book to book, even when entire pictorial cycles are based on the same, constantly repeated models, as in the case of the so-called Reichenau and Echternach schools.
4. Colours in the Codex Aureus of Echternach

The following reflections are based on the technological examination of the Codex Aureus Epternacensis, now in the Germanisches Nationalmuseum in Nuremberg, which was carried out by Doris Oltrogge and Robert Fuchs in 2006, as well as my own research on the use and the aesthetics of colours in manuscript illumination.6 The Codex Aureus, made around 1045 in the Benedictine monastery of Echternach, where it remained until the French Revolution, is one of the rare medieval Gospel books entirely written in gold ink (Oltrogge and Fuchs 2009; Grebe 2007; Kahsnitz [ed.] 1982).7 It is also one of the largest and most richly decorated manuscripts, with more than 50 large-format illuminated pages. These include a sumptuous doublepage depicting the Majestas Domini, while each Gospel is preceded by a sequence of equally lavish double-pages with the portraits of the Evangelist (fig. 1), full-page decorative initials, other decorative pages as well as four sequences of scenes from the life of Jesus. The texts of the Gospels themselves are not interrupted by scenic illustrations like in other Ottonian manuscripts. The codex is the work of at least eight illuminators, some of which executed only executed a single page. They must have been working together very closely and within a short period of time. In several cases, recto and verso of one and the same sheet of parchment are executed by different illuminators who possessed a distinctive style yet used nearly the same palette of colours. The technological examination of the manuscript has revealed a very rich and varied palette, in which nearly all the colours and hues available to 11th century scriptoria occur (Oltrogge and Fuchs 2009, 153162, 163-167). The colours red, blue or purple, for instance, are represented not by a single, but by two or more pigments. Among these we find many rare and expensive exotic pigments like lapis lazuli or kermes, which were actually used quite abundantly in this manuscript as compared to other Ottonian manuscripts. This observation leads to the hypothesis that the Echternach illuminators and their patrons explicitly wished to use as many and as costly colours as possible to provide
5. Translated from Halleux 1990, esp. p. 179: Cest peu peu, partie par partie, que sapprennent tous les arts. Lart des peintres est prcd par la confection des couleurs. Ensuite, que votre esprit tourne ses soins vers les mlanges. Alors, passez loeuvre, mais soumettez tout la rigeur, pour que ce que vous peindrez soit orn, et comme naturel. Ensuite, par bien des preuves dingniosit, lart augmentera loeuvre, comme le montrera ce livre. (Sensim per partes discuntur quaelibet artes/Artis pictorum prior est factura colorum/Post, ad mixturas convertat mens tua curas/Tunc opus exerce, sed ad unguem cuncta coerce/ut sit ad ornatum quod pinxeris, et quasi natum/Postea multorum documentis ingeniorum/Ars opus augebit, sicut liber iste docebit.) Theophilus, who treats colours always as pigments, shortly mentions the material importance of colours in the preface of his treatise, cf. Brepohl 1999. 6. Up to now, there are very few publications on the technical examination of manuscripts from the 10th and 11th centuries. Roosen-Runge 1967 focuses on English illuminated manuscripts from the 11th and 12th centuries. On Ottonian book illumination in Trier see Oltrogge and Fuchs 2005. On the Echternach Codex Caesareus in Uppsala see Fuchs and Oltrogge 1998. 7. Nuremberg, Germanisches Nationalmuseum, Hs 156142.
24
fig.1 codex aureus of echternach, st. john the evangelist, f. 112v. nuremberg, germanisches nationalmuseum
25
colour and meaning
fig.2a+b codex aureus of echternach, double-page with scenes of the passion and pentecost, f. 111v-112r. nuremberg, germanisches nationalmuseum
the maximum variety of colours for the Golden Gospels. Lapis lazuli, for instance, can be found as a distinguishing colour for the garments of the protagonists, but was used also for backgrounds, e.g. in the Pentecost scene (fig. 2B), where we would
26
expect a less expensive pigment like azurite. In the eyes of the contemporaries, the prodigal use of the most expensive pigment must have enhanced the exceptional character of the manuscript (Pastoureau 2001, 32-40).
27
colour and meaning
The most dominant colour in the codex is purple (fig. 3), a synonym of sumptuousness and costliness and symbol both of God and the Emperor, which occurs in different tones and shades both in the miniatures and the decorative pages (Muthesius 1997, 27-33; Steigerwald 1990; Fuchs and Oltrogge 2007, 82-98). The extensive use of purple enhances the religious significance of the book as an embodiment of Gods words. The illuminators did not apply the real Tyrian purple extracted from the murex shell (murex brandaris), but used less expensive pigments made from plants like different kinds of the Rocella (or auricella) or the Ochrolechia as well as Kermes vermilio, which were perhaps also better suited for the preparation of bodycolours (Oltrogge and Fuchs 2009, 155-159). Technological examination has revealed
fig.3 codex aureus of echternach, initial page of the gospels of st. matthew, f. 22r. nuremberg, germanisches nationalmuseum
fig.4 codex aureus of echternach, scenes of the infancy cycle, f. 19r. nuremberg, germanisches nationalmuseum
28
a preference for the darker Rocella purple, which was easier to procure, for the less prominent sections of the miniatures, whereas clothing is more often painted in the more costly Kermes purple, which had to be imported from the Mediterranean (Oltrogge and Fuchs 2009, 34-50). Thus, in the case of purple, the illuminators probably wished to enhance the more important parts of a figure or scene by using a more expensive pigment, again displaying an awareness of the material quality of colours (Oltrogge and Fuchs 2009, 34-50). Another level of this chromatic consciousness is revealed by the fact that many full-page miniatures, amongst others the double page depicting the Majestas Domini (f. 2v-3r , fig. 6A), the Evangelists as well as most of the miniatures in the cycle of
fig.5 codex aureus of echternach, third page of the infancy cycle with alterations by the master of the long nosed figures, f. 19v nuremberg, germanisches nationalmuseum
29
colour and meaning
fig.6a+b codex aureus of echternach, majestas domini double-page, f. 2v-3r nuremberg, germanisches nationalmuseum
the life of Jesus (fig. 4-5), were partially painted over at some point. This must have taken place in the workshop itself either during the making of the manuscript or immediately after the illuminations had been finished, but before the final binding
30
of the pages. The retouches were carried out by the same group of illuminators who, probably with one exception, had been responsible for the original or first phase of the work. Curiously enough, these alterations, which were done using the same col-
31
colour and meaning
ours and pigments as before, only concerned the backgrounds, the tone of the flesh and the hair of the figures, leaving the form of the heads and hands as well as the rest of the composition untouched (Oltrogge et Fuchs 2009, 34-35, ill. 57-72). In the case of the Majestas Domini double-page, the alterations can partly be detected with the naked eye (fig. 7-8). The figure of Christ enthroned, the angel symbolising St Matthew or the personification of Prudentia (fig. 9-10) each possess four eyes: one pair of eyes belongs to the actual face, while below this which appear two dark points that belonged to the original features. Investigation with infra-red light has revealed that the second illuminator completely covered the fleshy parts, but did not touch the original version which is entirely conserved.
fig.7 codex aureus of echternach, majestas domini, detail of the head of christ, f. 2v. nuremberg, germanisches nationalmuseum fig.8 codex aureus of echternach, majestas domini, detail of the head of christ (infrared reflectography), f. 2v. nuremberg, germanisches nationalmuseum
32
The alterations seem to be highly unusual and demand an interpretation. A practical explanation could be that we are faced with an early repair of defective portions or layers of paint which were corrupted or chipped. This, however, can be rejected in the view of the remaining, fully conserved original faces. A second hypothesis concerns the content of the miniatures. It states that the makers of the manuscript may have wished to change the character of a figure or a scene. There is, however, no substantial evidence for this explanation either. While the illuminators would frequently alter the tone of the flesh and some of the facial features, they would generally not change the posture of the head or the line of sight. One of the few exceptions is the Noli me tangere scene, where the painter corrected the orientation of the angels heads, which had originally looked straight ahead (fig. 11-12). By re-directing them to the right the illuminator enhanced the focus on Christ as the central figure of the scene, but did not completely change the meaning of the image. Similar changes do also occur with some of the minor figures which are part of the decorative framework and do not carry any particular meaning. This observation suggests another explanation. The alterations were probably carried out for aesthetical reasons and can tentatively be interpreted as aesthetical restorations. Interestingly enough, the illuminators did
fig.10 codex aureus of echternach, personification of prudentia (infrared reflectography), f. 3r. robert fuchs/ doris oltrogge, university of applied sciences, cologne
fig.9 codex aureus of echternach, personification of prudentia, detail from the majestas domini double-page, f. 3r. nuremberg, germanisches nationalmuseum
33
colour and meaning
not paint over entire figures, which would probably have been easier, but instead only retouched the fleshy parts, making sure that the renewed face fitted between the hairline and the neckline. In the case of the Majestas Domini and related pages, the rather long-nosed face-lifting did not really result in an improvement in appearance. However, as technological analysis has revealed, the more harmonious general features of the first version had been combined with a rather garish pink skin colour and blue hair as seen in the portrait of St John. Some minor figures preserve this colouring, and these give us an idea of the original facial tones: the ox symbolising St Luke (f. 2v), the grotesque masks in the frame of the decorative initial page of the Gospels of St John (f. 114r) and St Peter in the scene of the Doubting Thomas (f. 111v). Some of the trumpeting angels as well as St John the Evangelist (f. 112V) even retained their original blue hair (fig. 1). These examples suggest that the true reason for the alterations was probably to avoid violent colour contrasts within the images, and to create a chromatic harmony between several of the double-pages which had been executed by different illuminators. The new, brownish skin colour and black hair, for instance, do indeed match better with the colours of the clothing and backgrounds. Thus, the Codex Aureus underlines the high level of sensitivity of the Romanesque period towards both chromatic values and material virtues of colours and pigments.
fig.11 codex aureus of echternach, passion cycle, noli me tangere scene, detail: angel, f. 111v. nuremberg, germanisches nationalmuseum
fig.12 codex aureus of echternach, passion cycle, noli me tangere scene, angel (infrared reflectography), f. 111v. nuremberg, germanisches nationalmuseum
34
The astonishing insights into the production of the codex reveal a twofold aesthetic of colours, which had both the materiality of the colours or pigments and their chromatic interaction in view. Instead of focusing only on the symbolism of certain colours, which was the general supposition of research until recently, the makers of the manuscript actually gave priority to chromatic variation. Their aim seems to have been to enhance the status of the codex as an actual materialisation of the word of God by endowing it with the finest of colours both in the material and the artistic sense. Hopefully, the new interpretation proposed here of the employment of colours and the importance attached to chromatic values in medieval and especially in Romanesque book illumination may be supported by further technical and art historical research into other manuscripts in the nearby future.
Bibliography
Alberti, Leon Battista. 2000. De Statua. De Pictura. Elementa Picturae/Das Standbild. Die Malkunst. Grundlagen der Malerei. Btschmann, Oskar et al. (ed., transl.). Darmstadt: Wissenschaftliche Buchgesellschaft. Anzelewsky, Fedja. 1991. Albrecht Drer. Das malerische Werk. 2nd revised ed., 2 vol. Berlin: Deutscher Verein fr Kunstwissenschaft Bartl, Anna et. al. 2005. Der Liber illuministarum aus Kloster Tegernsee. Edition, bersetzung und Kommentar der kunsttechnologischen Rezepte. Stuttgart: Steiner. Bender, Beate. 1990. Color caelestis. Anmerkungen zur Farbe Blau im Mittelalter. Gerke, Hans (ed.). Blau. Farbe der Ferne. Heidelberg: Wunderhorn: 82-103. Brepohl, Erhard. 1999. Theophilus Presbyter und das mittelalterliche Kunsthandwerk. Gesamtausgabe der Schrift De diversis artibus. 2 vol. Cologne et al.: Bhlau. Decker, Bernhard. 1996. Drer und Grnewald. Der Frankfurter Heller-Altar. Rahmenbedingungen der Altarmalerei. Frankfurt/Main: Fischer Taschenbuch Verlag. Dittmann, Lorenz. 1987. Farbgestaltung und Farbtheorie in der abendlndischen Malerei. Eine Einfhrung. Darmstadt: Wissenschaftliche Buchgesellschaft. Fara, Giovanni Maria. 2002. Albrecht Drer lettore e interprete di Vitruvio e Leon Battista Alberti in uninedita versione di Cosimo Bartoli. Rinascimento, 42, 2002: 171-347. Fuchs, Robert and Oltrogge, Doris. 1990. Das Blau in der mittelalterlichen Buchmalerei. Quellenschriften als Basis naturwissenschaftlicher Farbuntersuchungen. Gerke, Hans (ed.). Blau. Farbe der Ferne. Heidelberg: Wunderhorn: 104-130. Fuchs, Robert and Oltrogge, Doris. 1998. Secreta colorum. Geheimnisse aus mittelalterlichen Skriptorien. Untersuchungen zur Echternacher Handschriftenproduktion. Tempus edax rerum. Le bicentenaire de la Bibliothque nationale de Luxembourg 17981998. Luxemburg: Bibliothque Nationale: 107-124.
35
colour and meaning
Fuchs, Robert and Oltrogge, Doris. 2007. Gold und Purpur zwischen Ideal und Werkstattpraxis. Segni e testo, 5, 2007: 31-98. Gaignebet, Claude. 1990/1991: Le sang-dragon du Jardin des Dlices. Ethnologie franaise, 20/21, 1990/1991: 378-390. Grebe, Anja. 2006. Albrecht Drer. Knstler, Werk und Zeit. Darmstadt: Wissenschaftliche Buchgesellschaft. Grebe, Anja. 2007. Codex Aureus. Das Goldene Evangelienbuch von Echternach. Darmstadt: Wissenschaftliche Buchgesellschaft. Gullick, Michael. 1995. A bibliography of medieval painting treatises. Brownrigg, Linda L. (ed.). Making of the medieval book. Techniques of production. Los Altos Hills, CA et al.: Anderson-Lovelace et al.: 241-244. Halleux, Robert. 1990. Pigments et colorants dans la Mappae Clavicula. Guineau, Bernard (ed.). Pigments et colorants de lAntiquit et du Moyen Age. Teinture, peinture, enluminure, tudes historiques et physico-chimiques. Paris: Editions du Centre National de la Recherche Scientifique: 173-180. Kahsnitz, Rainer (ed.). 1982. Das Goldene Evangelienbuch von Echternach. Codex Aureus Epternaciensis Hs 156142 aus dem Germanischen Nationalmuseum in Nrnberg. Faksimile and Commentary. Frankfurt/Main et al.: Fischer. Kutschbach, Doris. 1995. Albrecht Drer. Die Altre. Stuttgart et al.: Belser. Muthesius, Anna. 1997. Byzantine silk weaving AD 400 to AD 1200. Vienna: Fassbaender. Oltrogge, Doris. 2006. Rezeptsammlungen und Traktate. Die Vermittlung kunsttechnischen Wissens im Frh- und Hochmittelalter. Canossa 1077. Erschtterung der Welt. Geschichte, Kunst und Kultur am Aufgang der Romanik. Exhib. Cat. Paderborn. vol 1. Munich: Hirmer: 555-562. Oltrogge, Doris and Fuchs, Robert. 2005. Mit Infrarot und Rntgenstrahl. Ergebnisse naturwissenschaftlicher und kunsttechnologischer Untersuchungen des Egbert-Codex. Franz, Gunther (ed.). Der Egbert Codex. Ein Hhepunkt der Buchmalerei vor 1000 Jahren. Luzern/Darmstadt: Wissenschaftliche Buchgesellschaft: 189-216. Oltrogge, Doris and Fuchs, Robert. 2009. Die Maltechnik des Codex Aureus aus Echternach. Ein Meisterwerk im Wandel. Nuremberg: Verlag des Germanischen Nationalmuseums. Pastoureau, Michel. 1990. La couleur et lhistorien. Guineau, Bernard Guineau (ed.): Pigments et colorants de lAntiquit et du Moyen Age. Teinture, peinture, enluminure, tudes historiques et physico-chimiques. Paris: Editions du Centre National de la Recherche Scientifique: 21-40. Pastoureau, Michel. 1989. Vers une histoire sociale des couleurs. Pastoureau, Michel (ed.). Couleurs, images, symboles. Etudes dhistoire et danthropologie. Paris: Editions Le Lopard dOr: 9-68. Pastoureau, Michel. 2001. Blue. The history of a color. Princeton et al.: Princeton University Press.
36
Pfaff, Annette. 1971. Studien zu Albrecht Drers Heller-Altar (Nrnberger Werkstcke zur Stadt- und Landesgeschichte, vol. 7). Nuremberg: Stadtarchiv. Roosen-Runge, Heinz. 1967. Farbgebung und Technik frhmittelalterlicher Buchmalerei. Studien zu den Traktaten Mappae Clavicula und Heraclius. 2 vols. Munich/Berlin: Deutscher Kunstverlag. Rupprich, Hans. 1956. Drer. Schriftlicher Nachlass. Vol. 1. Berlin: Deutscher Verein fr Kunstwissenschaft. Schmidt, Peter. 1965. Goethes Farbensymbolik. Untersuchungen zu Verwendung und Bedeutung der Farben in den Dichtungen und Schriften Goethes. Berlin: Schmidt. Steigerwald, Gerhard. 1990. Das kaiserliche Purpurprivileg in sptrmischer und frhbyzantinischer Zeit. Jahrbuch fr Antike und Christentum, 33, 1990: 209-239. Tikkanen, Johan Jakob. 1933. Studien ber die Farbengebung in der mittelalterlichen Buchmalerei. Helsingfors: Centraltryckeriet et al. Wagner, Monika. 2001. Das Material der Kunst. Eine andere Geschichte der Moderne. Mnchen: Beck. Welchman, John. 1990. Aesthetics. Hope, Augustine Hope (ed.): The color compendium. New York: Van Nostrand Reinhold.
Biography
Anja Grebe studied Art History, History, and French Literature at the University of Constance (Germany) and Paris, La Sorbonne. 1995 M.A. Dissertation on Modern French Artists Books. 1996-1997 Research assistant at the University of Constance, Department of Art History. 1997-2000 Graduate school Medieval Writing Culture at the University of Mnster; PhD Dissertation on Art on the Edge: Book Design in the Burgundian Netherlands after 1470. Since 2001 Research assistant at the Germanisches Nationalmuseum, Nuremberg and Associate lecturer at the University of Erlangen, Department of Art History. Institutional affilation: Assistant professor of Medieval Art History at the Otto-FriedrichUniversity, Bamberg.
37
tude des rapports entre la couleur et le corps dans limage medievale: levangeliaire dotton iii, fin x e sicle
Resumo
O perodo que se estende entre o fim do sculo X e os finais do sculo XII uma poca de mutaes. por isso, que ele me parece interessante para estudar a emergncia de um cdigo da cor do corpo. A minha investigao actual articula-se em torno dos laos que existem entre a cor e os corpos na imagem medieval. Dou uma ateno muito especial sobre a cor do nu e a sua relao com a cor do vesturio que o cobre: ser que existe um lao entre estes dois espaos da cor, o da pele e o do vesturio? A cor do corpo, nu ou vestido, desempenha um papel no estatuto das personagens representadas? O pequeno nmero de manuscritos a cor que chegou at ns, o seu estado de conservao por vezes mau e a dificuldade de aceder s fontes, no parece , primeira vista, justificar a minha opo cronolgica. Contudo, estes documentos parecemme reveladores duma poca em plena mutao e so portadores de ndices dessas mesmas mutaes. As minhas pesquisas permitiram-me constatar que no haver ainda um sistema cromtico definido, mas tendncias. O que verdadeiro para um manuscrito no o para os outros. O estado transitrio do perodo que aparece por vezes em filigrana num manuscrito e a escassez de fontes directas ou indirectas sobre a cor das personagens aconselham uma grande prudncia nas minhas concluses. Na minha investigao sobre as relaes que existem entre a cor e o corpo representado, interrogo-me sobre a forma como se exprime iconograficamente a cor do corpo. Focalizo a minha ateno na presena ou no de tcnicas cromticas para traduzir a cor da nudez, da pele, dos cabelos, mas tambm na forma de representar o vesturio. Para poder responder, finalmente, questo de saber se a cor em geral, a cor do corpo em particular, desempenha uma funo no estatuto da personagem representada nas imagens. Com efeito, notei em certos documentos, um esforo para diferenciar, pela cor, a pele positiva da pele negativa e, para l da pele, o estatuto da personagem apresentada. Este esforo de classificao pode ir em certos casos, at dar ao corpo mais valorizado, o de Cristo, uma cor particular. De modo a responder s questes que me coloco, pareceu-me prudente centrar a ateno sobre o estudo dum manuscrito otoniano datado de finais do sculo X, o Evangelirio de Oto III. O manuscrito, produzido em torno de 997, um dos mais belos exemplos do atelier da abadia alem de Reichenau. O conjunto iconogrfico composto de 35 pinturas a pgina inteira, o que faz dele um dos mais importantes da poca. A riqueza de tons e a densidade da cor, o excelente estado de conservao das pinturas, a relao estreita entre a cor do corpo e o estatuto das personagens representadas, fazem desta obra um elemento chave no meu estudo sobre as cores do corpo. A gama cromtica do manuscrito caracterizada pela variedade das tintas, o jogo de matizes e o perfeito estado de conservao e saturao das cores. Ao olhar para este rico mostrurio de cores (nuancier), trs elementos atraram a minha ateno: a delicadeza do trabalho do artista para criar uma vasta gama de brancos, a presena de um matiz de vermelhos e castanho na cor do vesturio principal. E, enfim, a riqueza dos tons para a cor da pele. O Evangelirio de Oto III no cessa de levantar um grande nmero de questes. A cor aparece a codificada e pensada em funo dum contexto bem particular. Mas este estudo obriga a uma grande prudncia, pois este documento permanece uma excepo. Ele obriga-me a modelar pouco a pouco e sensatamente a minha anlise, pois o que vlido para este manuscrito no o para a maior parte dos integram o meu estudo.
palavras-chave
cor corpo imagem manuscrito ano 1000
39
Abstract
The period stretching from the end of the 10th century to the end of the 12th century is a time of many changes. This is why it seems interesting to us to study the emergence of a code of colour for the body. My current research is based on the existing links between the colour and the body in medieval imagery. I pay particular attention to the colour of the naked body and its relationship with the colour of the clothing that it covers: Is there a link between these two spaces of colour, which are the skin and clothing? Does the colour of the body, naked or dressed, play a role in the status of the various represented characters? The small number of manuscripts with colours remaining to this day, their state of conservation sometimes bad and the difficulty to access the sources does not seem, initially, to justify our chronological choice. However, these documents appear to us as an insight into a time in complete transformation and they are, at the same time, carrying the indications of these changes. My research identified that there was not a defined chromatic system yet, but some tendencies. What is true of one manuscript is not of others. The transitory state of the period that appears sometimes pen flourished initials in the manuscripts and the scarcity of the direct or indirect sources on the colour of the characters, forces the greatest of caution regarding our conclusions. In my search for existing links between the colour and the depicted body, I question the way in which the colour of the body is represented from an iconographic point of view. I concentrate my attention on the presence or not of chromatic techniques to translate the colour of nudity, of the skin, of the hair, but also to represent clothing. Ultimately, the question is to know if the colour in general, the colour of the body in particular, plays a role in the status of the represented characters in my pictures. Indeed, I have noticed that on certain documents an effort to differentiate, by using the colour, the positive skin from the negative skin and, beyond the skin, the status of the character represented in the scene. This effort of classification can, in certain cases, lead to giving to the most valued body, that of Christ, a particular colour. In order to answer the questions that we ask ourselves, it seemed judicious to focus our attention on the study of an Ottonian manuscript dating from the end of X th century, the Evangeliar of Otto III.1 The manuscript, made around 997, is one of the finest examples of the workshop of the German abbey of Reichenau. Its iconography consists of thirty-five full-page paintings, making it one of the most important iconographic programs of the time. The richness in the tinge and density of the colour, the excellent state of conservation of paintings, the close relationship between the colour of the body and the status of the characters represented makes this work a key element in our study on the colours of the body. The colour chart of the manuscript is characterized by the variety of colours, a set of shades and the perfect state of preservation and saturation of colours. In this rich colour chart, three points have caught my attention: the delicate work realized by the artist to create a wide range of white, the presence of shades of red and brown in the colour of the main clothing and, finally, the richness of nuances in the coloring of the skin. The Evangeliar of Otto III continues to raise many questions. The colour appears codified and thought in terms of a particular context. But this study requires the utmost caution, this document remains an exception. It forces me to adjust gradually and carefully my analysis, because what is valid for this manuscript is no longer valid for the great majority of theconfigurao do of my study. de Ana Catarina Sousa, Pedro Fialho Agradecimentos por ajuda na manuscripts texto e sugestes
key-words
colour body image manuscript year 1000
1. Munich, Bayerische Staatsbibliothek Clm. 4453, towards 997.
tude des rapports entre la couleur et le corps dans limage mdivale: lvangliaire dotton iii, fin xe sicle
m a r i e a s c h e h o u g - c l au t e au x cole Pratique des Hautes tudes, Paris
(France)
1. Munich, Bayerische Staatsbibliothek, Clm 4453. 2. Otton III meurt lge de 19 ans au nord de Rome. Ne laissant pas dhritier mle, la succession de lempire devient incertaine. Henri II, ayant entretenu par le pass des bonnes relations avec son cousin, sempresse prendre la succession, avant mme quune lection aie lieu. 3. Aprs la mort de son pre, en 985, Henri II devient alors duc de Bavire. Il pouse Cungonde de Luxembourg (978-1040) vers lAn Mil. Cest sous son rgne, en 1007, que la ville de Bamberg devient un vch.
La priode allant de la fin du Xe sicle la fin du XIIe sicle est une poque de mutations. Cest pourquoi, elle me semble intressante pour tudier lmergence dun code de la couleur du corps. Mes recherches actuelles sorientent autour du rapport qui existe entre la couleur et le corps dans limage mdivale. la couleur joue-t-elle un rle dans les diffrents statuts du corps reprsent? Afin de rpondre cette question, il ma sembl judicieux de focaliser lattention sur ltude dun manuscrit ottonien datant de la toute fin du X e sicle, lvangliaire dOtton III.1 Pour cette tude, je nai malheureusement pas eu accs au manuscrit original conserv Munich, ce dernier tant trs difficile daccs. Pourtant, consciente des difficults danalyse que cela engendre, la consultation du fac-simil ma permis de soulever un certain nombre de questions. De ce fait, la richesse des nuances et de densit de la couleur, lexcellent tat de conservation des peintures, le rapport troit entre la couleur du corps et le statut des personnages reprsents font de ce manuscrit un lment cl dans mon tude sur la couleur du corps. Lvangliaire dOtton III, ralis vers 997, est lun des plus beaux exemples de latelier de labbaye allemande de Reichenau. Il fait lobjet dune commande du jeune empereur Otton III (983-1002). Aprs la mort subite de ce dernier,2 le bel vangliaire revient son cousin loign et successeur Henri II (973-1024) qui, son tour, loffre au trsor de la cathdrale de Bamberg. 3 Lensemble iconographique est compos de trente-cinq peintures pleine page, parmi lesquelles vingt-neuf miniatures mettent en scne la vie du Christ et quatre miniatures sont consacres la figuration des vanglistes et de leurs symboles. Le programme pictural et christique de notre
41
colour and meaning
vangliaire est lun des plus importants de lpoque. Aprs les douze Canons de Concordance, le cycle dimages souvre sur deux peintures se faisant face, aux folios 23 v. et 24: les quatre provinces impriales, Sclavinia, Germania, Gallia et Roma, prtent hommage au jeune empereur Otton III. Ce dernier est reprsent en majest, entour de nobles, clercs et lacs. (Fig.1)
fig.1 folios 23 v. et 24 (briechle andrea, etc., das evangeliar kaiser ottos iii, bayerische staatsbibliothek, mnchen, clm 4453. augsburg, haus der bayerischen geschichte, 2002)
La gamme chromatique du manuscrit est caractrise par la varit des teintes, le jeu des camaeux et le parfait tat de prservation et de saturation des couleurs. Jai relev pas moins de cinquante-huit nuances diffrentes. Parmi ces nuances, lor a une place centrale: il est la couleur principale des fonds des peintures. Toutefois, il est intressant remarquer que la mise en scne de la majest impriale et les quatre dernires images du cycle christique, la Passion et la Rsurrection du Christ, ne sont pas figures sur un fond dor mais sur un fond polychrome. Cest comme si ces fonds de couleur rappelaient la nature terrestre du pouvoir imprial et la nature humaine du Christ. Au regard de ce riche nuancier, trois lments ont attir mon attention. Tout dabord, le dlicat travail que ralise lartiste pour crer une vaste gamme de blancs. Ensuite, la prsence dun camaeu de rouges et de marron dans la couleur des vtements principaux est intressante signaler. Enfin, la richesse des nuances dans la mise en couleur de la peau est un lment qui pose un grand nombre de questions. Le minutieux travail artistique pour reprsenter une riche gamme de blancs est un point intressant. La couleur blanche, lorsquelle est applique au vtement des personnages, nest pas une couleur pure, totalement blanche. La peinture blanche pure est seulement rserve aux rehauts de couleurs de la peau, des yeux et des plis des toffes. Le blanc vestimentaire est toujours un blanc bleut, ou bien un blanc
42
gristre, parfois mme violac. Ce blanc est la couleur la plus claire, gnralement rserve au vtement port le plus prs du corps: la tunique, la tunique du dessous ou la chemise, certains bas-de-chausses et les voiles des femmes. Par exemple, au folio 24, la tunique du vieux lac gauche dOtton III est dun blanc bleut, tandis que le linceul de Lazare, au folio 231 v., est dun blanc violac. Jai remarqu quil tait parfois difficile de diffrencier le blanc du gris. Le gris permettant denrichir la gamme des blancs, lide serait davoir une couleur claire pour le vtement le plus prs du corps. (Fig.2)
fig.2 folios 24, 60 v., 94 v., 97 v., 113 et 231 v. (dtails) (briechle andrea, etc., das evangeliar kaiser ottos iii, bayerische staatsbibliothek, mnchen, clm 4453. augsburg, haus der bayerischen geschichte, 2002)
fig.3 folio 116 v. (dtail) (briechle andrea, etc., das evangeliar kaiser ottos iii, bayerische staatsbibliothek, mnchen, clm 4453. augsburg, haus der bayerischen geschichte, 2002)
Lorsquun groupe est reprsent, aux tuniques blanches viennent sajouter des tuniques jaunes. Par exemple, au folio 116 v., les jeunes enfants qui entourent le Christ portent des tuniques blanc violac et jaune clair de manire intercale. En rythmant limage, ce jeu de couleurs semble donner vie au groupe et rendrait la scne moins monotone. (Fig.3) Mais lusage du jaune comme couleur associe au blanc peut aussi crer une tension au sein dun couple de personnages. Dans la scne du Baiser de Judas, au folio 244 v., le Christ et Judas portent tous les deux des tuniques claires. Pourtant, celle du Christ est blanc gristre tandis que celle de Judas est jaune clair. Leurs manteaux sont rouges, mais celui du Christ est rouge violac, plus fonc que celui de Judas, qui est rouge orang. Le contraste de couleurs apparat la fois faible et extrmement subtil. Par la couleur, lartiste na-t-il pas voulu traduire lambigut de ce baiser et, au-del du baiser, la complexit du geste de Judas? (Fig.4) Cette figuration dhabits qui ne sont pas rellement blancs nest-elle pas le reflet dune ralit vestimentaire, cest--dire limpossibilit de teindre les toffes en blanc?
43
colour and meaning
Le fait de reprsenter des tuniques aux diverses nuances de blancs et de jaune nest-il pas une manire dexprimer la couleur changeante des tissus non-teints? Les tissus blancs sont gnralement les tissus de couleur naturelle, ou qui ont subi des techniques dclaircissement comme le schage au soleil. Cest pourquoi, lpoque qui nous concerne, il est presque matriellement impossible davoir des toffes rellement blanches. Gnralement, celles-ci sont de couleur crue, ou bien gristre ou bruntre. Prcdemment, nous avions mentionn la prsence dun camaeu de rouges dans la palette vestimentaire des personnages. Le rouge est enrichi de diverses nuances, tirant parfois mme vers le marron. Dans certaines images, il est mme difficile de distinguer le rouge et le marron. Cela mamne penser que le marron, dans ce manuscrit-ci, fait partie de la palette des rouges. La diversit des nuances rouges et leur rcurrence sur le vtement semble imposer lide dune couleur fonce pour lhabit des personnages principaux. Au folio 24, lempereur Otton III est vtu dune luxueuse tunique rouge violac et dun manteau vert. La longue tunique du dessous est blanc bleut. On peut remarquer que la couleur fonce du vtement principal contraste avec la couleur claire du vtement le plus prs du corps. (Fig.5) De son ct, au folio 34 v., le Christ revt un manteau couleur rouge ou marron. Comme pour le vtement imprial, la couleur du manteau du Christ contraste toujours avec la couleur claire de sa tunique. une exception prs. (Fig.6) La seule scne o le Christ ne porte pas de couleur fonce, rouge ou marron, est celle de la Transfiguration, au folio 113. Il y est vtu dune tunique blanche et dun
fig.5 folio 24 (dtail) (briechle andrea, etc., das evangeliar kaiser ottos iii, bayerische staatsbibliothek, mnchen, clm 4453. augsburg, haus der bayerischen geschichte, 2002)
fig.6 folios 34 v., 97 v., 192, 231 v. et 251 (dtails) (briechle andrea, etc., das evangeliar kaiser ottos iii, bayerische staatsbibliothek, mnchen, clm 4453. augsburg, haus der bayerischen geschichte, 2002) > fotofig.6.jpg
44
4. Depuis lAntiquit, le rouge est la couleur par excellence. lpoque romaine, la plupart des pigments qui servent teindre en rouge pntrent mieux les fibres textiles. Les diffrentes gammes de rouges sont plus rsistantes que les autres couleurs. Les belles toffes sont donc gnralement de couleur rouge, dont les teintes sont extrmement diversifies. (Lire ce sujet le chapitre consacr aux couleurs et aux teintures dans louvrage de PASTOUREAU Michel, Couleurs, Images, Symboles, Paris, Ed. Le Lopard dOr, 1989, pp. 20-31).
manteau orang. La couleur claire de lhabit du Christ se ferait lcho de ce vtement dune blancheur clatante dont parlent les vangiles. (Fig.7) La richesse des nuances de rouge et de marron sur la tunique dOtton III et sur le manteau du Christ confirmerait le statut encore exceptionnel qua le rouge dans le monde occidental aux alentours de lAn Mil.4 Le travail ralis par lartiste sur la couleur de la peau mrite notre attention tant il est extrmement subtil. Une premire couche de couleur sature est pose sur lespace destin la peau. Pour la couleur de cette couche de base, jai relev pas moins de onze nuances diffrentes: beige moyen rostre, beige verdtre, beige jauntre, jaune verdtre, beige clair, beige rougetre, blanc verdtre, marron clair, verdtre, beige clair rostre et gris verdtre composent le nuancier de la peau. Ensuite cette couche de base est releve de reflets blancs et dombres de couleurs diverses. Jai remarqu que la couleur des ombres variait en fonction de la couleur de base de la peau. Par exemple, au folio 24, la peau des lacs figurs auprs de lempereur est beige, releve dombres marron rougetre, alors que la peau dOtton III, jaune verdtre, est rehausse dombres verdtres. (Fig.8)
fig.8 folios 15 v., 17 v., 18, 23 v., 30 v., 32 v., 60 v., 97 v., 113 et 149 v. (dtails) (briechle andrea, etc., das evangeliar kaiser ottos iii, bayerische staatsbibliothek, mnchen, clm 4453. augsburg, haus der bayerischen geschichte, 2002)
fig.9 folio 17 v. (dtails) (briechle andrea, etc., das evangeliar kaiser ottos iii, bayerische staatsbibliothek, mnchen, clm 4453. augsburg, haus der bayerischen geschichte, 2002)
Pourquoi retrouve-t-on tant de couleurs diffrentes pour la peau? Afin de rpondre cette question, je ferai appel sept cas de figure qui me paraissent significatifs: la peau dun vieillard indigent, celle de Satan, celle du lpreux, celle de Lazare, celle des quatre provinces impriales, celle de lempereur Otton III et, enfin, la peau du Christ. Au dbut de notre cycle dimages, au-dessus de lune des Tables de Concordances, au folio 17 v., nous remarquons un vieillard courb, tunique courte et pieds nus, en train de se chauffer les mains auprs dun feu. Sa peau est beige verdtre. Face lui, un jeune vigneron est aussi vtu de court et les pieds nus. Mais sa peau est marron clair. Ici, la couleur de la peau semble tablir une frontire entre deux ges, la jeunesse et la vieillesse. (Fig.9) Pourtant, un feuillet plus loin, au folio 18, un jeune homme et un vieillard ont t galement figurs au-dessus des Tables de Concordances. Ce sont des artisans. Leurs
45
colour and meaning
tuniques sont courtes et leurs jambes et leurs pieds sont recouverts. Et, surtout, ils ont tous les deux la mme couleur de peau, beige jauntre. La couleur de la peau ne cre plus seulement une distinction entre deux ges, mais elle semble tablir un cart social entre lunivers des paysans et le monde des artisans. Le vieillard qui se rchauffe les mains auprs du feu, vtu dune courte tunique, jambes et pieds nus, pourrait tre un indigent. La couleur de sa peau, beige verdtre, contraste avec celle des trois autres travailleurs reprsents: lhomme est non seulement un paysan, mais il est surtout un exclu. (Fig.10)
5. Sur laspect pjoratif dune couleur sature pour le corps, lire PASTOUREAU Michel, Couleurs, Images, Symbolique, Paris, Ed. Le Lopard dOr, Paris, 1989, pp. 95-97.
fig.10 folio 18 (dtails) (briechle andrea, etc., das evangeliar kaiser ottos iii, bayerische staatsbibliothek, mnchen, clm 4453. augsburg, haus der bayerischen geschichte, 2002)
Dans les trois scnes des Tentations du Christ, au folio 32 v., Satan a la peau beige rougetre, beaucoup plus sombre que celle du Christ, qui est beige clair. Son corps est recouvert dune toffe marron clair. La manire dont est pose ltoffe sur le corps du diable est particulire: elle cache une partie tout en dcouvrant lautre partie. Lambivalence vestimentaire, double dune couleur sombre de peau, soulignerait le caractre sournois et inquitant de Satan.5 (Fig.11)
fig.11 folio 32 v. (dtails) (briechle andrea, etc., das evangeliar kaiser ottos iii, bayerische staatsbibliothek, mnchen, clm 4453. augsburg, haus der bayerischen geschichte, 2002)
46
6. Sur le tachet de la peau, lire PASTOUREAU Michel, Couleurs, Images, Symbolique, Paris, Ed. Le Lopard dOr, Paris, 1989, pp. 92-95. 7. La lpre serait la plus visible des maladies de la peau. Elle reprsente ainsi larchtype de la maladie de peau. (Lire ce sujet TOUATI FranoisOlivier, Archives de la lpre: Atlas des lproseries entre Loire et Maine au Moyen Age, Paris, 1996). 8. Le lpreux ouvre la voix du salut par la charit quil appelle. (Lire TOUATI Franois-Olivier, Maladie et socit au Moyen Age. La lpre, les lpreux et les lproseries dans la province ecclsiastique de Sens jusquau milieu du XIVe sicle, Bruxelles, De Boeck Universit, 1998). 9. Lire, ce propos, larticle dALIBERT Dominique, Approche de liconographie politique autour de lAn mille, in Gerberto dAurillac, da Abate di Bobbio a Papa dellAnno 1000, Atti del Congresso Internazionale, Bobbio, Archivum Bobiense.
Le lpreux, au folio 97 v., est peint le corps courb, une corne attache en bandoulire autour de son torse. Sa peau est marron, seme de taches sombres.6 Cest un personnage singularis, par la couleur de sa peau et par le bruit quil fait avec sa corne. Et ainsi diffrenci, on peut mieux le mettre lcart, lexclure.7 Toutefois, je me demande si la figure du lpreux est aussi ngative que cela dans notre manuscrit. Non seulement le lpreux bnficie de la compassion du Christ, mais aussi il jouit dun traitement de couleur particulier: sa peau, aussi dvalorise soit-elle, nest pas de la mme couleur que celle de Satan.8 (Fig.12) Plusieurs feuillets plus loin, au folio 231 v., lorsque Lazare, couvert dun linceul, est ressuscit par le Christ, sa peau est gris verdtre. Cest comme si cette association du gris et du vert signifiait le mieux la pleur cadavrique. (Fig.13) Les personnifications fminines des quatre provinces impriales, au folio 23 v., se prtent galement un intressant jeu de couleurs de la peau. La couleur est intercale dune femme une autre. De cette manire, Roma et Germania ont la peau de couleur jaune verdtre, tandis que celle de Gallia et de Sclavinia est marron clair. Place de manire intercale, la couleur de la peau donne du rythme limage. Or, est-ce que cest un hasard si la peau de Roma, la province place le plus prs du pouvoir imprial, est de la mme couleur que celle de lempereur Otton III?9 (Fig.14) Au folio 24, lempereur Otton III, figur en majest, a la peau jaune verdtre. sa droite, les deux clercs ont aussi la peau jaune verdtre, tandis qu sa gauche, les lacs lont beige. La couleur de la peau semble rapprocher lempereur des clercs, le pouvoir imprial de lglise. Est-ce quelle ne traduit pas les enjeux idologiques de la politique impriale des ottoniens? La peau du Christ pourrait donner un dbut de rponse. (Fig.15)
47
colour and meaning
Jai relev trois nuances diffrentes pour la peau du Christ: beige clair, blanc verdtre et verdtre. Pourquoi ces diffrenciations? Jai remarqu que Jsus avait la peau beige dans toutes les scnes qui ont un rapport avec sa vie terrestre et publique. Je pense ici, aux scnes du Baptme, des Tentations, des enseignements, des gurisons et de la Passion. Jai galement observ quil avait la peau couleur blanc verdtre dans les scnes o la nature divine du Christ tait mise en avant, comme celle du Christ en majest ou bien de la Rsurrection. Enfin, dans la scne de la Transfiguration, la peau du Christ est entirement verdtre. Une couleur de peau particulire, unique dans le manuscrit, qui saccorde avec la couleur claire de son vtement. Comme si le verdtre tait la nuance de peau qui se rapprochait le mieux du divin. (Fig.16) quelques feuillets dintervalle, la majest impriale et la majest du Christ se font cho. Comme sil ny avait aucun intermdiaire entre lempereur et le Christ. La peau de lempereur et celle du Christ divin tendent toutes les deux vers le verdtre. une nuance prs: celle de lempereur est jaune verdtre tandis que celle du Christ est blanc verdtre. La couleur de leur peau se ressemble, mais nest pas la mme. Tout en faisant partie du monde spirituel, voire de la mouvance divine, Otton III nest pas
48
fig.16 folios 174 v., 34 v. et 113 (dtails) (briechle andrea, etc., das evangeliar kaiser ottos iii, bayerische staatsbibliothek, mnchen, clm 4453. augsburg, haus der bayerischen geschichte, 2002)
le Christ, lempereur nest pas Dieu. Serait-il hasardeux de dire que la couleur est un outil de pouvoir dans lidologie impriale? Lvangliaire dOtton III est extrmement intressant par le nombre de questions quil ne cesse de soulever. La couleur y apparat codifie et pense en fonction dun contexte bien particulier. Comme si sur ces deux espaces que sont la peau et le vtement il existait un chiasme de couleur du clair vers le fonc et du fonc vers le clair. La peau claire et le vtement fonc tant mieux mis en avant que la peau fonce et le vtement clair. Mais ltude de la couleur dans ce manuscrit force la plus grande prudence. Car au vu de lensemble de mon corpus, ce document demeure une exception. Il force moduler peu peu et prudemment mon analyse, car ce qui est valable pour ce manuscrit ne lest plus pour la grande majorit des manuscrits de mon tude. Pour la priode choisie, fin Xe fin XIIe sicles, en pleine mutation, la norme nest jamais de mise.
Biography
Je suis dorigine franco-vnzuelienne. Aprs des tudes dHistoire lInstitut Catholique de Paris puis lUniversit La Sorbonne-Paris IV, je suis en train de terminer un Doctorat en Histoire mdivale lcole Pratique des Hautes tudes (E.P.H.E.), Paris, sous la direction de Monsieur Michel PASTOUREAU. Le sujet de la Thse est: Les couleurs du corps nu et vtu dans les manuscrits enlumins des Xe-XIIe sicles: une tude des rapports entre la couleur et le corps dans limage mdivale. Publication: Article Ich halte Dich am Bart: Spiel der Hnde, Spiel der Farben in einem Ritual zwischen Mnnern, in Edgar BIERENDE, Sven BRETFELD et Klaus OSCHEMA, Riten, Gesten, Zeremonien, Trends in Medieval Philology, Volume 14, Walter de Gruyter, Berlin, New-York, 2008, pp. 69-78. Adresse e-mail: iotamarie@gmail.com
Crdits photographiques: BRIECHLE Andrea, FOX Angelika, MEYER Carla, Das Evangeliar Kaiser Ottos III: eine Handschrift zum Blttern, Bayerische Staatsbibliothek, Mnchen, Clm 4453. Augsburg, Haus der Bayerischen Geschichte, Bayerisches Staatsministerium fr Wissenschaft, 2002.
49
Resumo
Este artigo discute as cores e pigmentos utilizados numa das cpias dO bestirio do amor de Richard de Fournival, datado do 3. quartel do sc. XIII. A aplicao de uma certa paleta cromtica bem como de certas especificidades iconogrficas faz-nos pensar que este artista conhecia as tcnicas utilizadas na produo bem mais sofisticada da Cappella Regis sob o patronato de S. Lus: o Saltrio de S. Lus que se encontra na BNF, em Paris, e o ms designado por Saltrio de Isabella no Museu Fitzwilliam, em Cambridge.
palavras-chave
cor pigmento iluminura Richard de Fournival bestirio
Abstract
This paper considers the colours and pigments of a manuscript of the Bestiary of Love of Richard de Fournival from the third quarter of the 13th century. The particularities of the use of the chromatic palette in this manuscript, together with several iconographical devices, allow one to reflect on the knowledge, by the artist who decorated this manuscript, of the methods of work used in the most sophisticated products of the Cappella Regis under Saint Louis: the Saint Louis Psalter in the BNF of Paris and the so called Isabella Psalter in the Fitzwilliam Museum in Cambridge.
key-words
colour pigment illumination Richard de Fournival bestiary
Agradecimentos por ajuda na configurao do texto e sugestes de Ana Catarina Sousa, Pedro Fialho de Sousa, Justino Maciel, Felix Teichner e Heidi.
remarks on colours and pigments in the french court illumination of the 13 th century
x e n i a m u r atova
Professor Emeritus xmuratova@aol.com
1. X.Muratova, B. Roy, Le manuscrit T du Bestiaire dAmour de Richard de Fornuival. Edition critique et tude des enluminures, Ramsen, in print.
The aim of this paper is to present the study of colours, pigments and of the pictorial technique in an illuminated manuscript of the Bestiaire dAmours of Richard de Fournival, the so called manuscript T of this work (Switzerland, private collection)1. Rather modest in size (23 folios, of the dimension 17 x 12 cm), it is, nonetheless, one of the most important manuscripts of this famous work. It is not only one of the most ancient known manuscript of the Bestiary of Richard de Fournival (it can be dated from the 1260s-1270s, whereas the other 23 known manuscripts of this work date from between the 1270s and the 15th century ). Written in francien, the Ilede-France version of the Old French, it is, also one of the most ancient illuminated vernacular manuscripts known to us. Another important point is its relationship to the French court: the miniature of the Vulture, allegory of the persecution of a beloved by her lover in the ironical interpretation of this fine writer of the 13th century, presents the images of Saint Louis followed by his son-in-law Thibaud V, count of Champagne and king of Navarra (1235-1270), easily recognizable by their arms. It is probable that the manuscript (or its model) was commissioned by or for Thibaud of Navarra, and it cannot be excluded that the manuscript, the last miniature of which represents the God of Love, served as a wedding present. The miniatures of this manuscript give an exciting possibility for the close investigation of colours, pigments, of their deterioration and transformation, their reaction with the parchment and their modification visible on the reverse side of the leaves. This study shows us the proceedings employed by an artist who knew the methods
51
colour and meaning
of work and the repertory of forms used by an important group of artists which executed the French Royal Psalters of the 1260s, the Saint Louis Psalter, ms. 10525 of the BNF of Paris2 and the so-called Isabella Psalter, ms. 300 of the Fitzwilliam Museum in Cambridge3. It is difficult to say if he belonged himself to this group of artists; but in any case he was acquainted with their work. At the same time he was confronted with an unusual task: to ornate a profane, vernacular manuscript for the decoration of which he was not always provided with the iconographical devices of several scenes4. Beside that, the study of the technique of the execution of the manuscript gives rather an impression that the order was particularly urgent and that the work was executed in a hurry. Sixty-four small miniatures of the manuscript (the height of the majority of them takes between three or five lines of the text) are placed on a blue or a pink watercolour background but the tonality of these colours (especially a very deep, brilliant, transparent blue) as well as the technique of the transparent watercolour itself used for the background are closer to the chromatic devices and to the technique used by English illuminators or by those who had an English training or were strongly influenced by it. It is not an exception in the North French and Parisian miniature of the 13th century. But the use of the watercolour for the backgrounds can be hardly found in the majority of the French miniature production of this period. However, it is characteristic for full-page miniatures of the two Royal Psalters where the coloured parts of the background surround the central parts painted in gold. In the Bestiary manuscript, the slightly orange pink colour is also used in watercolour technique and covers the surface of the background in a rather irregular manner. This orange pink is very close to the same colour in the Saint Louis Psalter. In rare cases, however, the painter used a small amount of a binder or of a white in order to unify the tone. But his mixtures were too fragile and led to the deterioration of the some paintings. Perhaps, these mixtures were prepared too rapidly, the binders were used in a too precipitated manner or, what is also possible, the degree of the heating of the mixture was too high or too rapid. In his work, the artist transgressed several norms adopted in the French manuscript production of the 13th century. The alternation of blue and pink backgrounds is not regular, as in the majority of the French manuscripts where this kind of alternation constituted a system of personal marks in cases of the decoration of manuscripts of the a large size. The artist of the Ms. T demonstrated a surprising freedom regarding several norms which characterized the work of established workshops of illuminators and book-makers. His practices are those of the illuminators who were called to execute the decoration of small vernacular manuscripts, where many rules used in the ateliers specialized in the production of important Latin manuscripts were simply neglected and disregarded. It is also difficult to judge whether the designer and the painter who just coloured the silhouettes were the same person. It seems, in fact that the quality of the design,
2. R.Branner, Manuscript Painting in Paris During the Reign of Saint Louis, Berkeley and Los Angeles, 1977, pp.5, 132-137, 238-239;H. Stahl, Picturing Kingship. History and Painting in the Psalter of Saint Louis, University Park, Philadelphia, 2007; X.Muratova, B.Roy, op. cit.; X. Muratova, Officina di miniatori della Cappella Regis allepoca di San Luigi, Medioevo: Le Officine. I Convegni di Parma XII. Atti del Convegno internazionale di studi, Parma, 22-27 settembre 2009, a cura di A. C. Quintavalle, Milano, 2010, pp. 511-516. 3. S. Cockerell, A Psalter and Hours Executed Before 1270 for a Lady Connected with St. Louis, Probably His Sister Isabelle of France, London, 1905; S.Panayotova, The Psalter and Hours of Isabelle of France, The Cambridge Illuminations. Ten Centuries of Book Production in the Medieval West, Catalogue of the Exhibition, ed. by P. Binski & S. Panayotova, Cambridge, 2005, London-Turnhout, 2005, pp.178-180; X.Muratova, Officina, op. cit. 4. X.Muratova, Un episode de la pratique de travail des enlumineurs au XIIIe siecle: lutilisation des motifs bibliques pour illustration des ecrits profanes, Imagenes e promotores en el arte medieval. Miscellanea en homenaje a Joaquin Yarza Luaces, Barcelona (Bellaterra), 2001, pp. 545-554.
52
r e m a r k s o n c o l o u r s a n d p i g m e n t s i n t h e f r e n c h c o u r t i l l u m i n at i o n o f t h e 1 3 t h c e n t u r y
fig.1 the unicorn and the vultur, bestiaire damours of richard de fournival, switzerland, private collection, ms. t, fl. 10v and 23r, respectively.
vivid, light and rapid, in some cases is superior to that of the colouring but several deterioration of the pictorial layer and of pigments prevent us from making a definite judgment about the execution of the miniatures by one or by two artists. The colours and the pigments of the ms. T were studied essentially de visu, with a magnifying glass and a microscope. A physical-chemical analysis was attempted in a local laboratory but was not carried through to its end for several reasons having nothing to do with scientific research. No radiographical or microspecrometrical analysis was ever produced. Some physical-chemical results confirmed hypotheses made during my study of pigments with a microscope. In several cases the artist did not follow the rules of the utilisation of pigments set in the known manuals of medieval illuminators, which is by itself not an important exception in the work of the medieval illuminators who worked in the milieu dominated by an oral tradition and where the individual experience of each artist who transmitted it to his workshop, to his pupils and to his professional surrounding played the primary role. I dont think that the deteriorations of pigments in several miniatures are due to a bad knowledge of the technique; the manuscript was undoubtedly decorated by a fine professional artist, but I am inclined rather to think that he had too short a time for the execution of his order and that is why he could not respect the time needed for the rest (or suspension) necessary for the preparation of pigments and their mixtures. He may also have found himself in a particular situation, outside of a workshop, travelling, for instance, which obliged him to use the moyens de bord. Yellow: the frames, very simple, of miniatures are painted in a very pale yellow which imitates probably, the golden frames. The presence of the infinitesimally small particles of gold (which can be observed through the microscope and in several cases can be observed even with a magnifying glass) in this yellow pigment shows the use of the orpiment, the auripigmentum album obtained from sulphur of arsenic and used with parchment glue as a binder. Blue: the chromatic palette is distinguished by the use of various tones of blue. First of all, it is the deep blue sapphire colour of the backgrounds, in several cases strongly dissolved and clear; in some other cases it becomes blackened because of the oxidation of the ornamental motives painted on the surface of the backgrounds. In many cases the dissolved blue pigment has transgressed the lines of the silhouettes of animals and of personages, even if the artist seems to do everything to avoid this extension of the blue This blue pigment, cold and intense, has corroded the parchment and produced visible traces on the reverse of the sheets (where it gave clear green traces); thus, it included corrosive substances which have profoundly penetrated the texture of the parchment. This type of deterioration of the parchment shows the use of a rather active corrosive substance, the azurite, a mineral pigment based on the acetate or a carbonate of copper. The blue of garments and vestments is also intense, profound and a brilliant colour but it has not the transparency of the blue of the backgrounds. It is not the same pigment used in more concentrated proportions and in a more dense manner. This
53
colour and meaning
blue pigment has none of the corrosive proprieties of the blue of the backgrounds and did not attack the parchment. This blue was obtained from indigo. Thus, the palette of our artist has several blue pigments, rather close in their chromatic effect, but of different origin. In several cases one can observe the clearer, and more opaque blue due to the mixture with the white lead in various proportions. In several cases of the utilisation of this mixture one can observe small losses of the pictorial layer. There is also the blue-grey, very clear, and, finally, a very rare blue grey, transparent, very clear and cold. This colour has a slightly corrosive action: from the other side of the sheet it gives an effect of a dark yellow or a slightly orange colouring on the vestment of the Allegory of the Memory (Figure 2). This corrosive effect shows equally the use of a pigment containing the copper oxide, prepared probably on a base of the fusion of various substances and based on copper sulphate. One should note that this pale blue colour, transparent and cold, very rare and particular, has analogies with the pale blue which can be found in the Saint Louis Psalter (Figure 3) but in this manuscript a different or better cleaned pigment was used and it had no corrosive action on the parchment. A similar tone of the pale and transparent blue is used also in the Cambridge Isabella Psalter (Figure 4). A unique case where the blue colour which has a blue celestial tone, is that of the lapis-lazuli with which were painted the vestment and the banner of the King Louis IX in the miniature of the Vultur: in is interesting to note that in the same miniature the painter used an extremely mordant pigment to paint in blue the back legs of the horse of Thibaud, King of Navarra. Small touches of blue grey (in the neighbourhood with the touches of clear brown colour) were used for modelling the silhouettes of animals painted in beige or in clear yellow. This perfect knowledge of the use of the complementary colours and the taste for the modelling and the plastic quality in the treatment of forms can unfortunately be appreciated very rarely: this extremely delicate finish and its nuances were effaced because of the extensive use of the manuscript and the fragility of the technique and can be only guessed at in certain miniatures. The variety of the blue tones shows a rather refined taste but only more profound scientific analysis could help to conclude which were exactly the resources of the painter in the matter of pigments. In any case, the use of these resources was extremely masterly: already in the unique initial of the manuscript representing the Allegory of the Memory (Figure 2) one can find four tones of blue and five tones of pink-beige colours, some of them were produced thanks to the mixture with white lead and with other pigments. Green: as to the green, the dominant tone is that of a vivid and almost emerald brilliant green, very widespread in the medieval miniature and produced from the basic copper acetate, called in the Middle Ages Iarim or Iarin and a common name of which was a verdigris. It is used for the painting of silhouettes of animals, for vestments, trees and grass. As is well-known, it is one of the most aggressive pigments of the medieval painting which corrodes and penetrates the parchment and
fig.2 allegory of the memory, bestiaire damours of richard de fournival, switzerland, private collection, ms. t, fl. 1r (ou 25v?).
fig.3 saint louis psalter, paris, bnf, ms.10525, fl. 9r.
54
fig.4 isabella psalter, cambridge, fitzwilliam museum, ms.300, fl. 1v
fig.5 the tiger, bestiaire damours of richard de fournival, switzerland, private collection, ms. t, fl. 10r.
becomes visible on the reverse of the sheet. Several silhouettes painted in green preserve the rest of the modelling executed with the same colour, a little darker or a little clearer. Thus, it was a rather refined and skillful painting where the painter knew how to use the possibilities of each pigment which he worked with a great refinement and savoir-faire. However, one should note that normally this brilliant tone of green is not characteristic of Parisian painting of the 13th century, where the artists, who always used verdigris, preferred the tones of bottle green and of grass green and in general used the green pigment rather rarely, knowing its destructive properties (Figure 5). In fact, the use of green constitutes one of the sensitive points of the Paris illumination. On the contrary, in the English miniature, the green lavis (in fact the pigment is generally dissolved with water) is one of the preferred means of chromatic expression. Its extensive utilisation in our manuscript is another testimony of the training of the artist not in Paris but in the North of France, in any case, under the strong influence of practices adopted in English illumination. A similar bright and brilliant green is characteristic for the painting in both Royal Psalters. Pink, beige, orange, brown: the manuscript T has a very particular pink orange colour used for the backgrounds where it is put by means of a transparent brushwork painting and in a very irregular and unequal manner which creates the darker and clearer parts of the painted surface. It is not at all common for Parisian and North French illumination where the pink colour, called often the rose de Paris and obtained from a mixture with lead, covers the surface in a dense, opaque, equal and regular manner. Here it is a saffron colour, intense and brilliant. This orange pink, as well as the watercolour technique of its utilisation which gives an impression of the irregularity of the pink orange tone can be found only in the Saint Louis Psalter executed by the artists attached to the Cappella Regis, especially in the architectural parts of miniatures. Several other orange and beige colours, their tones and nuances were used for the silhouettes of animals, attributes and vestments. One of these beige-pink pigments has corrosive action on the parchment and is visible from the other side of the leaf (Figure 5), testimony of the use of pigments rich in iron oxide and other metals. This effect could be also produced through the use of the extract of fish gall. A clear pink colour of this manuscript is extremely fragile and inclined to deterioration: its surface is often covered with small spots. This deterioration is due to the fermentation of the clarea egg white probably heated too rapidly. In fact, the painter illuminating a vernacular manuscript, even if he knew the old traditions and rules, was no longer strictly attached to them; the work on the order obliged him to neglect several old rules and led sometimes to a too rapid and hurried preparation of colours. White is used to paint the silhouettes and vestments: it is an extremely white, limpid, brilliant and transparent colour obtained by the use of chalk. The drawing is perfectly visible through the its surface. The traditional lead white is, of course, also used rather widely.
55
colour and meaning
Black ink was employed for the drawing of the contours and parts of silhouettes. Grey is dissolved black ink. Very light grey, mixed with white lead was used for the modelling of faces and silhouettes. Faces are painted with touches of light grey and pink. Several silhouettes are of a dark grey colour. In several miniatures one can observe the degradation of the grey colour because of the use of an unadapted binder (or too hurried heating of the clarea which turned during the heating). Red is used only for the rubrics: it is a vivid and brilliant red obtained from natural cinnabar, red sulphate of mercury. Gold, silver and other metallic pigments are absent at a first glance. The gold particles, which can be observed in the yellow of frames and in some attributes, is due to the effect of the orpiment. It is probable however that the gold or its substitute was used for the decoration of ornamental scrolls on backgrounds of several miniatures. A small portion of such a scroll is visible in the lower part of the miniature of the Wild Ass. The disappearance and the oxidation of the scrolls (Figure 6) painted with metallic pigments or their substitutes shows that the essential rules of the work with these pigments were not followed by the artist. At the same time, these scrolls are so blackened and oxidized that their traces appear on the reverse of the folios. It is not an effect which could be produced in the case of the use of white lead for the painting of the scrolls. Besides that, white lead doesnt show through the parchment. The black traces left by the pigment used for the painting of the decorative scrolls reveal the use of a substitutive pigment for the gold, may be of silver which often attacks the parchment, shows through it and leave the black traces on the reverse of the folios. The almost total disappearance of this part of the decoration of our manuscript changed considerably the initial aspect of the manuscript. There is no doubt that in the beginning it produced an impression of a much more sumptuous object. It is difficult to make a definitive conclusion as to the pigments used in the ms. T without a new verification of chemical results and microspectoscopic analysis. However, observations concerning the colours and the chromatic palette permit one to make several observations and to propose a number of hypotheses. The choice itself of the chromatic palette, based on the tones of blue, green, grey, pale yellow, beige and pink orange is very particular. It is a choice which is not due to the limited means which were at the disposal of the artist. This choice is due to his training and to his personal inclination. Even if numerous elements of the decoration have disappeared or were modified with time, the analysis of pigments and the choice of the chromatic palette show a refined artistic taste. At the same time, it is not a palette characteristic of the Parisian workshops of the middle of the 13th century and of its third quarter. These workshops attached a great importance to a perfect balance of blue and pink colours retouched with a parcimonious and skillful use of red and green. In the production of the illuminated manuscripts in vulgar languages during the third quarter of the 13th century, this kind of approach characterizes especially the deco-
fig.6 the panther, bestiaire damours of richard de fournival, switzerland, private collection, ms. t, fl. 10v.
56
ration of books executed by artists working North of the capital, and independent in their chromatic tastes from Parisian norms. Thus, on one hand, the use of pigments and colours in this manuscript shows a rather free approach to the colour; on the other hand it demonstrates the knowledge of English illumination and of its chromatic devices. The chromatic palette of the ms. T, in its present state, in any case, reminds one of the favourite palette of English artists of the second and third quarters of the 13th century. At the same time, in relation with the majority of illuminated manuscripts of this period coming from workshops of the North and East from Paris, with their variegated colours, thick linear drawing and absence of modelling, the ms. T impresses one by the grace and vivacity of its linear drawing, freshness of the narration, refinement of modelling in painting and its remarkable choice of he chromatic palette. The disregard of traditional rules, indicated in the medieval manuals of painting, in the use of mixtures of pigments, is a feature which seems rather becoming to the work of artists who were not bound by the secular traditions of the professional production of Latin manuscripts. These artists began to specialize independently in the decoration of manuscripts written in vulgar languages. It is particularly interesting to note that in this case the artist based his work partly on the experience of artists who did not belonged to big Parisian workshops but who constituted a group of artists attached to the Capella Regis and working together on the decoration of Royal Psalters obviously intended for the French court.
Biography
Xenia Muratova is a Professor Emeritus of French Universities. She studied History of Art in Russia, in France and in Germany. Her work is dedicated to various problems of medieval art, especially to the study of the 12th and 13th centuries, of methods of work of the medieval artist, of the medieval illumination and of medieval bestiaries. She is the author of several books and numerous articles and essays on these topics, the most important of these publications are monographs: Masters of the French Gothic, 1988; The Medieval Bestiary, 1984; Le Bestiaire medieval, 1988; Storia universale dellarte. Arte del secolo XI, 2003; Storia universale dellarte. Arte del secolo XII, 2005. Xenia Muratova works also as art critic and writes on Modern and Contemporary painting as well as on the History of the history of Art. She is member of numerous associations and scientific societies in Europe and in America.
57
byza n t i n e re ci pe s a n d b ook illum ination
Resumo
Conhecem-se centenas de receitas medievais para a produo de cores, iluminuras e tintas, no Ocidente latino. Pelo contrrio quase nada parece ter sobrevivido do mundo bizantino que lhe foi contemporneo. H uma enorme lacuna entre os papiros de Estocolmo e Leiden, datados aproximadamente da terceira centria (c.), e os tratados ps-bizantinos como o Annimo 1566 de Jerusalm e, o mais conhecido, Hermeneia do sc. XVIII, escrito por Dionsio de Phourna. Tem sido muitas vezes assumido que os textos ps-bizantinos reflectem mais fielmente as tradies bizantinas, mas uma anlise rigorosa desses tratados tardios revela, afinal, em parte, influncias das tcnicas ocidentais ps-medievais. Portanto, no podem ser tomados sem alguma cautela como substitutos dos textos bizantinos perdidos. Alm disso, durante os 1000 anos que durou o Imprio bizantino, a arte bizantina no foi uniforme nem em estilo, nem textualmente. A tcnica da pintura varia entre scriptoria de diferentes regies e cronologias, novos pigmentos foram inventados (amarelo de estanho e chumbo, azul de esmalte) mudanas que os textos ps-bizantinos no referem. Do tempo bizantino, encontram-se publicadas apenas umas poucas receitas, sobretudo de tintas. Num projecto coordenado por Peter Schreiner (Universidade de Colnia) e Doris Oltrogge (Fachlochshule Koln), a pesquisa sistemtica em manuscritos catalogados poder ampliar este nmero para cerca de 80 receitas relacionadas com a produo de manuscritos. Os principais tpicos so tintas, crisografia, douradura, a produo de vermelho e de laca vermelha. Os primeiros textos datam do sc. XII, tendo muitos deles sido escritos nos scs. XIV e XV. At agora no sobreviveram tratados sobre tcnicas artsticas comparveis a Theophilus ou Cennino. H igualmente poucos vestgios de papiros da Antiguidade Tardia. Sendo a maioria das receitas novas e nicas, algumas tcnicas so comparveis s que so descritas nos textos do Ocidente, mas apenas possvel referir uma influncia directa nos manuscritos escritos por gregos exilados na Itlia do sc. XV. Algumas das tcnicas descritas podem ser observadas nos manuscritos bizantinos. Por outro lado, h ainda muito poucas receitas para a produo de pigmentos sintticos, mesmo quando as anlises cientficas mostram que alguns pigmentos sintticos importantes, tais como, o amarelo de estanho e chumbo foram usados na iluminura bizantina mais cedo que no Ocidente podendo ter sido inventados em Bizncio.
palavras-chave
bizncio receitas iluminuras crisografia pigmentos
59
Abstract
Hundreds of mediaeval recipes on colour production, book illumination and inks are known from the Latin West. On the contrary nearly nothing seems to have survived from contemporary Byzantium. There is an enormous gap between the late antique papyri in Stockholm and Leiden, dating from about the 3 rd century, and the postbyzantine treatises like the 1566 Anonymous in Jerusalem and the better known 18th-century Hermeneia, written by Dionysios of Phourna. It has often been assumed that the post-byzantine texts reflect rather faithfully byzantine traditions, but a close analysis of these later treatises reveals at least partly influences of postmediaeval Western techniques. Therefore, they cannot be taken without reservations as substitute for lost Byzantine texts. Furthermore, Byzantine art is not uniform neither in style nor in text during the 1000-year-period of the Byzantine Empire. The painting technique varies between scriptoria of different times and locations, new pigments were invented (lead-tin yellow, smalt) changes on which the post-byzantine texts give no information. From the Byzantine time, only some few recipes, mainly on inks, are published up to now. In a common project of Peter Schreiner (University of Cologne) and Doris Oltrogge (Fachhochschule Kln) a systematic survey on catalogued manuscripts could enlarge this number to about 80 recipes dealing with the techniques of manuscript production. The main topics are inks, chrysography, gilding, and the production of vermilion and of red lakes. The earliest texts date from the 12th century, most of them were written in the 14th and 15th centuries. So far no treatise on art technology comparable to Theophilus or Cennino has survived. There are also only few traces of the late antique papyri. Most of the recipes are new and unique. Some techniques are comparable to those described in Western texts but a direct influence can only be stated in manuscripts which were written by Greek exiles in 15th century Italy. Some of the techniques described can be observed in Byzantine manuscripts. On the other hand, there are still extremely few recipes for the production of artificial pigments, even if analysis can show that some important artificial pigments like lead tin yellow were used in Byzantine illumination earlier than in the West and were perhaps invented in Byzantium.
key-words
byzantium, recipes, book illumination, chrysography, pigments
byzantine recipes and book illumination

d o ris o ltro g g e
Cologne Institute for Conservation Sciences Cologne University of Applied Sciences D-50678 Kln (Germany)
1. For the most comprehensive overview see: Clarke 2001. 2. In this article Byzantine is strictly defined as a period term, i.e. the time before the fall of Constantinople in 1453. Texts written by Greek scribes before the middle of the 15th century in places outside the Byzantine Empire especially since the later 14th century in areas already under Turkish dominion or by exiles in Italy or elsewhere are considered as Byzantine. The major Greek texts on painting technique dating from the 16th century onwards belong to the postbyzantine period.
More than 400 mediaeval manuscripts dealing with the production of pigments and inks, with binding media, and with the techniques of painting and book illumination have survived from the Latin West.1 On the contrary nearly nothing seems to have come down to us from contemporary Byzantium. This is more than strange if we consider that the earliest known Latin source the Carolingian Lucca manuscript is clearly a translation from the Greek (Clarke 2001, 8). There are, however, Greek texts on art technology from Antiquity: information on pigments are found in the Book on stones written in the 4th century BC by Theophrast or in the medical treatise of Dioskurides, dating from the 1st century AD. Probably in the 3rd century AD the Leyden and Stockholm papyri were written, which transmit the most comprehensive collection of recipes for metallurgy, dyeing and alchemy (Halleux 1981). The art of writing is considered in 16 recipes for true and false chrysography (Trost 1991, 58-102). On the contrary, texts on art technology seem to be nearly non-existent in Byzantine times, that is in the period between ca. 500 and 1453 AD.2
The Postbyzantine tradition

Due to this lack of original sources from mediaeval Byzantium studies on Byzantine painting technique usually refer to the Hermeneia tis zographikis (the Treatise of Painters) written in 1730-1733 on Mount Athos by the painter-monk Dionysios of Phourna (Papadopoulos-Krameus 1909; Hetherington 1974; Bentchev 2004). This treatise is generally assumed to transmit rather faithfully the earlier Byzantine tradition. This is certainly true for large parts of the iconographical section: postbyzantine wall and icon painting depends largely on mediaeval models and so do Dionysios descriptions of the major scenes from the New Testament. However it has been stated that in some cases the iconography follows the more modern types of the Cretan school which was clearly influenced by Italian, namely Venetian paintings from the middle of the 16th century onwards (Bentchev 2004, 67). And also some of
61
co lo u r i n w r i t t e n m e d i e va l s o u rc e s
the technical recipes reflect the techniques of Italo-Cretan and Italian oil painting (Bentchev 2004, 40; 67). The technical section of Dionysios Hermeneia can be traced back to about 1566, the date proposed for two anonymous treatises which were copied in a 17th century manuscript in the Library of the Greek Patriarchate in Jerusalem (Bentchev 2004, 39). Anonymous I quotes as references the painters Panselinos, who most probably lived in the 14th century, and Theophanis the Crete, who worked in the second half of the 16th century on Mount Athos. Dionysios also mentions Panselinos, but not the name of the more recent painter, whose recipes he however copied (Bentchev 2004, 41). Besides, two other postbyzantine Greek treatises on painting technique are known. In 1720, the Ionian painter Panagiotes Doxaras (1662-1729) translated Leonardos Trattato della pittura into Greek, and in 1726 he wrote himself a treatise on painting Peri zographias (Bentchev 2004, 130). This treatise as well as Doxaras translation of Leonardos Trattato were used as references by the icon painter and engraver Christofor efarovi in his Instructions for young people desiring to learn the art of painting (Koinnou thesia t Hermeneia eis tous neous), a work written most probably in Church Slavonic but translated also into Greek (Bentchev 2004, 154s). efarovi is the most Western of these authors, besides Leonardos Trattato he used the German Kunst- und Werck-Schul of 1705 and Pictorius Lackir-Kunst of 1708. However, he also knew the Greek tradition of the Jerusalem treatises which was the major source for the Hermeneia of Dionysios of Phourna (Bentchev 2004, 156). But how Byzantine is this later Greek tradition? The translators of the Hermeneia have remarked that the text of Dionysios as well as that of the Jerusalem manuscript is full of unusual terms, words of Italian, Turkish and even German origin. So the German term (Goldfarbe gold colour) is used explicitly to explain a Venetian gilding technique considered as unusual to Greek traditions.3 Turzisms are not unusual in the late Byzantine period, especially in the parts of the Byzantine Empire which were already occupied by the Turks before 1453. Even more interesting are some of the Italianisms. In a recipe for the making of verdigris 4 the product is not called ios or ios chalkou as it should be in classical Greek but vardramon which clearly derives from the Italian verderame. Another recipe of the Jerusalem manuscript and the Hermeneia describes the production of a blue or red colour: 5 First a lye is made from quicklime and potash. Then wool shearings are put into the lye and cooked until the colour is extracted. Afterwards alum and a little glair are added to the filtered colour extract. Depending on the colour of the shearings the pigment obtained will be blue or red. The recipe describes thus the making of an indigo pigment from blue shearings and of red lakes from the shearings of scarlet cloth. Similar recipes have often been copied in Western manuscripts from the 14th century onward (Wallert 1991). Is this a Byzantine technique taken over by the Latins? Or was a Western recipe adopted in late or postbyzantine time by the Greek? The terminology used by the Jerusalem
3. Dionysios of Phourna, 34 (BENTCHEV 2004, 92). 4. Anonymous 35 (BENTCHEV 2004, 57); Dionysios 42 (BENTCHEV 2004, 95). 5. Anonymous 39 (BENTCHEV 2004, 58); Dionysios 45 (BENTCHEV 2004, 97).
62
6. The edition is planned for 2010 (Wien: sterreichische Akademie der Wissenschaften).
Anonymous and Dionysios points to the latter: the shearings are called tzimarsmata, a word so uncommon in Greek that it has lead to some confusion in the various translations of the Hermeneia. Even Bentchev who recognised the meaning of the recipe translates it inexactly as Stofflumpen (rags). The word clearly derives from the Italian cimatura (shearing), therefore most probably the recipe is a translation from the Italian. Whether this was done already in the Byzantine period or only by the Anonymous of the 16th century cannot be decided.
Byzantine texts on manuscript production

These few examples show that the Hermeneia of Dionysios of Phourna and the earlier Jerusalem Anonymous should not be taken without restrictions as sources for painting techniques of the Byzantine period. The immense gap between the 3rd century papyri and the postbyzantine treatises still remains open. But at least a small number of recipes have survived which give us some isolated information on art technology in mediaeval Byzantium. A few instructions for making vermilion are transcribed in alchemical manuscripts (Berthelot 1887). Some ink recipes have been published by Zerdoun Bat-Yehouda in 1983. A major finding was done by Peter Schreiner, former chair of Byzantine Studies at the University of Cologne, when he discovered the first known small treatise on manuscript production from the Byzantine Period, written in the early 15th century by Isidor of Kiev (Schreiner 1988, 116-125). The problem to understand the techniques described, often with ambiguous or unusual terms, led to an interdisciplinary project with Professor Schreiner as Byzantine philologist and myself as an expert for art technology. The aim of this project is not only the publication of the small text compiled by Isidor of Kiev, but a commentated edition of all available Byzantine recipes on the techniques related to writing and book illumination. 6 Purely alchemical treatises are not part of the project. For practical reasons, the search had to concentrate on catalogues so that libraries without or with very old catalogisation are not considered. Therefore new findings will be possible when more libraries are made accessible by useful catalogues. Nevertheless the number of recipes concerned with the art of manuscript production could already be increased to about 80, and furthermore, with Isidors of Kiev text there is a first indication for the existence of more or less systematic art technological recipe collections in Byzantium. The earliest texts date from the 12th century, most of them were written in the 14th and 15th century. Compared with the abundant Western tradition the number is still very meagre. It seems that art technology plays an extremely smaller role in the surviving Byzantine manuscripts than in their Western counterparts. It is difficult to decide what the reasons are. Was there less interest in the practical arts by the literate people in Byzantium? Or got the manuscripts simply lost after the fall of Byzantium because the Western collectors were more interested in luxurious codices or in literary and
63
scholarly pretensious texts than in recipe collections of doubtful literary quality and perhaps even bad Greek? At least, there are some arguments for the latter hypothesis: First, there was a tradition of technical texts in Greek Antiquity as documented in the papyri which must have been still well known at least in the early Byzantine period, when the group of recipe texts known as Mappae Clavicula was translated into Latin (Oltrogge 2006, 555). The two translations of this text the version of the late 8 th century Lucca manuscript with its numerous graecisms and the philologically more correct Mappae Clavicula version which became the standard for later Latin copies document the great interest of the Carolingian West in Greek technological literature which would be difficult to understand if this tradition would have been already interrupted in contemporary Byzantium. A second argument for the assumption that the scarce number of texts is mainly a question of survival are the few documents which have come down to us, especially the treatise of Isidor of Kiev.
Topics
The main interest of the scribes was their writing material, first of all iron gall inks. Another large group of recipes deals with chrysography, that is gold script; the recipes describe the production of gold ink, but also the application of gold leaf. Both techniques are also appropriate for illumination. Remarkably smaller is the group of colour recipes. The production of vermilion is either described in alchemical manuscripts or related to its use for writing. A number of recipes deals with the red lakes lachas and barzion. Instructions for other pigments are extremely rare, one for the refining of lapis lazuli, one for a green ink, another for dyeing bones with copper green. Information can also be found on glues for sizing paper, but also on casein glues for other purposes. Very few recipes do not belong to the art of manuscript production: oil varnishes, a recipe for an artificial steatite and a unique description of a pseudo-sgraffito. The recipes are found in rather different types of manuscripts. The most comprehensive recipe collection is that of Isidor of Kiev with 14 recipes (an extremely small number compared to Western manuscripts). Isidor begins with chrysography, gold inks and grounds for gold ink and gold leaf, he continues with iron gall inks, vermilion for the rubrication and starch glue for sizing the paper. At the end he returns to chrysography with three recipes for a gold ground made from brazil wood. The compilation was not written continuously, at a later date, Isidor added two recipes for chrysography on a blank space on fol. 2r, just after the iron gall inks. The interest of Isidor is clearly the interest of a scribe. Isidor, born about 1380/90 in Monembasia (Peloponnes) was highly esteemed for his erudition, his political and rhetorical qualities; he became monk, greek ambassador at the Conciles of Bale and Ferrara, Roman Cardinal and finally Unionist Patriarch of Constantinople. After the fall of Constantinople he had to exile to Rome where he died in 1463. Despite his
64
7. Oxford, Bodleian Library, ms. Canon. gr. 39.
brillant career he was an industrious copyist and compilator whose scholarly interests included theology, philosophy, rhetorics, natural sciences and as we see also the technological aspects of writing materials. (Schreiner 1996) The Codex Vaticanus graecus 914, which contains our small treatise, is a good example of Isidors occupations, the contemporary binding unites the recipe collection with rhetorical texts, medical excerpts, geometrical and metrological treatises and some alchemical notes by Hermes Trismegistos. Isidor must have collected recipes over a certain period, because he copied one of the recipes on iron gall ink and the instruction for making vermilion also on the last, blank leaf of an older manuscript with the works of Xenophon. 7 To these notes he added also another recipe for iron gall ink and for paper size which are not found in the Vatican codex. In other cases we are not so well informed on the compilators. Moreover are the recipes usually more isolated. Instructions for iron gall ink often stand on fly leafs or are added on blank spaces of pages with completely different content: patristic, rhetoric, law, literature. The recipes for pigments and chrysography are more often found in miscellanous manuscripts on nature sciences: medicine, astronomy, astrology and alchemy. Sometimes but not necessarily they are embedded in alchemical texts. But even in this context we find seldom more than five recipes in a sequence.
Recipes and workshop practice

Inks
As said before, one of the main topics are writing inks. There is one recipe for a gall nut ink, the other 22 instructions deal with iron gall ink. Most of them are very common: a mixture of gall nuts, vitriol and gum together with a liquid, which can be water, vinegar or wine. Variations are given by the quantities of the ingredients and the heating or not heating of the ink. Additions like nutshells or the fruits of cypresses occur in two recipes. On the whole, the group of iron gall ink recipes seems not to be very different from their Western counterparts, even if up to now no exact parallel could be found. But it is striking, that the Greek scribes were more interested in different colours of their inks. A number of recipes explicitly state that the ink should become black or light, and sometimes recipes of both types stand in a sequence. The black inks are always genuine iron gall inks without any additions of carbon black. Nevertheless carbon black inks or mixtures from iron gall inks and carbon black were used in some Byzantine scriptoria, especially in the Greek monasteries in the Arabian countries or in the border areas to the Muslim world. An interesting example is a New Testament, written in Syria in 1273 (Rome, Vatican Library, Cod. Borg. gr. 18). The iron gall ink which had first been used did not adhere well and was obviously too pale on some places. Therefore the scribe overwrote his text partly with another iron gall ink mixed with carbon ink. With the same ink he added the initial which was embellished with gold powder sprinkled upon the wet ink. This sprinkling technique as well as the carbon ink are characteristic for Arabian
65
scriptoria. From the Arabian countries also a number of recipes for carbon inks are known (Schopen 2004). Comparable recipes have also survived from Armenia, where carbon inks and mixtures with iron gall inks were common writing materials (Eliazane & Engel 2006). It is amazing that in the Liber Illuministarum and some other German manuscripts of the 15th century the production of a carbon ink is explicitly called incaustum graecum, Greek ink. (Bartl et al. 2005, 92). But up to now no written Greek source is known. And the practice to use a carbon ink for writing was restricted to very few scriptoria, the common writing ink of Byzantine scribes was the iron gall ink which is rather good documented in the recipes.
Chrysography and gilding

The next large group of Byzantine recipes concerns the gold writing, chrysography. This technique was highly esteemed in the Byzantine culture, golden and silver script was usually written by specialists, the chrysographers. It is therefore not amazing, that more than 35 recipes are related to this topic. They describe the production of gold inks, and the application of gold inks and gold leaf on grounds. Three different methods for the preparation of gold powder can be distinguished. The most simple is the mechanical grinding of gold leaf. Grainy additives like salt pulverise the tender gold leaf and can afterwards easily be diluted in water. But also viscous materials like honey, gum or glue are suited to disperse the gold powder. These methods are already described in the late antique Greek papyri. It is amazing that none of the Byzantine recipes is copied from these older texts, the method is comparable but the text of the recipes is new! In practice the grinding with salt must have been rather common because in a number of manuscripts with silver ink the presence of silver chloride could be proved. When silver is ground with salt it reacts easily with the natrium chloride to silver chloride. Gold on the other hand is inert so that the use of salt is more difficult to prove. In the Greek Psalter in Zrich, written in the 7th century, on some places pure natrium chloride could be analysed in the gold ink obviously the gold powder had not been washed carefully after grinding (Fuchs & Oltrogge 2007, 84). Another, more sophisticated method is the preparation of gold powder with the help of an amalgam. Gold leaf or gold filings are ground together with mercury, the amalgam is heated, the mercury evaporates and a very fine gold powder remains. Also this method was already described in the 3rd century papyri but again none of the Byzantine recipes is a direct copy from these ancient models. There is no information on the diffusion of this method in practice, to my knowledge residues of mercury have so far not been detected in any gold ink. But this may be also a question of analysis. Six recipes describe a third method to prepare gold powder: gold is mixed with sulfur and mercury, then heated at a rather low temperature. At first sight this recipe looks like a confusion of the amalgam method with the production of vermilion. The alchemical terminology of some of the recipes makes it even more conspicious. But a more detailed analysis gives sense to the instruction. The addition of sulfur to gold
66
8. The analysis was done with spectrophotometry with which it is not possible to distinguish between the lakes of the different scale insects, lac lake, Kermes, Polish or Armenian cochineals. 9. Fuchs, Oltrogge (forthcoming).
and the heating at low temperature is one possible process of cementation, a process to remove impurities from gold. Silver and copper which are often present as impurities in gold react at low temperatures with sulfur to silver respectively copper sulfide, gold does not react. The best results are obtained when the gold is already pulverised. This seems to be the reason for the addition of mercury, even if the description in the recipes is not always very clear. It seems that first a gold powder is produced with help of the amalgam and after the mercury has evaporated the gold powder is cementated with sulfur. Craddock (2000, 67) supposed that the cementation with sulfur was already known in Antiquity even if the first precise description is found in Agricola. Our six Byzantine recipes, all in 14th century manuscripts, add precious information on the diffusion of this technical knowledge.
Red lakes
About 20 recipes deal with the application of gold on a ground or with the preparation of this ground. Most often it is made from a red lake (lachas and barzion). The lachas can be produced from a material called lachas or from brazil wood (barzion). Some recipes mention, that both materials should give a red colour. The preparation of lachas from lachas is always described in a similar manner: the lachas is crushed, then heated with soda, alum and water. Afterwards the red liquid is filtered and dried. It can be cut to pieces and warmed with a little water for painting. The term lachas is derived from the hindi lakka, the name of the Indian lac insect. The red dye is embedded in a resinous matrix and cannot be extracted in pure water. On the other hand it can be easily extracted in alkaline solutions. A soda solution gives a deep red purple extract, from which a purple red lake can be precipitated with alum. This is the process described in the Byzantine recipes. We can therefore conclude, that the lachas material is the crude lac. The red lake produced with this method still contains some resins, so that the powder can be dispergated in water and painted without another binding medium. Crude lac was also known in the Latin West where it is documented as very expensive raw material for dyeing textiles and leather and for the preparation of red lakes for painting. It is possible that lachas was also a rather precious pigment in Byzantium. At least it is striking that the term lachas which originally meant a material could also be used to designate a red lake in general. The earliest example for the use of lachas in the sense of red lake is found in a 13th century recipe for a brazil lake. Brazil wood is the second base material for red lakes mentioned in Byzantine manuscripts. The colorant is extracted with glair, vinegar or red wine, and stabilized with alum. Interestingly both, the recipes for the lac lake and for the red lakes from brazil wood are never intended for illumination or other painting but always for gilding and writing. The red lake should be written as underlayer for chrysography or painted as ground for gold leaf. A red underlayer for chrysography can often be observed in Byzantine manuscripts (Mokretsova 2003, 217). It could be identified as a scale insect lake 8 for example in a patristic manuscript from the 11th century in Weimar (Herzogin Anna Amalia Bibliothek, ms. Q 741; fig. 1). 9
67
fig.1 weimar, haab, q 741, fol. 1r (11th century): the red lake used as underlayer for the gold ink in script and kephalion is made from scale insects. the yellow is lead tin yellow. (robert fuchs, cics, 2009)
fig.2 weimar, haab, q 743, fol. 39 (11th century): a red lake from scale insects is used as underlayer for gold leaf and to paint the red decorations. (robert fuchs, cics, 2009)
Grounds for gold leaf made from red lakes can be observed in initials or in the kephalia the ornamental head pieces in a number of Byzantine codices (Mokretsova 2003, 224). Certainly these red lakes could also be used in illumination, even if the recipes do not mention this. In an 11th century Gospel Book in Weimar a lake insect dye was used as rose colour in the decoration (Herzogin Anna Amalia Bibliothek, ms. Q 743; fig. 2). 10 In few cases it is not explicitly stated that gold should be laid on top of the red lake script and it is possible that the rubrication is meant. Rubrics in Byzantine manucripts are often written with vermilion but also with red lakes. There are still too few analysis to know if the one or the other material was prefered in certain periods or scriptoria. But at least in the 15th century there were scribes who used both, lakes and vermilion, for example, Isidor of Kiev in the Vatican codex gr. 914.
10. Fuchs, Oltrogge (forthcoming).
Gold grounds
Red lakes were not the only gold grounds mentioned in the recipes. There are also some other pigments, mainly red: vermilion, armenian bole, red ochre, but also yellow ochre. Again, they were thought both as underlayer for gold inks and for gold leaf. This is different from most of the Western recipes where usually coloured grounds
68
are only described for gold or silver leaf and not for metal inks. An exception is Theophilus (I, 29) who mentions a mixture of red lead and vermilion as underlayer for gold inks. This mixture on the other hand is not described in our Byzantine texts. In the Byzantine recipes only the gesso grounds for which two examples are known are restricted to gold leaf. They are coloured red-rose with either vermilion or armenian bold. Coloured grounds seem to have been typical for gilding in Byzantine manuscripts (Mokretsova 2003, 224). In the already mentioned Gospel Book in Weimar, yellow ochre is used as an underlayer for the gold leaf in the miniatures, whereas in the kephalia the gold leaf is put on top of a red scale insect lake, the same red lake is the underlayer for the gold ink of the script. Because of these differences in materials and techniques between the miniatures and the decoration of the script it is most likely that both were executed by different painters.
Pigments
Besides iron gall ink and gilding there are few recipes on proper colours, most of them dealing with the making and tempering of vermilion and with its use either as gold ground or as rubrication. As mentioned vermilion was used as one of the rubrication materials by Isidor of Kiev who also copied a recipe for grinding and tempering the pigment. Only one recipe describes the refining of lapislazuli. The pigment powder is mixed with resin, colophonium and linseed oil and than thrown into hot water. The process is repeated twice so that at the end three fractions are produced. This 14th century instruction is comparable to many Western texts even if up to now no exact parallel could be found. Lapislazuli has been proven in Byzantine illumination (Mokretsova 2003, 213), but also scribes like Isidor of Kiev wrote sometimes initials or did some decoration with it. Isidor does not transcribe a recipe for the refining. It is amazing that up to now no recipe for the making of verdigris could be found, a process already described in Greek in the 1st century by Dioskurides. Sure, there exist Byzantine copies of Dioskurides but these copies were probably done only for literary purposes. There is one recipe using verdigris, it is a translation from an Italian text, using also Italian terms like verderami, but in this case verdigris is not produced but put into vinegar to dye bones. The only Byzantine recipe to produce a green colour describes a mixture made from indigo and a yellow colorant. Besides the production of vermilion, no Byzantine recipe for any artificial pigment is known up to now. But certainly other artificial pigments were well known to Byzantine painters. In the already mentioned patristic manuscript in Weimar most of the pigments analysed are artificial: red lead, lead white, a copper green which presumably is an artificial pigment but amorphous and thus not identifiable in the X-Ray-diffraction and even lead tin yellow. This proves that in Byzantium the technique for the production of lead tin yellow was already well known in the 11th century about 150-200 years before the earliest occurences in the West. But no Byzantine recipe for this process is known.
69
Resume
It is possible to fill the gap between the late antique papyri and the postbyzantine Hermeneia at least with a small number of recipes. They concern mostly the art of writing: inks, chrysography, pigments for rubrication. The recipe are also useful to illuminators even if this is not explicitly stated. But they do not give the whole palette for illuminators and scribes. Other arts are occasionally touched. The information dates mostly in the Palaeologan period, that is the 14th and 15th century, no recipe was written before the 12th century. To the contrary to what one would expect direct copies from the papyri are extremely rare. The relationship to the Western recipes is very loose in the beginning, but in the 15th century we find translations and adaptations of Latin and Italian recipes easily to recognise because of their terminology. In some cases we can prove that the scribes were sitting in Italy as exiles from the already Turkish parts of Byzantium. Up to now no Byzantine recipe was found which could have served as a model for a translation into a Western language.
Bibliography
Bartl, Anna et al. 2005. Der Liber illuministarum aus Kloster Tegernsee. Stuttgart: Franz Steiner Verlag. Bentchev, Ivan. 2004. Griechische und bulgarische Malerbcher. Recklinghausen: Museen der Stadt Recklinghausen. Berthelot, M. 1887. Collection des Anciens Alchimistes Grecs. Paris: Georges Steinheil. Clarke, Mark. 2001. The Art of All Colours. Mediaeval Recipe Books for Painters and Illuminators. London: Archetype. Craddock, P.T. 2000. Historical Survey of Gold Refining. Ramage, Andrew, Craddock, P. (Ed.): King Croesus Gold. London: British Museum, 54-71. Eliazane, Gayane & Engel, Patricia. 2006. Ink Manufacturing Methods Used in Ancient Armenia. Buletyn Informacyjn Konserwatorw Dziel Sztuki. 17, 94-112. Fuchs, Robert & Oltrogge, Doris. 2007. Il salterio purpureo Zentralbibliothek Zrich, RP 1. Segno e testo 5, 31-98. Halleux, Robert. 1981. Les alchimistes grecs.Paris: Les Belles Lettres. Hetherington, Paul (Ed.). 1974. The Painters Manual of Dionysius of Fourna. An English Translation. Isleworth: Saggitarius Press.
70
Mokretsova, Inna et al.2003. Materials and Techniques of Byzantine Manuscripts. Moscou: Ministry of Culture of the Russian Federation. Oltrogge, Doris. 2006. Rezeptsammlungen und Traktate. Die Vermittlung kunsttechnischen Wissens im Frh- und Hochmittelalter. Stiegemann, Christoph and Wemhoff, Matthias (Ed.). Canossa. Mnchen: Hirmer, I, 555-562. Papadopoulos-Krameus, A. 1909. Denys de Fourna. Manuel diconographie chrtienne. St. Petersburg. Reprint Athen: Spanos. Schopen, Armin. 2004. Tinten und Tuschen des arabischen Mittelalters. Gttingen: Vandenhoek & Ruprecht. Schreiner, Peter (Ed.). 1988. Codices Vaticani Graeci. Cod. 867-932. Citt del Vaticano: Biblioteca Apostolica Vaticana. Schreiner, Peter. 1996. Literarische Interessen in der Palaiologenzeit anhand von Gelehrtencodices: Das Beispiel des Vaticanus gr. 914. Seibt, Werner (Ed.). Geschichte und Kultur der Palaiologenzeit. Wien: sterreichische Akademie der Wissenschaften, 205-219. Trost, Vera. 1991. Gold- und Silbertinten. Wiesbaden: Harrassowitz. Wallert, Arie. 1991. Cimature di grana: Identification of Natural Organic Colorants and Binding Media in Medieval Manuscript Illumination. Zeitschrift fr Kunsttechnologie und Konservierung 5, 74-83. Zerdoun Bat-Yehouda, Monique. 1983. Les Encres Noires au Moyen ge. Paris: ditions du CNRS.
Biography
Study of art history and archaeology in Gttingen and Bonn (PhD 1987 Die Illustrationszyklen der Histoire ancienne, 1250-1400). Since 1987 researcher for art technology and book illumination (University of Gttingen, since 1996 Cologne Institute of Conservation Science). She published widely on the painting technique and materials of medieval book illumination as well as on sources for art technology. She curates an online-database of art-technological sources of the Middle Ages and Renaissance <www.re.fh-koeln.de>. Among her recent publications are, together with Anna Bartl, Christoph Krekel and Manfred Lautenschlager Der Liber Illuministarum aus Kloster Tegernsee (2005); with Robert Fuchs Ein Meisterwerk im Wandel: Untersuchungen zur Maltechnik des Codex Aureus Epternacensis (2009). Forthcoming: Byzantinische TintenTuschen- und Farbrezepte. sterreichische Akademie der Wissenschaften, Wien 2011 (together with Peter Schreiner).
71
t he st ra s b ourg fami ly t e x t s : ori gi n al i t y and survival
Resumo
Durante muito tempo, os Manuscritos de Estrasburgo foram considerados as fontes mais antigas, em lngua germnica, para o estudo das tcnicas de iluminura no Norte da Europa. Perdidos em 1870, no fogo da biblioteca de Estrasburgo, o seu contedo tcnico s sobreviveu graas a uma transcrio feita por Sir Charles Eastlake, no sc. XIX. Vrios estudos como os de Ploss e, mais recentemente, os de Oltrogge tm destacado a existncia de textos com contedos semelhantes, agrupando-os sobre o nome de Famlia de Estrasburgo. A partir destes estudos, foi possvel definir claramente um corpo de manuscritos como pertencendo a esta famlia e nova evidncia textual tem vindo a ser descoberta. Os procedimentos e receitas artsticas descritas nestes manuscritos so, na sua maioria, dedicadas pintura e iluminura e, em especial, preparao de cores. Existe ainda um grande nmero de receitas que descrevem quais as combinaes mais adequadas de pigmentos e quais os ligantes a usar na iluminura. Graas a uma anlise filolgica e codicolgica, podemos propor, para esta famlia de manuscritos, no s uma rea geogrfica mas tambm uma data para a sua produo, entre 1400 e 1560. Atravs destes textos, temos assim, dentro de um quadro preciso, os dados para uma histria dos materiais e tcnicas. A descoberta de novos manuscritos e o seu tratamento, levou criao de uma base de dados. Inicialmente, esta centrou-se nos manuscritos escritos durante o perodo medieval, especialmente na Alemanha e nos pases limtrofes. At agora, mais de uma centena de manuscritos foram tratados e quatro mil receitas foram transcritas. Cada instruo foi codificada numa combinao especifica de ingredientes, reunidos de acordo com a ordem em que aparecem na receita. Graas a esta base de dados possvel analisar para cada ingrediente, a sua frequncia global ou a sua repetio no corpus de textos. Alm disso, podemos tambm observar para certas combinaes de ingredientes, i.e., nas receitas, qual a estrutura bsica, qual a sua frequncia no corpus e evoluo, de forma a perceber de que maneira uma receita foi sendo modificada ao longo do tempo ou por aco de certos factores externos. Uma comparao com um corpus maior de textos artsticos medievais provenientes de pases de lngua germnica permitir-nos-ia destacar a originalidade e a novidade de certos processos para a produo de cores descritos nos textos da famlia de Estrasburgo. Alm disso, tambm possvel relacionar a histria de um nmero de prescries e correlacion-las com tcnicas de mais ampla difuso.
palavras-chave
tcnicas de iluminura manuscritos receitas artsticas materiais base de dados
73
Abstract
For a long time, the Strasbourg Manuscript has been seen as one of the oldest German-language sources containing instructions on North European illuminating techniques. Lost in the 1870 Strasbourg Library fire, its technical content only survives in a nineteenth-century transcription made for Sir Charles Eastlake. Several studies like those of Ploss and, more recently, those of Oltrogge have highlighted the existence of texts with very similar contents, grouping them under the name of the Strasbourg Family. Since the appearance of these studies, a clearly defined corpus of manuscripts has been established as belonging to this family and new textual evidence has been discovered. The artistic instructions of these manuscripts are mostly dedicated to painting and illuminating and, especially, to the preparation of colours. A great number of recipes are also concerned with suitable combinations of pigments and specific binding agents used in illumination. Thanks to a philological and codicological analysis, we can propose both a geographical area in which the manuscripts of the so-called Strasbourg tradition were produced as well as a chronological range from 1400 to 1560. Through these texts, we thus have a survey of artistic materials and techniques within a precise framework. The discovery of new witness manuscripts and their treatment has necessitated the creation of a database. Initially, this database concentrated on manuscripts written during the medieval period, especially on those from Germany as well as from some border countries. Up until now, more than one hundred manuscripts have been treated and four thousand recipes have been transcribed. Each instruction has been coded in a specific association of ingredients, assembled according to the order in which they appear in the recipe process. Thanks to this database it is possible to examine for each ingredient its global frequency or repetition in the corpus of texts. Moreover, we can also observe in the associations of ingredients, i.e. recipes, the basic structure, their frequency in the corpus and their evolution, in order to perceive in which way a recipe has been modified over time or by other external phenomenon. Comparison with a larger corpus of medieval artistic texts emanating from Germanspeaking countries would allow us to highlight the originality and the newness of certain processes for the manufacture of colours described in the Strasbourg family texts. In addition, it is also possible to relate the history of a number of prescriptions to, and to correlate them with, more widely diffused techniques.
key-words
illuminating techniques manuscript artistic recipes materials database
the strasbourg family texts: originality and survival

a survey of illuminating techniques in medieval south germany
sy lv ie n ev e n
Universit de Lige, Lige (Belgium)
The Strasbourg Manuscript and its technical tradition

The evidence provided by, and the usefulness of, artistic recipe books for a better knowledge of artistic practices and materials has been debated by several authors for some decades (Ploss, 1962; Clarke, 2001; Oltrogge, 2005). This paper does not intend to present a summary of their characteristics and history, or even discuss their relevance when using them as a source for the historical study of artistic practices and materials. Instead, it will focus on a specific textual tradition of this kind of literature. For a long time, the Strasbourg Manuscript has been seen as one of the oldest German-language sources containing instructions on North European painting techniques. Its text was generally dated to the fifteenth century. However, for some specialists, such Sir Charles Eastlake, first director of the London National Gallery, the practices described may perhaps date from an earlier period (Eastlake, 1847, 126). Taking into account these characteristics, this manuscript was especially famous for containing recipes for the manufacture of oil media at a very early period. But, apart from these instructions, the manuscript is mostly dedicated to painting and illuminating and, especially, to the preparation of colours. A great number of recipes are also concerned with suitable combinations of pigments and specific binding agents used in illumination. This text is thus, at different levels, a precious witness for the
75
co lo u r i n m e d i e va l w r i t t e n s o u rc e s
illuminating practices in medieval times. Unfortunately, the manuscript was lost in the 1870 Strasbourg Library fire. However, the artistic recipes have survived in a nineteenth-century transcription made for Charles Eastlake. Since this date, several editions of the text have been published, firstly those of Eastlake, with the Material for a History of Oil Painting (Eastlake, 1847), those of Berger (Berger, 1897) and those of Borradaile (Borradaile, 1966). Like most medieval recipe books, the Strasbourg Manuscript results from compilation of older or contemporary texts. It thus shares some of its content with other books. Relevant studies like those of Emil Ploss (Ploss, 1962, 1964, 1971) and more recently those of Doris Oltrogge (Oltrogge, 2005) have highlighted the existence of texts with very similar contents, grouping them under the name of the Strasbourg Family. Since the appearance of these studies, a project has been initiated at Lige University in order to define a clearer corpus of manuscripts belonging to this family 1. For two years, new textual evidence has been discovered and the Strasbourg family currently corresponds to a corpus of sixteen manuscripts. Thanks to a philological and codicological analysis, it is now possible to propose not only a geographical area in which the manuscripts of the so-called Strasbourg tradition were produced, but also a chronological range from 1400 to 1560. The manuscripts of the Strasbourg family mainly originate from the south of Germany (Upper German). The three main dialects of this region (Franconian, Alemannic and Bavarian) can be identified in the different versions. Some of them present consistent feature errors suggesting transpositions of an Alemannic original 2. In the first instance, the opportunity of such a group would offer us not only a chance to reconstruct the text of the lost manuscript but also to compare different versions of a nearly identical text in order to visualize a structural basis, to study its recurrence and its evolution through time and the different members of the family, at each stage of the copying process.
1. This project forms part of my doctoral thesis, some results of which are presented in this paper. 2. For example, in the Amberger Malerbuch, we can observe an automatical diphthongization of alemannic monophthongs (sey instead of si(e) her) (ff. 219-220, recipe 15) or misunderstanding of south-western stat stands (south-east: stet) (f. 220, recipe 16). 3. Other arts have also been entered, such as metalwork, and the dyeing of textiles, leather, wood or bone.
The Strasbourg family database

The discovery of new witness manuscripts and their treatment has necessitated the creating of a database. Initially, it focussed on manuscripts written during the medieval period and especially on those from Germany and some border countries. More than 250 examples have been systematically recorded in a specific setting within the database, containing information about the title, the current location, place and date of origin, scribes or authors when possible, description of the other contents and additional information. As a second step, their content has been divided according to each of the recipes that they contain. Up until now, over 100 manuscripts have been treated and 4,000 recipes have been transcribed. The greatest part of these instructions relate to colours used in painting and illuminating 3. They have been recorded in a second dedicated file which is accessible from the first interface, dedicated to descriptions of the manuscripts.
76
DB_Recipes Key manuscript Recipe index Association Manuscript Key manuscript Recipe index Recipe number Title Subject Pagination Transcription Edition Adaptation Technique Ingredient Quantity Material Association Remarks
DB_Manuscripts Key manuscript Key edition1
DB_Codicology cl_ms Ms_Title Current siglum Old siglum Content Format Paleography Binding Foliation Quirings Watermarks Support quality Previous owner Previous location Decoration style Hands Marginal notes Examination date Examined by
Key manuscript Title Author Conservation Starting date Ending date Original date Original place State Language Pagination Artistic content Technique Other contents Description Remarks Edition
DB_Transition Link to DB_Recipes Link to DB_Glossary
DB_Bibliography Key edition Author Title Edition place Editor Revue Volume Pages Date Topic Keywords State Library Remarks
Link to DB_Recipes Link to DB_Glossary Transcription
DB_Glossary Key ingredient Ingredient Original language French Latin English German Dutch Synonym Denition Using Remarks
DB_Association Index_association
Index_association Global2 Count Number_association
fig.1 schema of the structure of the database
From these instructions, a complete index of ingredients has been set-up. This glossary file lists each ingredient and includes their original formulation, the actual appellation (in German, English and French) and a short description. For most of them, the glossary also mentions their scientific name. Futhermore each ingredient has been coded using an abbreviation corresponding to the first letter of its name and a consecutive number. The glossary file is linked to a list from which it is possible to select an ingredient according to its abbreviation. This technique allows to encode each recipe as a succession of abbreviations, which reflects the specific association and chronological intervention of ingredients in a given recipe. Thanks to this abbreviation technique, it is possible to retrieve the recipes linked to a specific ingredient or a specific preparation process.When creating these associations, we have borne in mind that some ingredients were not explicitly cited. This is frequently the case when the scribe makes an allusion to an ingredient prepared in a previous recipe. Next, each association has been recorded in another file of the database
77
which cross-references the number of recurrences and indicates the manuscripts where this specific association appears. The database has been cross-checked for integrity and consistency using random queries techniques. Thanks to subject classifications, queries can also be done by keywords for specific recipes, methods or materials. The global frequency and recurrence for each ingredient can be derived from the corpus of texts. Moreover, it is possible to observe, through factors such as frequency in the corpus, basic structure, and evolution, the way in which recipes were modified over time or by other external phenomenon.
Functions of the database

A survey of the materials
In parallel to scientific research and material analysis, information about the possible use of certain materials, ingredients and techniques can be obtained from the research of historical written sources. Our database enables access to this kind of information by allowing the researcher to find quickly and easily the technical instructions he needs. Searching with the database is carried out using keywords arranged in different thesauri that allow us to fine tune the result. We can thus combine the search of an ingredient, mentioned in a specific group of texts (Strasbourg Family, Heraclius, Mappae Clavicula, Theophilus,) with a specific technique (illuminating, painting, dyeing,). The database will count the number of recipes dedicated to this ingredient or involving it in a procedure. As an example, graph 1 presents pigments used in the Strasbourg family recipes that produce a red coulour. We note the frequent recurrence of cinnabar and Brazil wood mentioned in recipes dedicated to illumination and, less frequently, the use of gum lake but also Papaver Rhoeas L. species (poppies).
300
250
Number of recipes
200
150
100
50
0 Red Brazil wood Red paris Cinnabar Lake Poppies Pigments
graph 1 red colouring pigment found in the strasbourg family texts
78
The associations
Concerning associations, initial research was solely concentrated on the associations or recipes which are similar to those found in the Strasbourg Manuscript. This helped us not only to recognize the witnesses belonging to the Strasbourg family but also to demonstrate their membership. So far, 200 recipes have been found to be common to at least two manuscripts of the Strasbourg family. These similar recipes have been put in parallel as a means of comparing them.This comparative approach allows us to highlight the basic structure of certain recipes. Quite often, it corresponds to a small number of frequent ingredients associated together in a great number of recipes. As an example, one of the most repeated associations is a recipe employing copper and acetic acid contained in vinegar. This instruction usually serves to obtain a green copper pigment. This was a very well known procedure in the Middle Ages and has been detected in a great number of cases in the Strasbourg family texts. Moreover, it has been identified in a very consistent formulation,which involves an identical text. We also noted that more complex procedures, involving a certain number of ingredients, are repeated less frequently in the same form throughout the different manuscripts. Of the 254 recipes dedicated to the preparation of green copper pigment using copper and vinegar, 66 include tartar, 53 add honey and 30 involve salt. Moreover, amongst the 200 recipes common to at least two manuscripts of this tradition, many of the instructions are not isolated recipes but they are contained in different sequences. These sequences are often characterised by a more or less similar order and a specific artistic technique. These common sequences may be interpreted as the reflection of older and probably lost artistic manuals which perhaps served as a (partial) basis for the compilation of the manuscripts of the
fig. 2. diagram of the partial compilation process of manuscripts of the strasbourg family
79
Strasbourg group. When we concentrated solely on the Strasbourg Manuscript, we observed that it seems to derive partly from two distinct sequences. The first one appears to correspond to a treatise on illumination, which has been frequently copied over time and place and from which at least a small part is found in each manuscript of the family. The oldest manuscripts partly originated from this treatise on illumination and their recipes seem to respect more or less the same order. In the later manuscripts, the content has been reorganised, perhaps with a view to improving the manuscripts practical use by ordering their material by subject, media and so on but also by including new procedures dedicated to other artistic techniques. The second sequence from which the Strasbourg Manuscript may derive appears to be a more local treatise which is only otherwise found in the so-called Colmarer KunstKunst buch and the Bamberger Malerbuch and mostly dedicated to the mixing of colours and the preparation of media. Therefore, the Strasbourg Manuscript is largely the result of a text that was widely diffused and modified, as well as another more local one. These observations may be put in parallel with an examination of the recurrence or diffusion of recipe procedures within the Strasbourg family. For example, a great number of recipes are related to Brazil wood: most of them are dedicated to the preparation of this pigment; a great number are also dedicated to its application with a suitable media; others are concerned with the mixture of several pigments including Brazil wood. Graph 2 presents the relative percentage of recipes dedicated to the preparation of Brazil wood in the different texts of the Strasbourg family, arranged according to their date. We observe that the older ones contain a great number of recipes dedicated to the preparation of Brazil wood but that this number decreases more or less linearly, especially in the earliest manuscripts, except for the Colmarer Kunstbuch
90 80 70 60 50 40 30 20 10 0
) ) 35 ) 3) ) ) ) 2) 12 75 00 90 77 79 90 60 09 35 56 XV It (e N 2 nd -8 014 14 14 14 5803215 h) ) ) ) )
Relative percentage of recipes
64
50
47
47
45
50
53
5-
14
st
(1
46
(1
(1
(1
1(
(1
(1
(1
B1
B2
B3
Strasbourg Family Texts
graph 2 frequency of recipes dedicated to brazil wood in the strasbourg family texts
80
po
(c
(c
(c
(1
(B2). However, we have already noted that this manuscript is specifically dedicated to the mixing of colours and is less focussed on preparation. We have observed that the older examples are those which are closest to the Strasbourg manuscripts first sequence (fig.2). This sequence is thought to correspond to a lost treatise on illumination. It would thus seem plausible to find a greater number of recipes dedicated to preparation of Brazil wood. In the earliest examples, we have also seen that this sequence has been modified, there are fewer recipes in common and these manuscripts are also characterised by new additions to their content in the form of treatises on metalwork for example (with the exception of the Codex Palatinus Germanicus 489 H1 and the manuscript N2 which is a manuscript copy of the Illuminier Buch of Valentin Boltz von Ruffach).
Evaluating the modification of recipe process

Not all manuscripts belonging to the Strasbourg family have their entire contents in common with each other. Moreover, the degree of similarity is quite sometimes different between the examples. Thanks to the database, we can examine not only the recurrence and the diffusion of a recipe or a type of recipe within the family but also its modification through time or through the different examples of the family. When looking at the relation between the recipes of the Strasbourg texts, we can observe, on the one hand, a certain number of instructions which are exactly the same (or which perhaps differ only in the dialect or the use of some words). This implies both a similar procedure but also an identical text. On the other hand, some recipes are slightly different in terms of the modification of vocabulary, additions or suppression of information. Quite often, we observe that the change of recipe process is often due to the addition or suppression of an ingredient. So, as the recipe books evolved and were modified by adding new texts and procedures, the recipes themselves could be modified in their technical formulations during their transmission from one manuscript to another. Frequently, the copyist was free to add, to remove or to modify some ingredients and/or procedures. It is often stated that the main reason for this change is the role the text played as a technical and instructive text. It therefore seems likely that an author may, voluntarily, have corrected the text, or added information to it. However, a great deal of evidence suggests that these recipes were also assembled in book form to serve a more literary, non-practical purpose. This phenomenon could indeed be explained in several ways: it could be due to an attempt to improve or to diversify a previous formula; it could correspond to a quid pro quo, in which an unknown or expensive ingredient was replaced by a more well known one or a less expensive one; it may have been a voluntary reduction of the recipe text. In the later case, the most obvious parts of the recipe are not recorded, the copyist conserving only the essential part of the recipe (as in the case of an aide-mmoire). Furthermore, changes to the recipe may occur with a misunderstanding of the procedure. This could happen when the copyist was not a practitioner or when he was not able to translate and/or to transcribe an unreadable formula (Halleux, R., 1989).
81
Looking at the first sequence from which the Strasbourg Manuscript originated, the procedures are almost the same to the ones occuring in the older examples of the family. They seem to respect more or less the same order of recipes. In the later ones, the content has been reorganised and the recipes are more complex. If we focus more specifically on the procedure in itself within the texts of the Strasbourg family, and looking at instructions dedicated to the preparation of the Brazil wood, the database counts up to sixty five entries.
9
Recipe frequency
0
A6 A6 1 A6 C7 A6 A6 A6 A6 C7 C7 A6 W + + 1+ 1+ 2+ + + + + 1+ + + B3 2+ E2 C7
2+
C7
V4
A6
A1
2+
1+
2+
2+
B3
2+
2+
B3
V4
B3
B3
L9
L9
2+
B3
B3
2+
2+
B3
2+
2+
B3
B3
B3
B3
Basic recipes
graph 3 typical procedures for the preparation of the brazil wood and their frequency in the strasbourg texts
Graph 3 summarizes the basic procedures and their frequency in the family texts. The most repeated recipe dedicated to this procedure corresponds to an association of Brazil wood with potash lye made of ashes from different trees and alum (B32 + A12 + L9 + A6). Another frequently repeated recipe consists of the extraction of the colorant agent in Brazil wood through water and the addition of alum (B32 + W1 + A6). We find a fewer number of preparations that specify Brazil wood plus an undetermined lye (B32+L9+ A6) and, even less frequently, the use of Brazil wood and water (B32 + W1). In the Strasbourg family texts, we noted in particular two different ways in which the recipe is modified: Modification 1 (addition) Comparing the characteristic associations, we see that some basic recipes are still present as a sort of nucleus to which other ingredients may have been added. One of the simplest only consists of Brazil wood whose colorant is extracted by putting it in a lye (B32 + L9). The colour can next be fixed on alumine or alum. The addition of alum allows a beautiful red colour to be obtained (B32 + L9 + A6). Some other chemical agents can also be added in order to modify the pigments hue or charac-
82
B3
2+
1+
C7
E2
L9
L9
teristics. As an example, the addition of white lead serves to obtain a more opaque colour (B32+L9+A6+W5). Example 2 (substitution) In some other cases, ingredients have been replaced by other ones. As an example, in the later manuscripts of the family we find the typical association of Brazil wood extracted with water and fixed with alum (B32 + W1 + A6), or Brazil wood extracted in a lye and fixed with alum (B32 + L9 + A6); in the oldest ones, the lye may be replaced by urine ( B32 + U1 + A6) or vinegar (B32 + V4 + A6). Chalk can be added in order to modify the characteristics of the colour. Quite often, it is obtained from crushed egg shell. In the earliest ones, we found recipes where vinegar is mixed with beer or the lye is mixed with urine. Generally, these changes are due to technical or aesthetic reasons but sometimes, when comparing a certain number of recipes, we can conclude that the the name of an ingredient has probably been misunderstood and replaced by another substances name whose presence cannot be justified in the recipe process from a technical or aesthetic point of view. Thanks to the database, it is possible to identify the basic structure of a great number of recipes as well as study the differences and modifications appearing in the Strasbourg family texts. It is also possible to try to postulate in which ways the procedure has been modified, and, by highlighting similar or parallel recipes, to trace back the route of the artistic instructions, in a certain way, to reconstruct the life of a recipe.
Comparison with a larger group

This method makes sense in the current framework of my PhD thesis which seeks to demonstrate and to study the place and the originality of the Strasbourg Manuscript within the family. However, comparison with a larger corpus of medieval artistic texts emanating from German-speaking countries would allow me to highlight the originality and the newness of certain processes for the manufacture of colours described in the Strasbourg family texts.
Example of anthocyanin recipes

As an example, we have seen that the Strasbourg family texts convey recipes dedicated to the manufacture of red pigment or more exactly colorant obtained from poppies (graph1). The typical red colour of this species is due to the presence of anthocyanin colorant. The use of anthocyans is very frequent in the Strasbourg family texts. If we compare it to a larger group we see that more than sixty percent of the procedures including anthocyans come from the Strasbourg family texts. Graph 4 presents the proportion of recipes found in the database and employing several species of flowers or fruits for their anthocyanin colorant agent. The Strasbourg family texts are marked in red and the other encoded manuscripts are in blue. We observe the use of poppies and
83
60
50
40 Number of recipes
30
20
10
0 Papaver Rhoeas L. Centaurea Cyanus L. Vaccinium Myrtillus L. Anthocyanin species Sambucus Ebulus Sambucus Nigra
graph 4 number of recipes using anthocyanin colorant in the strasbourg family texts and in the other manuscripts contained in the database
cornflowers but also bilberry is also a lot more frequent in the Strasbourg familly texts than in other texts. Eighty percent of recipes that include poppies can be found in the Strasbourg family and more than seventy percent of recipes with cornflowers. Oppositely, the use of berries such elderberry or other Sambucus species is more frequent in the other texts.
Conclusion and perspectives

Up until now a comparative analysis has been performed for every recipe in common between the Strasbourg Manuscript and the other witnesses of the textual tradition in order to study the recurrence and evolution of the recipes. One may think that, through these texts, we have a survey of illuminating practices in a more or less clearly-defined framework. However, these manuscripts convey texts that could be qualified as living, since they have not always been the object of a simple copy but have been adapted and modified in several manuscripts. So, if their textual architecture is more or less stable over nearly one and a half centuries, the ingredients used in the recipes have been modified, and some procedures have been expanded with other ingredients. Several explanations may be suggesting for explaining these modifications. Moreover, the nature of certain variations or errors across the text can often tell us something about the author and the context of compilation. Let us take the example of substitution, which is a change in ingredients used in a recipe. On the one hand, substitution may be due to palaeographical problems that resulted in a word being misunderstood and thus being replaced by another well-known one; on the other hand, it may correspond to a deliberate technical improvement by the scribe of the recipe.
84
Finally, as illustrated through these few examples from a well known family texts, we would insist on the fact that a recipe or a manuscript should not be studied in isolation. Comparison with a larger corpus of medieval artistic texts coming from german language countries allows us to underline the originality and the newness of certain processes for the manufacture of colours described in the Strasbourg family texts. On another way, it is also possible to relate the history of number of prescriptions and to correlate them with more widely diffused techniques.
Bibliography
Berger, Ernst. 1897. Quellen und Technik der Fresko-, Oel- und Tempera Malerei des Mittelalters. Munich: Georg D.W. Callwey. Borradaile, Rosamund and Viola. (eds). 1966. The Strasburg Manuscript: A Medieval Painters Handbook Translated from the Old German. London: Tiranti. Clarke, Mark. 2001. The Art of All Colours: Mediaeval Recipe Books for Painters and Illuminators. London: Archetype Publications. Eastlake, Charles.L. 1847. Materials for a History of Oil Painting. London: Longman, Brown, Green and Longmans. Halleux, Robert. 1989. Recettes dartisan, recettes dalchimistes. Artes Mechanicae. numro 34 : 25-49. Oltrogge, Doris. 2005. The Cologne database for painting materials and reconstructions. M. Clarke, J.H. Townsend and A. Stijnman (eds), Art of the Past: Sources and Reconstructions, 9-15. London: Archetype Publications. Ploss, Emil. 1962. Ein buch von alten Farben.Technologie der Textilfarben im Mittelalter mit einem Ausblick auf die festen Farben. Heidelberg: Verlag Heinz Moos. Ploss, Emil. 1964. Ein Malerbchlein aus dem Bamberger Karmelitenkloster. 100. Bericht des Historischen Vereins fr die Pflege der Geschichte des ehemal. Frstbistums Bamberg: 331-346. Ploss, Emil. 1971. Das Amberger Malerbchlein. Zur Verwandtschaft sptmittelhoch-deutscher Farbrezepte. Festschrift Hermann Heimpel, 693703. Gttingen: Vandenhoeck und Ruprecht.
Biography
After a Masters degree in History of Art and Archaeology, I am now conducting research into historical artistic practices with a specific focus on the medieval period and, more precisely, the techniques of painting and illuminating. I am studying the written sources for art technological research and, notably artists recipe books. As part of my PhD, I studied the so-called Strasbourg Manuscript and the other witness manuscripts in relation to its texts. Dr. Sylvie Neven, Universit de Lige, PhD Student, Universit de Lige, Dpartement des sciences historiques, Histoire de lArt et Archologie, Quai Roosevelt, 1B, B 4000 Lige, Belgium, phone: +32 4 3665443, Sylvie.Neven@ulg.ac.be
85
Resumo
As cores da iluminura: receitas de Michelino de Besozzo e Antoine de Compigne. O Liber colorum do humanista francs Jean Lebgue, datado de 1431, contem, entre outros textos, receitas de cor recolhidas poucas dcadas antes por Giovanni Alcherio de Milo, algumas no Norte da Itlia e outras em Paris. A receita para a produo do azul de lpis-lazli fornecida pelo grande pintor Michelino de Besozzo, que parece ter utilizado profusamente este pigmento, o mais caro da paleta medieval, nos poucos e extraordinrios livros que iluminou em Milo e Veneza. Um conjunto de quatro receitas, com o ttulo de De coloribus ad illuminandum libros, foi ditado pelo iluminador francs Antoine de Compigne. A primeira receita destina-se a um processo de douramento, a segunda produo de laca de pau brasil, e as ltimas de dois tipos de verde de cobre. O vermelho de sinpia, o amarelo aafro e o lazurium esto igualmente contemplados. Dados biogrficos permitem-nos propor que o iluminador seja o dito Mestre de Policraticus, activo em Paris no ltimo quartel do sc. XIV.
palavras-chave
azul lpis lazli verde de cobre Jean Lebgue Mestre de Policraticuses
Abstract
The Colours of miniature paintings: recipes from Michelino da Besozzo and Antoine de Compigne. The Liber colorum of French humanist Jean Lebgue, dated 1431, contains among other texts colour recipes collected a few decades earlier by a Giovanni Alcherio from Milan, some in Northern Italy and the others in Paris. One, for the manufacturing of lapis-lazuli blue, came from the great painter Michelino da Besozzo, who seems to have made extensive use of this most expensive pigment of the mediaeval palette in the few exquisite books he illuminated in Milan and Venice. A set of four recipes, with the title of De coloribus ad illuminandum libros, was dictated by the French illuminator Antoine de Compigne. The first recipe is a gilding process, the second for brazil lake and the last ones for two kinds of copper-green, but sinopis red, saffron yellow and lazurium are also quoted. Biographical datas support an identification of the illuminator with the so-called Master of the Policraticus, active in Paris in the last third of the XIVth Century.
key-words
blue lapis-lazuli copper green Jean Lebgue Master of the Policraticus
les couleurs de lenluminure:
recettes de michelino da besozzo et dantoine de compigne

I ns Vil l el a- P e t i t
Bibliothque Nationale de France, Paris (France)
1. Une premire version de ce texte avait t prsente le 9 novembre 2006 au Sminaire sur les matriaux du livre mdival de lIRHT Institut de Recherche et dHistoire des Textes Paris. Il correspond deux chapitres de ma thse: I. Villela-Petit, La Peinture mdivale vers 1400 autour dun manuscrit de Jean Lebgue, dition du Liber colorum (Ecole nationale des Chartes, 1995), Le monde en bleu de Michelino da Besozzo et Frre Denis, Thierry de Flandre, Antoine de Compigne et quelques autres. 2. I. Villela-Petit, Imiter larc-en-ciel: la rgle des couleurs dans la Schedula diversarum artium de Thophile, Histoire de lArt, t. 39: La Couleur, octobre 1997, p. 23-36, pl. IV et V; et I. Villela-Petit, Copies, Reworkings and Renewals in Late Medieval Recipe Books, dans J. Nadolny d., Medieval Painting in Northern Europe: Techniques, Analysis, Art History, Studies in commemoration of the 70th birthday of Unn Plahter, Londres, 2006, p. 167-181. Sur les rceptaires, voir aussi: M. Clarke, The Art of All Colours: Mediaeval Recipe Books for Painters and Illuminators, Londres, 2001.
Le Liber colorum ou recueil de Jean Lebgue (Paris, BnF, lat. 6741) est un recueil de recettes techniques pour la prparation des couleurs compil au dbut du XVe sicle 1. Il inclut des textes plus anciens comme le trait de Thophile 2, mais aussi des recettes des environs de 1400. Parmi ces dernires, nous examinerons ici celles attribues lenlumineur parisien Antoine de Compigne, ainsi quune recette du peintre lombard Michelino da Besozzo, qui fut galement enlumineur. Certaines sries de recettes du Liber colorum reproduisent en effet des carnets de peintres et peuvent tre qualifies de recettes datelier 3. Quoique les modalits en soient assez variables, il faut entendre par l les recettes transcrites sur un calepin, un petit aide-mmoire ou vademecum du peintre conserv dans le fonds de latelier, au mme titre sans doute que les esquisses et les carnets de modle. De telles recettes datelier, consignes, recueillies ou recopies par le milanais Giovanni Alcherio dans les annes 1390, furent la base de la constitution de son recueil, copi par la suite par lhumaniste Jean Lebgue 4.
Le monde en bleu de Michelino da Besozzo

La recette qui suit fut obtenue Venise le mardi 4 mai 1410, de Michelino da Besozzo, le meilleur peintre du monde (Recueil de Jean Lebgue, fol. 39)... Du peintre Michelino da Besozzo nous ne possdons quune seule recette (n.117 dans le recueil), et encore celle-ci nest-elle pas trs originale puisquil sagit de lune des nombreuses recettes de purification du lapis-lazuli 5. Cest peu pour le plus excellent
87
des peintres (pictor excellentissimus inter omnes pictores mundi, selon Alcherio), que les Annales du Dme de Milan o il travailla aussi qualifient de pictor supremus, quun chroniqueur de lpoque, lhumaniste Uberto Decembrio, voit comme un nouvel Appelle 6 et que le grand historien de lArt Roberto Longhi nommait le Watteau du gothique international 7. Il faut dailleurs remarquer que lestime dont tmoigne Alcherio envers lui estime partage par ses contemporains est sans doute la cause premire qui le pousse consigner cette recette, somme toute banale, et en souligner la source illustre. Mme si la rencontre avec lartiste fut brve et peu enrichissante pour son projet de rceptaire, une indication de Michelino mritait dtre firement signale comme telle. Michelino Molinari, dont la famille tait originaire du village de Besozzo sur le lac Majeur, est dabord attest Pavie vers 1388, ce qui lui vaut le surnom de Michele da Pavia dans les Annales de la Fabrique du Dme de Milan o il est mentionn en 1404. Giovanni Alcherio le rencontra le 4 mai 1410 Venise 8 o Michelino aurait particip la dcoration du Palais des Doges et o il semble avoir sjourn quelques annes. A partir de 1420, il est nouveau cit dans les Annales de la cathdrale de Milan: Michelinus de Molinari de Besuccio, pictor supremus et magister a vitreatis 9, et est attest sur le chantier du Dme jusquen 1442. La mention dAlcherio nous renseigne donc sur une priode mi-parcours de sa longue carrire: il est alors en pleine possession de ses moyens et sa renomme sest propage de Lombardie jusquen Vntie. Sa seule uvre signe serait un peu postrieure. Il sagit dun Mariage mystique de sainte Catherine (Sienne, Pinacothque), peint sur panneau vers 1415-1420. Devant un fond dor o les auroles et les noms des saints personnages sont inscrits en relief a pastiglia 10, la Vierge drape dans un manteau dun bleu profond tient lEnfant Jsus sur ses genoux. Celui-ci se tourne vers la sainte au visage enfantin agenouille en manteau rose doubl de blanc. Derrire, un saint Jean-Baptiste en tunique bleumauve fait pendant un saint Antoine vtu de noir parements dor... A ct de cette palette sduisante, la recette de Michelino fait pauvre figure et les indications du recueil Lebgue sur sa palette sont maigres: un bleu..., du bleu et rien de plus!? Cependant, Michelino fut aussi un peintre denluminures trs recherch 11. Son uvre enlumin comprend notamment un livre dheures de la Bibliothque municipale dAvignon (ms. 111), le Sermo in exsequiis Johannis Galeatii ducis Mediolani de Pietro da Castelleto (Paris, BnF, lat. 5888), des Eptres de saint Jrme excutes Venise en 1414 (Londres, British Library, ms. Egerton 3266), ou encore un prcieux livre de prires dit tort heures Bodmer 12 (New York, Pierpont Morgan Library, ms. 944), qui nest pas comme on croit une commande vnitienne mais bien un manuscrit des Visconti. Dans ce dernier, parmi les scnes de la vie du Christ et les reprsentations des saints, un feuillet retient plus particulirement lattention. Il sagit de lillustration dune prire saint Luc, que Michelino da Besozzo a choisi de reprsenter non pas tant en vangliste, bien que son symbole soit figur ses pieds, mais en patron des peintres, ce qui en fait aussi une manire dautoportrait du peintre lui-mme (fol. 75v). Le saint y apparat en pied, vtu dun ample manteau diapr de vert et bleu, tenant un petit pinceau et mettant la dernire main un panneau
3. I. Villela-Petit, Recettes de couleurs et analyses scientifiques, esquisse dune confrontation devant luvre de Giovanni da Modena, Archives internationales dhistoire des sciences, n. 143, vol. 49, 1999, p. 269-280. Voir aussi: I. Villela-Petit, Alberto de Porcellis: une cole de calligraphie au Moyen Age, communication au Sminaire sur les matriaux du livre mdival, 16 juin 2005. 4. S.B. Tosatti, Trattati medievali di tecniche artistiche, Milan, 2007, chap. VII: Giovanni Alcherio e Jean Lebgue, Imprenditori artistici tra Milano e Parigi intorno al 1400, p. 129-148; et G. Ouy, Jean Lebgue (1368-1457), auteur, copiste et bibliophile, dans G. Croenen et P. Ainsworth d., Patrons, Authors and Workshops: Books and Book Production in Paris around 1400, Louvain, 2006, p. 143-171. 5. Texte latin et traduction sont donns en annexe. 6. I. Villela-Petit, Propositions pour Jean dArbois, dans La cration artistique en France autour de 1400, actes des XIXes rencontres de lEcole du Louvre (Paris Dijon, 7 au 10 juillet 2004), Paris, 2006, p. 315-344. 7. R. Longhi, Arte lombarda dai Visconti agli Sforza, Milan, 1958, prf. p. XXVIII; et R. Longhi, Il Tramonto della pittura medioevale nellItalia del Nord, dans Lavori in Valpadana dal Trecento al primo Cinquecento, Florence, 1973, p. 137-138. 8. Et non 1430 comme crit erronment dans le Dictionnaire des peintres, sculpteurs et graveurs de Bnzit, t.1, et lAllgemeines Lexikon der bildenden Knstler de Thieme et Becker, t.24. 9. Sur les vitraux de Michelino da Besozzo, voir le Corpus vitrearum Medii Aevi Italia, t.IV: La Lombardia, t.1: Le Vetrate del Duomo di Milano, Milan, 1986, p. 97-118. 10. Lor est pos sur une assiette fortement encolle et moule en relief. 11. L. Castelfranchi Vegas, La formazione e gli esordi di Michelino da Besozzo miniatore, Prospettiva, t. 83-84, 1996, p. 116-127; et G. Algeri, Un Boccacio pavese del 1401 e qualche nota per Michelino da Besozzo, Arte Lombarda, 1996, p. 40-51.
88
l e s c o u l e u r s d e l e n l u m i n u r e : r e c e t t e s d e m i c h e l i n o d a b e s o z z o e t d a n t o i n e d e c o m p i g n e
12. C. Eisler, The Prayer Book of Michelino da Besozzo, New York, 1981. 13. Cennino Cennini, Il libro dellarte, d. F. Frezzato, Vicence, 2004, chap. LXXVII LXXX.
de dvotion, tandis que ses godets de couleurs sont poss proximit. Le morceau de bravoure que constitue la diaprure du manteau, qui passe insensiblement dune couleur lautre, correspond ce quun autre recueil de recettes de son temps, le Libro dellArte de Cennino Cennini nomme cangianti 13. La reprsentation de saint Luc est donc pour Michelino loccasion de dployer toute la virtuosit de son art. Dans cette image emblmatique, le bleu le plus satur, pur lapis-lazuli nen pas douter, est celui dont le peintre fait le manteau de la Vierge du petit panneau. La couleur bleue et le pigment qui lui correspond y sont ainsi lhonneur. Lexemplaire officiel de lloge funbre de Jean-Galas Visconti, enlumin en 1403, est un autre exemple loquent de limportance de ce bleu. La page douverture (fol. 1) prsente une composition raffine avec un tableau, une lettrine et une bordure comprenant trente quadrilobes o alternent petits prophtes en buste tenant des inscriptions la gloire du dfunt et cus ses armes. Dans la scne principale, le duc de Milan, agenouill devant la Vierge, est couronn par lEnfant Jsus. Au-dessous, la lettrine Heu montre lauteur, laugustin Pietro da Castelleto, prononant loraison lors des funrailles. Effectivement, on y trouve du bleu; mieux mme, dans les deux cas, le bleu domine. Il est omniprsent. Dans la scne cleste, la Vierge porte sa traditionnelle robe bleue, les anges qui lentourent sont en bleu et vert, quelques Vertus aussi sont en bleu, et le fond tout bleu dcor lencre dor. Il y a ici autant de variations sur le bleu quon en peut souhaiter. Le rouge vif du manteau ducal, dun cimier et du cercle des sraphins tranche et forme comme un contrepoint la tonalit bleue gnrale, malheureusement caille par endroits. La lettrine, de mme, est toute en bleu et or. Ce bleu ou du moins une partie des diffrents bleus employs est vraisemblablement du lapis-lazuli, comme le suggrait dj la recette de Michelino. Mais on peut distinguer au moins trois tons dans la scne cleste. Le bleu de la Vierge qui trne au milieu attire lattention par sa densit. Le bleu du second cercle danges est moins soutenu, et celui qui habille les Vertus manifestement dsatur, ce qui lui donne une nuance plus terne et plus ple quaccentuent les lignes blanches marquant le pliss. Ces trois tons de bleu pourraient correspondre ceux que laissent les trois eaux successives de lavage du lapis broy et malax avec une pte qui en retient les impurets, trois eaux de lavage que mentionnent dordinaire les recettes, mme si celle de Michelino voque une gradation plus subtile, partant de la meilleure qualit de lapis, celle obtenue lors du premier lavage de la pte, puis un nombre indtermin de tons obtenus par lavages successifs en augmentant progressivement la temprature de leau (gradatim) jusqu ce que la pte ne dgorge plus de pigment du tout. Le principe cependant est toujours le mme et lon se rfrera dautres recettes du recueil Lebgue pour avoir les prcisions voulues sur ces diffrentes varits de poudre de lapis. La premire eau, et donc la portion de pigment recueillie la premire, fournit un excellent bleu, dit bleu parfait (recette n. 114) ou encore fleur du bleu (n. 111). Excellent ou parfait doivent sentendre comme le plus pur, le plus riche en pigment, donc le plus satur, ce qui correspond la fois visuellement et symboliquement au ton rserv la Vierge. La seconde et la troisime eau don-
89
nent encore un pigment de qualit, mais moins pur et de plus en plus ml de ces impurets qui se rencontrent ncessairement dans la roche, des traces blanchtres de calcite notamment 14. Le bleu ainsi produit est alors moins vif et ml de blanc, donc ncessairement moins bon pour une poque et une esthtique qui valorisent puret et richesse du matriau, densit, vivacit et saturation de la couleur 15. En labsence danalyses physico-chimiques sur le manuscrit, on ne peut cependant exclure lemploi dautres pigments bleus (indigo, azurite) pour obtenir les valeurs souhaites, valeurs qui sont une faon de crer une hirarchie dans limage entre centre et priphrie, et entre les personnages reprsents formant cette pyramide cleste o trne la Madone. Dautre part, la plus ou moins grande puret du pigment, en loccurrence du lapis-lazuli, nest pas le seul facteur de modulation. Le choix et la proportion du liant 16 (gomme arabique, blanc duf, colle liquide ou huile de lin, comme lindique le n. 118), dventuels ajouts (dautres bleus ou de blanc), le nombre et la nature des sous-couches (le lapis ntant pas trs couvrant), les glacis, voire la simple proximit dautres couleurs peuvent jouer et modifier notablement la rsultante. Il est toujours intressant dobserver les rapports quentretiennent les couleurs entre elles dans une image donne. Ici bleu virginal et rouge de royaut se compltent et se rehaussent lun lautre, tandis quon remarque une quivalence entre le bleu dlav et le rose, le violac tendre, le vert ou le mordor des robes des Vertus qui forment la cour cleste, comme des demoiselles dhonneur de la Vierge. Ces tons doux marquent dans la hirarchie un degr en dessous du couple bleu-rouge satur (lapis et vermillon?). Mais encore, bleu et vert sont associs, soit que le bleu la robe de Marie se double de vert avec un liser dor, soit quil ait t ombr de glacis verts comme les armures bleues des anges placs de profil ou de trois-quart, dans leur ct perdu, cach aux yeux du spectateur suppos se trouver en face et en contrebas de ce monde divin. Le vert (un vert de cuivre sans doute) fait donc office de bleu dsatur ou de seconde catgorie, rserv aux doublures des vtements, mais aussi dombre du bleu, et en complte lusage. Dans la lettrine cependant, la nuance dominante nest pas la mme. Le bleu en est trs sombre et quelque peu us aussi, rendant sa lecture plus difficile. Hormis les visages et les mains des personnages, ce bleu-noir est la seule couleur de la scne funbre, tandis que lencadrement est dun bleu plus clair avec pigraphie, lui-mme entour dor bruni. Les figures sont vtues du ton le plus sombre, peut-tre un vritable noir, couleur de deuil. Le rideau du fond est un ton en dessous, bleu nuit. Et dans le toit en btire rsonne un simple bleu pur. On discerne donc une opposition colore entre monde terrestre o prend place la scne du deuil et monde cleste au bleu vif o le dfunt est accueilli. Deux valeurs symboliques sont donc possibles pour la couleur: bleu-noir de la souffrance terrestre lapis, indigo, ou quelquautre pigment ml de noir de charbon?, et bleu satur de la joie cleste pur lapislazuli? Cest--dire, un couple bleu-nuit, bleu-jour; bleu triste, bleu joyeux; bleu teint, bleu clatant, qui lie les deux scnes: mort et rsurrection bleues. En sus dune prfrence du peintre et du got gnral du gothique international pour lemploi du lapis-lazuli, le choix si manifeste de cette couleur dans le Sermo in
14. J. Plesters, Ultramarine Blue, Natural, dans A. Roy d., Artists Pigments, A Handbook of their History and Characteristics, vol. 2, Washington, 1997, p. 37-54. 15. M. Pastoureau, Les couleurs mdivales: systmes de valeurs et modes de sensibilit, dans Figures et couleurs, Paris, 1986, p. 35-49, la p.38: cela est caractristique de la sensibilit mdivale: une belle couleur est une couleur franche, lumineuse et sature; et M. Pastoureau, Bleu, histoire dune couleur, Paris, 2000. 16. Cl. Coupry et M.-T. Gousset, Les manuscrits mdivaux vus par laser La Recherche, t. 205, 1988, p. 1524-1526: Les spectres Raman des chantillons bleu clair [de manuscrits de Corbie du XIIe sicle o le seul bleu rencontr est le lapis] nindiquent pas la prsence de pigment blanc (cruse, craie, gypse, kaolin). La technique utilise pour ces variations de teinte est donc la dilution plus ou moins grande du bleu dans le liant et non ladjonction dun autre pigment.
michelino da besozzo, mariage mystique de sainte catherine (vers 1420), pinacothque de sienne
90
17. Sans doute aprs le conclave qui vit llection de lantipape Jean XXIII (17 mai 1410 29 mai 1415).
michelino da besozzo, livre de prires bodmer, new york, pierpont morgan library, ms. 944, fol. 75v : saint luc
exsequiis a bien sr des raisons symboliques. Cest dabord la couleur de la Vierge qui limage est en quelque sorte ddie et sous la protection de laquelle vient se placer le duc. Cest aussi la couleur du ciel et, assimile au noir, de la mort, qui sont les sujets mmes de lenluminure. Le bleu est encore une des couleurs hraldiques des Visconti de Milan, dont le blason (dargent la guivre dazur hissant de gueules) et les pennons, alterns avec ceux dautres possessions de Jean-Galas (duchs de Pavie, Vertus, comts de Bologne, Pise, Sienne et Prouse), parsment le champ de limage. Le bleu de la guivre se sera, pour ainsi dire, propag emphatiquement jusqu constituer le ton dominant de limage. Enfin, le faste dploy dans ce luxueux manuscrit la gloire du duc et de son lignage justifiait lemploi des matriaux les plus prcieux, parmi lesquels le lapis. Le bleu de lapis-lazuli rpondait donc au mieux toutes ces modalits. Sil peut paratre hasardeux didentifier un pigment lil nu, la ralisation dun manuscrit de cette qualit, confie lun des meilleurs peintres de son temps, requrait en tout cas des pigments de prix. Le lapis-lazuli purifi par le procd des trois lavages en fait partie: le premier vaut son pesant dor, le second son poids dargent et le troisime est bon pour les sous-couches, dit encore une recette (n. 349). La question serait plutt de savoir sil est le seul bleu employ. Nous avons relev plusieurs tons. Sont-ils obtenus avec des poudres de diffrentes qualits? en variant les proportions de liant et de poudre? en mlant le lapis et dautres pigments, blancs ou noirs? en employant le lapis-lazuli pour les parties principales (la Vierge) concurremment avec dautres pigments bleus (Vertus, fond)? On ne saurait extrapoler partir dune seule recette que ctait-l le seul bleu de Michelino, mme si la vogue de la couleur au gothique international et le renom du peintre pouvaient amener celui-ci lemployer plus quun autre. Nous avons simplement, recueillie par Alcherio, du meilleur des peintres la plus prise de ses couleurs.
Antoine de Compigne, enlumineur Paris

Le jeudi 8 aot de lan de la Circoncision de notre Seigneur Jsus Christ 1398, Giovanni Alcherio a crit et recopi Paris, chez lenlumineur Antoine de Compigne, homme dun grand ge, daprs les paroles de celui-ci qui toute sa vie durant, comme il a dit, avait prouv lensemble des recettes qui suivent, savoir les recettes pour lenluminure. Et par la suite, en dcembre 1411, le mme Giovanni revenu de Lombardie cest--dire de Bologne o la Curie Apostolique tait nouvellement assemble 17 Paris depuis plus dun an dj, les corrigea en plusieurs endroits suivant plusieurs informations quil avait reues depuis de plusieurs livres authentiques portant sur le sujet et par dautres moyens, et mit au net ce qui suit. (Recueil de Jean Lebgue, fol. 87) Le jeudi 8 aot 1398, quelque dix ans avant sa rencontre avec Michelino Venise, Giovanni Alcherio, sjournant alors Paris, avait rendu visite lenlumineur Antoine de Compigne in domo suo. Dj g, antiquus homo, celui-ci rsidait au quartier
michelino da besozzo, frontispice de leloge funbre de jean galas visconti par pietro da castelleto (1403), paris, bnf, lat. 5888, fol. 1
91
Saint-Sverin, quartier de nombreux peintres o il devait avoir aussi son atelier, et peut-tre maison et atelier ne faisaient-ils quun. Une inscription lapidaire autrefois en lglise Saint-Sverin confirme la ralit historique du personnage, autrement connu essentiellement travers le recueil Lebgue. Lglise de cette vaste paroisse de la rive gauche tait en chantier depuis le milieu du XIIIe sicle. En 1347, le pape Clment VI avait dailleurs accord des indulgences aux gnreux donateurs qui aideraient sa reconstruction. Or, le second pilier du premier bas ct sud de la nef portait encore avant la Rvolution une inscription en cuivre indiquant quil avait t lev en 1414 aux frais de la succession dAntoine de Compaigne, enlumineur, et de sa femme Odette 18. Le texte en tait le suivant: Les excuteurs de feux Antoine de Compigne [ou Compaigne], enlumineur de pincel, et de Oudette sa femme ont fait faire ce pilier du rsidu des biens desdits dfunts lan M. CCCC. XIIII 19. Lenlumineur devait donc jouir dune certaine aisance matrielle la fin de sa vie. Il dcda une dizaine dannes aprs la visite dAlcherio qui remarquait dj son ge avanc (1414, terminus ante quo), et fut probablement enterr dans lancien cimetire de lglise, actuel jardin du presbytre. Les dbuts de sa carrire remontent en effet au milieu du XIVe sicle et il est dj attest trente ans avant cette rencontre, puisquil figure, mais comme libraire, au nombre des artisans du mtier du livre qui, en tant que jurs de lUniversit de Paris, furent exempts du guet par un privilge du roi Charles V en date du 5 novembre 1368 20. Antoine stait ensuite spcialis dans lenluminure: illuminator librorum, nous dit la notice du recueil Lebgue (n. 297). Nayant pas de carnet de recettes, cest par oral quil transmit son savoir-faire au visiteur. Il sagit de quatre recettes prises sous la dicte et compltes par la suite par Alcherio qui les intitule De coloribus ad illuminandum libros: or bruni, rose de bois brsil et deux verts de cuivre, constituant une palette des plus rduites mais dautant plus intressante pour nous que les circonstances dans lesquelles ces recettes furent recueillies attestent que ces pigments-l taient encore fabriqus dans latelier alors que dautres, la cruse par exemple, sachetaient chez lapothicaire. Dans la premire (n. 298), outre le parchemin et le papier, supports habituels de lenluminure, sont aussi mentionns les panneaux de bois blanchis la craie, car le procd (ici la pose de la feuille dor) est le mme. De telles planchettes apprtes, runies en carnet par des lanires, servaient souvent de carnet dessin. Quelques-uns se sont conservs tel le carnet de buis de Jacquemart de Hesdin 21 (New York, Pierpont Morgan Library, M.346), dat vers 1385-1400. Une assiette rigide compose de craie, docre et de blanc duf, suivant la mthode traditionnelle, permet la feuille dor quon y pose de supporter le brunissage. La recette de rose 22 (n. 299) indique les mmes supports (parchemin, papier, panneaux de bois apprts). La laque de brsil fixe sur une base de craie et dalun est employe avec de la gomme arabique. Elle peut servir en peinture et pour dessiner, mais aussi comme encre rose. Suivent deux verts tous deux faits base de vert de gris, bien quaux proprits inverses. Lun (n. 300) est transparent, a peu de substance, sans corps dit le texte, et ne doit servir quen une sorte de glacis. Il reste terne, obfuscatus, et les couleurs sur lesquelles il serait pos transparatraient. Mais le principal dfaut de ce vert est
18. Daprs A. Boinet, Les difices religieux mdivaux et Renaissance, Paris, 1910 (collec. Les richesses dart de la ville de Paris). 19. Abb Lebeuf, Histoire de la ville de Paris, t.1, 1883, p. 102, cit par H. Martin, Les miniaturistes franais, Paris, 1906, p. 212. Je remercie Monsieur Franois Avril de mavoir signal cette rfrence. 20. P. De Winter, The Grandes Heures of Philip the Bold, Duke of Burgundy: The Copyist Jean LAvenant and his Patrons at the French Court, Speculum, 1982, t. 57, n. 4, p. 786-842. 21. Ph. Lorentz, Les carnets de dessins, laboratoires de la cration artistique, dans Paris 1400 Les arts sous Charles VI (1380-1422), Paris, Muse du Louvre, 2004, p. 304-306. 22. I. Villela-Petit, Brsil et autres rouges: dix recettes de laque mdivales, Techn Revue du Laboratoire de Recherche des Muses de France, n. 4: La couleur et ses pigments, 1996, p.68-73; P. Roger, S. Vandroy et I. Villela-Petit, Les laques de brsil dans lenluminure mdivale: reconstitution partir de recettes anciennes, Studies in Conservation, t. 48, n. 3, 2003, p. 155-170.
92
23. G. Banik, Green copper pigments and their alteration in manuscripts or works of graphic art, dans B. Guineau d., Pigments et colorants de lAntiquit et du Moyen Age, Paris, 1990, p.89-102; H. Khn, Verdigris, dans A. Roy d., Artists Pigments..., p. 131-147. 24. I. Villela-Petit, Histori de blanc et de noir: la tradition du portrait dencre dans lenluminure parisienne des XIVe-XVe sicle, paratre dans Les rapports des arts monochromes la couleur, actes du colloque de juin 2009 au Centre dEtudes Suprieures de la Renaissance de Tours. 25. I. Villela-Petit, Palettes compares: quelques rflexions sur les pigments employs par les enlumineurs parisiens au dbut du XVe sicle, dans M. Hofmann et C. Zhl dir., Quand la peinture tait dans les livres: Mlanges en lhonneur de Franois Avril, Berlin, 2007, p. 382-391.
matre du policratique de charles v, chronique de bernard gui (aprs 1384), besanon, bm, ms. 677, fol. 13 : le songe de nabuchodonosor
dtre trs corrosif 23 et de gter les autres pigments. Les superpositions sont donc viter, y compris lhabituelle prparation blanche. Ce vert particulier sans autre liant que le vinaigre semployait mme le parchemin ou le papier pour colorier des formes pralablement traces lencre noire. La recette concerne donc uniquement lenluminure et relve dune technique moins coteuse et moins prcieuse que la peinture en pleines couleurs, celle du dessin colori. Dite dessin dencre, cette technique fut couramment employe par les enlumineurs parisiens de la fin du XIVe sicle 24, ainsi dans lexemplaire de ddicace de LApparition de Jean de Meun dHonor Bovet destin vers 1398 la duchesse dOrlans, Valentine Visconti (Paris, BnF, fr. 811). Les couleurs y sont en nombre trs rduit: un lavis dencre pour les draps, quelques rehauts de rose pour les carnations, dor pour les couronnes et, pour le sol, un vert ou un jaune quasi sans substance, sans densit, comme ce vert n. 300. Rose, vert et or, ce sont l les couleurs du De coloribus plac sous le nom dAntoine de Compigne. Lautre vert, tout au contraire, est dulcis et corpulentus (n. 301). Il peut donc servir aussi bien sur papier et parchemin que sur toile ou panneau de bois blanchi, car il ne contient pas de vinaigre, et le vert de gris est cens tre tempr par les sucs dherbes ou de fleurs qui entrent dans sa composition. Li avec une eau de gomme arabique soigneusement purifie, il sert dencre ou de peinture, semploie aussi bien en glacis quen sous-couche, et lon peut galement poser lor par dessus. Ainsi passe-t-on ce vert en rehauts sur une couche de vermillon, de lapis-lazuli, de rose brsil ou toute autre, ce qui nest dailleurs pas sans rappeler la technique des cangianti quaffectionnait Michelino da Besozzo. Toutes ces recettes sont assez prcises dans lindication des proportions respecter, de la dure, de la consistance (par exemple, au n. 298, la consistance rechercher pour lassiette de lor ou putrefactio), de leffet visuel (transparence, opacit) et des incompatibilits entre pigments (n. 300). Les deux dernires en particulier contiennent un prambule relatif la nature, aux proprits, lusage, aux dfauts et aux qualits respectives des verts de cuivre. Sagit-il dun topos, dajouts dus au compilateur Giovanni Alcherio daprs des ouvrages de rfrence (libri autentiqui), ou du reflet de lexprience authentique du peintre? Les trois la fois peut-tre. La palette dAntoine de Compigne peut stendre aux autres couleurs cites: or bruni, rouge vermillon (ici sous le nom de sinopis), rose de brsil, bleu de lapis-lazuli (lazurium), vert-de-gris inconsistant, vert-de-gris consistant et jaune safran, soit sept couleurs auxquelles il faut encore ajouter le blanc de cruse 25. Et, bien quincomplet, ce De coloribus rdig sur les indications de lenlumineur est moins pauvre quil y parat puisquil fait allusion aux superpositions de teintes, technique courante mais rarement cite dans les recettes puisquelle relve dj de la phase de mise en uvre du pigment. Une couche de vert viendra se superposer une couche de vermillon, modulera une sous-couche bleue, ombrera un rose... Autant de nouvelles nuances dont il faut crditer lexprience du matre et qui enrichissent sa palette. Jusqu il y a peu, on ne pouvait gure en dire davantage sur la manire dAntoine de Compigne, aucun manuscrit ne lui tant attribu. Tout au plus pouvait-on le compter au nombre des enlumineurs parisiens de la seconde moiti du XIVe sicle (je
93
le rapprochais en 1995 du groupe aux Boqueteaux). Jai cependant propos en 2004, sans avoir alors le loisir dtayer mon propos 26, de reconnatre en lui un enlumineur de ce cercle baptis Matre du Policratique de Charles V daprs le manuscrit BnF, fr. 24287. La carrire de lartiste, reconstitue par Franois Avril 27, prsente en effet des concidences troublantes avec la biographie dAntoine de Compigne. On peut mettre lune et lautre en parallle (voir le tableau ci-aprs). Bien que fragmentaires, les donnes disponibles concordent et permettent mon sens didentifier notre artiste avec le Matre du Policratique. Les nombreux manuscrits enlumins par celui-ci illustrent au mieux la mise en uvre des recettes du petit De coloribus ad illuminandum libros inclus dans le Recueil Lebgue, notamment cette technique du dessin dencre rehauts de lavis qui semble tre la destination ordinaire du vert translucide dAntoine de Compigne.
26. I. Villela-Petit, Matres peintres et enlumineurs: identits incertaines, dans Paris 1400..., p. 203. 27. Fr. Avril, Le parcours exemplaire dun enlumineur parisien la fin du 14e sicle: la carrire et luvre du matre du Policratique de Charles V, dans B. Fleith et F. Morenzoni d., De la saintet lhagiographie. Gense et usage de la Lgende Dore, Genve, 2001, p. 265-282.
94
Antoine de Compigne
Matre du Policratique
Attest partir de 1368 entre 1398 et 1414
En activit de 1366 1403
Libraire-jur de lUniversit de Paris (1368)
Enlumine un Dcret pour lUniversit de Paris
Demeure au quartier Saint-Sverin, comme lenlumineur Perrin Rmiet (et le libraire Regnault du Montet)
Reoit commande des abbayes voisines, Sainte-Genevive (1380) et Saint-Victor (1392); collabore avec Perrin Rmiet
Un des enlumineurs parisiens les plus considrs la fin de sa vie, rang aux cts des plus grands artistes dans le recueil Lebgue
Travaux denluminure pour Charles V, Louis dAnjou, Philippe le Hardi, Jean de Berry et Louis dOrlans
Reoit la visite du milanais Giovanni Alcherio en 1398
Travaux pour Pasquino Capelli, conseiller du duc de Milan (vers 1390-1395) et pour Valentine Visconti (1398)
Encore actif la fin du XIVe sicle et fait alors rfrence
Un des rares enlumineurs importants actifs Paris cette poque o les matres pucelliens ne sont plus et ceux de la gnration suivante dbutent peine
Assez fortun pour offrir un pilier lglise Saint-Sverin sur sa succession
Un atelier trs productif, dont on peut supposer quil tait rentable
Antiquus homo en 1398
Signes de snescence de son style vers 1400
Palette: vert pour le dessin dencre, safran, azur, encre rose, or, cruse, ocre, vermillon; assiettes, superpositions
Dessin dencre, enluminures en pleines couleurs, bordures; verts, jaune diaphane (safran?), lapis-lazuli, rose brsil, or, vermillon, bleu-gris, etc.
95
Recette de Michelino da Besozzo

Item in eodem exemplari sic erat scriptum: Hoc sequens experimentum hujusmodi, in Veneciis, die martis IIII maii, anni 1410, a Michelino de Vesucio, pictore excellentissimo inter omnes pictores mundi. De mme se trouvait dans le modle la mention suivante: La recette qui suit fut obtenue Venise le mardi 4 mai 1410, de Michelino da Besozzo, le meilleur peintre du monde. n. 117 Azurium sic fit: Recipe libram unam lapidis lazuli et tere bene in lapide porfirico. Postea ablue ipsum cum aqua clara, deinde desica et reduce ipsum in pulverem. Pastillum sic fit: ad libram unam pulveris lapidis, pone libram unam picis Grece, oncias II vernicis liquide, onciam I masticis; ponantur in olla rudi oncias III olei communis, idest lini vel olive, et boni, et fac bullire, et tunc masticem et vernicem pulverizatam pone in oleo et bene moveas cum ligno. Et cum videas resoluta, pone piscem pulverizatam et permitte parum bullire donec omnia fuerint bene incorporata. Postea cola per pannum in aqua frigida et mancetur manibus unctis oleo communi, et postea pulver lazulli incorporetur super lapidem cum dicto pastillo et optime, et dimittatur per tres dies in dicto pastillo. Postea extrahatur azurrum de pastillo hoc modo: misceatur cum baculo in aqua calida parum plusquam tepida, et taliter teneatur quousque aliquid exiverit. Si vero non exiret, ponatur aqua magis calida, et sic gradatim, mittendo aquam calidiorem et miscendo donec aliquid exiverit. Ultimo ponatur aqua quando magis fervet, et extracto toto azurro et separato ab aqua et sicato, fiat lexivium fortissimum et ponatur azurrum in planis vasis et superius ponatur lexivium, sicut nosti, ut exeant immondicii pastilli, quo purgato, dulcifica cum aqua clara, etc. 117 Le bleu se prpare ainsi: Prenez une livre de lapis-lazuli et broyez-le bien sur une meule de porphyre. Puis lavez-le deau claire, faites scher et rduisez en poudre. La pte se prpare ainsi: pour une livre de poudre de lapis, prenez une livre de poix grecque, deux onces de vernis liquide, une once de mastic; mettez dans une marmite neuve trois onces de bonne huile commune, cest--dire de lin ou dolive et faites bouillir, ajoutez alors dans lhuile le mastic et le vernis en poudre et remuez bien avec un bout de bois. Lorsque vous le verrez dissout, ajoutez la poix en poudre et laissez lgrement bouillir jusqu ce que tous les ingrdients soient bien incorpors. Puis filtrez dans leau froide travers un tissu et malaxez avec les mains ointes dhuile commune, puis incorporez soigneusement sur la meule la poudre de lapis-lazuli dans cette pte et ly laissez reposer trois jours. Le bleu sextrait ensuite de la pte de la faon suivante: mlangez la pte avec un bton dans de leau plus chaude que tide et continuez ainsi tant que de la couleur en sortira. Mais sil nen sort pas, mettez de leau un peu plus chaude et ainsi par degrs, ajoutant de leau toujours plus chaude et mlangeant jusqu ce que la couleur sorte. En dernier lieu, ajoutez de leau bouillante, puis une fois le bleu entirement extrait, spar de leau et sch, faites une lessive bien forte, mettez le bleu dans des rcipients fond plat et versez dessus la lessive, comme vous savez, pour que sortent les salets de la pte amollie par leau claire dont vous laviez purg, etc.
96
Recettes dAntoine de Compigne

De diversis coloribus in sequenti tractatur, et primo, modus prohemii: n. 297 Anno circoncisionis domini Jesu Christi 1398 die Jovis octavo augusti, Johannes Alcerius scripsit et notavit in Parisius in domo Anthonii de Compendio illuminatoris librorum, antiqui hominis, a verbis que ipse Anthonius sibi dixit, et qui omnia que secuntur tentaverat toto tempore vite sue, ut dixit, de coloribus scilicet ad illuminandum libros, sequencia capitula. Et postea anno 1411 de mense decembris, idem Johannes qui jam per plusquam annum reversus fuerat a partibus Lombardie, videlicet a Bononia, ubi erat curia apostolica noviter unita, correxit in pluribus partibus ea, secundum plures informationes quas inde postea per plures libros autentiquos de talibus narrantes, et aliter habuerat, et rescripsit ea ad nettum ut sequitur. Trait de diverses couleurs, et dabord, en guise dintroduction: 297 Le jeudi 8 aot de lan de la Circoncision de notre Seigneur Jsus Christ 1398, Giovanni Alcherio a crit et recopi Paris, chez lenlumineur Antoine de Compigne, homme dun grand ge, daprs les paroles de celui-ci qui, toute sa vie durant, comme il a dit, avait prouv lensemble des recettes qui suivent, savoir les recettes pour lenluminure. Et par la suite, en dcembre 1411, le mme Giovanni revenu de Lombardie cest--dire de Bologne o la Curie Apostolique tait nouvellement assemble Paris depuis plus dun an dj, les corrigea en plusieurs endroits suivant plusieurs informations quil avait reues entre-temps de plusieurs livres authentiques portant sur le sujet et par dautres moyens, et mit au net ce qui suit. n. 298 Ad ponendum aurum super diversis quod burniatur, et de diversis cautelis utendis super hoc illuminando libros: Ad ponendum aurum in papiro, in pergameno seu carta, et in tabulis ligneis creta alba dealbatis quod aurum burniatur seu poliatur, accipe gersam seu cretam albam et modicum ocre de ru, per tertiam partem quantitatis crete et totum simul subtilia, et tere cum aqua clara magis spissum quam poteris, idest cum pauca aqua, super lapidem equalem durum cum molleta lapidis similiter. Postea pone ipsum colorem, qui aliter tempera vel assisia auri dicitur, in conchilla aut in scutella figuli vitriata aut in vase vitri. Et cum operari vis, accipe de ipso in conchilla alia parviori quantum vis et modera ipsum cum claro ovi spongiato ad rationabilem molliciem seu liquiditatem pro pingendo aut scribendo de ipso. Et si habes tempus cum temperaveris, dimittas inveterari per plures dies vel septimanas ipsam temperam, quia melior erit putrida quam recens. Postea de ipsa scribe, pinge et protrahe que vis et ubi vis, et dimittas siccari. Postea sis in loco recluso cum aurum vis ponere et elige tempus ydoneum ut supra dictum est. Et habitis ydoneis loco et tempore et remediis, ponas aurum in locis carte vel papiri quibus ipsum colorem vel assisiam posuisti, et super trahe et premendo primo leviter, postea fortius burnissorem, scilicet dentem apri vel equi, et polias tantum dictum aurum quod adereat colori et lucidum fiat ut supra jam dictum est. Ideo cum aurum poni vult, color talis remansus de alia positione auri alias facta melior est, dum ex interpollata visitatione deductione et ovi aut aque interpositione conservatus sit in debita liquiditate, ita quod ad totalem siccitatem vel nimiam putrefactionem et alterationem deductus non sit.
97
298 Pour poser de lor brunir sur diverses matires, et des nombreuses prcautions prendre pour enluminer les livres: Pour poser de lor sur papier, parchemin et panneaux de bois blanchis la craie, de manire que lor soit bruni ou poli: prenez de la craie blanche et un peu docre de ru pour environ le tiers de la quantit de craie, broyez le tout en poudre fine avec de leau claire le plus densment que vous pourrez, cest--dire avec peu deau, sur une meule de pierre dure et plane au moyen dun pilon galement de pierre. Puis mettez cette couleur autrement appele assiette de lor dans une coquille ou une cuelle dargile vernisse, ou dans un rcipient de verre. Et quand vous voudrez vous en servir, versez-en volont dans une plus petite coquille et dlayez de blanc duf battu jusqu la consistance ou la fluidit requises pour en peindre ou en crire. Et si vous avez le temps, lorsque vous laurez dtrempe, laissez cette assiette vieillir plusieurs jours ou semaines, parce quelle est meilleure gte que frache. Puis crivez, peignez et dessinez-en ce que vous voudrez et o vous voudrez, et laissez scher. Soyez en un endroit ferm lorsque vous voudrez poser lor et choisissez un temps favorable, comme il est dit plus haut. Et une fois runis le lieu, le temps et les moyens adquats, posez lor aux endroits du parchemin ou du papier o vous aurez mis de la couleur ou assiette, frottez dessus en appuyant dabord lgrement puis plus fort le brunissoir, cest--dire la dent de sanglier ou de cheval, et polissez suffisamment lor pour quil adhre lassiette et quil brille, comme il a dj t dit. Cest pourquoi, lorsque vous voulez poser lor, lassiette restant dune prcdente dorure est meilleure, pourvu quelle ait t conserve la fluidit voulue par une surveillance renouvele pour la remuer et ladditionner duf ou deau, de manire quelle ne soit amene ni se desscher compltement, ni trop se gter et saltrer. n. 299 Ad faciendum rosam: Ad faciendum rosam pro operando in carta et in papiro et in ligneis tabulis creta dealbatis, accipe brixillium rasum subtiliter cum cutello vel cum vitro et liga in subtili pecia lini non stricte, sed late et fluctuanter. Et sic ligatum, pone in vase figuli vitriato novo ad temperandum in lixivio aut in urina hominis ebriatoris potantis forte vinum, et si urina sit vetera tanto melius. Et si non possis habere talem, accipe lessivium fortissimum et pone de creta alba in ipso lessivio, cum dicta pecia in qua est braxillium et per quantitatem de tribus vel quatuor vicibus quantitatis brixilii ad pondus et etiam sicut inspiciendo melius videbis convenire plus et minus secundum bonitatem brixillii. Postea pone de alumine glacie crudo, pisto in pulverem, quod sit tantum quantum est quartum dicte crete vel circa, et ante plus quam minus, et misceas hec omnia insimul dimittendo semper ligatum in dicta pecia dictum brixillium et dimittant sic per horam unam vel circa. Postea ponas vas ad ignem non lignorum sed carbonum et bulliant non nimis fortiter et per spacium quarte partis hore vel minus, ita quod solum alumen fondatur. Postea de ipso vase tollatur dicta pecia brixillii et exprimatur et extorqueatur fortiter ut color de ipsa totaliter exeat in eodem vase. Postea tollatur ipse color ita calidius ab igne et ponatur super lapidem crete concave vel super laterem de terra etc. ad hoc, quod urina seu lessivium intret in lapidem subito et color ipse remaneat ibi inspissatus et semisiccus. Postea facias ex toto siccari ad solem, deinde eleva ipsum colorem, que rosa est cum cutello a lapide vel latere, et repone servando pro usu.
98
Et cum de ipsa operari vis, accipe de ipsa quantum vis et subtilia, idest tere super lapidem durum et planum cum aqua gummata que fit per duas partes gummi arabici fusi in tam pauca aqua, quod pene coperiatur ipsa gumma cum in ipsa ponitur aqua, et colati postea per telam lineam, et per tertiam partem sit aqua clara insimul cum dicto gummi fuso et colato. Et de ipsa aqua gommata ipso modo facta distempera dictam rosam ad debitam molliciem et operaberis de ipsa que volueris, tam scribendo quam pingendo ac protrahendo. 299 Pour faire du rose: Pour faire un rose employer sur parchemin, papier et panneaux de bois blanchis la craie, prenez du brsil finement rp avec un couteau ou avec du verre, et lier-le dans un morceau de lin sans serrer, mais de manire lche et flottante. Et ainsi attach, mettez-le dans un rcipient neuf dargile vernisse tremper dans de la lessive ou de lurine divre amateur de vin fort, et tant mieux si lurine est vieille. Mais si vous ne pouvez en trouver de cette sorte, prenez une lessive trs forte et mettez-y de la craie blanche avec le tissu contenant le brsil, soit en poids une quantit trois quatre fois suprieure celle de brsil ou plus ou moins selon ce qui vue dil semblera mieux convenir et suivant la qualit du brsil. Puis ajoutez de lalun de glace cru et rduit en poudre, qui fasse autant ou peu prs que le quart de la craie, et plutt plus que moins, mlangez tout cela ensemble en gardant toujours le brsil attach dans son tissu et quils reposent ainsi environ une heure. Puis mettez le rcipient au feu non de bois mais de charbons et quils ne bouillent pas trop fort pendant un quart dheure ou moins de manire ne faire fondre que lalun. Otez ensuite du rcipient le tissu contenant le brsil, pressez-le et tordez-le avec force pour en faire compltement sortir la couleur; puis retirez du feu le rcipient avec la couleur ainsi chauffe et versez-la sur un morceau de craie creux ou sur une brique de terre, etc. de sorte que lurine ou la lessive soit aussitt absorbe par la pierre et que la couleur y reste compacte et demi sche. Faites-la entirement scher au soleil, puis enlevez de la pierre ou de la brique cette couleur rose laide dun couteau, et conservez pour lusage. Quand vous voudrez vous en servir, prenez-en volont et affinez, cest--dire broyez-la sur une meule de pierre dure et plane avec de leau de gomme faite pour deux tiers de gomme arabique dissoute dans juste assez deau pour que la gomme en soit presque recouverte quand vous y verserez leau, et filtre ensuite au travers dune toile de lin, et pour un tiers deau claire ajoute cette gomme fondue et filtre. Et de leau de gomme ainsi faite dtrempez le rose jusqu la consistance voulue et vous pourrez en faire ce que vous voudrez, tant pour crire que pour peindre ou dessiner. n. 300 Ad faciendum viride corrosivum absque substancia seu corpore: Ad faciendum viride in substancia clarum et non corpulentum, idest substanciam non habentem, ut verbi gratia clarus atque sine substancia est color safrani, idest croci qui non cooperit alios colores pro ejus subtilitate, claritate et raritate, qua alii colores apparent per medium ipsum et ex hoc ipse pro raritate sua, ut et dictus color viridis remanet obfuscatus, et nil vel minimum apparet, neque multum apparere potest super alios colores. Sed ipse color viridis non est dulcis sicut est dictus color croci, ymo ex sua natura est acer et corrosivus, et taliter quod destruit et rodit alios colores si ponatur super ipsos, vel ipsi super ipsum, et hoc pro viride eris qui in ipso ponitur, et est talis conditionis. Et ponitur in carta et in papiro.
99
Accipe viride eris et modicum de feve vini sicca et fina, que dicitur in latino tartarus et in gallico gravella, et subtilia et tere super lapidem durum et planum insimul que dicta sunt cum aceto. Postea omnia que in carta et in papiro protrahere vis, protrahe, ac vacuum, videlicet per lineas de colore scilicet nigro, postea de ipso colore viridi sic facto ut dictum est, colora ad libitum ea que ut dictum est protraxeris. Et nota quod super ipsum colorem viridem ut dictum est, nullus alter color debet poni neque ipse super alios nisi solum super cartam albam vel papirum, et non super colorem aliquem album artificiatum seu pictum, quia ipse color viridis illo modo factus est fortis seu acer et pro sua acritudine destruit alios colores ut supra jam dictum est. 300 Pour faire un vert corrosif sans corps ni substance: Pour faire un vert transparent et sans corps par nature, cest--dire dpourvu de substance, comme, pour prendre un exemple, le jaune de safran, transparent et sans substance, qui na aucun pouvoir couvrant du fait de sa tnuit, de sa transparence et de son peu de matire au travers de laquelle les autres couleurs se voient; cest pourquoi, en lui-mme, comme le vert, il reste teint du fait de son peu de matire et ne se voit pas ou peu, et pos sur dautres couleurs il ne peut pas se voir beaucoup non plus. Mais ce vert nest pas inoffensif comme le safran, au contraire il est par nature acide et corrosif de sorte quil dtruit et ronge les autres couleurs sil est pos dessus ou dessous, et ce cause du vert de gris quil contient: telle est sa nature, et on lutilise sur parchemin et sur papier. Prenez du vert de gris et un peu de lie de vin sche en fine poudre qui se dit tartre en latin et gravelle en franais, broyez-les ensemble et rduisez-les en poudre sur une meule de pierre dure et plane avec du vinaigre. Puis dessinez tout ce que vous voulez sur le parchemin et le papier et laissez blanc lespace entre les traits tracs au noir, puis de la couleur verte ainsi prpare, remplissez votre guise ce que vous aurez dessin. Notez bien quon ne doit recouvrir ce vert daucune autre couleur, comme on la dit, ni le poser lui-mme par dessus dautres, si ce nest seulement sur le blanc du parchemin ou du papier, mais non sur quelquautre couleur blanche fabrique ou peinte, parce que le vert ainsi fait est corrosif ou acide, et du fait de son acidit il dtruit les autres couleurs, comme indiqu prcdemment. n. 301 Ad faciendum colorem viridem cum corpore et non corrosivum: Ad faciendum colorem viridem dulcem et corpulentum, pro operando in pergameno, in papiro, in telis et in tabulis ligneis dealbatis, accipe viridem eris seu arani et succum herbe que dicitur in gallico flamma et ipsum succum herbe cola per telam lineam et cum ipso tere super lapidem viridem suprascriptum, addendo aliquantulum de aqua gommata. Postea ipsum pone in conchella vel in scutella figuli vitriata, et distempera cum dicta aqua gummata et cum dicto succo ipsius herbe. Et dicta aqua gummata debet fieri de gummi arabico lucido et collata, ne cum infusum sit gummi in ipsa, adsint in ipsa ulle pallee, terra vel alie turpitudines. Et postea de ipso colore viridi scribe, protrahe et pinge que vis. Et nota quod succus rute esset melior quam suprascripte herbe ad ponendum in dicta compositione dicti viridis coloris. Et alii sunt qui ponunt succos quarumdam aliarum herbarum. Et color suprascriptus est talis qnod potest super ipso pingi cum aliis coloribus, et super ipso poni aurum etc., sicuti posset fieri super sinopide vel super lazurio, vel super rosa et aliis similibus, quia ibi non est acetum, et acritudo viridis eris mitigata est dicto succo herbe.
100
301 Pour faire une couleur verte non corrosive et qui a du corps: Pour faire une couleur verte inoffensive et consistante employer sur parchemin, papier, toile et panneaux de bois apprts, prenez du vert de gris et du suc de lherbe dite clmatite flammule en franais, filtrez celui-ci au travers dune toile de lin et broyez avec sur une meule le vert susdit en ajoutant un petit peu deau de gomme. Puis versez le vert dans une coquille ou une cuelle dargile vernisse, et dtrempez de cette eau de gomme et du suc de cette herbe. Et leau de gomme doit tre faite de gomme arabique transparente et purifie afin que, lorsque la gomme sy sera dissoute, ne se retrouvent dans leau aucune paillette, trace de terre ou autres impurets. Puis crivez, dessinez et peignez ce que vous voulez de cette couleur verte. Remarquez que le suc de rue serait meilleur que lherbe susdite pour entrer dans la composition de ce vert. Dautres y mettent du suc de certaines autres herbes. Et la nature de cette couleur est telle quon peut peindre dessus avec dautres couleurs ou poser de lor, etc. comme on peut le faire sur le cinabre, lazur, le rose et autres semblables, car elle ne contient pas de vinaigre, et lacidit du vert de gris est attnue par le suc dherbe.
Biography
Ins Villela-Petit is an Art Historian and, since 2003, a Curator at the Cabinet des mdailles et antiques in Paris Bibliothque national de France. She has published two books: Le Brviaire de Chteauroux (2003), and Le Gothique international Lart en France au temps de Charles VI (2004). Her research work concerns mainly three fields: Colour recipes and painting technics, Illuminated manuscripts, and the Arts of the XIVth and XVth Centuries. Personal website: www.i-villela-petit.fr
101
Resumo
A Crnica Geral de Espanha, cdice conservado na Academia das Cincias de Lisboa a partir de 1879, a cpia do Quatrocentro da Crnica de 1344 de D. Pedro Afonso, conde de Barcelos. Abundantemente iluminado, mostra-nos uma cor e uma tcnica de aplicao devedora da tradio local propagada por tratados e formulrios, se bem que seja claro uma nova linguagem e um gosto internacional. A notcia que se encontra sobre o termho do Cho de Bellotas remete-nos para as diversas produes da regio, entre as quais o azougue de que extrado muyto vermelhon et muy bo. Esta informao importante do ponto de vista do estudo da cor e, particularmente, da cor na iluminura. De facto, permite relacion-la com a abundncia do vermelho na iluminura portuguesa, a partir dos primeiros exemplos, como ns viemos a verificar desde 1986, com Lusa Alves, a quem eu presto aqui a minha homenagem, e, ainda, revelar uma tradio corroborada por formulrios e tratados. O vermelho , sem nenhuma dvida, a cor mais abundante no fundo de Alcobaa, como tambm, no de Santa Cruz de Coimbra e Lorvo; e o colorante que, quase sempre, entra na sua composio conhecido por vermelho. N O livro de como se fazem as cores, tal como no Mappae Clavicula, encontra-se a receita do vermelho que ns vemos utilizada na iluminura portuguesa e na rubricao. A interessante coincidncia entre os dois formulrios mostra-nos uma tradio desenvolvida em Santa Cruz de Coimbra, em que o Mappae Clavicula est registado num inventrio de livros emprestados, do primeiro quartel do sc. XIII (1218), ainda que em Alcobaa no se encontre nenhuma meno de receiturios de pintura, assunto que no mereceu a ateno da sua Livraria, o que poderia significar a pouca importncia que se dava cor mas, tambm, uma maneira de produzir e aplicar as cores de acordo com uma tradio local. Esta tradio do vermelho pode seguir-se, mais tarde, no Breve Tratado de Illuminao, escrito por um monge da Ordem de Cristo, em Tomar, e no seu contemporneo, Arte da Pintura Symmetria e Perspectiva, de 1615, escrito pelo dominicano Filipe Nunes. Mas, esta notcia sobre uma matria prima to abundante e boa pode ser uma das justificaes para que o vermelho seja a cor das imagens do rei representadas na Crnica, sendo lcito pensar que ele continua a ser, nas Espanhas, a cor principal, no acompanhando o triunfo do azul na Europa medieval, a partir de incios do sc. XIII.
palavras-chave
crnica geral de espanha vermelho cor vermelha imagem do rei
103
Resume
La Crnica Geral de Espanha, codex appartenant lAcadmie des Sciences de Lisbonne, depuis 1879, est une copie du Quatrocento de la Crnica de 1344 de D. Pedro Afonso, comte de Barcelos. Abondamment illumine, elle nous montre une couleur et une technique de son application redevable de la tradition locale propage par des traits et des formulaires, bien que soit clair un nouveau langage et un got international. La notice quon y trouve sur le termho do Cho de Bellotas rapporte les diverses productions de la rgion, parmi lesquelles lazougue dont est extrait muyto vermelhon et muy bo. Cette notice est importante du point de vue de ltude de la couleur et, notamment, de la couleur dans lenluminure. Du fait, il permet de la rapporter avec labondance du rouge dans lenluminure portugaise, depuis les premiers exemples, comme nous sommes venus vrifier depuis 1986, avec Luisa Maria Alves, a qui je prt mes homages, et, encore, rvler une tradition corrobore par des formulaires et traits. Le rouge est, sans aucun doute, la couleur la plus abondante aux fonds dAlcobaa, comme, aussi, ceux de Santa Cruz de Coimbra et Lorvo; et le colorant que, presque toujours, entre dans sa composition est celui qui est connu par vermillon. Dans O livro de como se fazem as cores, telle que dans le Mappae Clavicula, on trouve la recette du vermillon que nous voyons utilise dans lenluminure portugaise et dans la rubrication. Lintressante concidence entre les deux formulaires nous montre une tradition dveloppe Santa Cruz de Coimbra, o le Mappae Clavicula est rapport dans un inventaire de livres prts, du premier quart du XIIIme sicle (1218), bien que Alcobaa on ne trouve aucune mention de formulaires de peinture, sujet qui na pas mrit lattention de sa Librairie, ce que pourrait signifier la petite importance quon prtait la couleur mais, aussi, une faon de produire et appliquer les couleurs conformment une tradition locale. Cette tradition du vermillon pour crire peut se suivre, plus tard, au Breve Tratado de Illuminao, crit par un moine de lOrdre de Christ, Tomar, et au son contemporain, Arte da Pintura Symmetria e Perspectiva, de 1615, crit par le dominicain Filipe Nunes. Mais cette notice sur une matire-premire aussi abondante et aussi bonne, peut tre une des justification ce que le rouge soit la couleur des images du roi reprsentes dans la Crnica, tant licite de penser quil continue tre, dans les Espagnes, la couleur principale, en naccompagnant pas le triomphe du bleu en Europe mdivale, partir des dbuts du XIIIme sicle.
mots-cls
crnica geral de espanha vermillon couleur rouge image du roi

h o r c i o p e i x e i ro
Instituto Politcnico de Tomar, Tomar (Portugal)
1. Do termho do Chao das Bellotas. O termho do Chao das Bellotas parte per o termho dAariz e jaz antre ho ouriente. e jaz contra o sseptentrion de Cordoua. E Aariz he uylla em que moraron os barboros. E en seu termho he o monte em que h o uyeiro de que sacam o azougue e dally o leu pra todallas partes dEspanha. E dally sacam muyto uermelhon e muy bo que n saben outro t bo se n aquelle que tragem da terra dultra mar. E na demais desta terra n h outras aruores sen azinheiras. E por esta rrazon o chamon o Chao das Bellotas. E som t doces e t saborosas que n as h tanto Espanha. E em este chao iaz a cidade de Bued que he muy antiga cidade. (fl. 15r.) 2. La production rfre vient mentionne en divers auteurs anciens. Le cinabre dsign par cinnabar ou minium tait des pigments les plus priss dans lAntiquit. En citant Thophraste, Pline signale ses locaux dorigine en Espagne. Vitruve, aussi, rapporte ce pigment pour rappeler sa faible rsistance la lumire et aux agents chimiques atmosphriques, cause, selon lui, de lextraction du mercure argentum vivum en lui provocant la perte de la vis vitalis. Voir Plnio XXXIII, 119; Vitrvio VII, 8-9; Luciana COLOMBO, I colori degli antichi. Fiesole, Nardini Editori, 1991, cap. IV. 3. Pl. XXXIII, 119. 4. La louange de lEspagne, terre admirable, prdestine par Dieu, o non seulement des mer-
La Crnica Geral de Espanha, manuscrit appartenant lAcademia das Cincias de Lisbonne depuis 1879, est une copie du XVme sicle de la Crnica de 1344 de D. Pedro Afonso, Comte de Barcelos. Le sujet de cette intervention est une brve notice, trouve au folio 15r de ce manuscrit. Dans les premiers chapitres, lAuteur fait une description des fastes lgendaires des premiers temps dEspagne; au chapitre XII commence la description de sa gographie, de Cordoue ciga; et, au chapitre XXXIII, est, ainsi, dcrit le territoire de Cho das Bellotas et de Ariz.: Du territoire du Cho de Bellotas: Cho de Bellotas confine avec le territoire dAriz et le Nord de Cordoue. Ariz est une ville habite, jadis, par les barbares. On y trouve une colline o est situe une mine de mercure, export vers toutes les rgions dEspagne. On y extrait, beaucoup de vermillon, de trs bonne qualit, qui on ne connait un autre aussi bon que celui ramen de la terre doutre-mer. Sur ce territoire les seuls arbres qui existent sont des chnes verts, raison pour laquelle cette terre sappelle Cho de Bellotas terroir des glands. Ces glands sont les plus doux et savoureux de toute lEspagne. L se trouve la ville, trs ancienne, de Bued. 1 Ce que cet auteur du XIVme sicle nous raconte tait connu depuis lAntiquit 2: la richesse en mercure du centre-sud de la Pninsule Ibrique. Les territoires dAriz et de Cho de Bellotas se situent au nord-ouest de Cordoue et, daprs la Crnica, confinent avec Feriz ou Constantina, non loin dAlmadm, ou il y a beaucoup de gisements de mercure. La terre doutre-mer, que le texte cite comme zone dorigine dun vermillon de pareille qualit, tait, daprs Pline, en Mditerrane orientale, comprenant phse (o on trouvait un cinabre trs pris par sa couleur rouge vif et dextrme puret) et la Colquide, en Mer Noire 3. Cette notice est trs significative dans le contexte seigneurial de la Crnica, car lauteur cherche rehausser limportance des Espagnes, vu son Histoire et ses richesses naturelles. 4 Au Prologue il indique lobjectif atteindre avec son texte: prserver la mmoire des exploits des trs nobles et sages barons, dont la ligne descend du grand Hercule de Grce. (fl 2r); ensuite il fait, aussi, lloge des terres, pleines de toute sorte de richesses. En Espagne existent plusieurs noblesses que ne
105
peuvent tre racontes; cest pour cette raison que les anciens que sont vnus, en premier lieu, les peupler, lont beaucoup prise, vues les bonts quils y ont trouv. (fl. 10) 5 En Espagne on trouve beaucoup de rivires et de sources, bons airs, hautes montagnes, larges valles et plaines, ou ce trouve tout ce que lhomme peut dsirer. Ainsi certains ont dit que lEspagne est lgale du paradis de Dieu 6 (fl. 10). Elle est riche en poisson, fruits, pain, lait, btail, gibier et vins, et riche aussi en plomb, tain, cuivre, argent vif mercure fer, or et argent. Et cela grce au grand nombre de mines que sy trouvent, particulirement dans la partie occidentale. 7 (fl. 11). Par a, quel serait le grand prince ou puissant seigneur ou trs fort gentilhomme que naimerait pas possder ces terres? 8 (fl. 11). Cest dans ce contexte quensuite sont dcrites les bonts et noblesses des diffrentes villes et ses territoires, dont celui qui est lobjet de cette notice. Mais, au del des loges sur la terre et ses lignes, on peut trouver dautres indications dans cette intressante information, nommment sur: Labondance de vermillon, ce que suggre quil tait frquemment utilis en peinture et en enluminures; Une justification pour le fait que le rouge continue tre la couleur principale, mme quand ce ntait plus le cas dans les rgions dEurope.
veilles naturelles natraient, mais aussi des hommes courageux, vient en Isidore de Sville (Laude Spaniae) et continue dans lhistoriographie hispano-arabe, tant insre, a partir dici, dans la Crnica do Mouro Rasis et dans la Cnica de 1344. Voir Antnio REI, O Louvor da Hispnia na Cultura Letrada Medieval Peninsular. Das suas origens discursivas ao Apartado Geogrfico da Crnica de 1344. Dissertation doctorale en Histoire Culturelle et des Mentalits Mdivales, prsente la Faculdade de Cincias Sociais e Humanas, Universidade Nova de Lisboa, Lisbonne, 2007, et aussi, sous le mme titre, revue Medievalista on line, anne 5, n. 6, 2009. 5. Em Espanha h muytas nobrezas, as quaaes nom podem seer comtadas; e porende os antigos que a comearam a pobrar muito a prearam por as bondades que en ella vyron. (fl 10) 6. Les images de la Crnica, surtout lornementation des marges, avec scnes champtres, bucoliques, de chasse, de pche, sont le rflexe de labondance e de la douceur du lieu. 7. metaees de chumbo, destanho, de cobre, dargem vyvo, de ferro, de ouro e de prata. E esto por grande multidom de vyeiros que h Espanha e especialmente em as partes de Ocidente. (fl. 11) 8. qual seerya aquelle grande prncipe ou senhor de grande poder ou muy forte baron que n fora contento de seer senhor de tal terra? (fl. 11) 9. Cette dissertation, ayant le titre: Missais iluminados dos sculos XIV e XV Contribuio para o estudo da iluminura em Portugal, je lai prsente en 1986, tant soutenue lanne suivante; elle nest publie que partiellement. 10. Lire: BNL Bibliothques Nationale de Lisbonne: BPMP Bibliothque Publique Municipale de Porto ; ANTT Archive Nationale Torre do Tombo. 11. Voir la dissertation que jai prsente comme preuves publiques au concours pour professeur coordinateur, Um olhar sobre a iluminura do Apocalipse de Lorvo. Tomar, 1998. Voir, aussi, Animalia et aliae bestiae: Representaes do
1. Lusage du vermillon en enluminure

Depuis quelque temps dj, nous avons vrifi, propos des enluminures du fond dAlcobaa de la Bibliothque Nationale de Lisbonne, que la plus abondante et simple initial colorie, ne prsentant dautres lments distinctifs que la taille et la couleur, apparat, en plus de la moiti des manuscrits des XIVme et XVme sicles, avec le rouge en alternance avec dautres couleurs vert, bleu et violette et que, pour un quart des manuscrits, le rouge est la seule couleur utilise. Donc, est de signaler la prsence dominante du rouge. Au moment ou ltude de la couleur des enluminures connait un nouveau lan au Portugal ce que je salue avec beaucoup de plaisir et satisfaction, permettez-moi de, ce propos, rappeler brivement les premiers essais en laboratoire, faits parmi nous et, en mme temps, dvoquer la mmoire de quelquun que nous a dj quitte et fut pionnire en ces domaines. En 1984, dans le cadre dassistance llaboration dune thse de matrise prsenter au Dpartement de Histoire de lArt de lUniversit Nouvelle de Lisbonne, Lusa Maria Alves, chef du laboratoire du au temps Institut Jos de Figueiredo, utilisant les mthodes pour lanalyse des peintures, initia un travail que, malheureusement, na pu avoir une dimension plus significative. Les critres dchantillonnage taient peu slectifs. On cherchait une vision gnrale sur les matriaux des couleurs des enluminures et les techniques dapplication, pour un ensemble de manuscrits du
106
bestirio no Apocalipse de Lorvo. In Animalia Presena e Representaes. Lisbonne, Ed. Colibri, 2002, p. 79-99. 12. Luisa Maria P. A. ALVES, Alguns aspectos relativos ao estudo dos materiais que entram na composio de alguns cdices iluminados dos sculos XIV e XV in 2. Congresso Nacional de Bibliotecrios, Arquivistas e Documentalistas. Coimbra, Liv. Minerva, 1987, p.439-465. 13. Les anciens appelaient minium secundarium ou cerusa usta loxyde salin de plomb que, autant le moine Thophile que lauteur anonyme du De arte illuminandi, de mme que Cennino Cennini et la plus part des auteurs de traits, prfrent considrer comme minium proprement dire. Cf. Franco BRUNELLO, De arte illuminandi e altri trattati sulla tecnica della miniatura medievale. Vicenza, Neri Pozza Editore, 1992, pp. 211-212 e 234-235. Aussi labb Frre Joo de Jesus Maria, du monastre bndictin de Santo Tirso et professeur de pharmacie et directeur du Laboratoire Pharmacien du Jardin Botanique, propos de la description du sang de dragon, dans un trait pharmacien indit du XVIII sicle, Histria Pharmaceutica das Plantas Exticas () 1777 , p. 209 e s., dit que celui ci ne se doive pas confondre avec le cinabre minral, chamado dos gregos minio; nome que hoje em dia unicamente se confere quella operao feita de chumbo de que se d notcia na Pharmacopeia Dogmtica. (appel chez les grecs minium; nom donn aujourdhui uniquement lopration faite de plomb de quon donne notice dans la Pharmacopeia Dogmtica.) Cette uvre correspond aux deux premiers tomes, dj dits, tant la part indite, le tome III, une adjonction. 14. Voir, Adelaide MIRANDA, et alii, A cor na iluminura portuguesa Uma abordagem interdisciplinar, in Revista de Histria da Arte, n.5, Instituto de Histria da Arte, 2008, p. 228-245. Lampleur de cet tude, comprenant centaines dobservations, ne peut pas tre compar avec la limitation de lessai de caractrisation de la couleur que jai entrepris dans ltude rfr sur lenluminure de lApocalipse de Lorvo. Malgr tout, je registre, avec satisfaction, la concidence
XIVme et XVme sicles, tout en partant de livres de recettes, parmi lesquelles O Livro de que como se fazem as cores das tintas todas para aluminar os livros. Une slection fut faite, de vingt trois manuscrits, la majeure partie datant des XIVme et XVme sicles, appartenant au fond dAlcobaa de la BNL, au fond de St Cruz de la BPMP, au ANTT 10, la Bibliothque et au Muse de la Cathdrale de Viseu et la Biblioteca Pblica e Arquivo Distrital de vora; au total on a couvert les diffrentes couleurs utilises pour lenluminure et lcriture en 112 essais. Quelques annes plus tard, en 1997-98, semblable recherche, plus limite, fut faite ayant pour base le Livre de lApocalypse de Lorvo 11. Les rsultats furent publis pour la premire fois dans le Deuxime Congrs National de Bibliothcaires, Archivistes et Documentalistes, en 1987. 12 Plusieurs pigments sont utiliss pour obtenir le rouge, mais cette recherche a rvl que le pigment qui est presque toujours prsent cest le vermillon, minium pour les anciens, qui donnaient ce nom au cinabre, ou vermiculum (vermillon sulfure rouge de mercure) 13. En regardant les rsultats provisoires prsents dans ltude A cor na iluminura portuguesa, du groupe pluridisciplinaire de lUNL (que heureusement organisa cette confrence), on peut aussi vrifier lutilisation de ce pigment dans les deux manuscrits rfrs (lApocalipse de Lorvo et le Livro das Aves) et la prsence du rouge dans presque touts les autres sept documents 14. Lampleur de cette tude, avec centaines dobservations, comme nous avons vu dans ces deux jours, se situe dans une chelle compltement diffrente de mes modestes et presque solitaires recherches. Je registre, quand mme, quelque concidence danalyse, comme celle la. Dans O Livro de que como se fazem as cores, ainsi que sur le Mappae Clavcula on trouve cette recette de vermillon utilis dans lenluminure portugaise et dans la rubrication, au moins depuis lApocalypse de Lorvo, du final du XIIme sicle, aux manuscrits dAlcobaa des XIVme et XVme sicles 15. La procdure est la mme pour les deux livres de recettes, bien que les proportions de soufre et de mercure soient diffrentes: Un de souffre pour cinq de mercure, dans O livro de como se fazem as cores; deux pour un, dans le Mappae Clavicula (un par deux dans le Theophile). Cette concidence, fort intressante pour dautres recettes aussi nous rvle une tradition, dveloppe, peut-tre, Santa Cruz, de Coimbra, ou le Mappae Clavicula est cit dans une liste de livres prts, du premier quartier du XIIIme sicle (1218), transcrit au manuscrit St. Cruz 34/43 16. La mme tradition ce trouve aussi dans les autres scriptoria, bien que, on ne trouve plus aucune mention de recettes de peinture. Alcobaa, la seule recette technique que jai trouve dans le fond de manuscrits dcrit la faon de coller (solfar) le parchemin 17. Ici, dans ce priode, la couleur navait pas limportance du sicle prcdent, regardant le soin technique et la diversit de la production, comme Adelade Miranda a dmontr ce matin. Cette tradition de lusage du vermillon pour crire, se retrouve plus tard dans le Breve Tratado de Illuminao, crit par un frre de lOrdre du Christ, de Thomar, et dans son contemporain Arte da Pintura Symmetria e Perspectiva de Filipe Nunes, Dominicain, texte de 1615. 18
1 07
2. La signification de lusage du vermillon

Brivement, dans ltude effectue, on a vrifi que, pour la priode du XIVme et XVme sicles, 19 dans les diffrents fonds, il y a une pauvret de la palette, la concidence des formules et procdures utiliss avec des livres de recettes, lutilisation systmatique de certains pigments, faciles obtenir, et des procdures peux labores, comme par exemple la dorure. 20 Le rouge, centre de lancien systme ternaire des couleurs, qui a pour ples le blanc et le noir 21, est tenu, dans tout lOccident, depuis les temps Protohistoriques, comme la premire de toutes les couleurs, la couleur par excellence, avant lclatement de ce vieux schma au XIIIme sicle 22. Limportance donne cette couleur, dans les scriptoria portugais (et aussi dans les ateliers de peinture), pourra signifier la permanence de vestiges de lancien systme de couleurs que, daprs Michel Pastoureau, commence se dcomposer entre les sicles XI et XIII, pour donner place au systme plus linaire quon utilise encore aujourdhui. 23 Mais, au de la de la facilit dobtention, il faut aussi tenir compte le cot, relativement bas. Au livre des dpenses du Couvent du Christ Thomar 24, de la premire moiti du XVIme sicle, on trouve une srie de prix des pigments pour les enlumineurs et pour les peintres, ce que peut-tre utilis pour faire une comparaison de cots relatifs. On peut admettre que ce rapport se maintient, grosso modo, dans des priodes antrieurs. Ainsi, le vermillon cotait entre 120 et 53 Ris la livre (arrtel), selon la qualit du produit, alors que pour le bleu, la valeur va de 400 Ris la livre (arrtel) 1050 lonce (28, 69 grammes) 25, prix du bleu fin. Le prix du mercure (azougue) tait de 100 Ris la livre (arrtel). Selon le mme registre de dpenses, lenlumineur Antnio de Hollanda tait pay de 6000 Ris pour chaque frontispice, 500 Ris pour chaque lettre vignette, 80 Ris pour lettre simple en indigo (anil), bleu et or, et 40 Ris pour lettre en noir. On voit, donc, que le prix du bleu courant 26 tait, plus ou moins, trois fois et demie plus cher que celui du vermillon, alors que le bleu fin (peut-tre du lapis-lazuli) tait 140 fois plus cher. Le bleu tait le plus cher des pigments indiqus. Il suffisait Antnio de Hollanda denluminer une lettre vignette pour pouvoir acheter quatre livres de vermillon, mais seulement un peu plus dune livre de bleu courant; et il lui faudrait un frontispice pour soffrir une once de bleu fin. Labondance de vermillon explique, probablement, non seulement son cot relativement bas, mais aussi son usage frquent. Revenons la Crnica Geral de Espanha. On ne pourra y parler ni de pauvret de palette, ni de contentions, tel que dans dautres manuscrits du XVme sicle 27. Le recours aux commandes royales, de que ce manuscrit est exemple, pendant le XVme et le XVIme sicle, introduit une certaine opulence et une mise au got international dans la production de manuscrit. Bien que pour la Crnica on ne dispose daucune tude en laboratoire pour les pigments et agglutinants, on peut essayer, avec toute prudence, pour ce que jai dnomm de deuxime style dornementation 28, un certain rapport avec des manuscrits de la mme poque, o on trouve lemploi didentiques
danalyse relativement a ce manuscrit. 15. Voir O livro de como se fazem as cores, cap. XV: Para fazeres vermelhon, toma cinco libras de fujativo . ides. azogue e pe-no en a arredoma o tigela grande vidrada e toma a libra de pedra de enxofre been miuda e deita-lhe o poo do enxofre poco e poco sobre o arjen vivo ata que seja been encorporado, e toda via mecendoo con pee de co con sua pele e sa laa ata que se torne o fogo como cinza e depoes que asi for mortificado deitao en duas olhas novas que sejam feitas como aredomas enjas de joso e estreitas encima e no fique por encerar de elas seno un furaco pequeno por o saya o umor e pors as olhas sobre o fogo en suas fornalhas e baras ben con o baro e poen a tijela en cima dos purados e quando o fumo vires que sal vermelho e no feder mete dentro no furaco un espeto delgado e si algua cosa se apegar ao espeto tira as olhas do fogo e leixo esfiar e depoes que for frio quebranta as panelas e ajaraa o bermelho feito .e por este peso faras quanto vermelhon quisieres fazer e a a tera doazogue poye 5 libras do enxofre e a 5 libras da azogue a libra de enxofre equilheres ental guisa o fogo que no se queme edalhe fuego temperado nen bivo nen manso e te se per ventura sese quemar o vermelho quebranta as olhas e moiyo e encorporao e misturao con otro peso dazogue e de enxofre e poeno en otras olhas e faze como dito e: e para been mentes nos fumos como saen asi e nunca eraras. (Revista da Faculdade de Letras. Lisboa, S. 3 (4), 1960,(transcription provisoire de Jos Ramos); voir aussi le Mappae Clavicula, la premire recette: De vermiculo. Si vis facere Vermiculum, accipe ampullam vitream et lini deforis de luto, et sic accipe unum pondus vivi argenti, et duo pondera sulfuris albi aut crocei coloris, et mitte ipsam ampullam super iij. aut iiij. petras, et adhibe ignem in circuitu ampulle ex carbonibus, ignem tamen lentissimum, et sic cooperies ampullam ex parvissima tegula: et, quando videris fumum exire ex ore ampullae blavum, cooperi: et quando exierit fumus crocei coloris, iterum cooperi: et quando videris exire fumum rubeum quasi vermiculum, sic tolle ignem, et habeas vermicullum optimum in ampulla. Au chap. CV on voit un autre recette trs pareille celle du Composi-
108
tiones ad tingenda De compositione cinnabarin Tolles ydroargiris (mercure) mundi partes ii et sulfuris vivi partem i et mitte in ampullam, sine fumo, et lento igni, decoquens, facies cinnabarin, et lava utiliter. (Thomas Phillips, Mappae Clavicula; manuscript traetise on the preparation of pigments, and on various processes of the decorative arts practised during the Middle Ages. London, 1847, p. 7). Dans un autre trait, le Manuscrit de Bologne Libro dei colori (moiti du XVme s.), crit en lItalie, o, tel quen Espagne, abondait le mercure, on peut voir deux recettes dont les concidences avec le Mappae Clavcula sont videntes: Ad faciendum cinabrium Tolli una parte de argento vivo et doi parte de solfo giallo e necto e bem macinato poi pone omne cose in una bocia et incoprila legiermente cum luto de sapientia poi la pone in lo fornello et dalli da prima lo foco ligireo et copre la boca della bocia cum una tegola, e quando tu vederai lo fumi giallo continua lo foco per in fino che vedrai uscire el fumo rosso o vermeglio alora tolj via lo foco et quando ser freddo troverai bello cinapro. Ad idem alio modo. Habeas unam ampollam vitream lutata de luto sapientie usque ad summum collj deinde recipe partes duas sulforis albi et bene triti et partem unam argenti vivi postea pone in ampulla sopradicta et fac de carbonibus ignem lepidissimum et circa eam cum quatuor lapidibus et ponem ampullam desuper et coperi eum cum tegula et sepe discoperias et quando videbis fumum lividum coperi dummodo videbis exire fumum rubeum tunc tolle ab igne quod factum erit. (M Merrifield, p. 326. Cap. VII, n. 182 e 183 Secreti Per Colori.) 16. Voir Antnio CRUZ, Santa Cruz de Coimbra na cultura portuguesa da Idade Mdia. Porto, 1964, p. 203. 17. Voir lOrdinrio do Ofcio Divino, BNL, Alc. 63, fl. 147v. 18. Breve Tratado de Ilvminao composto por hum religioso da ordem de Xp. (...). Bibliothque Gnrale de lUniversit de Coimbra, Col. Jardim Histrico, vol XXXVII, ms. n. 344. Voir au chap. IX,, part 1.re, une recette pour faire
motifs et de procdures dusage de la couleur: lettres phytomorphiques, ramures avec feuilles dacanthe e lancoles que se prolongent, en tige, vers les zones priphriques, mais aussi une volution de la palette depuis la priode antrieure et, principalement, le perfectionnement des procdures, crant un plus large chromatisme. 29 Les images de la Crnica, dessins et colores, comme on la dj dit, au XVme sicle, nous fournissent une donn intressante, sur les couleurs utilises pour les reprsentations du roi, la figuration la plus importante. Tout dabord on trouve lor comme fond en toutes les pages, lettres ornes et miniatures, ce qui est rare dans lenluminure portugaise de cette poque ; cela la transforme en ouvrage dapparat, o la filigrane, lornement le plus courant, disparat compltement. Dautre part, lenvers du reste de lEurope, la couleur principale pour les vtements royaux est le rouge, deux exceptions prs, ou le bleu domine: le roi dcd Alphonse VI (fl. 289r.) et le roi enfant Henri I de Castille (fl. 285r.), le voyant, dans la scne de lacclamation, sur une estrade recouvert dans drap rouge. 30 Ce langage des images, incorpores dans la Crnica un moment diffrent de celui de lcriture, 31 introduit un nouveau discours, ou le roi est prsent et domine, jouant le rle principal. Cest en lui qui se fonde la lgitimit, renversant, ainsi, le ton seigneurial sous-jacent au texte de la Crnica. a veut dire que la faon de reprsenter le roi, nest pas dpendent dune raison graphique ou dune combinaison chromatique, par exemple, la loi de lalternance: mais son image est construite avec une intentionnalit, montrant clairement ses attributs. Nous allons vous prsenter lexemple le plus significatif. Limage du folio 185v, reprsente le roi Jacques II dAragon (1291-1329), frre de la reine Sainte Isabelle, pouse de D. Dinis. La figure, un buste, est inscrite dans linitial D, construite avec des motifs architecturaux, en fond dor piqu. Le roi est reprsent de front, avec des cheveux coups court, mode la plus rpandue dans la Crnica 32, avec une longue barbe en boucles et peign en deux; coiff dune riche couronne, il est habill en rouge. Cette image du roi, avec la force du portrait, est la plus labore de toute la Crnica, ce que relve limportance, dans ce manuscrit, de liconographie des rois dAragon. Avec de belles proportions et un aspect solennel, nous sommes tents de y voir lallgorie de la Cour Impriale (Corte imperial XIVme sicle), ouvrage existant la bibliothque du roi D. Duarte, ou se dcrit la figure de lEmpereur le Christ visage vnrable, inspirant lamour et la crainte, une barbe abondante, divise en deux au menton, des yeux beaux, simples et clairs, avec une couronne de pierres prcieuses sur la tte. Ctait, on le voyait, le plus parfait des hommes. 33 Le traitement spcial de la tte du roi, la partie la plus noble du corps, montre que le roi est la tte du royaume, tout comme le Christ lest par rapport lglise. Cette image peut, donc, voquer la grce spciale du roi une sorte de dualit, comme le Christ, que vient de la grce divine 34 mise en vidence par lor, la couleur de la divinit, et par le rouge, la couleur de lhumanit. Dautres exemples de liconographie royale pourraient tre indiqus, dmontrant la prfrence pour le rouge, comme cest le cas du clbre portrait de D. Joo I, au Muse National de LArt Ancien de Lisbonne, dont la datation du XVme sicle fut,
109
rcemment, mise en cause 35, ou limage de D. Manuel, comme roi David, du Missal Rico de S. Cruz de Coimbra 36. Or et rouge paraissent tre attributs royaux, couleurs prsentes aussi dans les armoiries des rois portugais. De la simple notice, du XIVme sicle, quon a lu dans un manuscrit crit et enlumin au XVme sicle, on peut dduire, grce quelques donns et une certaine fantaisie, que le vermillon, abondant dans la Pninsule Ibrique, continue dtre une source privilgi et plus bon march pour obtenir ce rouge lumineux quon trouve dans notre enluminure et que va garder ici, pendant tout le Moyen Age, le statut de color principalis. 37
(aparelhar) le vermillon crire. Filipe NUNES, Arte da Pintura Symetria e Perspectiva. Composto por Philippe Nunes natural de Vila Real. Ed. Fac-simil a partir de ld. de 1615, avec introduction de Leontina Ventura. Porto, Ed. Paisagem, 1982. Voir ici les fls. 65 e 66: Tomase hum pucaro nouo, e nelle se bota o enxofre. E o azougue partes iguais, e depois se barra muito bem que no saya o bafo fora, e posto ao fogo at que se encorpore ha cousa com outra por espao de sinco ou seis horas. 19. Voir la dissertation cite (note 9), Missais iluminados dos sculos XIV e XV []. On a utilise les tudes ralises au Laboratoire Centrale de lInstitut Jos de Figueiredo, publies, aussi, par Lusa Maria P. A. ALVES, O.c., note 1.re. 20. En Alcobaa, la dorure est trs imparfaite quant la prparation et quant aux rsultats obtenus. Alors on peut penser que dans ce scriptorium on prtait peu dattention a cette procdure, dispensable dans la notation du texte, au contraire des traits de recettes, tel que le Mappae Clavcula ou le O Livro de como se fazem as cores, o on peut trouver une large quantit de recettes ddies cette procdure. 21. Thomas de Verceil, dans sa thorie mystique de la couleur, dfinit Dieu comme blanc et rouge, lumire e chaleur: Indicibiliter se candidum et rubicundum ostendit (Deus) [...] ita ut nihil aliud sit sua luciditas quam sua igneitas, nec minus igneitas quam luciditas Cit. em E. de BRUYNE, La esttica en la Edad Media. Madrid, Visor, 1994, p. 130. Au Portugais on continue a utilis les mots corar et corado (coloratus) pour indiquer laction de donner une teinte ou de rougir ou quelque chose color de rouge, vestiges de limportance du rouge comme couleur par excellence. 22. Michel PASTOUREAU, Figures et couleurs tudes sur la symbolique et la sensibilit mdivales. Paris, Le Lopard dOr, 1986, p. 37. 23. Cf. M. PASTOUREAU, Couleurs, images et symboles tudes d`histoire et anthropologie. Paris, Le Lopard DOr (1988), p. 23.
le roi jacques ii daragon crnica geral de espanha academia das cincias lisbonne, ms. 1 azul, fl. 185v.
Biographie
Email: hpeixeiro@ipt.pt / hpeixeiro@gmail.com N en 1945, a fait la licence en Histoire (Faculdade de Letras da Universidade de Lisboa, 1976) et la maitrise en Histoire de lArt (Faculdade de Cincias Sociais e Humanas da Universidade Nova de Lisboa 1986). Professeur Coordinateur au Instituto Politcnico de Tomar les dernires 22 annes de a carrires, actuellement est en retraite. Dans les prochaines annes est proviseur de ltudiant au mme Institut. Ses tudes sur le livre manuscrit sont centrs sur lenluminure portugaise aux XIVme et XVme sicles.
110
24. ANTT, Livro 23, Despesa das obras do convento de Tomar, feita por Fr.Gaspar, fora da empreitada de Joo de Castilho, Tomar, 1533-1539. Indites. Fl. 137-185. Cf. Sylvie DESWARTE, Les Enluminures da la Leitura Nova. 1504-1552 tudes sur la culture artistique au Portugal au temps de lHumanisme. Paris, F.C. Gulbenkian, 1977, p. 211. Lauteur transcrit les payements faits Antnio de Holanda. Voir, aussi, Maria Amlia P. S. CASANOVA, A pintura de Gregrio Lopes em Tomar sob o mecenato de Frei Antnio Lisboa. Lisbonne, 2002, thse de maitrise, sous presse. Ici lauteur collige la documentation complte relative aux contrats avec les enlumineurs et les dpenses faites dans les achats des matriaux. 25. L arrtel fait 0,459 kilo et seize onces et lonce fait 28,69 grammes. 26. Latelier du peintre Gregrio Lopes (c. 1485c.1550) utilisa, pour le bleu, lazurite, et pour le rouge on employait toujours le vermillon et frquemment locre rouge, la garance et le kerms. Voir Lusa Maria ALVES et Vitor SERRO (Coord.), Estudo da Pintura Portuguesa Oficina de Gregrio Lopes. Lisbonne, Instituto Jos de Figueiredo, 1999, pp. 179-232. 27. Voir les manuscrits produits dans lentourage de la cours dAvis: Leal Conselheiro e o Livro da Ensinana de Bem Cavalgar B. Nacional de Paris, Fonds portugais, n. 5 (1433-38); Vida e Feitos de Jlio Csar Escorial, Monasreio de San Lorenzo, Q-I-37, (1446-85); Ordinrios do Ofcio divino Alc. 62 (1475) et Alc. 63 (1483); Missal Cisterciense Alc. 459 (2. moiti do XVme s.); Livro da Virtuosa Benfeitoria cd. 9, Real Ac. de Hist. de Madrid il a appartenu au fils de D. Pedro, le conntable (ca. 1430), et un autre exemplaire de la Biblioteca Municipal de Viseu, cofre n. 14 (ca. 1430); Vida de Cristo de Ludolfo de Saxnia Alc. 451-453 (1440-50). Dans lAlc. 451 on voit, au fl. 56v. Ata aqui fez o scripvam del Rey (Jusquici il tait fait par lcrivain du Roi) et au fl. 57r lindication de quil en a achev dcrire et relier Fr. Bernardo de Alcobaa, 1445-1446.
28. Le deuxime style on peut le caractriser par les feuillages et les ramures. Plus simple et retenu, ce programme privilgie la lettre initiale, quelques fois se prolongeant sur les marges, en excluant toute sorte de figuration. Cest ce modle qui on voit aussi dans dautres manuscrits sortis du mme atelier royal, tel que le Leal Conselheiro et ceux qui sont dsigns dans la note prcdente, o cest possible voir lidentit de motifs et procds. 29. Les nuances appliques sur le tom de base ont lobjectif dobtenir non seulement les lumires (lumina) et lombre (tractus), comme il vient dcrits dans le De diversis Artibus du moine Thophile, mais, surtout, le model, avec videntes influences de la technique de la peinture. On assiste, nommment, au retour du jaune (tain et plomb) et llimination du contour en noir. 30. Les autres reprsentations du roi, trouves la Crnica, son les suivantes: fl.155r. Le roi avec manteau rouge, assis, dans la scne du taureau; fl. 160r Le roi maure de Tolde avec son manteau rouge prenant sa femme chrtienne; fl. 182 r Le roi Pierre III dAragon avec manteau carlate et vert; fl.185v. Le roi Alphonse III dAragon shabille en rouge et bleu; Jacques II dAragon shabille en rouge; fl. 205 r Rencontre amoureux de D. Urraca et El Cid; fl. 266r La reine Urraca shabille en rouge et son amant le comte de Lara, en carlate; fl. 287 v Tte de roi dans un fond rouge; fl.318r. Le roi Alphonse X, le sage, shabille en rouge carlate. 31. Sur ce sujet jai crit un petit texte, en train dtre publi la Revista de Histria da Arte, de lInstitut dHistoire de lArt, Facult de Sciences Sociales et Humaines de LUniversit Nouvelle de Lisbonne, insr au Project Imago, dvelopp dans le mme Institut: Imagem e tempo Representaes do poder na Crnica Geral de Espanha. L je pose lhypothse de la Crnica pouvoir tre enlumine pendant le rgne de D. Duarte ou, mme, dans la priode de la rgence do son frre, lInfant D. Pedro.
32. La mode des cheveux coups courts, rass dans la nuque et sur les oreilles, tombant en frange sur le front, semble tre initi en France vers 1420. Voir QUICHERAT, J., Histoire du costume en France depuis les temps les plus reculs jusqua la fin du XVIIIe sicle. 1875-77, p. 256. On peut voir que les figures reprsentes dans le polyptiques de La Vnration de S. Vincent, du Muse National de LArt Ancien de Lisbonne, uvre contemporaine ou presque de les images de la Crnica, ont les cheveux coups de la mmes faon, nommment le personnage que la tradition a identifi comme lInfant D. Henrique, mais que probablement sera son frre, le roi D. Duarte. Le costume a, aussi, des lments communs. 33. Voir Mrio MARTINS, A Corte Imperial. Alegorias e smbolos e exemplos morais da literatura medieval portuguesa. Lisbonne, Ed. Brotria, 1980, p. 208. 34. WIRTH, Jean, Limage mdivale. Naissance et dveloppements (VIe- XVe sicle). Paris, Mridiens Klincksieck, 1989, p. 211. 35. Voir Jos Alberto Seabra CARVALHO, O retrato de D. Joo I Reviso crtica. Revista de Histria da Arte, n. 5, 2008, p. 67-75. 36. Voir Horcio A. PEIXEIRO, Um missal iluminado de Santa Cruz, in A Luz do mundo Oceanos, 26, Avril-Juin, 1996, pp. 60-69. 37. Je remerci Manuel Guedes Vieira de la traduction en franais de ce texte.
111
Resumo
A escolha do ligante para as tintas da iluminura foi um dos aspectos tcnicos mais importantes a ter em conta no passado, uma vez que este pode interagir de forma diferente com os pigmentos e contribuir de forma decisiva para o efeito visual. O temperar dos pigmentos para a iluminura ocupa uma parte significativa da maioria dos tratados medievais sobre tcnicas artsticas, sendo descrito com detalhe tanto a sua preparao como as misturas e aplicaes. Este artigo estuda e organiza a informao relativa aos ligantes referidos em importantes fontes histricas e textuais, com o objectivo de esclarecer se a escolha de um certo ligante estava associada ao tipo de pigmento ou ao efeito visual final; e tambm, de fornecer um instrumento til de apoio documentao necessria s anlises em laboratrio e prpria interpretao dos dados analticos.
palavras-chave
ligante iluminura tratados medievais investigao das fontes
Abstract
The choice of the binding media in manuscript illumination was one of the most important technical aspects as each medium interacts differently with pigments and optical results can be quite different. Most medieval treatises on art technology dedicate extensive parts on tempering pigments for illumination, explaining with details their preparation, mixtures and use. The paper will study and organize the information regarding binding media quoted in these important historic textual sources, with the aim to clarify several technical issues concerning the choice of a binder in relation to the pigment to be used or the wanted final appearance of colours; and second, to provide a useful tool for the documentary support of laboratory analysis and for the correct interpretation of analytical results.
key-words
binding media manuscript illumination medieval treatises source research
binding media in medieval manuscript illumination: a source research

st e fa n o s k ro u sta l l i s
stefanos.kroustallis@gmail.com
Introduction
Manuscript illumination is an aqueous painting technique and therefore needs a binder to keep the pigment particles together, facilitate its application with a reed, quill or brush, and also improve its adherence to the surface of parchment, the writing material par excellence in European Middle Ages. The choice of the binding medium was very important because each medium interacts differently with pigments and can change its optical properties. Thus, medium preparation and mixture with pigment was one of the essential parts of the whole process of manuscript illumination. A good example is the scheme of painting materials and techniques in medieval treatises on art technology, where tempering is as important as pigment identification, preparation and mixtures. What is more, the anonymous author of the late eleventh century treatise De clarea warns that to make something beautiful it is necessary to prepare materials properly, in his case the different ways to prepare glair (Thompson 1932: 15). Medieval treatises on art technology are one of the most important sources for the study of binding media in manuscript illumination (Clarke 2001). Illumination was a major sumptuary art, as illuminated manuscripts were gifts of high esteem due to the wealth of materials and the symbolic value of images. For this reason, it is not surprising that illumination is one of the techniques that constantly appear in these treatises, either as specific technical process or as part of pictorial arts. The study and organization of this information can, first, clarify several technical issues concerning the choice of a binder in relation to the pigment to be used or the wanted final appearance of colours; and second, can provide a useful tool for the documentary support of laboratory analysis.
113
Binding media
The principal binding media used in manuscript illumination were clarified egg white or glair (clare, albumen, glarea, albugine ovi); gums, such as gum arabic (gumma): and glues, such as fish glue (ichtyocollon), casein glue (glutine casei) or parchment size (cola pergamena). In most cases binders were applied alone, but depending on the technique or pigment, mixtures were also prepared in different proportions.
Glair
Glair (from the late Latin clarea, from Latin clarus, clear) is the settled liquid of the egg white froth. There are three ways to prepare glair that, ultimately, are the three ways to make the egg white in to foam. First, by whipping egg whites (verberata, fracta, percutita) with a wooden whisk 1 or spoon; second, by pressing and squeezing them with a sponge (cassata, spongiata); and, third, by passing them through a wool or linen cloth. A survey of medieval treatises points out that until the twelfth century whipping was the only method used. The reason is given by the anonymous author of De clarea treatise who states that the use of a sponge or a filter could contaminate glair either with grease and dirt of hands or with impurities of the sponge or cloth and the result would be a weak and brittle binding media (Thompson 1956: 15). When Theophilus in his Schedula diversum artium 2 refers to glair it is always the one prepared by whipping (Hawthorne and Smith, 1979: 31-38). However, it seems that since the thirteenth century the sponge method gained currency and the references to clara ovi spongiata become more common 3. In very few cases it is recommended the use of a funnel-shaped wet linen filter to prepare glair (Heraclius treatise, Merrifield, 1967: 233). Glair beating was a laborious and complicated process, mostly because any failure would mean the loss of its adhesive power. De clarea text is quite explicit: containers should be very clean without grease and brass vessels shouldnt be used, as glair can turn greenish (Thompson 1932: 17). The last comment is interesting because it shows also how everyday experience interferes with artistic praxis: copper containers stabilize egg white foam and does not settle, something desired for culinary purposes but not for artistic ones (McGee 2004: 102-103; Perego 2005: 511). Moreover, if egg white is not well beaten pigments could not be tempered easily, states again our anonymous author (Thompson 1932: 19). Glair consists mainly of water (circa 88%; Colombini and Modugno 2009: 167) and proteins (with hydrophilic and hydrophobic amino acids) and a strong beating that creates foam means that proteins become completely denaturized and, once settled, glair can be mixed with water and become a binding medium for manuscript illumination. Again the author of the De clarea treatise draw special attention to the fact that if glair was not well beaten it would behave as if it was glue, so pigments would not run well from the pen of the scribe and the colour would appear unsightly on parchment (Thompson 1956: 19). The proof that glair was well whipped was that foam could remain adhered to the container without running. Then glair was left to rest with the container tilted, so when the
1. At the De clarea treatise we have a very good description and even a drawing of this wooden whisk (f. 2r). At the De arte illuminandi a brush (pinzellum situlare) was also used to whip egg (Brunello 1992: 89). An anecdotic case is described in an Italian recipe from the XIII century where glair was the binding media for vermilion applied with quill bene rocta con la spongna o con la scopa et con lartifiggio delfico (Tossati Soldano 1978: 142). 2. Current research considers that the treatise is a composite text and the author (or authors) was a compilation; see Clarke, M. (2011) Reworking Theophilus: adaptation and use in workshop texts, in: A. Speer et al.(eds.) Die Schedula diversarum artium ein Handbuch mittelalterlicher Kunst? (in series Miscellanea Mediaevalia). BerlinNew York: W. de Gruyter. 3. For example, Experimenta de coloribus (Merrifield 1967: 56); De coloribus diversis modis (Merrifield 1967: 265); Liber diversarum arcium (Libri 1849: 765; Clarke 2011 1.22.1); Libro secondo de diversi colori where the author recommends to use four sponges bound together (Wallert 1995: 40, 42); or the fifteenth century Bolognese manuscript (Merrifield 1967: 466).
114
b i n d i n g m e d i a i n m e d i e va l m a n u s c r i p t i l l u m i n at i o n : a s o u r c e r e s e a r c h
4. Old glair was a very common varnish in manuscript illumination and bookbinding as it forms resistant and elastic films, although its permeability and solubility decreases with ageing (Colombini and Modugno 2009: 238). For example, Theophilus recommends old glair to varnish gilded areas (Hawthorne and Smith 1979: 37).
foam would become liquid it could be easily collected. In that moment it was really important to control environmental conditions as low temperatures could freeze glair and high ones could dry it. Finally, glair was kept in a glazed clay pot or even in the shell of the egg because according to a widespread medieval belief, the natural container of substances was the best to conserve them. It seems that old glair was more appreciated than the new one. The author of De coloribus faciendis recommends to use a three or four days old glair as a binder for the mixture of red lead and vermilion, because colour would look shine with a short of varnish brilliancy (Merrifield 1967: 142) and the author of De coloribus naturalia used old putrefied glair to give shine and strength to blue pigments (Thompson, 1935: 139). Glair could be modified with water to improve pigments fluidity, once tempered. Theophilus in his treatise recommends beating egg white with water in summer and without it in winter, so it would never become too dry or too aqueous (Hawthorne and Smith, 1979: 36). Glair was an optimal medium for miniature painting, free flowing and easily applicable, but there were also some inconveniences in its use as a binding media, besides the above-mentioned difficulties. The main disadvantage was the formation of bubbles while pigments were tempered (spumositates de coloribus), which was really annoying to illuminators. The recommendation of almost all authors to avoid this problem was to add earwax (ceroti auricule) in the mixture, especially if it was going to be used with azurite blue and vermillion (Liber diversarum arcium, Libri 1849: 747; Clarke 2011 1.3.19). Another problem was the fact that if glair stayed too long mixed with pigments it could damage them. It seems that, again, blue pigments were the problem, as many authors warn not to work more than a day with the pigment tempered with glair, because becomes darker and its beauty gets spoiled (Livro de como se fazen as cores, Blondheim 1930-1: 82; De coloribus faciendis, Merrifield 1967: 134; Bolognese manuscript, Merrifield 1967: 410). As glair reduces the natural saturation of colours, sometimes pigments had to be varnished after drying 4. Another inconvenience is that, unlike egg yolk and gums that can be used more or less immediately, glair needs to be prepared a few hours before it can be used, as it has to sit and the older it is, the better it gets. Many of these difficulties were resolved with the mixture of adhesives. For example, as glair dries fast, yolk was added and, at the same time, the intensity and brilliancy of yolk was combined with the durability of glair (Bolognese manuscript, Merrifield 1967: 410).
Egg yolk
Egg yolk (vitello ovi) was the foremost medium for panel painting in late medieval and early Renaissance Europe. But, according medieval treatises on art technology, it was rarely used alone as a binder in manuscript illumination. As the author of De clarea stated, the reasons was that it had not a good adhesive power, as it is more greasy than glair, pigment surface cracked and made spots (Thompson 1932: 73). But as egg yolk left pigments brighter, he also recommended mixing it with glair and taking advantage of the characteristics of both media. The use of this mixture is also found in the Marciana manuscript, where putrefied glair and yolk is specially recom-
115
mended for colours which have no body (Merrifield 1967: 610). The anonymous author of the twelfth century Hi sunt omnes colores also quotes the use of yolk as a binding media for illumination, together with glair and cherry gum (Novk 1996: 77). Finally, pigments tempered only with yolk clot on the brush, so only a small amount could be applied each time. Therefore, the amount of water in the tempering process would be really important, as it could influence the transparency of the colour. The preparation process was really easy: egg yolk was removed from its sac by making a hole very carefully with a thorn or a needle and it was mixed with a drop of water (De coloribus et artibus romanorum, Merrifield 1967: 235).
5. The term albam here is not used in the sense of something opaque white, but in the sense of something bright and transparent; for example, a common adjective for water is also albam. That is why the author of De Arte Illuminandi recommends to chose always a gum that is lucidisima or albam et claram (Brunello 1992: 99). 6. It seems that Theophilus recommendation to use local products, equally good and less expensive, rather than imported ones is quite applicable in this case (Hawthorne and Smith 1979: 12).
Gums
Gum arabic was one of the most common binding media for pigments in manuscript illumination and the exclusive medum in writing inks. As the author of the Liber diversarum arcium states (Libri 1849: 766; Clarke 2011 1.23.1), this gum was imported from Arabia and there were commercially available three kinds, according colour: the best quality was the white one (albam 5) and of a lower quality was the yellow (citrinum) and the pink (subrussum). The gum was very easy to prepare and use: it was commercialized in solid lumps, which were powdered or soaked in water, until they were completely dissolved. Then it was tested with fingers: if they stuck together tightly, then it was well done, if not, more gum should be added. Finally, it was filtered with a linen cloth and mixed with pigments. The main advantages of gum arabic are its high solubility in water, its good adhesive power, and the fact that gives intensity to colours, as it saturates pigments. The disadvantage is its brittleness, and normally it had to be emulsified with natural plasticizers like honey. However, we must take into account that during Middle Ages gum arabic was also a generic term to refer to any kind of gum used as a binding medium in painting techniques. In the case of manuscript illumination cherry and plum gums (the local ones 6) were frequently used. It is obvious that access to true gum arabic depended on commercial routes to and from Northeast Africa (Senegal still is the main gum Arabic producing country), something that not always was possible. Yet in the sixteenth century the famous Spanish physician Andrs Laguna, stated that in his time the ordinary gum arabic in drugstores unworthily have such a name [...] as it was born here, among us, from plums, pear, cherry and almond trees (Laguna 1556: 87). That is why, in this case, the commercial aspect was very important as, according Prosperus Alpini, the fact that the gum came directly from Egypt or Arabia then was the original, as they dont have plum, cherry or other gum trees (Alpini 1592: f. 5r). But the fact that it was very common to commercialize local fruit trees gums under the name of gum arabic does not mean necessarily that artists were tricked into using a gum of lower quality. It is well known that in Middle Ages the appearance of a substance prevailed over any other property and a gum that was a good binder in manuscript illumination could be perfectly considered as gum arabic. For example, in the text Hi sunt omnes colores we read that pigments for books were tempered with cherry gum, because it could be used either with water or wine and it could maintain
116
its strength even for a year (Novk 1996: 77). A century later, the author of the treatise De coloribus naturalia exscripta et collecta, gives a recipe to prepare an artificial gum arabic, using also cherry gum (Thompson 1935: 143). In the Liber de coloribus illuminatorum sive pictorum the author tells us that the gum from plum tree was also a good binder (Thompson 1926: 287). Cherry gum gives great transparency and brightness to colours, but it is only partially soluble in water (60% soluble fraction and 40% insoluble one; Perego 2005: 338). In the above mentioned text Hi sunt omnes colores cherry gum was prepared either by softened in water a whole night or by boiling it with water or with wine (Novk 1996: 77). Also Theophilus recommendation that cherry and plum gums should be left in sun, during summer, or close to fire, during winter, is probably due to this low solubility of both gums (Hawthorne and Smith 1979: 33). A prolonged boiling can transform them into a substance similar to gum arabic (Perego 2005: 340). The advantage of using cherry gum in tempering pigments for illumination is that it is less viscous than gum arabic, although cherry gum films chips easily if used alone and for that reason it has to be emulsified also with honey or fig tree sap. It seems that also almond tree gum was used a binder, but less frequent than cherry or plum tree gums. For example, in the text Tractatus qualiter quilibet artificialis color fieri possit almond gum was mixed with glair in a gilding process with gold leaves (Thompson 1934-5: 467-8). Gum tragacanth (adragante, draganto) appears in De arte illuminandi as a binder in manuscript illumination but its use is anecdotic (its adhesive power is less than other gums) and only as an additive to other gums or glair (probably as an emulsifier) for the tempering of blue pigments (Bolognese manuscript, Merrifield 1967: 410) or in parchment gilding (De arte illuminandi, Brunello 1992: 45). The use of tragacanth gum is problematic due to, first, the fact that it is only partially soluble in water and it is one of the most viscous vegetables gums; and, second, it gives mat pictorial films (Perego 2005: 334-335).
Glues
Glues of animal origin were used since Antiquity in art techniques (Pliny, Historia Naturalis: 28.236), particularly in woodworks, sculpture, panel or wall painting, textiles. Animal glues derive from collagen, a protein present in skins, bones, and connective tissue. Various types of animal glues are available, according to animal or the part used. In illumination techniques parchment size (cola pergamena) and fish glue (cola piscium, ichtyocola) are the ones that most often appear in medieval treatises on art technology, mainly in gilding, chrysography, tempering certain colours or as additives to glair or gum arabic. Recipes for its preparations are very common in medieval treatises like De arte illuminandi (Brunello 1992: 93), Livro de como se fazen as cores (Blondheim 1930-1: 82) or De coloribus faciendis (Van Acker 1972: 190). Parchment size was considered as the highest quality of all animal glues and it was prepared by boiling parchment clippings, until broth got concentrated (Heraclius treatise, Merrifield 1967: 230); then it was strained through a piece of cloth and allowed to cool. Parchment size was used mostly in gilding on parchment. Parchment size does not
117
spread so easily as gums or glair, so they are easier to use when precision is needed but shrinkage is considerable and plasticizers should be used (Horie 2005: 143). Fish glue (or isinglass when it was prepared from swim bladders) was prepared in a similar way, by boiling skins or bones in water, although for the purest form the swim bladders were used. The most famous fish glue was the one made from the sturgeon fish (Heraclius treatise, Merrifield 1967: 192), but other fishes were also used, like pike or eels (Il libro del arte, Brunello 2002: 148-9; Schedula of Theophilus, Dodwell 1961: 29). It seems that fish glue was a common binder in manuscript illumination on parchment, as the eighth century manuscript Compositiones ad tingenda quotes (Hedfors 1932: 33), although in latter treatises fish glue was used normally in gilding techniques, chrysography and argyrography (Experimenta de coloribus, Merrifield 1967: 56). Fish glue is transparent, with good adhesive power and it does not darken as do other animal glues (Colombini and Modugno 2009: 168) and its solutions resist gelling in room temperature (it can melt even at 6 C while parchment size has to be kept warm during use, as it needs 30-50 C; Horie 2005: 143). Fish glue is too sensible in the presence of salt (Perego 2005: 220), so it is incompatible with pigments like salt verdigris. Cheese glue (or casein glue) was another animal protein glue based on casein and obtained from milk or cheese, mixed with an alkali 7 (Horie 2005: 144). If the alkali is lime the adhesive is highly water resistant (Gettens and Stout 1966: 8).
7. See for example the recipe for cheese glue in De coloribus faciendis (Van Acker 1972: 180) or in Il libro dell arte de Cennino Cennini (Brunello 2002: 151-2). 8. Francisco Pacheco and Antonio Palomino, the well known Spanish painter and writer, quotes that by adding branches and leaves of fig tree to parchment size, the glue remained liquid and did not need heating (Pacheco 2001: 451-2; Palomino 1724: 80). However, still it is not very clear the purpose of the use of fig tree sap and alternative explanations can be found like fungicide replacing the use of vinegar or because it clots and holds together egg temple (Villarquide Jevenois 2004: 392).
Additives
Additives to modify the properties of binding media and facilitate their preservation were two very important aspects in medieval workshop praxis. That is why a series of substances were added in the tempering mixture. Increase the flexibility of the adhesive film once dried was a key point, as the turning of book pages could deteriorate pigments. Honey and sugar was the most employed additives to prevent binders becoming brittle (Borradaile 1966: 59). According the author of De arte illuminandi an illuminator should have always prepared a water of honey or a water of sugar as an additive to glue or glair (Brunello 1992: 100-103). But the same author also warn not to put too much honey otherwise pigments will get spoiled as they will not dry easily (Brunello 1992: 81). Fig tree sap also provided flexibility and, moreover, due to latex, could increase adhesive power and water-proof the painting film 8 (Blondheim 1930-1: 80; Libri 1849: 765; Clarke 2011 1.22.1B; Pomaro 1991: 120, Merrifield 1967: 475). Only rarely fig tree sap was used alone as a binder, as in a case of yellow colour made by orpiment and sulfur (Experimenta de coloribus, Merrifield 1967: 96). In animal glues wine and vinegar (alcohol) could prevent the formation of a gel at room temperature. Vinegar was added to gum arabic (aceto gumato) when used with colorants in order to control pH and consequently the tone of the colour (Pomaro 1991: 115). Small quantities of vinegar added to animal glues facilitate their solubility as an agent against gelification (Perego 2005: 216-7). Finally to prevent from mould and insects, arsenic, camphor, clove, myrrh or even orpiment (in glair) were added (Brunello 1992: 97; Wallert 1995: 42; Borradaile 1966: 29; Merrifield 1967: 676).
118
The tempering of pigments

The basic technical principle that always should be followed in the tempering of pigments and colorants was that for illumination they should be prepared in order to be applied more than one time, in order to create light and shadows effects. Theophilus in his treatise De diveris artibus made it quite clear in stating that in a book all pigments should be applied twice, first very thinly, then more thickly; but only once for letters (Hawthorne and Smith 1979: 38). In this regard it is understandable the choice of gum Arabic as the only ink binder: an adhesive that technically is stronger than glair (good covering) but weaker than animal glues (less crackling possibilities) and, moreover, saturates pigments increasing their intensity. On the contrary, in the illumination of manuscripts technical aspects such as the necessary drying time of the pigment or its opacity were more important. The same Theophilus argues that if someone wanted to spread up his work, then he had to use cherry or plum tree gum, as it dries faster than the others (Hawthorne and Smith 1979: 33). And Heraclius in his De coloribus et artibus romanorum recommends to use instead of oil yolk, otherwise orpiment will never dry (Merrifield 1967: 235). Cherry or plum gums were added to other binders to control fluidity, as we see in Mappae clavicula (Phillipps 1846: 223) or in De coloribus faciendis (Van Acker 1972: 195). The technique where both glair and gum Arabic were used as binding medium allowed a slow and careful work, as the pigment could run easily from the pen (Liber diversarum arcium, Libri 1849: 746; Clarke 2011 1.3.17A) and tiny strokes could be employed for details, as colour would be bright and opaque enough. Finally, the anonymous author of the Strasbourg manuscript warns about the necessity to control the relationship between binding media (glue) and pigment (vermilion) for a satisfactory result (Borradaile 1966: 23-24). Manuscript illumination was a sumptuary art, thus aesthetic criteria had also to be taken into account. Probably the most important aesthetic aspect was the fact that intense and brilliant colours were synonymous with richness and beauty. We can see this in several medieval illuminated manuscripts where scenes were developed in intense gold or blue backgrounds, with detailed representations of clothing, jewellery, furniture, etc. Technically this means that the artist had to work with pigments and binders that give intensity and lustre with opaque or semi-opaque pictorial layers. In this sense it should be interpreted the reference in the small treatise on book illumination that precedes the main text of Mappae clavicula (a later addition commonly referred to as De coloribus et mixtionibus) that all colours on parchment should be spissi et clari (Phillipps 1846: 188). Such a preoccupation is very common in treatises on art technology. For example, the author of De clarea states that vermilion tempered with yolk will be very bright and that it could be used to dye low quality parchments, so it looked like purple has been employed (Thompson 1932: 71). Centuries later, painter Pierre Lebrun and author of the so called Brussels manuscript (1635 AD) quotes that gum is used in illumination because gives lustre and brilliancy to colours (Merrifield 1967: 784). Some technically interesting results of a survey on medieval treatises on art technology regarding pigments and their binders in manuscript illumination will now be detailed.
119
Casein glue was used traditionally for woodwork (Hawthorne and Smith 1979: 26) or as an additive to earth plasters or as binding media in the a secco wall painting technique. In manuscript illumination its use in not so frequent and appears almost exclusively in the preparation of the pigment folium on parchment (De coloribus faciendis, Van Acker 1972: 180; Liber diversarum arcium, Libri 1849: 757; Clarke 2011 1.13.3-4), probably because folium changed colour according to acidity or alkalinity of its environment (rubeum, purpureum or saphireum) and casein does not denature in strong alkaline solutions. Theophilus used folium purpureum without any tempering, but he warned that once illumination was finished, the whole area had to be varnished with old glair (Hawthorne and Smith 1979: 40). Casein glue of a high pH values is incompatible with pigments that contain aluminium (Perego 2005: 166) so it could not be a binding media for any lake pigment. Normally lead white and verdigris were used on parchment diluted with wine or vinegar (Liber de coloribus illuminatorum sive pictorum, Thompson 1926: 293; Mappae clavicula, Phillipps 1846: 189). But, as master Peter of St. Omer stated in his De coloribus faciendis, this was right only for artificial greens and whites prepared without salt 9 (Van Acker 1972: 177). Well known is Theophilus recommendation not to use salt verdigris on books (Hawthorne and Smith 1979: 38) because it is too reactive, and probably this is the reason of master Peters previous recommendation. Nevertheless, it is also equally common to find recipes where lead white and verdigris were tempered with glair or egg yolk (Liber de coloribus illuminatorum sive pictorum, Thompson 1926: 289; De diveris artibus, Hawthorne and Smith 1979: 38; De arte illuminandi, Brunello 1992: 105; Il libro del arte, Brunello 2002: 101) 10. The reason of this could be the fact that yolk is a fat medium and glair, once dry, is impermeable to air and humidity; therefore, they would be good binders for reactive or artificial pigments like the above mentioned lead white and verdigris or orpiment and vermilion (Experimenta de coloribus, Merrifield 1967: 234; Liber diversarum arcium Libri 1849: 752; Clarke 2011 1.7; Bolognese manuscript, Merrifield 1967: 502) that could interact easily both with their environment or with nearby pigments. Gum arabic was used as a binder for verdigris on parchment or paper mostly in the not corrosive version, where the pigment was mixed with the juice of gladiolus (De coloribus diversis, Merrifield 1967: 286). It has already been mentioned that yolk gives intensity to colours, that is why azurite and vermilion were also tempered like this. However Vasari in his Introduzione alle tre arti (Pittura, cap. V) comments that earlier painters were tempering blues with animal glue because yolk turned colour to greenish hues (Brunello 2002: 151). Another reason is given by the author of the Liber diversarum arcium who explains that colours tempered with a mixture of egg yolk and glair would flow better from pen or quill (Libri 1849: 746; Clarke 2011 1.3.17A). It seems that there was also a tendency to temper some vegetable colorants, like indigo or saffron, and lake pigments (like brazil) with glair (De coloribus faciendis, Van Acker 1972: 181; Experimenta de coloribus of J. Le Begue, Merrifield 1967: 54; De arte illuminandi, Brunello 1992: 113, 127). A possible reason could be the fact that normally alum was used in their preparation. Technologically there is a certain af-
9. But the fact is that the same author contradicts himself as in the first recipe on salt verdigris he recommends the use of wine or vinegar. 10. Copper salt-based pigments like verdigris can react with proteinaceous material, as egg yolk and animal glues (Easteaugh et al 2004: 135).
120
11. Although we will only refer to binding media for parchment gilding, the same materials were also used to gild other surfaces (silver, tin, copper, wood, leather) and in other techniques (silver writing, gold leaf imitations, etc.).
finity in the use of alum and glair, as both were the principal ingredients for alumine zuccarino, a base for lake pigments and, moreover, alum was distempered always with glair (Easteaugh et al 2004: 12). Alum can form a complex with egg white, presenting a neutral pH and stabilizing it (Cunningham 1995: 298). Another reason could be the fact that pH of glair close to neutral (around 7,4; Perego 2005: 511) and it could not alter the tone in animal or vegetable colorants sensitive to pH changes.
Gilding media
Gilding was one of the most important techniques in manuscript illumination 11. Gold was applied in fine gold leaf or in powder and, according to the desired results, the surface was polished or not. In this case, workshop practises are more clear. For example, Theophilus in Schedula, or the anonymous compiler of Compositiones ad tingenda, write that when gold leaves were applied directly on the parchment glair should be used (Hawthorne and Smith 1979: 31; Hedfors 1932: 23, 31), as does the anonymous author of the Bolognese manuscript, but in this case fig tree sap is added (Merrifield 1967: 462); whereas Alcherius in his treatise De coloribus diverisis modis not only recommends parchment size, but also quotes that glair is more rigid and less flexible and that gilding could fall off from parchment or paper (Merrifield 1967: 269). For gesso mordants, gum or glue was used more frequently, because as the author of the Bolognese manuscript writes, glair can crack it; but he also gives a recipe according to the German manner where gesso and white clay were tempered with glair and fig tree sap (Merrifield 1967: 446, 474). Cennini also tempers gesso mordant in gilding on parchment or on paper with well beaten glair (Brunello 2002: 196-7). To apply gold powder Theophilus recommends parchment size or fish glue for gold powder either for writing (chrisographia) for use in gilding of illuminated manuscripts. But he warns, firstly to pay special attention in fish glue preparation as if it is left too thick during boiling gold will flake, and secondly to apply glue moderately otherwise gold leaf will lose brilliance and it will be spoiled (Hawthorne and Smith, 1979: 36). On the contrary, master Peter of St. Omer in his treatise De coloribus faciendis writes that gum arabic is excellent for gilding on parchment (Van Acker 1972: 192). In this binding media selection process other technical criteria were also taking into account, such as the type of parchment to be gilded: for parchment made of sheeps skin several a mixture of glair and plum tree gum or gum arabic was specially recommend, as this type of parchment is too fat and a stronger binding media might be needed (De coloribus faciendis, Van Acker 1972: 192). The anonymous author of the Liber illuminatorum sive pictorum adds also that such gilding method should be carried out in a damp place, especially in warm weather, otherwise burnished gold will get spoiled (Thompson 1926: 305).
Incompatibilities
The presence of tannins in the tempering process could be a great inconvenience. As was mentioned earlier, most authors when referring to gum arabic advised to use
121
the most transparent one. This was really an important aspect as the most coloured gums contain tannins, due to long contact with bark (Perego 2005: 336) and tannins can react with iron ions (iron based pigments or pigments where iron can be present as impurity) and the result is a dark coloration that could easily spoil the final colour. Moreover, parchment size and fish glue in the presence of tannins become insoluble (Perego 2005: 217). Gum Arabic is also incompatible with gelatin, mostly is presence of salt (Perego 2005: 336) and probably this is the reason why the mixture of parchment size with gum Arabic is not so frequent in the tempering of pigments.
12. The fact that these texts were written by practising artists indicates more a traditionalism of workshop practices than a mere copy of old contents.
Conclusions
Glair and gum Arabic were the main binders for pigments in medieval illumination techniques and they were used continuously throughout Middle Ages and it is imposible to establish any chronological criteria for their use. For example, there is a very common assertion that until the fourteenth century the widely used binding medium for illumination was glair and after was replaced with gum Arabic, due to aesthetic criteria, like the preference for bright and vivid colours since that century. However, medieval treatises on art technology do not corroborate such a hypothesis, as parchment size, glair and gum Arabic are mentioned constantly in texts from XV, XVI and even XVII century 12 (Bolognese, Marciana, Paduan or Brussels manuscripts, Merrifield 1967: 408, 610, 664, 786). In short, the use of glair or gum in the tempering process is a technical aspect that, together with the nature of pigments, determines the election of the two possible grosso modo illumination techniques as described by the Spanish painter and writer Pacheco in his Arte de la Pintura (1649). The author states that in his days in manuscript illumination the colour of parchment was used as light with transparent pigments and subtle tones (so gum arabic or one of its mixtures should be used), unlike what happened with what he calls the old temple where opaque and intense colours were used, closer to oil painting (so glair or one of its mixture were more suitable) (Pacheco 2001: 454-5). This optical behaviour of binders once tempered with pigments was well known and much employed by medieval illuminators as a technique to achieve different tones and hues in the same work. We have several examples of this use. The anonymous author of De clarea quotes that to control glair and water in the tempering process was very important as the artists could do things as he wished, glossy or mat (Thompson 1932: 75). In the Bolognese manuscript we read that for body colours blue should be tempered with animal glue or parchment size and vermilion with glair and fig tree sap; but if they supposed to be used in capital letters or decorative motives, for blue gum arabic or glair was more appropriate and for vermilion glair and yolk (Merrifield 1967: 408, 500). In the same way, the author of the Livro de como se fazen as cores advises to use gum for a dark tone in blue, or glair for lighter one (Blondheim 1930-1: 75). In this sense we have to interpret also the fact that there is no clear criterion on the
122
13. Francisco Pacheco (1649) and Vicente Carducho (1633) describe the aguadas de colores as the appropriate type of painting executed on paper (Pacheco 2001: 452; Carducho 1633: 132). The almost exclusive use of gum arabic is also found in Salmons Polygraphice (1685: 95, 447); Jenner`s A Book of Drawing, Limning, Washing or Colouring of Maps and Prints (1652: 20-1); and Flibiens Des principes de larchitecture, de la sculpture, de la peinture. (1690: 621).
use of a specific binder with each pigment, as most of them were tempered one way or another depending on the necessities of the illuminator or even on workshop (or school) traditions: Egg yolk increase intensity and gives depth; glair without varnish gives a flat mat finish; gum Arabic gives a glossy appearance; cherry gum makes colours brighter. That is why the binder should be the last ingredient to add in order to control the development of the tone, and when a gum or glue was used in the grinding process normally the pigment was washed to remove the excess of adhesive, and then it was tempered. Cennino Cennini makes a very interesting comment when he writes that for illumination on paper only gum arabic should be used (Brunello 2002: 198). Such an observation shows that technical limitations were equally (or even more) important in artistic praxis: paper is more flexible and absorbent than parchment, thus washing techniques and gum arabic as binder are more appropriate. The progressive use of paper as support for writing and illumination meant that, since the sixteenth century onwards, gum arabic appeared almost exclusively in printed treatises of painting techniques 13. And consequently it was felt that the gum arabic was the binding media par excellence even in medieval illumination. Probably the only valid conclusion we can draw is that artistic praxis was strongly influenced by traditions (local or international) and workshop experience. Distempering of pigments is a good example, as it is very common to find in medieval treatises expression like do what from your experience seems better to you (De arte illuminandi, Brunello, 1992: 105) or according to the choice of the artist and the nature of the work which is to be done (Experimenta de coloribus, Merrifield 1967: 110). A representative case is the use of fish glue as a binder for all pigments, recommended in the oldest treatises on art technology in the Occident such as the Compositiones ad tingenda (Hedfors, 1932: 33) and Mappae clavicula (Phillipps 1846: 218). Both works show a strong Byzantine influence and it seems that the use of fish glue can also relate to this tradition, as animal glues were very common in icon painting. Two centuries later things changed as in the small treatise on book illumination that precedes the main text of Mappae clavicula (the De coloribus et mixtionibus) where we read that all colours on parchment should be tempered with glair (Phillipps 1846: 188). As we have seen, medieval illuminators used binding media and pigments according to their specific characteristics and properties and according the illumination technique they thought appropriate in order to carry out their work. Moreover, illuminators guild rules obliged the use of specific pigments in different parts of illumination as, for example, azurite in pictures and only lapis lazuli in capital letters. This technical and historic aspects should be always taken into account, first, because many aesthetic studies and laboratory analysis dedicated to medieval illumination, are based on exceptional pieces, luxury books for wealthy patrons, that by no means could be representative of the common practise in manuscript illumination; and second, because the place where a sample was taken for analysis becomes really important as results can vary (different binders and pigments in capital letters, backgrounds, foliage, etc.) and conclusions might be wrong or over generalized.
123
Bibliography
Alpini, P. (1592): De plantis Aegypti liber. Venice: Franciscum de Franciscis Senesem (impr.). Borradaile, V.; Borradaile, R. (1966): The Strasburg Manuscript. A Medieval Painters Handbook. London: Alec Tiranti. Brunello, F. (1992): De arte illuminandi. Vicenza: Neri Pozza. Carducho, V. (1633): Dilogos de la Pintura, Origen, Esencia, Definicin, Modas y Diferencias. Madrid: Fr. Martnez (impr.). Clarke, M. (2001): The art of all colours. London: Archetype. Clarke, M. (2011) The Medieval Painters Methods (The Montpellier Liber Diversarum Arcium). London: Archetype Publications. Colombini, M.P. and Modugno, F. (2009): Organic Mass Spectrometry in Art and Archaeology. Chichester: Wiley. Cunningham, F.E. (1995): Egg-product pasteurization, in Stadekman, W.J. and Cotterill, O.J. (eds.), Egg science and technology. New York: Food Products Press: 289-315. Eastaugh, N; Walsh, V.; Chaplin, T.; Siddall, R. (2004): Pigment Compendium: A Dictionary and Optical Microscopy of Historic Pigments. Oxford: Elsevier Butterworth-Heinemann. Flibien, A. (1690): Des principes de larchitecture, de la sculpture, de la peinture. Paris: La Veuve and fils Jean Baptiste Coignard (impr.). Jenner, T. (1652): A Book of Drawing, Limning, Washing or Colouring of Maps and Prints. London: M. Simmons (impr.) Hawthorne, J.G.; Smith, C. (1979): Theophilus. On divers arts. New York: Dover. Hedfors, H. (1932): Compositiones ad tingenda musiva. Uppsala: Almqvist & Wicksells Boktryckeri-AB. Horie, C.V. (2005): Materials for conservation. Organic consolidants, adhesives and coating. Oxford: Butterworth-Heinemann. Laguna, A. (1556): Pedacio Dioscorides Anarzabeo, acerca de la material medicinal y de los venenos mortiferos traduzido de lengua griega en la vulgar castellana e illustrado con claras y substanciales Annotaciones y con las figures de numerosas plantas exquisitas y raras. Salamanca: Mathias Gast (impr.). Libri, M.; Ravaisson, F. (1849): Catalogue Gnral des Manuscrits des Bibliothques Publiques, vol. 1. Paris: Imprimerie Nationale. McGee, H. (2004): On Food and Cooking: The Science and Lore of the Kitchen. New York: Scribner. Merrifield, M. (1967): Original treatises on the arts of painting, v. I. New York: Dover Publications. Novk, A. (1996): Hi sunt omnes colores. Text from the 12th century from the Library of St Peters Monastery in Salzburg a XI 4, fol. 241, Technologia Artis, 4: 77-9.
124
Pacheco, F. (2001). El arte de la pintura. Madrid: Ctedra (edicin, introduccin y notas de Bonaventura Bassegoda i Hugas). Palomino de Castro y Velasco, A. (1724): El Museo Pictrico y Escala Optica, t. II. Madrid: Viuda de Juan Garca Infanzn (impr.) Perego, F. (2005): Dictionnaire des matriaux du peintre. Paris: Belin. Pomaro, G. (1991): I recettari del Fondo della Biblioteca Nazionale Centrale di Firenze. Milan: Editrice Bibliogrfica. Salmon, W. (1685): Polygraphice: or the arts of drawing, engraving, etching, limming, painting. London. Thompson, D.V. (1926): Liber de coloribus illuminatorum sive pictorum from. Sloane Ms. No. 1754, Speculum, I: 280-307. Thompson, D.V. (1934-5): Medieval color-making: Tractatus qualiter quilibet artificialis color fieri possit from Paris B.N., Ms. Latin 6749, in Isis (23): 456-468. Thompson, D.V. (1935): De coloribus naturalia exscripta et collecta, from Erfurt, Stadtbcherei, Ms. Amplonius Quarto 189 (XII-XIV century), Technical Studies in the Field of the Fine Arts III (3): 133-45. Thompson, D,V. (1956): The materials and techniques of medieval painting. New York: Dover. Tosatti Soldano, B. (1978): Miniature e vetrate senesi del secolo XIII. Genova: Universita di Genova (Collana Storica di Fonti e Studi, 25). Van Acker, L. (1972): Petri pictoris carmina, in Corpus Chistianorum Continuatio Mediaevalis, XXV. Turnhout: Brepols: 145-246. Villarquide Jevenois, A. (2004): La pintura sobre tela. I. San Sebastin: Nerea. Walert, A. (1995): Libro secondo de diversi colore e sise da mettere a oro, in Wallert, A.; Hermens. E.; Peek, M. (eds.), Historical Painting Techniques, Materials and Studio Practice. Los Angeles: Getty Trust Publications.
Acknowledgments
I would like to thank Mark Clarke for his comments and his open-handed exchange of knowledge.
Biography
Dr. Stefanos Kroustallis is a researcher into historic art materials and techniques, especially in medieval sumptuary arts. He is a graduate in history (University of Athens, Greece), in conservation and restoration (Escuela Superior de Conservacin y Restauracin de Bienes Culturales of Madrid, Spain) and he received his doctoral degree (Complutense University, Madrid) on medieval art technological source research. Currently he is working on developing a Thesaurus data base on art materials and techniques for museum cataloguing.
125
Resumo
Em trabalhos anteriores, estudaram-se os pigmentos azuis e verdes, de manuscritos franceses, por microscopia Raman. Este artigo, para alm de incluir os resultados mais relevantes que entretanto foram publicados, prope uma cronologia para o uso dos pigmentos azuis, ndigo, lpis lazli e azurite. No foi possvel a identificao dos pigmentos verdes em manuscritos dos scs. X-XI, por microscopia Raman, devido forte emisso de fluorescncia que domina os espectros. So ainda discutidos outros pigmentos verdes, analisados noutros manuscritos, sugerindo-se a possibilidade de se tratar de verdigris, dado o seu teor em cobre.
palavras-chave
identificao anlise fsico-qumica pigmentos azuis pigmentos verdes
Abstract
Blue and green pigments in French manuscripts have been investigated in previous works by Raman microspectrometry. Including the most significant published results, this report reveals a chronological use for the blue pigments, indigo, lapis-lazuli and azurite. Raman microscopy was unable to identify green pigments in X-XI manuscripts, as spectra are dominated by a strong fluorescence. Different green pigments, identified in other manuscripts, are considered. Because of their copper-content, the hypothesis of verdigris is suggested.
key-words
identification analyse physico-chimique pigments bleus pigments verts
a la recherche des pigments

cl au de co upry
Ingnieur honoraire, Centre National de la Recherche Scientifique (France)
1. UMR CNRS Universit Paris VI Pierre et Marie Curie.
A la fois matire et beaut, les pigments illuminent de leurs couleurs les manuscrits. Pour le copiste, le peintre, lartiste, comptent, avant tout, leur teinte, leur clat, leur stabilit, souvent leur prix, parfois leur symbolique, rarement leur nature. Au contraire, lhistorien va sintresser particulirement cette dernire caractristique. Produit naturel ou prpar? Sil sagit dun produit naturel, do vient-il? Est-il de provenance locale ou a-t-il ncessit un long voyage? Par quelles routes commerciales, traditionnelles ou de trac rcent? Pour un produit fait de main dhomme, par quel(s) procd(s), partir de quels composs? Les changements observs dans le choix des pigments utiliss sont dun grand intrt et les raisons peuvent en tre multiples: produits nouveaux ou artisans novateurs, amlioration ou changement dans les techniques, effets de mode A la varit des produits, sajoutent les possibilits de leur mise en uvre dans des mlanges aux tons subtils. Une recherche sur la nature des pigments dcorant les manuscrits apparat ainsi comme riche dinformations pour lhistoire des techniques, du commerce, des pratiques artistiques. Soulignons que les conclusions ont un sens lorsque que les manuscrits sur lesquelles elles sappuient sont parfaitement dats et localiss. La collaboration entre historiens de lart et analystes est indispensable pour lintgration des rsultats dans une problmatique historique. Notre propos nest pas de parcourir lventail exhaustif des produits et des palettes utiliss, mais de dresser un bilan simple sur les pigments de deux couches colores, bleues et vertes, qui tmoigne de lapport et des difficults de ces tudes. De mme, un nombre restreint de manuscrits, tudis pour la plupart au Laboratoire de Dynamique, Interaction et Ractivit 1 (LADIR), est prsent, les plus significatifs pour tayer notre dmarche. Plusieurs approches, correspondant des niveaux diffrents dinformation, permettent dtudier les pigments dans les manuscrits mdivaux, Lobservation lil nu puis sous microscope est une premire tape indispensable alors que son intrt est souvent sous-estim. Lapprciation de la forme et de la taille des grains de pigments fournit des lments de caractrisation. Les produits naturels minraux, obtenus par broyage de minraux, de dimensions suprieures au micromtre, prsentent des
127
t h e m at e r i a l s o f c o l o u r
formes anguleuses alors que les grains du mme produit, obtenu par synthse, sont souvent de forme arrondie et de dimensions infrieures au micromtre. La couleur peut tre diffuse ou concentre dans des particules discrtes. Ainsi, lorpiment est aisment reconnaissable par ses longues aiguilles jaune brillant. Ces premiers indices peuvent tre confronts la liste des produits potentiels, fournie par le dpouillement des rceptaires mdivaux, mais de probabilit dutilisation inconnue. Une autre approche consiste procder une exprimentation: un pigment est prpar, associ un liant et dpos sur une feuille de parchemin (Roosen-Runge 1967). La teinte obtenue est ensuite compare celle de manuscrits mais similitude de teinte ne signifie pas identit des pigments. Une technique photographique innovante (Isacco 2008) a t exprimente pour lidentification de certains pigments de miniatures indiennes, base sur ltude simultane de clichs en lumire infrarouge, naturelle et ultraviolette. Cependant lobservation, lexploitation des textes, la reconstitution des produits fournissent des suggestions, non des certitudes (Porter 1995).
2. Travaux raliss en collaboration troite avec Marie-Thrse Gousset, ingnieur la Bibliothque nationale de France. Quelle trouve ici lexpression de nos remerciements.
Approche analytique
Pour identifier avec certitude les pigments, il est ncessaire davoir recours lanalyse. Le choix de la technique analytique est guid en fonction de certains critres, dont le plus important est le respect de louvrage. Les techniques physico-chimiques simposent alors comme moyen dinvestigation. Parmi celles-ci, lanalyse par microspectromtrie Raman (Coupry et al. 1996) prsente des caractristiques particulirement intressantes: totalement non-destructive, elle ne ncessite aucun contact entre lappareil de mesure et la zone analyse. Les donnes obtenues, le spectre Raman, permettent didentifier, sans ambigut dans la trs grande majorit des cas le compos analys, par comparaison avec des spectres de rfrence. Nanmoins, deux causes peuvent empcher lidentification: lobservation Raman ne permet pas de caractriser des bandes avec une intensit suffisante ou lorsquun spectre est obtenu, il ne figure pas dans la bibliothque de rfrences. Lanalyse est faite sur une surface micromtrique (de lordre de 10 m2), correspondant la taille des grains de pigments, sur un micro-prlvement ou directement in-situ sur le manuscrit. Les deux modes exprimentaux ont des avantages et des inconvnients. Les tudes faites au LADIR 2 ont t ralises essentiellement sur des micro prlvements, le mode in-situ nayant utilis que dans le cas dun feuillet isol. Les micro chantillons sont prlevs des emplacements choisis avec soin, la fois sans risque pour le manuscrit et reprsentatifs de la couche picturale, comme la dcharge dune lettre sur le feuillet en vis--vis.
Pigments bleus
Les pigments bleus participent avec une grande frquence au dcor des manuscrits do la possibilit dune large enqute sur des manuscrits dorigine, dpoque, dimportance trs varies. Seuls, les manuscrits ayant fait lobjet danalyses ont t
128
3. Communication prive.
retenus. Ces pigments ont fait lobjet de recherches sur un groupe de manuscrits crits au XIIe sicle labbaye de Saint-Pierre de Corbie (Guineau et al. 1986). Une partie importante de la production de son scriptorium est conserve la Bibliothque nationale de France, ce qui a permis dtudier cette production sur une dure dun sicle. Le bleu est rare dans les premiers manuscrits, trs frquent dans les derniers; ple ou fonc, il peut avoir une teinte grise. Les prlvements ont t choisis de faon explorer au mieux lventail des situations: diffrentes teintes, lettres histories ou dcoration secondaire, pour les lettres filigranes, dans le filigrane ou le corps de lettre, simples lettrines. Quelque soit la localisation du prlvement, celui-ci montre sous microscope la prsence de grains microscopiques dun bleu intense et satur au sein dune matrice blanche. Quelque soit le grain bleu analys, un seul produit a t identifi, le bleu de lapis-lazuli, alumino-silicate de sodium polysulfur. Ce pigment minral est extrait dune pierre semi-prcieuse, le lapis-lazuli, dont le gisement dAfghanistan tait dj exploit lpoque mdivale comme en tmoigne Marco Polo dans le Livre des Merveilles. Aucun pigment blanc na t mis en vidence, les teintes claires sont obtenues par dilution dans le liant. Par contre, les teintes grises rvlent un mlange de grains trs bleus et de particules noires, identifies comme tant des particules de noir de carbone. Ce rsultat atteste de faon formelle la prsence de ce pigment bleu en Picardie ds le XIIe sicle. Elle a t confirme dans une zone gographique plus large, la Bourgogne, dans le scriptorium de labbaye de Cteaux 3. Prsent dans les manuscrits aux riches couleurs du dbut du sicle, le bleu de lapis-lazuli y a t galement identifi dans la production du milieu du sicle, au dcor monochrome en accord avec llan asctique suscit par saint Bernard, qui na cependant pas affect le choix du pigment. Lenqute se poursuit par ltude dun scriptorium sur une priode longue (Coupry 1999). Celui de labbaye de la Trinit de Fcamp (Normandie) a eu une production importante, bien conserve et bien documente. Fonde en 658, labbaye fut dtruite en 842; la vie religieuse y est restaure avec larrive de chanoines avant 990, suivie par celle de moines clunisiens en 1001. Quatre abbatiats vont se succder jusquen 1139 avec des cycles de grande activit intellectuelle et artistique, en relation avec les centres anglais et les abbayes normandes voisines, dont le MontSaint-Michel. Lensemble des manuscrits de Fcamp de cette priode conservs la Bibliothque nationale de France a t pris en compte. On observe une utilisation du bleu avec des frquences diffrentes, jamais dans plus de dix ouvrages, une seule fois dans quelques-uns jusqu trs abondamment dans certains. Neuf manuscrits, jugs comme reprsentatifs de ces diffrentes priodes, ont t slectionns en privilgiant labbatiat de Jean dAlie (1028-1078) en raison de son intrt artistique et historique. Comme pour ltude des manuscrits de Corbie, les prlvements ont t raliss dans les diffrents types de lettres et de teintes, les rubriques, ainsi que dans les diffrentes mains des manuscrits. Sur lensemble des manuscrits, deux pigments bleus ont t identifis par leurs spectres Raman qui les diffrencient sans aucune ambigut. Un des pigments, lindigo,
129
est un produit colorant bleu, naturel pour les poques qui nous intressent, utilis pour teindre les textiles. Dorigine vgtale, il est obtenu partir de plantes de trs nombreuses familles dont le pastel des teinturiers, indigne en France, nomm gude au nord et pastel au sud, cultiv en Normandie et Picardie au Moyen ge, ou Toulouse dont il fit la richesse aux XVe et XVIe sicles. La molcule responsable de la couleur est lindigotine. Lindigo nest utilis que dans un seul manuscrit de Fcamp, dat de la fin du Xe sicle, lexclusion de tout autre pigment bleu, quelque soit la teinte de la couche picturale, bleu franc ou bleu-vert. Lobservation ne met pas en vidence des grains dautres couleurs, qui expliqueraient ces variations, attribues alors au liant et/ou une mise en uvre diffrente. Le bleu de lapis-lazuli est le pigment bleu prsent dans le dcor de tous les autres manuscrits. Tout oppose ces deux produits naturels, leur origine et leur structure, organique et vgtal pour lun, minral pour lautre, produit localement ou de provenance lointaine, et, malgr labsence de documents comptables, de prix que lon peut supposer diffrents. La prsence de pigment bleu peut relever dun choix du copiste, comme le montrent les manuscrits lat. 2253 et 3776. Les deux mmes copistes ont particip leur ralisation, lun utilise frquemment le bleu, lautre jamais. Ces rsultats tablissent avec certitude la prsence de bleu de lapis-lazuli en Normandie-Picardie pour la priode des XI-XIIes sicles. Il importe alors de considrer des manuscrits dautres origines. La page de garde dun recueil des coutumes et franchises de Narbonne, dat de 1221, crit et conserv Narbonne, est dcore denluminures de qualit mdiocre. Les fonds bleus des vtements, plus sombre pour le manteau de la Vierge que pour la tunique de saint Jean, sont raliss avec un mme et seul pigment, lindigo, mais avec des concentrations diffrentes (Cardon 2000, 111). Un tout autre pigment bleu a t identifi dans des manuscrits gographiquement proches et postrieurs. Ainsi, dans un sacramentaire du XIVe sicle de lvch de Carcassonne (sud de la France), une splendide enluminure se dploie sur deux feuillets en vis--vis. La robe du Christ prsente trois tons de bleu, pour lesquels lanalyse (Cardon 2000, 163) identifie un seul pigment, un produit minral naturel, lazurite. Les principaux gisements de ce carbonate basique de cuivre se rencontrent en France, en Italie, Chypre et surtout en Allemagne do son qualificatif dazur dAllemagne. Les teintes claires ne sont pas obtenues par dilution mais par ajout dun pigment blanc, le blanc de plomb. Plus tardivement, dans une Gographie de Ptolme du XVe sicle 4, la surface des ocans est peinte avec de lazurite, ce qui tmoigne du large usage de ce pigment. Des publications dun trs grand intrt permettent dtendre notre enqute vers des manuscrits princiers, dcors par des artistes au renom prestigieux soit du dbut soit de la fin du XVe sicle. La palette du Livre dHeures du marchal Boucicaut, ralis au dbut de la dcennie 1410, a t analyse (Villela-Petit et al. 2003) par spectromtrie dabsorption par rflexion diffuse dans le visible et par spectromtrie de fluorescence X, deux techniques complmentaires non-destructives. Le bleu de lapis-lazuli claire de son clat les vtements de la Vierge et des saints tandis que lindigo est rserv pour les personnages peu nobles, toujours en mlange avec le
4. Paris, BnF lat. 4801, Communication prive.
130
5. Evangelia, BnF latin 270 f.106v. 6. Sanctorale, BnF latin 1205 f.92v. 7. Vita sancti Wandregisli, BnF Nal 18315 f.24v.
blanc de plomb. La prsence de ces pigments a t confirme dans deux autres manuscrits dcors attribus au Matre de Boucicaut, le Livre des proprits des choses, de Barthlemy lAnglais, et le Brviaire de Chteauroux, dans lequel il intervient aux cts du Matre de Bedford et du Matre dOrose, avec des produits bleus similaires pour les trois enlumineurs. A la fin du sicle, la production de Jean Bourdichon stend sur plus de trente ans. Trois manuscrits couvrant ltendue chronologique de son activit montrent par spectromtrie Raman (Trentelman et al. 2008) la prsence de bleu de lapis-lazuli dans tous les feuillets analyss. Mais lartiste a aussi recours un mlange faisant intervenir lazurite, principalement dans le manuscrit ralis au dbut de sa carrire, avant quil devienne peintre officiel de la cour de France. Notons que lindigo na t observ dans aucun des trois manuscrits. Un rapprochement peut tre fait avec la stratigraphie des couches picturales de la chapelle des Moines Berz-la-Ville (Bourgogne, France). Cet difice, construit pour labb Hugues de Cluny dans la deuxime moiti du XIe sicle, prsente des repeints XIV-XVe sicle. Le pigment de la couche originelle est le bleu de lapis-lazuli, celui de la couche postrieure est lazurite. Il est ncessaire dtendre nos investigations vers des poques antrieures afin de prciser lintroduction du bleu de lapis-lazuli en Europe occidentale. Un manuscrit crit labbaye Saint-Germain dAuxerre (Bourgogne, France), le Commentaire dHaymon dAuxerre sur Ezchiel, prsente sur le feuillet 2r une scne ddicatoire, labb Helric prostern devant saint Germain. Les deux personnages sont vtus de bleu (Coupry 1990), bleu vif pour lun et bleu sombre pour lautre, identifis comme tant du bleu de lapis-lazuli pour le saint patron de labbaye et de lindigo pour labb. Le choix de ces deux pigments souligne la hirarchie des personnages et la prsence en France vers lan mil du bleu de lapis-lazuli, dj identifi Fcamp est confirme. Trois manuscrits du nord de la France tmoignent du choix des pigments bleus disponibles: les deux premiers, copis Corbie au IX e sicle, ont t dcors avec du bleu de lapis-lazuli 5 pour lun et de lazurite 6 pour lautre, le troisime, dat du VIIIme sicle, la t avec de lindigo 7. A ce groupe de trois produits, va sajouter de faon exceptionnelle un autre pigment, identifi seulement dans quelques manuscrits. LEvangliaire de Charlemagne, dit aussi Evangliaire de Godescalc, du nom du scribe qui la ralis, est un ouvrage prcieux, crit vers 781-783 en lettres dor et dargent sur parchemin pourpr et magnifiquement illustr. Il contient en particulier la premire reprsentation de la Fontaine de Vie au symbolisme fort. Le manuscrit a fait lobjet dune tude trs complte (Roger 2007). Lanalyse des couches picturales bleues montre la prsence dindigo dans lensemble de louvrage lexception de la Fontaine de Vie dont les paons sont peints avec du bleu gyptien. Ce produit fabriqu en Egypte ds 3000 BC est le grand pigment bleu du monde mditerranen antique et disparat de la palette des artistes une date encore imprcise, estime vers les VII-VIIIes sicles. Actuellement ce pigment a t identifi par le mme chercheur dans deux autres manuscrits. Par contre, la Fontaine de Vie des Evangiles de Saint-Mdard de Soissons, manuscrit luxueux ralis vers 805, donc
131
quasiment contemporain, est ralise avec du bleu de lapis-lazuli dont ce serait une toute premire attestation dutilisation. Les recherches sur un autre manuscrit de grande importance, les Evangiles de Lindisfarne, dat de 715-721 et conserv la British Library (Londres), illustrent les difficults et lintrt de lidentification des pigments. La premire tude ralise en 1956 identifia par comparaison visuelle deux pigments bleus: lindigo et le bleu de lapis-lazuli, ce qui repoussait lutilisation de ce dernier au dbut du VIII me sicle. En 2004, ltude fut reprise (Brown et al. 2004), qui montra la seule prsence de lindigo, quelque soit la nuance de la couche bleue, et infirma lutilisation de bleu de lapis-lazuli. Ce rsultat souligne la fragilit dune identification fonde sur une observation visuelle. A partir de ces rsultats, il est possible de tirer un premier bilan, certes complter et prciser en fonction de nouvelles analyses et de nouvelles identifications. Le schma suivant propose lutilisation chronologique des trois principaux pigments bleus dans le dcor des manuscrits occidentaux, tablie sur un nombre restreint de manuscrits analyss.
prsence atteste de pigments bleus dans des manuscrits mdivaux deurope occidentale
Il ne faut pas ngliger lutilisation possible de pigments locaux. Ainsi larinite, minral argileux qui doit son nom sa couleur bleue, se trouve dans quelques gisements des Pyrnes. Il a t identifi en peinture murale proximit de son lieu dexploitation dans quelques sites des XI-XII es sicles, dont le logis abbatial de la grande abbaye de Moissac (Daniel 2008). A-t-il pu tre utilis galement pour des manuscrits? La recherche reste faire. Lidentification de ces produits conduit des problmes spcifiques. Pour lindigo, la caractrisation de lespce vgtale, comme pour lazurite, la dtermination du gisement dexploitation, semblent irralisables avec les possibilits techniques actuelles. Nous avons tent de diffrencier par spectromtrie Raman les diffrents bleus de lapislazuli (Torrs-Bourdel 2003) soit sur des chantillons provenant de manuscrits soit sur des chantillons minraux. Dans une premire dmarche, nous nous sommes intresss la couleur, due la prsence de deux chromophores, des ions polysulfures, dont on peut valuer par spectromtrie Raman les proportions relatives. La dcomposition et lajustement des bandes principales des chromophores montrent des diffrences
13 2
qui suggrent des variations dans les environnements lectroniques des ions, dont linterprtation est poursuivre. Dautre part, le spectre Raman obtenu avec de nouvelles conditions exprimentales montre des bandes attribues des ions polysulfures plus condenss, ce qui ouvre galement des perspectives de recherche prometteuses.
Couches picturales vertes

Ltude des couches vertes dans la dcoration de manuscrits rpond une problmatique trs diffrente: elle est centre sur la dtermination de leur nature dans les neuf manuscrits de Fcamp, dj prsents. Cette couleur, moins prsente. que le rouge et le bleu, y occupe nanmoins une place importante, dans diffrents types de lettres, de la plus labore la plus simple, avec une grande diversit de tons. Ainsi les verts peuvent tre dcrits comme vert clair franc, vert fonc terne, vert olivtre, vert vif dune nuance meraude, sans que nous ayons corrl ces termes des donnes colorimtrique. Certaines couches sont corrosives vis--vis du parchemin, dautres scaillent. Cette grande diversit recouvre-t-elle une diversit de produits? Ceci a conduit effectuer un grand nombre de prlvements, plus de trente, rpartis au mieux pour tenir compte de ces diffrentes observations. Or, quelque soit lchantillon et les conditions exprimentales, lobservation Raman a toujours t masque par un phnomne de trs grande intensit, la fluorescence, interdisant daccder la moindre identification. Un comportement identique a t rapport pour des manuscrits dorigine varie et antrieurs trs souvent au XIIIe sicle. Le problme nest donc pas limit un scriptorium et acquiert, de ce fait, une porte plus gnrale pour la connaissance des pigments verts. Lors dune recherche prcdente (Coupry 2007) par microspectromtrie Raman sur deux manuscrits coptes, les pigments verts de lun dentre eux navaient pu tre identifis, tandis le spectre Raman des grains verts du second manuscrit tait la combinaison des spectres de lindigo et dun pigment jaune, lorpiment, la couleur verte observe tant la combinaison des couleurs bleue et jaune. De faon analogue, lobtention dune teinte verte par le mlange Bleu/Jaune a t observe dans de nombreux manuscrits. Plusieurs combinaisons sont possibles et refltent les possibilits dapprovisionnement en lun ou lautre des constituants. Indigo/orpiment est identifi ds 715-720, dans les Evangiles de Lindisfarne (Brown et al. 2004). Si le mlange orpiment/azurite, chimiquement instable, na pas t identifi, dautres combinaisons ont t observes, comme bleu de lapis-lazuli/orpiment (Villa-Petit et al. 2003) ou le pigment de synthse jaune de plomb et dtain associ lazurite (Vandenabeele et al. 1999) ou au bleu de lapis-lazuli (Burgio et al. 1999). La composition peut tre extrmement variable, indigo/orpiment avec des proportions diffrentes, mlanges binaires jaune/vert ou bleu/vert, mlanges ternaires, pour offrir une trs grande varit de couleurs (Cennini 1978, 32). Lanalyse lmentaire sur des verts de manuscrits de Fcamp montre la prsence de cuivre, ce qui limine lhypothse de terres vertes, aluminosilicates de fer, potassium et magnsium. Plusieurs sels de cuivre peuvent tre utiliss comme pigments: les
133
spectres Raman du carbonate basique, la malachite, ou de sulfate permettent gnralement leur identification, ainsi pour les verts dans des manuscrits de Jean Bourdichon (Trentelman et al. 2008) qui ny utilise pas le mlange bleu/jaune. Les sels le plus frquemment cits dans les traits de peinture mdivaux sont les actates, souvent regroups sous le terme de vert-de-gris. La littrature technique, antique et mdivale, est riche de nombreuses recettes de fabrication, bases sur laction de vinaigre sur des feuilles de cuivre. Les ractions sont lentes et complexes, dpendant de nombreux facteurs, plusieurs produits peuvent tre obtenus, chimiquement proches et pouvant voluer dune forme lautre. Ils sont verts, bleus, turquoise La composition du pigment est souvent non stchiomtrique. Pour modifier la teinte ou les proprits du pigment, les recettes proposent des composs additionnels: le vinaigre pour modifier la teinte, le safran comme anti-oxydant, le savon ou le miel pour stabiliser le produit. Lajout de jus de cerise ou de baies de fleurs sans rpondre des buts clairement tablis ne fait quaugmenter la complexit du mlange pigmentaire. Linteraction avec les liants protiniques base duf doit tre prise en compte dans la structure du produit final ainsi que la dgradation au sein de la couche picturale mme. Alors que les spectres Raman des diffrents actates de synthse sont connus (Chaplin et al. 2006), ils nont t que trs difficilement et trs rarement observs dans les couches picturales vertes de manuscrits. La cause principale est la trs forte fluorescence, inexistante dans les composs purs, et attribue au compos complex. En consquence, la prsence de cuivre dans un pigment vert, en labsence dune identification formelle (carbonate ou sulfate), ne peut conduire qu proposer lhypothse dun vert-de-gris. Il sagit dune identification par limination. Mais tous les pigments verts non identifis ne sont pas du vert-de-gris comme le montre ltude dun manuscrits italien du XVe sicle (Bruni et al. 1999). Le spectre Raman ne permettait pas didentifier le pigment au cuivre. Cest le recours la microspectromtrie IRFT par rflexion qui a permis de caractriser la malachite. Ces quelques exemples montrent lavance trs diffrente des recherches sur les pigments. Les pigments bleus principaux dans les manuscrits de lEurope mdivale sont connus et facilement identifiables, cest la chronologie et les aspects artistiques, conomiques et sociaux de leur utilisation qui ouvrent des perspectives intressantes dans le domaine historique. La nature et la structure des pigments verts, drivs des actates de cuivre, sont identifier, ce qui permettrait dlucider leur mode de prparation. La constitution des mlanges bleu/jaune et la chronologie de leur utilisation nont pas fait lobjet de synthses. La palette verte recle toujours bien des inconnues.
13 4
Bibliographie
Brown K.L. and. Clark R.J.H 2004. The Lindisfarne Gospels and two other 8th century Anglo-Saxon/Insular manuscripts: pigment identification by Raman microscopy. Journal of Raman Spectroscopy, 33: 4-12 Bruni S., Cariati F., Casadio F., and Toniolo L. 1999. Identification of pigments on a XV century illuminated parchment by Raman and FTIR microspectroscopies. Spectrochimica Acta Part A 55: 1371-1377. Burgio L., Ciomartan D.A., and Clark R.J.H. 1997. Pigment identification on medieval manuscripts, paintings and other artefacts by Raman microscopy: application to the study of three German manuscripts. Journal of Molecular Structure. 405: 1-11. Cardon D. 1999. Teintures prcieuses de la Mditerrane. Carcassonne/Terrassa. Cennini C. 1978. Le Livre de lArt ou Trait de la Peinture, 1437. Trad. V. Mottez, Ed. 1978 (F. de Nobele, Paris) Chaplin T.D., Clark R.J.H., and Scott D.A. 2006. Study par Raman microscopy of nine variants of the green-blue pigment verdigris. Journal of Raman Spectroscopy, 37: 223229. Coupry C. 1990. Etude des pigments Manuscrit Paris BN latin 12302 in LEcole carolingienne dAuxerre, ed. D. Iogna-Prat, C. Jeudy et Lobrichon (Beauchesne, Paris) 119. Coupry C. and Brissaud D. 1996. Application in Art, Jewelry and Forensic Science in Raman Microscopy, ed. Turrell G. and Corset J. (Academic Press) 421. Coupry C. 1999. Les pigments utiliss pour lenluminure Fcamp aux XIe et XIIe sicles in Manuscrits et enluminures dans le monde normand (X e-XVe sicles), ed. P. Bouet et M. Dosdat (Presses Universitaires de Caen) 69. Coupry C. 2007. Approche analytique du dcor de deux manuscrits coptes. Orientalia Lovaniensia Analecta. 163: 199-208. Daniel F., Laborde B., Mounier A. et Coulon E. 2008. Le pigment darinite dans deux peintures romanes du Sud-Ouest de la France. ArchoSciences: 32. http://archeosciences.revues.org/987. Guineau, B., Coupry, C., Gousset, M.T., Forgerit, J.P., Vezin, J. 1986. Identification de bleu de lapis-lazuli dans six manuscrits peintures du XIIe sicle provenant de labbaye de Corbie. Scriptorium, XL: 157-171 Isacco, E. 2008. Les pigments des miniatures indiennes. LAsiathque Paris.
135
Porter C. 1995. You cant tell a Pigment by Its Color. In Making the Medieval Book: Techniques of Production, Anderson-Lovelace Ed. 111. Roger P., 2207. Etude des couleurs et de la pratique picturale. Art de lenluminure, lvangliaire de Charlemagne, 20: 46-66. Roosen-Runge, H. 1967. Farbgebung und Technik frhmittelalterlicher Buchmalerei. Berlin. Torrs- Bourdel A. 2003.Etude de la lazurite dans des roches et des pigments de lapislazuli par spectromtrie Raman. Diplme dEtudes Suprieures, Universit Bordeaux3. Trentelman K. and Turner N. 2009. Investigation of the painting materials and techniques of the late-15 th century manuscript illuminator Jean Bourdichon. Journal of Raman Spectroscopy, 40: 577-584. Vandenabeele P., Wehling B., Moens L., Dekeyzer B., Cardon B., von Bohlen A. and Klockenkmper R. 1999. Pigment investigation of a late-medieval manuscript with total reflection X-ray fluorescence and micro-Raman spectroscopy. Analyst, 124: 169-172. Villela-Petit et Guineau B. 2003. Le Matre de Boucicaut revisit. Palette et technique dun enlumineur parisien au dbut du XVe sicle. Art de lenluminure, 6: 3-34.
Biographie
Ingnieur en retraite du Centre National de la Recherche Scientifique (Laboratoire de Dynamique, Interactions et Ractivit, auparavant en charge de lquipe Matriaux des Objets du Patrimoine). Thse en Chimie Physique (Universit de Bordeaux) et Licence dHistoire (Universit de Paris IV Sorbonne). Activit scientifique: Rsonance magntique Nuclaire; Spectromtrie vibrationnelle, principalement microspectromtrie Raman. Principaux centres dintrt: Pigments dans les manuscrits et les peintures murales; Matriaux contemporains, plastiques (identification) et verres (structure et altrations); Colorants textiles. Recherches en collaboration avec: Bibliothque nationale de France, Centre des Peintures Murales Romaines, Centre dEtudes Mdivales dAuxerre, Muses du Louvre, Arts Dcoratifs, Carnavalet, de la Mode et du Costume, Manufacture des Gobelins,; Archologues et historiens du CNRS. Distinction: Cristal du CNRS claude.coupry@orange.fr 20, rue Clisson 75013 Paris (France)
136
Resumo
A nossa abordagem esttica, filosfica, psicolgica e outras da histria da arte comeam no olhar e apreciar as cores da iluminura. No entanto, como ser descrito neste artigo, as cores que hoje vemos nem sempre correspondem s aplicadas pelo artista. O ignorar destas alteraes cromticas pode induzir a srios erros na interpretao da sua inteno original. A especulao histrico-artstica deve ser, assim, precedida por uma avaliao sobre qual seria o aspecto original das cores. Uma observao mais cuidadosa pode evidenciar perda ou alteraes de cor. Estas podem ser bastante dramticas em amarelos, nomeadamente nos corantes orgnicos utilizados no passado, em manuscritos. Este artigo focar-se- pois no caso destes amarelos desaparecidos. A investigao em Histria e tcnicas de produo artstica oferece trs formas de avaliar qual o aspecto das cores originais: (i) iconografia e harmonia visual das cores, (ii) anlises cientficas de tintas originais, (iii) livros de receitas de artistas medievais. Conclu-se que, a conscincia destas alteraes cromticas bem como da aplicao dos mtodos de anlise aos materiais como ferramenta interpretativa, enriquecer o olhar do historiador da arte.
palavras-chave
iluminuras alterao cromtica perda de cor histria e tcnicas de produo artstica fontes para os materiais e tcnicas da arte
Abstract
All our aesthetics, philosophy, psychology, and other art history starts from looking at and enjoying the colours of manuscripts. However this paper demonstrates that colours observed today are not always those applied by the artist. Ignorance of this alteration results in serious errors of interpretation. Art-historical speculation must be preceded by the evaluation of original appearances. Close visual examination reveals areas of fading or discolouration. This is severe for yellows, especially for organic colorants commonly used in manuscripts. This paper therefore concentrates on the example of lost manuscript yellows. Technical art history offers three ways to evaluate the original appearance of manuscript colours: (i) iconography and visual harmony of colour, (ii) technical analysis of surviving paint, (iii) mediaeval artists recipe books. It is concluded that art historians can benefit from an awareness of possible colour alterations, and from an awareness of the application of scientific analysis as an interpretive tool.
key-words
manuscript illumination discoloration fading technical art history art technological source research
Acknowledgments
This paper derives from a presentation at the conference Medieval colours: between beauty and meaning (Universidade Nova de Lisboa: Instituto de Histria da Arte, Instituto de Estudos Medievais, Departamento de Conservao e Restauro, 2009). I would like to thank the organisers Maria Joo Melo and Adelaide Miranda for their kind invitation to participate, and for their generous hospitality. Figures 2, 3, 4 are reproduced na configurao do texto e sugestes de Ana Catarina Sousa, Pedro Fialho Agradecimentos por ajudaby permission of the Master and Fellows of Corpus Christi College, Cambridge. Figures 5, 6 are reproduced by permission of de Sousa, Justino Maciel, Felix Teichner e Heidi. ***STA CRUZ***. Figures 1, 7, 8 are by the author.
colours versus colorants in art history: evaluating lost manuscript yellows

mark clarke
University of Amsterdam, Amsterdam (The Netherlands) mark@clericus.org
Introduction
All our aesthetics, philosophy, psychology, and other art history starts from looking at and enjoying the colours of manuscripts. It is perhaps less obvious that all scientific (chemical and physical) analysis also starts from the observation of the colours: analysis depends on the observation that there is something distinctive present that requires analysis. However as this paper will demonstrate the colour observed today on an artwork such as a manuscript is not always the original colour as chosen, composed, and applied by the artist. Ignorance of this alteration results in serious errors of interpretation. Interpretations of the iconography and symbolism of colour, or deductions concerning the appreciation and comprehension of colour by mediaeval artists and their contemporary audience, or attempts to group manuscripts by stylistic use of colour, cannot be correct if the colours that interpretations are based on are not those that were originally intended. Similarly for scientists, failure to suspect the presence of some hidden material may result in failure to apply suitable methods of analysis. Clearly all theorising must be preceded by the evaluation of the original appearance and likely original material composition.
The loss of mediaeval colour

Colours can change their appearance. Colours can change their colour. Colours can change their intensity and saturation. Colours can disappear almost completely. The
139
vast majority of mediaeval colour has disappeared: clothes, wall-hangings, wallpaintings, enamels, jewellery, street signs, and scenery for pageants. Even fine art paintings have not survived well: they have invariably undergone periodic restorations, sometimes with fairly drastic reconstruction of losses. But even when the original paint is not lost extreme changes of colour occur: for example increased yellowing due to aged varnish is familiar, as is the loss of yellow glazes (due to fading or over-cleaning) resulting in unnatural blue foliage. Compared with such losses manuscripts have survived comparatively well. Unlike clothes they do not wear out; unlike jewellery or precious metalwork they are not broken down and re-modelled as tastes change. Manuscripts are relatively safe containers for text and image: illustrations in a closed book are protected from light fading, from touch, and from atmospheric pollutants, and parchment (itself chemically stable) has alkaline surface treatments that help to counteract the acidity of modern air. Nevertheless, much colour in manuscripts has become altered or lost colour, and this alteration and loss needs to be thought about analytically. This paper considers the example of yellows: firstly because with yellows the effects of colour loss may be either dramatic or subtle, i.e. either obvious or almost undetectable, and secondly because many interesting Portuguese examples exist.
Colours versus colorants

Analytical thinking about colour (whether by an art historian or a chemist) must begin with an appreciation of the difference between a colour and a colorant, that is, between a colour and the colouring material making that colour. To talk of pigments is to talk of physical materials, and to talk of colours is to talk of a property of those materials. Colorants and colours do not have a one-to-one correspondence. Any colour can be made by more than one pigment or combination of pigments. For example the plant yellow weld (Reseda luteola) and the artificial pigment lead-tin yellow have identical colours (Fig. 1a). Perhaps more surprisingly any one pigment may exhibit more than one colour depending on its preparation method (Fig. 1b). Knowing this provides a powerful tool for the study of manuscripts. It has been shown by physical-chemical analysis that specific colours were made in different ways by different artists or ateliers, and in different regions and periods (Clarke 2001b). Identifying pigments can thus help clarify provenance and authorship. A knowledge of materials derived from chemical analysis is therefore not only interesting to a scientist or a conservator-restorer, but also to a book-historian or art historian, and can inform an art historian far more profoundly than a simple examination of colour ever can. Two examples illustrate the difference between considering colours and considering colorants.
fig.1 (a) one colour prepared from two different materials, (b) different colours prepared from one material (rocella tinctoria)
Examination of colours
Much work has been done grouping manuscripts stylistically based on patterns of colour use. However, any given colour can be made in a variety of ways. Presum-
140
c o l o u r s v e r s u s c o l o r a n t s i n a r t h i s to r y : e va l u at i n g l o s t m a n u s c r i p t y e l l o w s
1. Unpublished results, Trinity College Dublin: Raman spectroscopy found red lead, indigo, orpiment, carbon and iron gall ink, gypsum, lichen purple (Rocella tinctoria) and some unidentified translucent yellow/brown.
ably within a workshop the choice of method would be consistent. It follows that it is not enough to group manuscripts based on similar appearance: the materials should be the same too. Similarity in materials strengthens an attributed grouping, and inconsistency in materials weakens it. However this independence of colour and pigment is not always appreciated. It is extremely common for people to attempt to identify pigments by eye, using simple visual examination. Typically the appearance of a sample observed on a manuscript is compared with samples and reconstructions of known composition. These reconstructions are made based on mediaeval artists recipe books, of which hundreds survive (Clarke 2001a). This methodology of comparing unknowns (the materials on manuscripts) with knowns (historically accurate reconstructions) is essentially valid, and is used today in chemical analysis of manuscript pigments. The problem lies in the use of visual examination to do the comparison. It is an inadequate tool, and any attempt to identify pigments by eye is doomed to failure, partly for the reasons outlined above (that one colour may be made from alternative colorants and that one colorant may exhibit different colours), and partly because, with age, colours may degrade beyond recognition. This may easily be demonstrated by comparing the results of visual pigment identification and chemical analysis. Often categorical statements as to which pigments are or are not present have been made by visual examination, which subsequent chemical analysis has shown to be incorrect. For example, British Library MS Arundel 155 was examined by two independent art historians, who stated in the firmest terms that ultramarine was not present; yet chemical analysis found it. Similarly a number of authors have used visual examination to identify the pigments on the Book of Kells and the Lindisfarne Gospels. Comparison of their results reveals very little agreement (such disagreement alone should alert us to a problem with visual analysis), and recent reliable analysis has shown all of them to be to some extent wrong (Clarke 2004 a, b). 1 Many of these authors expressed caveats regarding the limitations of visual examination, but the limitations of analytical techniques then available (which required unacceptably large samples to be removed) meant visual examination was used faute de mieux. Chemical analysis has shown such visual examination to be largely worthless, and it should not be practiced today.
Examination of colorants (pigments)

Visual naming of colours has been shown to be unreliable, with inconsistencies solved only by chemical analysis of the colorants. One might ask: why might chemical analysis of colorants interest anyone but a chemist? Consider two examples where looking at colours was not helpful, but where pigment analysis produced useful arthistorical evidence. The first example results from a study of pigments in 100 Anglo-Saxon manuscripts (Clarke 2004b). Two particular manuscripts from c.980 AD share text and illustrative programme. One had a secure provenance of Canterbury Christ Church, and it
141
had been suggested on stylistic grounds that the other might also originate there. The pigment ultramarine (lapis lazuli) found on both. This was only introduced c.1000, and to find it so early was most unusual, so its presence in both manuscripts strengthened the link between them. In the second example stylistical research into the Lochorst Bible suggested it was illuminated by two teams of artists. Team A worked in an archaic traditional style, whereas team B (probably the Zweder Masters) were more naturalistic and illusionistic. Team A consistently used ultramarine, while Team B consistently used azurite. Thus pigment analysis confirmed and reinforced the proposed division of labour.
2. We might consider a third attitude: the symbolic, magical, and spiritual value. In jewelled objects one stone would certainly be chosen over another for its symbolic value, according to the lapidary texts. I have, however, found no evidence that such values were ever attributed to pigments.
Mixtures and imitations

Any one colour may also be made from many different mixtures of pigments. From c.300 AD onwards treatises survive recording hundreds of recipes for imitating expensive colours such as ultramarine or Tyrian Purple (Clarke 2001a). These substitutes can be completely convincing to the eye. For example, it has often been asserted that Tyrian purple (derived from shellfish) was used to dye parchment pages, but analysis indicates this is extremely rare compared to purples made from other ingredients and from mixtures. Similarly mediaeval recipes explain how red brazil-wood dye could be added to blue azurite to imitate the more purple hue of the costly pigment ultramarine. Other recipes describe manufacture of a tin-based compound called mosaic gold which contains no gold but can look remarkably like it.
Attitudes to inexpensive imitations

There were various mediaeval attitudes to substitution and imitation. It is useful to consider these when evaluating whether (and why) certain yellow colorants were used. One attitude was that to use valuable materials was desirable in itself. The use of gold, precious stones, and ultramarine to highlight iconographically important elements in a picture is well known. Another example is the choice of mosaic gold or real gold, or ivory black (an expensive pigment) or bone black (indistinguishable and inexpensive). Alternatively, skill could be considered as more important than the use of expensive materials. (This, perhaps, was the choice made by the stylistically more progressive Team B in the Lochorst Bible when they used the latest illusionistic techniques but inexpensive azurite; a more flexible use of material conventions consistent with stylistic liberties.) Theophilus seems to have favoured this attitude when he warned not to disparage any thing just because your native soil has spontaneously and unexpectedly produced it for you and asks, why would you despise these as cheap local products and travel over land and sea to procure foreign ones that are no better and are perhaps of less value (Prologue to Book I). When analysis demonstrates the use of local products, plants or minerals, this itself becomes a useful provenanceing tool. This preference for final appearance over cash-value is one reason for the popularity of imitations and substitutes. 2
142
These attitudes created reasons to choose one material over another (given otherwise similar colours and working properties). These attitudes would have been present in different proportions in different individuals. One might contrast for example Suger with Bernard of Clairvaux. Consequently sometimes an artist might have chosen to use a genuine expensive material, whereas an imitation may have been considered appropriate for a different object or by a different artist. That sometimes there were genuine expensive materials used, and sometimes not, is in itself interesting. The use of intrinsically expensive materials is an important indicator of the attitude of the artist and the patron to a book. And yet to the naked eye it is not always clear. Clearly if chemical analysis can discern between expensive and inexpensive materials this is helpful in indicating which attitude or attitudes applied to which objects or to which periods, regions, artists or patrons. This is directly relevant to the study of manuscript yellows because, as we shall see, imitation gold was common.
Sophistication of the palette

Of course there are other reasons to use pigment mixtures, not just to imitate precious materials. Not all colours can be obtained from a single pure natural material. Until very recently there was, for example, a great shortage of green pigments. For symbolic or diagrammatic work this was not problematic, since a face or a landscape could be adequately modelled with one or two crude colours, but for more sophisticated work, e.g. to be more realistically representational, the shortage of subtle colours was a problem. The problem was overcome in two ways. The first method used to increase the range of available colours was mixing pigments. It is often stated that mediaeval artists, having made such efforts to obtain pure colouring materials, did not want to adulterate them, and that in consequence they did not mix pigments. This is simply not true. Mediaeval artists recipe books contain thousands of prescriptions for mixtures (Clarke 2011), and chemical analyses have shown a great variety of mixed pigments, notably those mixed to produce greens (Clarke 2001b). Another solution was layering, where a thin or transparent layer of one colour was used on top to modify another. These combinations allow for more sophisticated or more representational art, and indeed, around the Eyckian turning point, c.1420, we find more and more mixtures and transparent over-layers; this was certainly done in manuscripts, not only in panel painting or in oil painting (Clarke 2011). Another method of increasing the gamut of available of colours was to harness the colours of plants. There are many recipes based on plant extracts, either to make lake pigments, or simply to colour white pigments with the juices: Take yellow flowers, and grind, and express the juice, and temper white lead with this juice, and dry. And temper it again, and dry, and repeat thus a third time (Glasgow MS Hunter 110, f.40r). These flower extracts are often specified to be for use in carta, i.e. on paper or parchment. However, even in a relatively well-protected environment as a closed book, organic pigments based on plant material are very prone to change colour and fade.
143
Lost manuscript yellows

Careful examination of manuscripts reveals areas where colour only remains as faint transparent traces. Figure 2 shows transparent infills in orange, red-pink, yellow and blue. The transparency clearly seems to indicate a faded organic colour. Similar discolouration and fading of red and yellow is visible in Figure 3. Once alerted to such fading, one can search for it.
fig.2 a, b, c cambridge, corpus christi college ms69 (england, 8th century), f.20r, f.14r, ms144 (england, 9th century) f. 13r
fig.3 a, b cambridge, corpus christi college ms69, f.1r (england, 8th century)
14 4
3. This does not reproduce well in print.
Why are lost manuscript yellows so particularly interesting and important? One reason is that so often yellow was used either to represent gold, or to imitate gold, or to be a substitute for gold, and in any picture golden objects are of undoubted iconographic importance. There are a great many recipes for imitation gold, and indeed some of the most ancient artists recipes we have are for imitation chrysography. The most common ingredient of recipes for imitation gold is saffron. The other reason for concentrating on yellow is that its colour changes and fades away so dramatically; indeed the changes to yellows are often the strongest indication and the strongest argument that colours in manuscripts have changed at all. Examples of degraded yellow are found in the earliest manuscripts. On the exhibited pages of British Museum papyrus EA10470 (Egyptian, 1320-1200 BC) the yellow pigment orpiment and the orange pigment realgar (two forms of an arsenic mineral) have both become white (having converted to a third form of the mineral, arsenolite); the unexhibited pages retain their colour. Orpiment is unusual in this respect, because it is a mineral pigment, which are usually stable. Plant-based pigments, on the other hand, all tend to fade very badly, and yellow plant-based pigments fade worst of all. My attention was first drawn to this phenomenon while studying the earliest mediaeval manuscripts from the British Isles. In the Cathach of St Columba (Dublin Royal Irish Academy) on f.48r is a zoomorphic initial that only the closest inspection shows is infilled with a colour not quite the same as the parchment. 3 Clearly in this and in many other early manuscripts the paint is extremely thin, or is a liquid dye or stain rather than a solid pigment bound in a medium. This phenomenon is not confined to English examples, nor to very early examples, and one cannot dismiss it as the result of primitive materials and techniques. If we concentrate on yellows, then fading phenomena can be still dramatic in the highest quality twelfth century Romanesque manuscripts. Figure 4 shows a page from the Bury Bible, a de luxe manuscript. The rubric is composed of primary colours, but compared with the other colours the yellow is dull,
fig.4 a, b cambridge, corpus christi college ms2, f.7r (england, 12th century)
145
weak, and translucent. Here surely we can infer that this buff or yellow-brown colour must originally have been intended to be a strong colour as well, most probably a bright yellow imitating gold. This phenomenon appears in Romanesque manuscripts from all over Europe. I first noticed it on English examples, and was most interested to see it again in the Torre do Tombo and the Biblioteca Nacional de Portugal. Examples of this same alternation of strong colours with a drab buff were noted from Lorvo (e.g. MS L16 f.7v, f.32), and from Sta. Cruz (e.g. MS1 f.2 and MS17 f.171). In some cases we can deduce that some faded yellows were deliberately transparent. Often a transparent yellow has been used much as one might use a yellow fluorescent pen today, to highlight a rubric (Fig. 5a). That this yellow was intended as a highlighter is confirmed by this unambiguous red used similarly (Fig. 5b). The red is carefully painted around the letters, whereas the yellow is painted all over, suggesting it was always transparent.
4. Further examples of faded and discoloured yellows (shown at the conference) include: Cambridge Corpus Christi College: MS256 (Italy, 6th), MS69 (England, 8th), MS193 (France, 8th), MS411 (Channel, c.1000), MS183 (England, 11th), MS4 and MS94 (England, 12th), MS82 (France, 13th), MS394 (England 14th), MS395 (Catalan 15th). Cambridge University Library: MS Ll.1.10 (England 9th), MS Ii.6.32 (England 10th), MS Ff.1.23 (England, 11th). Glasgow University Library: MS Hunter 404 (Italy, 10th, e.g. f.25r). Engelberg Stiftsbibliothek: Cod. 5 (Switzerland 12th). Portuguese Romanesque examples include Lorvao L16, Santa Cruz MS1, MS17, 43 and 76, and Biblioteca Nacional de Portugal MS Alcobaa 333 and MS360.
fig.5 (a) sta. cruz ms58 (12-13th century) f.1 (b) f.140
The use of a transparent yellow is confirmed in other Portuguese manuscripts, where it is used in combination with patterns of dots (Fig. 6). This use of transparent yellow is confined neither to Portugal nor to this period, and examples might be multiplied indefinitely, from 6th to 15th century, from England, France, Italy, central Europe and Iberia. 4 Transparency in aqueous paint media such as those used on manuscripts is important as it implies the use of an organic pigment, and organic pigments are those most prone to fading.
Evaluating losses
Visual examination and comparison clearly reveals areas of fading or discolouration, especially severe for yellow; indeed, in some cases it is not obvious that an area was coloured at all. Technical art history offers three ways to evaluate what the original appearance of manuscript colours may have been: (i) iconography and visual harmony, (ii) technical analysis of surviving paint, (iii) mediaeval artists recipe books.
fig.6 sta cruz ms40 (13th century) f.30
146
5. Degraded orpiment, arsenolite, is also invisible to Raman. 6. Cheryl Porter, personal communication 2009. 7. Some years ago I suggested that the use of Infra-Red lasers might overcome the fluorescence problems that have until now hindered analysis of organic pigments by Raman spectrosopy. Perkin-Elmer have now lent an Infra-Red FT-Raman instrument to the British Library, who in 2010 will examine certain Anglo-Saxon manuscripts where I identified faded yellow organics. Preliminary results using reconstructions seem promising. (David Jacobs, BL, personal communication 2009.)
Iconography and visual harmony

One must re-consider images where one might expect yellow for iconographical reasons. There are few motifs in mediaeval European art where the iconographic meaning of yellow is sufficiently robust and consistent to deduce a lost yellow, but possible useful motifs include heraldic devices (specified in unambiguous terminology: or should be yellow), or representations of golden objects and haloes. Even without iconographical clues we can deduce when the colours must be wrong by loss of visual harmony, as for example in Fig. 4 where the colour of the red, blue and green letters is intense, thickly applied, strong, dark and saturated, but the brownyellow letters are dull and insipid. Especially in these Romanesque examples it seems clear that the pale-brown colour is incongruous, and a strong yellow would surely have made better pictorial sense alongside the other strong colours. In cases where a colour is barely distinguishable from the colour of parchment (e.g. the yellow in Fig. 3 b) we may conclude that it would have been decoratively valueless, and thus is surely degraded.
Technical analysis
Unfortunately chemical analysis is unusually problematic for faded yellows. For organic pigments the currently favourite technique for manuscript analysis Raman spectroscopy is unsuitable. 5 The best method of analysis of organic colours is High-Performance Liquid Chromatography, but it requires samples; these only need to be very small, but they do need to be not too degraded. Sometimes when a yellow colorant degrades it can be identified by its degradation products (although not in the case of saffron). Saffron has been detected using Fourier-Transform Infrared spectroscopy, even in mixtures, but only when it has survived in good condition. Despite these problems there have been successes. The Wollaton Antiphonal (University of Nottingham MS250), for example, is a high-quality English manuscript c.1420, painted with a very full palette. Recent HPLC analysis identified several plant-based pigments, including yellows from broom (Genista tinctoria), sometimes with weld added. All are found on a base of either chalk or lead white, and in one remarkable case, on the synthetic inorganic pigment lead-tin-yellow. The organic yellows were also found mixed with verdigris. Of particular interest is that it has been suggested that this book was made in the English region of East Anglia, where broom is a local product. 6 The application of other analytical techniques to manuscripts is being developed, including Direct Injection Mass Spectrometry (requiring only tiny samples) and fluorescence spectroscopy (requiring no samples). 7
Recipes
So, in summary, analysis can be very helpful, but is difficult for yellows. Fortunately, when analysis is not able to provide the answer, we have another clue. There survive a considerable number of mediaeval treatises containing artists recipes, including for organic yellow pigments. The majority of these use saffron (Crocus sativus). Audemar (c.1300) stated saffron was produced in France (but that it was not good),
147
that it was imported from Spain and Italy, and the best from Sicily [read: Cilicia i.e. Turkey] (Clarke 2001a, #2790). The 11th century specialist treatise De clarea (On glair) devotes a large proportion of its short length to saffron (Clarke 2001a, #140). Other yellows suggested by recipes include weld, safflower (Carthamus tinctorius), buckthorn (Rhamnus cathartica), blackthorn (Prunus spinosa), aloes, fustic (Rhus cotinus), and the gall of an ox or fish. These were usually mixed in glair (egg white), although hide glue was also suggested, e.g. by Le Begue in 1431 (Clarke 2001a, #2790). Alternatively organic colorants could be added to inorganic white substrates such as egg-shell, lead-white, chalk and gypsum. The prevalence and wide diffusion of recipes for organic yellows means that we should be looking for traces of them. (The prevalence of recipes for mosaic gold after c.1400 suggests analysis of what appears to be gold is also needed.) It is satisfying that recipes in the 15 th century Portuguese Livro de como se fazem as cores (Clarke 2001a, #2950) seem to correspond well with chemical analysis of Portuguese manuscripts, suggesting that its recipes for yellows may well explain these faded areas. Knowledge of the recipes suggests lines of enquiry: in reconstructions of saffron and safflower recipes, plant fibres are visible (Fig. 7), so it would be worthwhile trying to find them (although a careful illuminator may have excluded them).
fig.7 reconstructions (left) safflower, (right) saffron
Microscopic examination of the lettering shown in Fig. 4 showed the paint medium to be thick, almost colourless brown, and to exhibit craquelure. This is what would be found if it had originally been a yellow transparent colorant in a thick glue medium. (Reconstructions showed that saffron in a glue or gum medium needs to be applied thickly.) If indeed the Bury Bible lettering was originally made with saffron (or similar) then all the colours of the lettering would have originally been equivalent in terms of strength and saturation, and thus more visually coherent. This effect has been reconstructed in Figure 8 by digitally increasing the degree of colour saturation. Surely this makes more decorative sense.
148
fig.8 digital reconstruction with enhanced yellow
Recipes supply the surprising information that the problem of faded yellow pigments is not confined to faded yellow colours. This is because saffron was not just used in yellow pigments, but was also an additive to other pigments. Recipes document that the addition of saffron to verdigris was common in Europe (and Persia), and analysis confirms this. (The stated reason was to improve colour, although saffron also reduces corrosion of pages by verdigris.) Saffron was added to azurite to make green, and even to improve orpiment. Consequently loss of yellow organic colorants can cause colour change in even non-yellow areas. Other organics were added to other inorganics too, and so although inorganic pigments such as verdigris, minium or azurite may be identified by reliable and robust techniques, nevertheless the colours may not be original (Clarke 2011).
Mediaeval responses to colour loss

It is unclear how much durability of pigments was understood by mediaeval craftsmen. Recipes warn that certain pigments are chemically incompatible and will discolour each other if mixed, and other recipes specify that some pigments (notably organics) were suitable for manuscripts but not suitable for other supports (such as panels and walls). It is not clear how soon colour loss occurred, whether it was noticed, or whether repairs were carried out. Financial documents record mediaeval restorations of wall- and panel-paintings, but regarding manuscripts are largely silent. Accounts from Merton College Oxford c.1500 itemise painting saffron onto fore-edges apparently now lost. Examination shows improvements to manuscripts certainly were carried out: not really restorations but rather re-decoration projects. Pigment analysis is useful in finding anachronistic materials that document this. British Library Add. MS 40618 is an Irish eighth century manuscript, but includes some pages of late tenth century illustrations that have been pasted in (f. 22v, f.49v). One might easily spot these added pages by their later style, but ultramarine is also found on the original eighth century initials (23r and 50r); clearly these original initials must have been re-touched while the book was undergoing improvement
149
(Clarke 2004b, which see for another example). Whether lost yellows were repainted remains to be determined.
Conclusions
Clearly there has been a loss and change of colour in mediaeval European manuscripts, due to the degradation and fading of colorants, and in particular there has clearly been a particularly severe loss of yellows, especially transparent yellows. This yellow loss is not confined to areas originally coloured yellow, but also in certain areas such as verdigris green. Furthermore many non-yellow pigments contained an organic component, which has probably altered in colour too. Since what is seen now is clearly not always what was originally there, the appearance of manuscripts must be re-assessed. In the case of transparent yellow, the contrast with the parchment is often negligible, and even allowing for some darkening and yellowing of the parchment, if it had always been this colour it would have made no decorative sense. The buff or brown colour in Romanesque manuscripts has in the past been accepted as a bona fide colour. It does contrast somewhat with parchment, and is legible, but, since the other colours are usually all strong and bright, a subtle or drab yellow-brown seems out of place, and so I suggest that it would make more decorative sense if it had originally been golden-yellow. It seems, from inspection of surviving traces of colour, and from mediaeval recipes, that there was an extensive use of vivid organic yellows, which are now lost. Clearly art historians can benefit from an awareness of possible colour alterations. It may help to explain, for example, why so little gold was used in Portuguese manuscripts in this period, compared with elsewhere in Europe. In addition to being vehicles for the transmission of texts and images, manuscripts are also archaeological artefacts, and one must constantly bear in mind this physicality. Chemical analysis is helpful to establish provenance, to identify intrusive anachronistic elements, and to deduce original appearance. This is a golden age for the study of the physical properties of art works in general and manuscripts in particular, with increasingly sensitive analytical instruments available, and a renaissance in the philology of artists recipe books. But to establish what was conventional usage for different periods, regions, or ateliers, more work is needed to build up a substantial and statistically significant corpus of analyses (similar to that extant for easel paintings). This must be collaborative, as the best analyses are done by physicists and chemists working closely with conservators, librarians, and art historians.
150
Acknowledgments
This paper derives from a presentation at the conference Medieval colours: between beauty and meaning (Universidade Nova de Lisboa: Instituto de Histria da Arte, Instituto de Estudos Medievais, Departamento de Conservao e Restauro, 2009). I would like to thank the organisers Maria Joo Melo and Adelaide Miranda for their kind invitation to participate, and for their generous hospitality. Figures 2, 3, 4 are reproduced by permission of the Master and Fellows of Corpus Christi College, Cambridge. Figures 5, 6 are reproduced by permission of BPMP. Figures 1, 7, 8 are by the author.
Bibliography
Clarke, M. (2001a) The Art of All Colours: Mediaeval Recipe Books for Painters and Illuminators. London: Archetype Publications. Clarke, M. (2001b) The analysis of medieval European manuscripts Reviews in Conservation 2: 3-17. Clarke, M. (2004a) Really dont trust your eyes to identify manuscript pigments! Gazette du livre mdival 44: 50-53. Clarke, M. (2004b) Anglo-Saxon Manuscript Pigments Studies in Conservation 49: 231-244. Clarke, M. (2011) Mediaeval Painters Materials and Techniques: The Montpellier Liber diversarum arcium. London: Archetype Publications.
Biography
Mark Clarke (1962) trained in England in conservation and conservation science, with an interdisciplinary doctorate on medieval manuscript paint. He has been a researcher in technical art history and art technological source research, at the University of Cambridge, the Institute Collectie Nederland, the Fitzwilliam Museum, and FOMAMOLF. He specialises in the interdisciplinary study of historic artists paint, combining technical analysis, art history, historically accurate reconstructions, and written sources. He has made a particular study of mediaeval artists recipe books. He co-founded the International Council of Museums (Conservation Committee) working group on Art Technological Source Research. He is currently working on the early history of oil paint at the University of Amsterdam, and is an invited Fellow of the VLAC Institute of Advanced Study of the Royal Flemish Academy. University of Amsterdam, Rijksmuseum Ateliergebouw, Hobbemastraat 22, 1071 ZC Amsterdam, The Netherlands. mark@clericus.org
151
Resumo
A iluminura romnica que chegou at ns encontra-se intimamente ligada ao mundo monstico, particularmente aos mosteiros de So Mamede do Lorvo, So Pedro de Arouca, Santa Cruz de Coimbra e Santa Maria de Alcobaa. neste universo cultural, rico de referncias religiosas, polticas e artsticas que nos propomos estudar a cor. Como formas de construo da cor, entendemos a procura de certas combinaes e padres bem como de efeitos visuais particulares. Procedeu-se quantificao das principais cores presentes, atravs do mapeamento da cor, obtendo-se assim a distribuio das reas relativas para o azul, vermelho, verde, amarelo e laranja. O trabalho de investigao levado a cabo por uma equipa interdisciplinar permitiu assim lanar novas hipteses em relao s opes estticas e simblicas dos iluminadores e datao dos manuscritos, respectivamente do Apocalipse do Lorvo e os De avibus. Procuramos ainda dar incio a uma narrativa que nos permita chegar ao significado da cor nos cdices medievais (scs. XII e XIII), detectando as contribuies das trs culturas medievais que forjaram o Portugal moderno hebraica, muulmana e crist.
palavras-chave
iluminura medieval cor anlises cientficas apocalipse do lorvo de avibus
Abstract
The Romanesque illumination is intimately linked to the monastic world; particularly to the So Mamede do Lorvo, So Pedro de Arouca, Santa Cruz de Coimbra and Santa Maria de Alcobaa monasteries. It is in this cultural universe, rich in religious, political and artistic references that we propose to study colour. By exploring how colour was created and applied in specific patterns and for especial visual effects, we aim to contribute for a better understanding of its symbolic and social meaning. Colour distribution was quantified by mapping the relative areas of the main colours, blue, red, green, yellow and orange. The interdisciplinary research that was carried out has put forward new hypotheses in relation to aesthetic and symbolic options of the illuminators and to the dating of the manuscripts, Lorvo Apocalypse and De Avibus respectively. Finally, we started a narrative that will allow us to unveil the meaning of the colour in medieval codices, detecting the contributions of the three medieval cultures which forged Portugal, that is the Hebrew, Muslim and Christian cultures.
key-words
medieval illuminations colour scientific analysis lorvo apocalypse de avibus
the colour of medieval portuguese illumination: an interdisciplinary approach

maria j oo m e lo (co o r d i n ato r ) cata rin a mi g u e l , a n a c l a ro, r i ta ca stro
Departamento de Conservao e Restauro and Requimte, Faculdade de Cincias e Tecnologia, UNL, Caparica (Portugal)
M a r i a a d e l a i d e m i r a n da ( co o rd in ator ), a n a l e mos
Instituto de Estudos Medievais and Instituto de Histria da Arte, FCSH, UNL, Lisboa (Portugal)
Preamble
In 2004, we started an interdisciplinary project on the colour of medieval Portuguese illuminations, within which an interdisciplinary team 1 was created. The experimental design as well as the modus operandi were developed and fully tested 2. Two important and dated manuscripts were selected as case studies, Apocalypse and The Book of Birds (De Avibus of Hugh of Fouilloy) from So Mamede do Lorvo monastery. The study of colour was continued in a project that included the scriptoria of Santa Cruz de Coimbra and Santa Maria de Alcobaa as well as a larger and representative selection of manuscripts 3. In this paper, the findings of these two projects will be presented.
v. s o l a ng e f. m u r a l h a
VICARTE: Vidro e Cermica para as Artes, FCT, UNL, Caparica (Portugal)
j oo a . lo pe s
REQUIMTE, Departamento de Qumica, Faculdade de Farmcia, Universidade do Porto, Porto (Portugal)
a nt nio p e r e i r a g on a lv e s
Departamento de Qumica, Instituto Tecnolgico e Nuclear, Lisboa (Portugal)
1. The team includes researchers from Art History, Conservation and Chemistry; computer scientists have recently joined the team. 2. An interdisciplinary approach to the study of colour in Portuguese manuscript illuminations, POCTI/EAT/33782/2000. 3. The identity of Portuguese medieval manuscript illumination in the European context, PTDC/EAT/65445/2006.
1. The Romanesque illuminated manuscripts in Portugal and the Alcobaa, Lorvo and Santa Cruz monasteries
The Romanesque illumination is intimately linked to the monastic world; particularly to the So Mamede do Lorvo, So Pedro de Arouca, Santa Cruz de Coimbra and Santa Maria de Alcobaa monasteries 4. The two last monasteries present the more homogeneous and consistent collections of our Romanesque scriptoria and, at that
15 3
time, were central for the formation of the Portuguese nationality, assuming artistic protagonism in the kingdom. It is in this cultural universe 5, rich in religious, political and artistic references that we propose to study colour 6. Inheriting of ancient traditions, the new religious orders, Benedictines, Augustinians and Cistercians, will integrate in their scriptoria important treatises for the artistic techniques, such as the Mappae Clavicula in the library of Santa Cruz de Coimbra 7 and On divers arts very likely present in Alcobaa. The monks artistic creativity is present in the architectural space they inhabited as well in the liturgical arts, particularly in the illuminated manuscripts. Colour appears as a fundamental element in the organisation of the Codex, creating sense and beauty. By exploring how colour was created and applied in specific patterns and for especial visual effects, we aim to contribute for a better understanding of its symbolic and social meaning in medieval Portuguese illuminations in these monasteries. We will approach this subject within an international context, namely considering the leading French monasteries, with which the Portuguese monks were connected, see Figure 1. The monastic libraries of Alcobaa and Santa Cruz were produced in their scriptoria, through the circulation of manuscripts between the monasteries which they were associated with, and together with donations made from or orders made to the most prestigious productive centres at that time. A short presentation of the monastic medieval libraries, and the monasteries in which they were established, is important for having a better understanding of the origin, purpose and circulation of medieval manuscripts.
4. Miranda, M.A. 1996. A Iluminura Romnica em Santa Cruz de Coimbra e Santa Maria de Alcobaa. PhD Dissertation. Lisboa: Universidade Nova de Lisboa; Miranda, M.A. 1996. A Iluminura de Santa Cruz no Tempo de Santo Antnio. Lisboa: Edies Inapa; Nascimento, A.A., Miranda, M.A. (coord.) 1999. A Iluminura em Portugal: Identidade e Influncia (do sc. X ao XVI): catlogo da Exposio. Lisboa: Biblioteca Nacional de Portugal; Miranda, M.A., Lemos, A., Claro, A., Miguel, C. and Melo, M.J. 2008. A Iluminura Portuguesa, uma abordagem interdisciplinar. Revista de Histria de Arte FCSH. 5: 228-245. 5. Mattoso, J. 2009. Os antepassados dos navegadores in Naquele tempo. Ensaios de Histria Medieval. Lisboa: Temas e Debates/Crculo de Leitores, p.240: (...) investigaes recentes acerca da cultura dominante em Coimbra e nas cidades portuguesas do Sul mostram que os conhecimentos dos Morabes estavam ainda suficientemente vivos imediatamente a seguir conquista crist e que exerceram uma influncia profunda sobre os clrigos destas mesmas cidades durante o sculo XII. Ora Lorvo situa-se precisamente junto a Coimbra e era um centro importante da cultura morabe. 6. Pastoureau, M. 1996. Couleur, images, symboles. Paris: Le Lopard dOr, sd.p.47: La symbolique des couleurs est une notion floue dont on use et abuse. Il ny a pas de symbolique des couleurs envisage hors du temps et de lespace, mais seulement de multiples systmes de la couleur ou dans un contexte donn, prcis, dat, localis, les couleurs prennent en charge tel ou tel rseau de significations. 7. Peixeiro, H. 1998. A Iluminura do Apocalipse do Lorvo. Provas pblicas para professor coordenador. Tomar: IPT/EST/DTAG/ACG, p.28. 8. A documentao do Liber Testamentorum toma por referncia a comunidade monstica do Lorvo. esta, certamente mais antiga que o documento de 857 (...). Efectivamente, comprovando ele que a comunidade monstica recebe doaes nessa data, h que admitir a sua existncia anterior in Nascimento, A.A. 2007. Liber Testamentarum do Mosteiro de Lorvo, in Monarquia y Sociedad en el Reino de Len. De
fig.1 portuguese medieval monasteries: so mamede do lorvo, santa cruz de coimbra and santa maria de alcobaa, and the french monasteries which they were associated with (mother abbeys)
154
t h e c o l o u r o f m e d i e va l p o r t u g u e s e i l l u m i n at i o n : a n i n t e r d i s c i p l i n a r y a p p r o a c h
Alfonso III a Alfonso VII. ed. J.M. Fernndez Catn, Len, 315-339. 9. While both scholars agree with the fact that the date mentioned in the ms. 777, must be corrected, Aires Augusto do Nascimento proposes 857 and Fernndez Catn 907 or 917. Please see also Mattoso, J. 2009. Recension to the paper by Nascimento A., Fernndez Cton, J. (ed). Liber testamentorum coenobii laurbanensis (estdios). Lon: Centro de Estudios y Investigacin. in Mediavalista. 7. Lisboa: FCSH: Enquanto A. Nascimento defende a correco do evidente erro de data do referido documento de 777 para 857, Fernndez Catn considera esta opinio impossvel e prope 907 ou 917, mas inclina-se mais para 917 (pp. 467-475). 10. Mattoso, J. 1996. A vida religiosa dos beneditinos portugueses durante o sc. XIII in Religio e Cultura na Idade Mdia Portuguesa. Lisboa: Imprensa Nacional-Casa da Moeda, 2. ed., 167-168: Provavelmente ignoraremos para sempre at que ponto eram verdadeiras as acusaes feitas aos monges de Lorvo. No podemos, porm, deixar de notar que o fim do sc. XII neste mosteiro uma poca de prosperidade e de vitalidade religiosa, como atestam as cpias de livros a feitas entre 1183 e 1189. Alm disso, sabemos que os bens do mosteiro provocavam as ambies do bispo de Coimbra, contra as quais os monges se queixavam amargamente por volta de 1200.
The So Mamede do Lorvo monastery provides us scarce documental evidence on the date it was founded. A recent publication on Chartularium Laurbanense (Cartulary of Lorvo), also known as Livro dos Testamentos do Lorvo, coordinated by Aires Nascimento and Fernndez Catn, proposes 857 8 and 917 9 for its foundation. This monastery had an important role in the period before the Portuguese Kingdom was founded, and was involved in the Iberian resistance to the introduction of the monarchism from Cluny 10. Its collection is heterogeneous and from 1206, the date from which it starts to follow the feminine Cistercian 11, we believe it will no more have a scriptorium, even though it continues to enrich its library. For the medieval period, the collection has 18 manuscripts, of which two have colophon, particularly significant for the history of illumination in Portugal the Lorvo Apocalypse (1189) (Figure 2) and the De Avibus of Hughes de Fouilloy (1183-1184) (Figure 3). These manuscripts can be found in the National Archives (Arquivo Nacional da Torre do Tombo) 12. Santa Cruz de Coimbra, which dates back to 1131, is one of the most important monasteries linked to the construction of Portugal as an independent kingdom. The communitarian life started the following year and its importance was reinforced by the fact it was the first royal pantheon. Donations from the crown and nobility were in the origins of the monastery that will follow the regular Canons of Saint Augustine and the guidelines of St. Ruf of Avignon, depending on the Holy See due to the conflicts with the Cathedral of Coimbra 13. We know that the scribes substituted the Visigoth by the Carolina script between 1054 and 1072 in the documents of the chancellery 14, being the writing of the Codex close to this date. The first Codex is dated of 1139, Homilirio Santa Cruz 4 (Appendix 1), even though there are doubts of its production in this monastery 15. Until the first quarter of the thirteenth century, Santa Cruz has approximately 36 illuminated manuscripts from which a significant number will have here its origin. The collection has a total of 99 codices, conserved
fig.2 lorvo apocalypse (1189), ff. 43, 115, 172v, 209v and 210. photo antt
155
fig.3 de avibus (book of birds) from lorvo monastery (1183-1184), ff. 5, 6, 16, 25 and 50v. photo antt
in the Public Library of Porto 16. It is one of the monasteries that have received more attentions from scholars 17. Santa Maria de Alcobaa is a Cistercian monastery founded in 1153, the year in which St. Bernard died. Its foundation is also regulated through a Carta de Couto given to the white monks. Only few codices are dated, but the fact the construction of the monastery only started in 1179, shows that the scriptorium would have started to operate lately. It possesses a set of approximately 458 volumes from the Middle Ages, of which 160 illuminated ones from the last quarter of the 12th and first quarter of the 13th century, which are kept in the Biblioteca Nacional de Portugal 18.
11. Cocheril, M. 1978. Routier des Abbayes Cisterciennes du Portugal. Paris: Fondation Calouste Gulbenkian, p.195. 12. The illuminations of these two codex are available at http://digitarq.dgarq.gov.pt 13. Nascimento, A.A. 1998. Vida de D. Telo in Hagiografia de Santa Cruz de Coimbra. Ed. critic, trad. e estudos introdutrios. Lisboa: Colibri, 81-82. 14. Azevedo Santos, M.J. 1994. Da visigtica carolina. A escrita em Portugal de 882 a 1172. Lisboa: FCG/JNICT, 277-278. 15. Nascimento, A.A. 2002. Liber Testamentorum cit. N.C. Borges, Arte monstica em Lorvo, sombras e realidade Das origens a 1737. Lisboa. 16. Vilares Cepeda, I. and Ferreira, T.D. 2001. Inventrio dos Cdices Iluminados at 1500. Vol.2. Lisboa: Imprensa Nacional Casa da Moeda. 17. Martins, A.A. 2003. O Mosteiro de Santa Cruz de Coimbra. Sculos XII-XV. Lisboa: Centro de Histria Universidade de Lisboa; Gomes, S.A. Gomes. 2000. Limine Conscriptionis: Documentos, Chancelaria e Cultura no Mosteiro de Santa Cruz de Coimbra: Sculos XII-XIV. Coimbra: Polimage Editores. 18. Ferreira, T.D. and Santana, A.C. 1993. O tratamento documental de manuscritos ao servio
2. Characterization of the manuscripts from Alcobaa,Lorvo, Santa Cruz collections (12th-13th centuries)
As far as the contents of the libraries of Alcobaa and Santa Cruz are concerned, there is, in both, a strong predominance of works related to the Patristics (documents written by the Fathers of the Church transmitting the Christian thoughts of the 3rd to 8th century). Although in Alcobaa the Greek Patristics were highlighted, it was however almost absent in Santa Cruz. In Alcobaa we found a significant number of Fathers from the East. Amongst them, Origen, John of Damascus, St. Ephrem, St. Cyprian, St. Athanasius, Eusebius of Caesarea and John Chrysostom. In Santa Cruz, we only have Origen and Eusebius of Caesarea with the Ecclesiastic History, which was justified by the presence of the Bernardin spirituality in the formation of Alcobaa monastery. As far as the Latin Patristics is concerned, the resemblance
156
da investigao: a experincia da Biblioteca Nacional. Cadernos BAD, p. 1. 19. Twelfth century encyclopaedias, contemporary for the ms. studied. 20. In Santa Cruz there is only one copy from Papias.
between both libraries should be stressed. Saint Augustine and St. Gregory in Alcobaa, and Saint Augustine and Saint Isidore, in Santa Cruz (library in which we also find a copy of Saint Ambrose, St. Gregory, St. Bede and Cassian), are the most representative authors, being significantly higher in number in Alcobaa (for example, Saint Augustine appears about twenty times). The encyclopaedias are the second group of works, being present Saint Augustine with De doctrina christiana, Saint Isidore with Etimologias, St. Bede with De Rerum Naturam and Hugh of Saint Victor with Didaskalion. In both monasteries, the encyclopaedias from Honorius Augustodunensis and William of Conches are absent 19. Peter Lombard, a polemic author whose work is considered by some to be one of the most important for the study of dialectics and by others a compiler, is largely represented in Alcobaa, however it is absent from Santa Cruz. In the libraries of both monasteries, there are also texts from theologians and homileticians, such as St. Bernard and Hugh of Saint Victor, with the latter being an important author as he represents the speculative mystics. The De Sacramentis, existing in Alcobaa as well as in Santa Cruz, is a large Summa Theologica. However, contrary to what expected, St. Bernard is not the most represented author, even though his works spread rapidly through the library of Alcobaa. Here, we can also find one of the first works of Saint Anselm, Cur Deus Homo, of mystic character, but three of his fundamental works are absent: the Monogium, the Proslogium and the De Veritate. Another well represented group in Alcobaa, though less significant than in Santa Cruz, is the Hagiography, the narration of miraculous and exemplary achievements of the Saints. These manuscripts constitute one of the most important remaining hagiographic collections from the 12th century, being an important link in the history of the first Cistercians legendaries. Writings in Law are absent from Santa Cruz, while present in Alcobaa but with an insignificant number, apart from Bucardo and Bernardo Papiense, the Canons of the Fourth Council of the Lateran, the Livros do Uso de Cister, the Cistercian Carta Caritatis and Consuetudines (Charter of Charity). The Decretum Gratiani, not present in the library but whose existence is known in Clairvaux, Pontigny and Fontenay, is considered as one of the lost manuscripts from the library. We highlight the Grammar works linked to teaching 20, which are well represented through Papias, and also with Alexander of Villedieu, Garnier of Rochefort (Cistercian monk of Clairvaux) and an Alegoric Dictionary of Pedro Cantor. We emphasize also the presence of the Bible, the Book, whose existing copies, due to the importance given in terms of dimension and care to its execution and ornaments, show the importance they had in this library. It is also worth mentioning that the two Santa Cruz bibles present only the Old Testament, with the New Testament being absent in this library. Finally, there are the liturgical books, which are quite significant in the constitution of these two libraries, due to their number and quality, the Psalters and Missals. At the end of the twelfth century, the Portuguese monastic cultural context was spiritually propitious to the De Avibus of Hughes de Fouilloy, a text that was meant to be a behaviour model for the lay brotherhood community, explaining to them
157
the mysteries of incarnation using birds as examples. The three monasteries had a copy of the Book of Birds 21. The Lorvo monastery, as previously referred, unlike the two other monasteries, present a very heterogeneous collection 22 as well as a few number of manuscripts. Within this project were studied the manuscripts dating from the Iberian monachism period of the monastery until the Cistercian period, which include two Lectionaries (one Temporal and one Santoral), one Martirology, one Psalter, a manuscript of Saint Augustine, Enarrationes in Psalmos (1183), one Gradual, one Livro das Calendas and two dated manuscripts, the Book of Birds (1183/84) and Lorvo Apocalypse (1189) 23. The manuscripts selected for study, produced during the 12 th and the first half of the 13th century, are available as Supplementary Material 24, together with the folia object of molecular characterization. Thirteen manuscripts were chosen from the Santa Cruz collection, fourteen from Alcobaa, and eleven from the Lorvo collection. This selection was based not only on the chronological period established, but also on the richness of artistic and chromatic features. Many of the manuscripts are only tentatively dated,Tables AP1-AP3, and only 6 24 present a colophon in which a date was originally included by the author.
21. Miranda, M.A. et al. 2010. On Wings of Blue: The history, materials and techniques of the Book of Birds in Portuguese scriptoria. in Afonso, L.U. (ed.) The Materials of the Image. As Matrias da Imagem, Lisboa: Campo da Comunicao, 181182. 22. The involvement of the Bishop of Coimbra, D. Pedro Soares, in the expulsion of the Benedictine monks from the Lorvo monastery, prompt us to consider the possibility that they would have carried with them some of the codex and left others, possibly some mss. that were necessary for the religious service as De Avibus and the Lorvo Apocalypse. 23. The Lorvo Apocalypse is part of a large collection of manuscripts known as Beatus, a name derived from the Beatus of Libana, a monk who lived at the turn of the eighth to the ninth century, in Libana and wrote a Commentary to the Apocalypse of Saint John, in an environment of belief in the end of the world, accentuated by the closeness to the end of the millennium and an Iberic Peninsula that was converted to the Islam. It is believed that the original manuscript would have been illuminated and that it would communicate the visual message of the events that occurred since the Revelation of Christ up until the moment of the reconstruction of New Jerusalem. In these manuscripts text and image are intrinsically interconnected. 24. Lorvo 5, 43 and 50; Alcobaa 410; Santa Cruz 4 and 27. Supplementary material at www. dcr.fct.unl, section Events Medieval Colours Conference. 25. This in part explains our engagement in the critical edition of The book on how to make colours that is presented in the Varia section of this number. 26. The BPMP manuscripts were studied in a Molab mission, in the framework of the Eu-ARTECH European project. www.eu-artech.org/files/MEDMAN-UserReport.pdf 27. Melo, M.J., Claro, A. 2010. Bright light: microspectrofluorimetry for the characterization of lake pigments and dyes in works of art. Accounts of Chemical Research. 43(6): 857-866.
3. Colour: molecular characterization

We approach the study of the materials and techniques of medieval illuminations from three perspectives: art history, molecular analysis of the original artworks and historical accurate reconstructions 25. This section will describe the results of the molecular analysis in Portuguese medieval paint colours. In this research, we aim to unveil the complete paint formulation, including the proportions of which colourants, binders, fillers and other additives were used. With this knowledge, it will be possible to reproduce, in the laboratory, a medieval paint that will be as durable and luminous as the colours that we still find in medieval illuminated manuscripts; we will also be better prepared to preserve this illuminated heritage.
3.1 Colourants
Missions to Torre do Tombo (ANTT-DGARQ), Biblioteca Pblica Municipal do Porto (BPMP) 26 and Biblioteca Nacional de Portugal (BNP) were carried out during November 2007, January 2009 and May 2009, respectively. The experimental design and the analytical techniques are described in Appendix Experimental Design 27. From the study of all three monasteries manuscripts, the medieval Portuguese molecular palette was proposed Figure 4 28. This included the best colourants available during medieval times: vermilion, red lead (minium), orpiment, lac dye (and possibly other red dyes), lapis lazuli, indigo, white lead, carbon and bone black 29 as well as a synthetic copper green named bottle green. Although included in Figure 4, azurite and malachite were very seldom used 30. In the Portuguese illuminations of the
158
28. Miranda, M.A., Claro, A., Lemos, A., Miguel, C. and Melo, M.J. 2008. A Iluminura Portuguesa, uma abordagem interdisciplinar. Revista de Histria de Arte FCSH. 5: 228-245. 29. Cabral, J.M.P. 2006. Histria Breve dos Pigmentos. 4. Das Artes da Idade Mdia (1. parte). Qumica-Boletim da Sociedade Portuguesa de Qumica. 103: 33-44. 30. They will become later the colour for blue and green [Moura, L., Melo, M.J., Casanova, C., Claro, A. 2007. A Study on Portuguese Manuscript Illumination: The Charter of Vila Flor (Flower Town), 1512. Journal of Cultural Heritage. 8: 299-306]. 31. Cabral, J.M.P. 2001. Histria Breve dos Pigmentos. III Das artes grega e romana. QumicaBoletim da Sociedade Portuguesa de Qumica. 82: 57-64. 32. Miguel, C., Claro, C., Melo, M.J. and Lopes, J.A. 2009. Green, blue, greenish blue or bluish green? Copper pigments in medieval Portuguese Illuminations. Sources and Serendipity Testimonies of Artists Practice, London: Archetype: 33-38. 33. In this text carmine describes a dark red, obtained from an organic red dye such as lac dye (laccaic acids A, B, etc.) or cochineal (kermesic and carminic acid). 34. Melo, M.J. 2009. History of natural dyes in the ancient Mediterranean world. Handbook of Natural Colorants, T. Bechtold and R. Mussak (eds), Chichester: John Wiley & Sons 3-18; Seixas de Melo, J., Melo, M.J., Claro, A. 2006. As molculas da cor, na Arte e na Natureza. Qumica-Boletim da Sociedade Portuguesa de Qumica, 101: 44-55. 35. Strolovitch, D. 2005. Old Portuguese in Hebrew Script: convention, contact, and convivncia. PhD Dissertation. Cornell University. 116-184; Blondheim, S.1928. An old Portuguese work on manuscript illumination. Jewish Quarterly Review. 19: 97-135; Moreira de S, A. 1960. Revista da Faculdade de Letras, 4: 210-223.
fig.4 colourants and colour patterns in portuguese medieval illuminations
twelfth and thirteenth century, lapis lazuli was the colour for blue and was usually used as a pure pigment 31. Dark blues were obtained by adding indigo to lapis lazuli (Santa Cruz) or by using indigo (Lorvo). A deep saturated green, named bottlegreen, found ubiquitously in all manuscripts, was always applied as a single colour 32. Reds were obtained with vermilion that we found as a pure pigment in the lettering or mixed with chalk or red lead (Lorvo) for more extensive painting. Dark reds or carmine colours 33 could be obtained both with an organic dye as lac dye 34 (Lorvo) or by adding an organic dye to vermilion (Santa Cruz). The whites and lights were applied using white lead, and for the black colour, both carbon black and bone black were employed. Red lead was used for the orange colour, namely for typical visual effects, together with pink and white, as depicted in column 2 and 4 of Figure 4. Orpiment was less often employed, having been found in the Lorvo Apocalypse, where it was extensively applied, in the Bible Sta Cruz 1 and in the Homilirio Sta Cruz 4. Other shades of yellow and brownish colours were also found in the studied manuscripts, but until now their characterization was not possible (please, see also next section). Pink and grey colours were consistently applied in the Alcobaa manuscripts and will be discussed in the section Colour construction and meaning. Alongside the molecular analysis of the artworks, paint historical reconstructions were produced based both on the analytical data and the recipes found in the medieval treaties The book on how to make colours 35. These form an important reference database, that will allow a better understanding of the degradation mechanisms and to study the most appropriate conservation strategies.
159
3.2 Binders, the invisible side

Binding media, the invisible component of a paint colour, is necessary for the cohesion and adhesion. It influences the perception of colour and it may also play a key role in colour changes over time as well as in the stability of the colour paint. Therefore, its characterization becomes important, not only to fingerprint the specificities of the paints formulations of Portuguese medieval illuminations, but also to understand the degradation problems that are now visible in some manuscripts. Binders can be applied as a single binder (e.g., egg yolk), in mixtures of two or more polymers (e.g., parchment glue and glair) as well as more complex formulations where several additives are present in however minor amounts, but could still contribute to the paints film performance and permanence. To address this issue, chemometrics methodologies were used to compare a considerable number of infrared spectra of medieval paint colours, unveiling multivariate hidden patterns and by comparison with a reference database, to characterize the binders present. The infrared spectra showed that the binder was protein based and therefore the reference database included infrared spectra for the most probable medieval proteineaceous binders: egg white, parchment glue, egg yolk and casein glue 36, Figure 5. A set of 19 micro-samples (six blue paint micro-samples and 13 red paint micro-samples) from eight Lorvo manuscripts 37 was analysed. Vermilion red and lapis lazuli blue were chosen because they are colours ubiquitously present in the medieval illuminations 38. To quantify the binder composition, a multivariate curve resolution (MCR) analysis was performed; this algorithm was able to break down the infrared spectra to its pure binders: mainly a collagen binder together with egg white. It was not yet possible to ascertain if the binders were applied as a mixture or as binder and varnish, or a combination of both. To further test these results, antigen-antibody essays are currently in progress.
fig.5 scores plot of a pca model for red (l) and blue paints (l). the grey circle represents the 95% of confidence limit. binders projected onto this model: parchment glue (n), egg white (n), egg yolk (s) and casein glue (s)
3.3 Colour degradation over time

When studying medieval illuminations, colour changes over time must be taken into account as they may not only dramatically transform the original colour paint, but also our perception of the overall colour construction and its meaning. As previously discussed, faded shades of yellow colours are yet to be characterized. It is possible that some of those faded achromatic yellow paints were originally bright yellows. However, evidence must be found to corroborate this hypothesis, as one must also consider that many organic dyes (including blues, reds and yellows) may have evolved to an achromatic yellow- brown stimulus upon degradation 39. Other colour changes that may give rise to misleading interpretations are the darkening of white lead and red lead, as these may be interpreted as shading. However, under microscope, the degradation is clearly visible, see bellow. These degradation phenomena are not frequent in the studied collections, but are present in several manuscripts as Lorvo Apocalypse, Bible Sta Cruz 1 and Etymologies, Alc 446, Figure 6. The most dramatic changes were observed in the extensive darkening of the orange colour used in Apocalypse, Figure 7, and in the loss of the ubiquitous bottle green in all
36. Principal Components Analysis (PCA) was applied in the 2840 to 3000 cm-1 spectral window, a region less vulnerable to pigments and fillers and therefore a good infrared window to discriminate the proteinaceous binding media. 37. The Book of Birds, Lorvo Apocalypse, Lorvo 3, 12, 13, 15, 16 and 50. 38. Lapis lazuli does not absorb in the selected protein region and vermilion does not absorb in the entire spectral range. 39. See also the paper by Mark Clarke, in this volume, pag. 138.
160
40. Miguel, C. Le vert et le rouge: A study on the materials, techniques and meaning of the green and red colours in medieval Portuguese illuminations. Ongoing PhD Thesis (SFRH/ BD/44374/2008).
fig.6 alcobaa 446 (13th century), f.96v evidencing signs of degradation on the white lead on the flesh paint. image bnp (please see text for more details)
the codices. Projects were launched aiming to unveil the degradation mechanisms 40, as this knowledge is fundamental for the development of stabilization procedures. Relevant aspects of this on-going research will be next presented.
161
Orange 41
The red lead from Lorvo Apocalypse changed from being a bright orange to a dark brown, following three general degradation patterns: a very textured surface where a pastel yellow colour and red grains were observed, a thin black-grey layer over the orange and, finally, a thicker layer of a metallic appearance, Figure 7. To further understand the degradation mechanisms, based on the fact that in the orange paints the presence of orpiment in ca. 1-4% and vermilion below 2% (%wt) was detected 42, a set of experiments were designed to test the influence of extenders, pH and the two other pigments used 43. The results obtained indicated that lead sulphide 44 (PbS) was one of the main degradation products 45. This concurs with the results obtained in situ and from Raman analysis on a micro-sample of a degraded area in f. 118, which confirmed galena as the main degradation product of red lead 41.
41. Miguel, C., Claro, A., Gonalves, A. P., Melo, M. J. and Muralha, V.S.F. 2009. A study on red lead degradation in the medieval manuscript, Lorvo Apocalypse (1189). Journal of Raman Spectroscopy, 40: 1966-1973. 42. Based on the quantification of As, Hg and S by micro-EDXRF and molecular fingerprint by microRaman. 43. The experiments were followed by microRaman, micro-EDXRF and XRD, for more details please see 41. 44. Also known as galena. 45. The general reaction pathway could be described by the general reactions described in eq (1) and (2): (1) 2Pb3O4 + As2S3 3PbS + (AsO4)2Pb3
3 (2) 2AsO4 + 3H2O As2O3 + O2 + 6OH
46. Colour coordinates in the Lab* system, L*=35.75 a*=-9.99 b*=7.58 (D65 illuminant and 10. observer angle).
fig.7 details from lorvo apocalypse for orange degradation: no visible degradation (angel, f.118), degradation (beast, f.158), and severe degradation that may be misinterpreted as a paint shadow (dragon, f.153v). two macro photos evidence it
Bottle-green32
The bottle-green paints are also suffering from extensive degradation, and even though this has not affected the brightness or the intensity of the deep green 46, it has however affected the invisible component of the paint: the binder. Bottle-greens are being lost due to a weakening in the cohesion and adhesion of the support. The bottle-green has a glassy fractured appearance under the microscope and under polarized light it appears as a non-crystalline structure. Historic reconstructions were performed to further characterize this paint, using several copper salts as the source for colour. The experimental design was based on the material analysis of the greens found in manuscripts, namely by -FTIR, pointing to the existence of a copper proteinate. The reproduction of the bottle-green colour concludes that copper sources play an important role and, from all tested compounds, only verdigris, previously dissolved in a 50 year aged homemade vinegar, proved to be adequate. Work is currently in progress on the framework of a PhD Thesis 40, to further elucidate both its synthesis and degradation mechanisms.
162
47. As already described, the study of colour in the Santa Cruz monastery collection was carried out in the framework of a Molab mission and it was possible, with the aid of a portable fluorimeter, to acquire emission spectra and to measure fluorescence lifetimes in situ. As indigo is easily identified both by fluorescence data as well as by Raman spectroscopy it was possible to conclude that the dark blues were obtained mixing lapis lazuli with indigo. With regard to the carmine or dark red colour, the presence of emission was detected but it was low and no match was found in the database. However, it was possible to conclude that the dark red was obtained again by mixing an inorganic pigment, vermilion, with this dye. The detection of these mixtures for the dark blue and red colours is, in itself, important evidence for the characterization of the Santa Cruz palette, as it was not observed in the manuscript collections of the two other monasteries, Lorvo and Alcobaa.
4. Colour construction and meaning

In all the collections, as already described, colours are applied almost as a pure colourant in a proteinaceous tempera. In a considerable set of manuscripts, matiz and highlights are also used to give volume and to create a greater visual impact, Figure 4. Exceptions have been found in the construction of some dark reds and blues, obtained by mixing an inorganic pigment with this dye 47; in the pinks and greys, obtained by adding white lead to red or black (or to dark blue indigo); in vermilion, which, is sometimes mixed, when applied on a large surface, to a red lead (minium) with no alteration of its hue, or in the use of fillers, such as calcium carbonate. None of the mixtures mentioned represents an obstacle for the colours visual impact, being luminous and with unique hues (red, green, blue, yellow), where the contrasts red/blue or red/green are frequently used. This contributes to the effect of radiance of light and colours which streams from the Portuguese medieval illumination.
4.1 Santa Cruz

The manuscripts dating from the middle of the 12th century, such as Santa Cruz 58 and 17, were probably the first to be produced and present a palette consisting of yellow, red and green. Another palette with clear dominance of blues, yellows, greens and reds would later replace. The colours are opaque, saturated and bright and they are applied homogeneously in backgrounds, using predominantly the contrasts yellow/green/red, which would be replaced in the end of the 12th century by blue/red, such as in the Psalter Santa Cruz 27 or the Bible Santa Cruz 1. Figure 8
fig.8 details from santa cruz manuscripts. from top to bottom: santa cruz 20 in ff. 152, 173v, 128v and 92; santa cruz 1 in ff. 314, 338, 197, 206v, 364v, 362v, 364 and 362v,
163
4.2 Alcobaa: matiz and volumetry

It has been revealed that the religious community of Alcobaa knew the new processes of manuscript painting, delivered through prescriptions and treatises presenting a Byzantine influence. This influence is visible in Italy, as well as in Catalunya, Bourgogne and Champagne. Besides Saint Bernards prohibitions 48 on image and colour, the matiz technique generalized in the Clairvaux scriptorium allowed to develop volume and a diversity of shades in one letter, mostly at an ornamental level. In Alcobaa, the monks only applied monochromatism as an exception and used this technique in letters full of volume and exuberance. The ornament presents a more restrictive palette in the first manuscripts, diversifying latter in the liturgical manuscripts. This can be verified in the set of Alcobaa missals 49. Figure 9
48. Lucet, B. 1974. La codification cistercienne de 1202 et son volution. Roma: Editions Cistercien. Litterae unius fiant et non depictae (Letras de uma s cor e no figuradas). 49. BNF, Alc 249, 251-53, 255-59. 50. Miranda, M.A., Claro, A., Lemos, A., Miguel, C. and Melo, M.J. 2008. A Iluminura Portuguesa, uma abordagem interdisciplinar. Revista de Histria de Arte FCSH. 5: 232. 51. Gage, J. 2009. Color and Culture. London: Thames and Hudson, 61-64. 52. In this counting are not included canon tables, schema and marginal illustration. 53. Wirth, J. 1999. Limage a lpoque romane. Paris : Du Cerf, p.389 : La faon de placer les couleurs est particulirement intressante. Dans lensemble, elles forment des bordures et ne couvrent quexceptionnellement un champ, laissant le parchemin nu sur la plus grande partie de sa surface. De part et dautre de larche, Hugues dispose trois poutres adjacentes dans le sens de la longueur, formant des bordures successives dingale paisseur. La plus forte est lextrieur et celle du milieu est la plus fine, contracte entre les deuz autres. Dans la partie suprieure de larche, correspondant au temps de la loi naturel, le vert est lextrieur, la pourpre au milieu et le jaune lintrieur. Dans la seconde partie, correspondant la loi crite, on a successivement le jaune, le vert et la pourpre. Enfin dans la troisime partie, celle de la grce, on a la pourpre, le jaune et le vert.
fig.9 details from alcobaa manuscripts. from top to bottom: alcobaa 419 in ff. 1v and 91v, alcobaa 426 in f. 252 and alcobaa 446 in f. 96v
4.3 Lorvo
Even though the Lorvo collection is the most heterogeneous, it includes two manuscripts that present relevant characteristics in the national and international context, having, for this reason, a monographic treatment. They are Lorvo Apocalypse and the Book of Birds, which will be later discussed. Besides the restricted palette of red, orange, yellow and black colours, Figure 10, the particularities of Lorvo Apocalypse in the context of Romanesque illumina-
fig.10 colour mapping of lorvo apocalypse by relative areas
164
54. The etymology of carmine is described by Kristol, A.M. 1978. Color. Les langues romanes devant le phnomne de la couleur. Zurique: ditions Francke Berne, p 170: carmim, formation hybride partir dun lment de provenance arabe et dun lment latin (<qirmizi x minium). (...) quant / cramoisi/ (<qirmizi rouge de kerms, cest--dire rouge de cochenille aussi) dont lintroduction est postrieure dun sicle environ celle d/ carlate/ et de /carmin/ (premire attestation 1298), ceste le terme rouge cochenille qui designe le rouge fonc, pourpre, rouge violtre. 55. Anil is synonymous for indigo and it was commonly used in Portuguese until the first decades of the twentieth century. 56. Lombard, M. 1978. Les Textiles dans le monde musulman: du VIIe au XIIe sicle. Paris: Mouton Editeur, p.252. 57. These red dyes could be obtained from animal or plant sources, for more details please see (Melo, M.J. 2009. History of natural dyes in the ancient Mediterranean world. Handbook of Natural Colorants, T. Bechtold and R. Mussak (eds), Chichester: John Wiley & Sons 6-7). 58. On Chapters 13 and 14 (this one is incomplete, with no reference for the colour source). For producing rose colours, other recipes are present based on Brazilwood on chapters 8, 9, 27 and 44. 59. Castro, I. 2010. Notas sobre a lngua do Livro de como se fazen as cores (ms. Parma 1959) in Afonso, L.U. (ed.) The Materials of the Image. As Matrias da Imagem, Lisboa: Campo da Comunicao, p.87: escrita deste texto em aljamia hebraica, percebe-se a existncia, no sc. XV, de uma comunidade judeo-portuguesa bem integrada no tecido nacional, com elevada competncia de literacia tanto em hebraico como em portugus. 60. Strolovitch, D.L. 2010. Old Portuguese in Hebrew script: beyond O livro de como se fazem as cores, in Afonso, L.U. (ed.) The Materials of the Image. As Matrias da Imagem, Lisboa: Campo da Comunicao, 29-43. 61. Castro, I. 2010. Notas sobre a lngua do Livro de como se fazen as cores (ms. Parma 1959) in Afonso, L.U. (ed.) The Materials of the Image. As Matrias da Imagem, Lisboa: Campo da Comuni-
tion in Portugal are also due to the fact it is the only one with an iconographic programme 50. The exclusive use of theses colours makes us search for its meaning; the predominance of yellow (50% of the area), materialized in the orpiment, provoke a luminous effect, which may be related to an aesthetic of light common to the Christian and Arabic worlds, where it is linked to the divine 5. The symbolic and eschatological character of the text and image in the literary tradition of the commentary on Apocalypse, and the need to spread a message in the political and religious context of the peninsular Christian expansion, would have conducted the artist of the Lorvo Apocalypse to create an iconographic programme which is conveyed in 88 52 images. He opted for the use of contrasting and luminous colours applied in the backgrounds and transparent bodies, which seems to exalt the spirituality present in the text. The Portuguese illuminators would use this painting process for a long time. There is a predominance of the use of drawing, being colour applied in the backgrounds 53. This practice had a large widespread in the Western Christian manuscripts, namely by the Norman (Mont Saint Michel, Fcamp Abbey) and Southern France communities (Albi, Limoges and Moissac) and also the Saxon world. In the other manuscripts, we can say that, generally speaking, we observe an extensive use of carmine 54, directly applied on the parchment, like in Diagrama da Pomba (Book of Birds, f.6), or on the minium, to darken it and create an effect of relief (incidire). Moreover, it is mixed with white lead to create a scale from dark to light (matiz), Figure 4. This organic red, lac dye in the Book of Birds, could be a characteristic of the Portuguese medieval illumination and, together with blue 55 indigo, was a colourant used in luxury fabric dyeing. By that time, both colourants were integrated in the extensive net of trade routes created by the Arab conquest 56. The use of white lead for the white colour is also characteristic. The contemporary European production presents many common characteristics, such as the use of lapis lazuli for blue, of vermilion for red and the presence of minium in the palette. The latter being applied pure, as already described, or in a matiz from pink, dark red and very dark red (or black) 57 Figure 4. It is important to mention that, in the Book on how to make colours 35, two recipes are described to obtain a lac dye carmine 58 as well as one for a parchment glue tempera. It is worth to stress that Besides Ivo Castro 59, a Portuguese expert in linguistics, also Devon Strolovitch 60, author of one of the recent critical editions of this manuscript, states that this manuscript is the product of a cult and socially integrated Hebrew community. Our critical edition of the science involved also shows that some recipes reflect complex technological processes but, generally, allowed us to obtain the colours described 25. So, we may conclude that there are evidences that, in the 15th century, the Hebrew community had the necessary knowledge to prepare a medieval colour paint and, even more important, this know-how was part of a more ancient tradition as discussed in a recent linguistic analysis 61. Further investigation in the technological aspects of the Book on how to make colours will probably allow us to reach conclusions on the contribution of this community to the colour of the Portuguese illumination.
165
cao, p.87: a co-presena de camadas lingusticas de diferente cronologia mantm a possibilidade de fontes textuais mais antiga. 62. Kristol, A.M. 1978. Color. Les langues romanes devant le phnomne de la couleur. ditions Francke Berne. 63. Melo, M.J. 2009. History of natural dyes in the ancient Mediterranean world. Handbook of Natural Colorants, T. Bechtold and R. Mussak (eds), p.7. 64. Possibly this vocabulary was brought by the Arabs into the Peninsula, but it derives from the Persian word for the blue stone lapis-lazuli, lzaward. Kristol, A.M. 1978. Color. Les langues romanes devant le phnomne de la couleur. ditions Francke Berne, 237- 238. 65. Lack of published molecular data on the ms. collections contemporary to ours.
fig.11 colour analysis for the book of birds of the portuguese monasteries (lorvo, santa cruz and alcobaa) and the french abbey clairvaux. reds used for lettering, being part of the illumination, are shown slightly out
4.4 General conclusions

From our investigation of colour of medieval Portuguese illumination, it emerges that the materials used in its production, play an important role in a global history. The cultural importance of the materials of colour is well presented in our language, where many terms used to designate colour come from the materials source, that is, medieval colourants. Thus, taking for example, the red and blue, the Portuguese etymology is for both directly related to the source producing the colour 62. In the case of red, the small parasite from which dye was extracted, vermiculum in Latin 63; in the blue, to the semiprecious stone lapis-lazuli from which the most coveted colour of the Middle Ages was obtained 64. Thus, one of the aims of this investigation would be to compare the materials used in the creation of colour of the Portuguese medieval illumination with some of the used ones in other important European centres, such as France and Spain. This comparative analysis is however not yet possible 65, and in order to overcome this obstacle, we decided to launch a project on colour mapping, which includes quantifying relative proportions of the area occupied for each colour
166
66. Manuscript preserved at Troyes Library, BM 177. 67. As proposed by Pastoureau, M. 2002. Bleu. Paris: ditions du Seuil, 67-68; Gage, J. 2009. Color and Culture. London: Thames and Hudson; and Brussatin, M. B.1995. Oro e azzurro. Lezione sui colori, Veneza: Libreria Editrioce Cafoscarina, p 13. 68. Based on the findings of Rmy Cordonnier and discussion with Patricia Stirnemann.
in the illuminations. In a first phase, we will quantify the areas of the main colours used in the Portuguese manuscripts: the red, carmine, orange, blue, green, black and yellow. In the Portuguese case, it is possible to associate a molecular palette to these colours. We tested the potential of this approach with the Books of the Birds from the Portuguese monasteries, comparing them with the one of the Clairvaux monastery 66, Figure 11. In a first analysis, it was possible to verify that the distribution of colours of the codex of Alcobaa is identical to the one of Clairvaux, as expected, considering the similarity of its iconographic programme. One of the most interesting and unexpected results which emerged from the analysis of the Book of Birds from Lorvo is the evidence of the amount of red being identical to the blue one, as if those two colours existed in its contrast. We emphasise the low proportion of green in this manuscript, which was increasing from the Santa Cruz manuscript to the Alcobaa at the expense of the blue. Will these data allow us to propose a review in the dating of theses codices towards a more recent dating of the Lorvo manuscript? In other words, the progression of blue 67 (and the regression of green) corresponds to a change of paradigm, the blue ascending to the main colour in the medieval palette, taking for itself, the meaning of spiritual colour? This is a question to which only an analysis on a larger number of manuscripts could answer; an analysis that is actually going on. However, it is possible to add that this hypothesis is in agreement with the new dating of the Clairvaux manuscript, 1170, proposed by Patricia Stirnemann and Rmy Cordonnier 68.
5. Outlook
One of the often-neglected components in the study of medieval illumination is what it is made from: paint. As if the materials, their process and technology involved were not also a cultural product, so many times paradigmatic. In the technological revolution started in the Middle Age, which coined the western modern civilization, the technological production is a cultural product with a high impact. With the present interdisciplinary project, we hope to give contribution to a cultural history of medieval colour, intrinsically global. The work of the investigation conducted by the interdisciplinary team has put forward new hypotheses in relation to aesthetic and symbolic options of the illuminators and to the dating of the manuscripts, Lorvo Apocalypse and De Avibus respectively. Moreover, we tried to start a narrative that integrates the contributions of different cultures, which coexisted in the Romanesque Portugal. We have integrated diverse evidences which, even apparently dispersed, allowed us to forecast a consistent pattern. For example, two of the most important colours of the medieval palette, the red and the blue, were traded and/or processed by the Arabs or Jews. Thus, lapis lazuli and lac dye (dark red) were imported from Afghanistan and India; the vermilion, though it could be imported from the Almaden mines as a mineral, became a production of Arabic sciences, that is, obtained by (al)chemical synthesis, combining mercury and sulphur. This recipe is clearly described in detail in the Hebrew-Portuguese Book
1 67
on how to make colours 69. In the case of, it is in the meaning of the colour itself that we can find evidences of the Islamic culture influence, consistent with the fact it was copied in Lorvo, located in a region that was an important centre of the Mozarabic culture, from a Mozarabic codex of the 10th century. Within the current on-going project Colour in medieval illuminated manuscripts: between beauty and meaning 70, we wish to deepen this lead and seek for new evidences which will allow us to unveil the meaning of the colour in medieval codices, detecting the contributions of the three medieval cultures which forged together Portugal, that is the Hebrew, Muslim and Christian cultures. Finally, and as a way to share our findings with the interested audience, in addition to hands-on workshops of production of medieval illumination that we frequently organize since 2007 71, we will also explore the modern computer interaction technologies 72, with the main focus on children and young publics. We intend to achieve an engaging, intuitive and easy to use interactive system. This installation will simulate the illuminations creation process in the medieval period, addressing several aspects from the materials origin and production methods to the painting process. It will also show users the historical and social context of that time and reveal the meanings of the colours and images depicted.
69. The new date for this ms. 15th century, was recently proposed by the team of FLUL that studied the treaties on the art of painting produced in Portugal from Medieval Ages to 1850. It is probable that some of the processes described in the manuscript reflect older traditions. As Ivo Castro could establish, based on linguist analysis, the first version could date form the 13th14th centuries. Ins Villela-Petit was one of the first scholars to propose a new date for the ms. For more details please see, Lus Afonso (p14) e Ivo de Castro (p93-94) in Afonso, L.U. (ed.) 2010. The Materials of the Image. As Matrias da Imagem, Lisboa: Campo da Comunicao. As well as paper by Maria Joo Melo and Catarina Miguel in Varia section of this volume, pag. 290. 70. PTDC/EAT-EAT/104930/2008. 71. As an example, the workshop that was organized in the framework of the International Day of Museums, to a public ranging from 7 to 77 years old (please see http://www.cvtv.pt/ imagens/index.asp?id_video=295 and http:// www.cvtv.pt/imagens/index.asp?id_video=302). These workshops are based on the step by step [Melo, M.J., Miranda, M.A. (coord.), Claro, A., Lemos, A. and Miguel, C. 2007. descoberta da cor na iluminura medieval com o Apocalipse do Lorvo e o Livro das Aves, published online at http://www.dcr.fct.unl.pt/step-by-step]. 72. See paper on Varia section of this volume, pag. 294.
Bibliography
Afonso, L.U. 2010. New devellopments in the study of O Livro de como se fazem as cores das tintas in Afonso, L.U. (ed.) The materials of the Image. As matrias da Imagem. Lisboa: Campo da Comunicao. Azevedo Santos, M.J. 1994. Da visigtica carolina. A escrita em Portugal de 882 a 1172. Lisboa: FCG/JNICT. Blondheim, S.1928. An old Portuguese work on manuscript illumination. Jewish Quarterly Review. 19: 97-135. Brussatin, M. B.1995. Oro e azzurro. Lezione sui colori. Veneza: Libreria Editrioce Cafoscarina. Cabral, J.M.P. 2001. Histria Breve dos Pigmentos. III - Das artes grega e romana. Qumica- Boletim da Sociedade Portuguesa de Qumica. 82: 57-64. Cabral, J.M.P. 2006. Histria Breve dos Pigmentos. IV. Das Artes da Idade Mdia (1. parte). Qumica- Boletim da Sociedade Portuguesa de Qumica. 103: 33-44. Castro, I. 2010. Notas sobre a lngua do Livro de como se fazen as cores (ms. Parma 1959) in Afonso, L.U. (ed.) The Materials of the Image. As Matrias da Imagem. Lisboa: Campo da Comunicao, p.87.
168
Clarke, M. 2001. The Analysis of Medieval European Manuscripts. Reviews in Conservation, 2:3-17. Clarke, M. 2001. The Art of All Colours: Medieval Recipe Books for Painters and Illuminators. Dorchester: Archetype Publications. Cocheril, M. 1978. Routier des Abbayes Cisterciennes du Portugal. Paris: Fondation Calouste Gulbenkian. Coupry, C. 1999. Les pigments utiliss pour lenluminure Fcamp, XI-XII me sicle in Bouet, P. and Dosdat, M. (eds) Manuscrits et enluminures dans le monde normand (X-XVme sicles). Caen: Presse Universitaires de Caen, 69-79. Ferreira, T.D. and Santana, A.C. 1993. O tratamento documental de manuscritos ao servio da investigao: a experincia da Biblioteca Nacional. Cadernos BAD. Lisboa: APBAD. Gage, J. 2009. Color and Culture. London: Thames and Hudson. Gomes, S.A. 2000. Limine Conscriptionis: Documentos, Chancelaria e Cultura no Mosteiro de Santa Cruz de Coimbra: Sculos XII-XIV. Coimbra: Polimage Editores. Kristol, A.M. 1978. Color. Les langues romanes devant le phnomne de la couleur. Zurique: ditions Francke Berne. Lombard, M. 1978. Les Textiles dans le monde musulman: du VII e au XIIe sicle. Paris: Mouton Editeur. Lucet, B. 1974. La codification cistercienne de 1202 et son volution. Roma: Editions Cistercien. Martins, A.A. 2003. O Mosteiro de Santa Cruz de Coimbra. Sculos XII-XV. Lisboa: Centro de Histria-Universidade de Lisboa. Mattoso, J. 1996. A vida religiosa dos beneditinos portugueses durante o sc. XIII in Religio e Cultura na Idade Mdia Portuguesa. Lisboa: Imprensa Nacional-Casa da Moeda, 2. ed. Mattoso, J. 2009. Recension to the paper by Nascimento A., Fernndez Cton, J. (ed). Liber testamentorum coenobii laurbanensis (estdios). Lon: Centro de Estudios y Investigacin in Mediavalista. 7. Lisboa: FCSH. Mattoso, J. 2009. Os antepassados dos navegadores in Naquele tempo. Ensaios de Histria Medieval. Lisboa: Temas e Debates/Crculo de Leitores. Melo, M.J., Miranda, M.A. (coord.), Claro, A., Lemos, A. and Miguel, C. 2007. descoberta da cor na iluminura medieval com o Apocalipse do Lorvo e o Livro das Aves. Published online at http://www.dcr.fct.unl.pt/a-descoberta-da-iluminuramedieval. Melo, M.J. 2009. History of natural dyes in the ancient Mediterranean world in Bechtold, T. and Mussak, R. (eds). Handbook of Natural Colorants. Chichester: John Wiley & Sons, 3-18. Melo M.J., Claro A. 2010. Bright light: microspectrofluorimetry for the characterization of lake pigments and dyes in works of art. Accounts of Chemical Research. 43(6): 857- 866. Miguel C., Claro C., Melo, M.J., Lopes, J.A. 2009. Green, blue, greenish blue or bluish green? Copper pigments in medieval Portuguese Illuminations. Sources and Serendipity Testimonies of Artists Practice. London: Archetype: 33-38.
169
Miguel, C., Claro, A., Gonalves, A. P., Melo, M.J. and Muralha, V.S.F. 2009. A study on red lead degradation in the medieval manuscript, Lorvo Apocalypse (1189). Journal of Raman Spectroscopy, 40: 1966-1973. Miranda, M.A. 1996. A Iluminura Romnica em Santa Cruz em Santa Cruz de Coimbra e Santa Maria de Alcobaa. PhD Dissertation. Lisboa: Universidade Nova de Lisboa. Miranda, M.A. 1996. A Iluminura de Santa Cruz no Tempo de Santo Antnio. Lisboa: Edies Inapa. Miranda, M.A., Claro, A., Lemos, A., Miguel, C., Melo, M.J. 2008. A Iluminura Portuguesa, uma abordagem interdisciplinary. Revista de Histria da Arte. 5: 228-245. Miranda, M.A., Lemos, A., Miguel, C. and Melo, M.J. 2010. On Wings of Blue: The history, materials and techniques of the Book of Birds in Portuguese scriptoria in Afonso, L.U. (ed.) The Materials of the Image. As Matrias da Imagem. Lisboa: Campo da Comunicao. Moreira de S, A. 1960. Revista da Faculdade de Letras, 4: 210-223. Moura, L., Melo, M.J., Casanova, C., Claro, A. 2007. A Study on Portuguese Manuscript Illumination: The Charter of Vila Flor (Flower Town), 1512. Journal of Cultural Heritage. 8-299:306. Nascimento, A.A. 1998. Vida de D. Telo in Hagiografia de Santa Cruz de Coimbra. Ed. critic, trad. e estudos introdutrios. Lisboa: Colibri. Nascimento, A.A., Miranda, M.A. (coord.) 1999. A Iluminura em Portugal: Identidade e Influncia (do sc.X ao XVI): catlogo da Exposio. Lisboa: Biblioteca Nacional de Portugal. Nascimento, A.A. 2002. Liber Testamentorum cit. N.C. Borges, Arte monstica em Lorvo, sombras e realidade Das origens a 1737. Lisboa. Pastoureau, Michel. 1996. Couleur, images, symboles. Paris: Le Lopard dOr. Pastoureau, M. 2002. Bleu. Paris: ditions du Seuil, 67-68. Peixeiro, H. 1998. A Iluminura do Apocalipse do Lorvo. Provas pblicas para professor coordenador. Tomar: IPT/EST/DTAG/ACG. Seixas de Melo, J., Melo, M.J., Claro, A. 2006. As molculas da cor, na Arte e na Natureza. Qumica-Boletim da Sociedade Portuguesa de Qumica, 101: 44-55. Strolovitch, D. 2005. Old Portuguese in Hebrew Script: convention, contact, and convivncia. PhD Dissertation. Cornell University. 116-184; Strolovitch, D.L. 2010. Old Portuguese in Hebrew script: beyond O livro de como se fazem as cores, in Afonso, L.U. (ed.) The Materials of the Image. As Matrias da Imagem. Lisboa: Campo da Comunicao, 29-43. Vilares Cepeda, I. and Ferreira, T.D. 2001. Inventrio dos Cdices Iluminados at 1500. Vol.2. Lisboa: Imprensa Nacional Casa da Moeda. Wirth, J. 1999. Limage a lpoque romane. Paris: Du Cerf.
170
Biographies
Ana Claro obtained her PhD in Conservation and Restoration in 2009 from the Universidade Nova de Lisboa, Lisbon, Portugal. She was a trainee assistant at this University (2005-2009) and joined the Associate Laboratory REQUIMTE-CQFB (FCT-UNL) in 2004 and the Medieval Studies Institute (FCSH-UNL) in 2005. Her primary research focus is the study of materials applied in illuminated manuscripts. She worked at the Getty Conservation Institute as a Visiting Post Doc (2009). Currently she is doing research on Asian organic colorants as Post Doc at Centre for Overseas History (CHAM-UNL) and Hercules laboratory vora University. Ana Lemos is graduated in History-Art History by Faculdade de Letras, Lisbon University. In 2006 she integrated the research team supervised by Maria Joo Melo and Maria Adelaide Miranda, working on the study of colour in Portuguese Medieval Illuminations. In 2009, Ana Lemos got a Master degree by Faculdade Cincias Sociais e Humanas, New University of Lisbon, with the thesis Um Novo olhar sobre o livro de Horas de D. Duarte. Currently, she is working on her PhD thesis entitled Catlogo dos livros de horas iluminados de origem francesa nas coleces pblicas portuguesas (primeira metade do sculo XV). Antnio Pereira Gonalves is graduated in Chemical Engineering by Instituto Superior Tcnico (IST), Technical University of Lisbon, and PhD in Chemistry (IST). He has been focusing his research on solid-state chemistry, Materials Science, High Temperature Chemistry and Crystallochemistry. Antnio Pereira Gonalves is currently a Principal Researcher at Instituto Tecnolgico e Nuclear (ITN), and is responsible for the High Temperature Synthesis and Crystal Growth and for the Films Deposition Laboratories. Catarina Miguel studied Chemical Engineering at Instituto Superior Tcnico (IST), Technical University of Lisbon. After a final internship at the Portuguese Institute for Conservation and Restoration (IPCR) in 2004, she joined Maria Joo Melo research group (DCR-FCT-UNL) on the study of colour in Portuguese Medieval illuminations. Presently, she is focusing her studies on art and technological source research, in the framework of her PhD thesis Le vert et le rouge: A study on the materials, techniques and meaning of the green and red colours in medieval Portuguese illuminations. Joo A. Lopes is graduated in Chemical Engineering by Instituto Superior Tcnico (IST), Technical University of Lisbon, and PhD, also by IST. He is a Researcher at Requimte, and member of the Physical-Chemistry group of Faculdade de Farmcia, Oporto University. His research interests include Chemometrics, Process Analytical Chemistry, Vibrational Spectroscopy, Process Systems Engineering and Biospectroscopy. Joo A. Lopes current research activity is focused on the development and application of chemometrics and data-mining tools through the principles of process systems engineering. Main research interests consist on on-line, in-situ, non-destructive and automatic systems for monitoring. Maria Adelaide Miranda has graduated with a degree in History from Faculdade de Letras, Lisbon University. MA and PhD degrees, Faculdade de Cincias Sociais e Humanas
171
(FCSH), New University of Lisbon. Associated professor of Medieval Art History at the Department of Art History (FCSH-UNL, Lisbon). Her research interests in Medieval Art include the study of colour, iconography and manuscript illuminations, having coordinated interdisciplinary projects such as The colour of medieval illuminations and Imago. She is a fellow researcher at Instituto de Estudos Medievais and associate researcher at Instituto de Histria da Arte as well as member of the editorial board of the journals Medievalista on line (IEM) and De Arte (Univ. de Len). Maria Joo Melo obtained her PhD in Physical Chemistry, in 1995, from New University of Lisbon. In 1999, after a Post Doc at ICVBC-CNR in Florence, she joined the Conservation Unit at the New University of Lisbon, where she is responsible for the C&R scientific laboratory. Since 1999 she has also been researcher at Requimte. Her research interests include the colour of medieval illuminations and conservation of Modern Art, namely the study of the mechanisms of photodegradation in polymer systems and colour paints. Other areas of interest are Color in Art and Nature and Semiochemistry. Rita Castro studied Art History, Chemistry and Conservation at the New University of Lisbon (Caparica). 2009, M.A. Dissertation on Chinese hand-painted wallpaper (18 th century). Since 2010, she carries out research in medieval illumination manuscripts, particularly in the assessment of parchment and in the study of the meaning of color. Currently, she intends to develop these subjects in a PhD in Conservation and Restoration. Solange Muralha obtained her PhD in Physical Chemistry in 2005, from the University of Lisbon. The same year she started her research in the field of Conservation Science at the University College of London, with special emphasis on Islamic illuminations. After this thrilling experience with collaborations with the Institute of Archaeology in London and the Vitoria and Albert Museum, she joined the Conservation Unit at the New University of Lisbon. The research interests follow the colour in medieval illuminations field but also provenance studies on artistic glazed tiles. The author acquired a proficiency in several analytical techniques, especially Raman spectroscopy.
172
Appendix 1
Experimental Design
The global approach and experimental design set-up for a comprehensive study on colour production for medieval Portuguese illuminations will be briefly described. Missions, where the equipment is transported to the institutions where the manuscripts are preserved, were prepared after a careful selection of the manuscripts by the Art history experts. From these manuscripts a relevant number of folia were analyzed in order to ensure statistical relevance. MicroRaman, microFTIR, microXRF and microXRD are powerful complementary techniques and were used for the characterization of colorants in medieval manuscripts. Analytical techniques based on emission fluorescence have been recently experimented for the identification of dyes and lake pigments with rewarding results, and were used to study the organic colourants in Portuguese medieval illuminations. During the missions, the first screening is carried out by microEDXRF, which indicates the possible colorants and extenders present and allows a first quantification of these elements, moreover its 70 m enable us to obtain data that is representative for the distribution, in the manuscript, of a certain paint colour. This was performed by selecting in each folio, and for each colour, five representative areas and three points in each colour area enabled to have relevant and representative data, allowing to have statistical value that can also be used in other analytical techniques, such as chemometrics. To quantify mixtures even more acquisitions were needed. An estimate time to analyse a 10x10 cm illumination with the entire palette applied (Figure A2.1) , reveals that five hours would be needed to assure its representativeness. For instance in a blue colour, 11 micro-EDXRF analyses, 11 micro-Raman and 1 microsampling would take at least 50 minutes. Regarding the number of illuminations by manuscript and the real time to analyse them, one must choose between a folio by folio analysis or selecting representative illuminations. MicroRaman, which allows for high spatial resolution (1- 5 m spot) and enables the diverse paint components to be excited separately as well as emission fluorescence techniques are used to address specific points, such as the molecular characterization of an inorganic pigment or a dye, respectively. Together with the spectroscopic investigation, paints are also analyzed by optical microscopy which allows to understand how the final colour is built up (e.g., by layers or by mixture), to detect possible degradation phenomena and to sample the colour paints that will be subjected to a more detailed characterization in the laboratory, as regards the colorants, binders and additives. Designed micro-sampling presents several advantages, namely it keeps to a minimum the handling of the manuscript and allows for more detailed studies in the laboratory, without the time constraint that in situ mission implies. Typically, in the laboratory a sample is first analyzed by microFTIR, allowing for binder characterization and to gain an insight into the full paint formulation. If a dye is present, microspectrofluorimetry may be carried out first, as it requires no contact with the sample and employs a low-intensity radiation. MicroRaman and microDXR may be used to address specific points; with both methods, sample destruction may occur.
fig. A2.1
173
Resumo
A grande experincia do Instituto em anlises no-invasivas para a caracterizao de pigmentos, tintas de escrever e corantes permite a identificao das paletas utilizadas pelos iluminadores e, por vezes, dos prprios scriptoria. Caracterizam-se ainda tratamentos e intervenes passadas bem como se avaliam os mtodos e produtos mais adequados para a sua conservao. Neste trabalho sero apresentados os resultados obtidos, por anlise molecular (microRaman) e elementar (XRF e PIXE), no estudo de trs manuscritos, o Pontificale 492 (sc. XIII, Museo Diocesano, Salerno, Italia), a Bibbia Amiatina (scs VI-VIII, Biblioteca Medicea Laurenziana, Florena, Italia) e o ms Piana 3.207 (sc. XIII, Biblioteca Malatestiana Cesena, Italia).
palavras-chave
pigmentos xrf raman pixe ouro musivo
Abstract
The huge experience of the Institute in non destructive analysis and characterization of pigments, inks, colours and dyes allows to recognize the palettes used by miniaturists, sometimes distinguishing the scriptoria themselves, identify posterior treatments and additions and assess the proper methods and products to be used in conservation. We present the results obtained by means of non destructive spectroscopic molecular (Micro Raman) and elemental (XRF and -PIXE) techniques in three case study, the Pontificale 492 (13th century, Museo Diocesano, Salerno, Italy), the Amiatinus Bible (7th-8th century, Biblioteca Medicea Laurenziana, Firenze, Italy) and the ms. Piana 3.207 (13th century, Biblioteca Malatestiana Cesena, Italy).
key-words
pigments xrf raman -pixe mosaic gold
illuminations: secrets, alchemy and conservation in three case studies

m a r i n a b i cc h i e r i
Istituto centrale restauro e conservazione patrimonio archivistico e librario, Roma (Italy), Head of Laboratory of Chemistry, marina.bicchieri@beniculturali.it
mich el a mo n t i
Istituto centrale restauro e conservazione patrimonio archivistico e librario, Roma (Italy), Researcher of Laboratory of Chemistry, michela.monti@beniculturali.it
g i ova n n a p i a n ta n i da
Istituto centrale restauro e conservazione patrimonio archivistico e librario, Roma (Italy), Researcher of Laboratory of Chemistry, giovanna.piantanida@beniculturali.it
a r m i da s o d o
Istituto centrale restauro e conservazione patrimonio archivistico e librario, Roma (Italy), Researcher of Laboratory of Chemistry, and Universit di Roma Tre. Department of Physics, sodo@fis.uniroma3.it
Achieving highly accurate information on miniatures composition allows to customize conservation and restoration strategies aiming at the best preservation with the least impact. Cultural heritage objects represent complex and multifaceted problems and conclusions drawn from observations with the naked eye are likely as faulty as those drawn from limited diagnostics. Moreover, the use of destructive techniques as many traditional methods actually are is to be avoided, since samples are often unique, priceless specimens. In order to work out reliable, unambiguous information from non destructive methods a truly scientific approach, based on multidisciplinary analyses, laboratory simulations and proper statistics, must be combined to the knowledge of history, art and of course alchemy that guide the illuminators hand. From the experimental point of view, a rich literature upon the use of elemental (XRF, LIBS, EXAFS, -PIXE ...) and molecular (ATR-FTIR, DRIFT, FT-Raman, MicroRaman...) spectroscopies can be found. In the light of previous experiences (Bicchieri et al. 2001, 169), the combination of XRF, -PIXE and Micro Raman was considered the proper setup choice. XRF spectrometry is often (Malzer et al. 2004, 229; Cesareo et al. 2004, 703) applied to pigments and dyes analysis: information can be achieved in reasonable times (minutes). Unfortunately, such information is only qualitative on paper and parchment substrates, since elements lighter than Al (Si, in some cases) give a fluorescence response below the detection limit, so it is impossible to detect organic substrates even though the organic matrix contributes to the global spectrum. At the same time, the X-rays penetration depth increases inversely to the atomic weight of the target element and a sheet of paper or parchment is completely crossed. Quantitative XRF analysis requires, in fact, that the thickness of the sample could be considered as
175
infinite, i.e. greater than the average thickness crossed by the used XRF radiation. Calibration of the measuring arrangement and the use of reliable calibration model that must be representative of the matrix and target element morphology (particle size distribution, heterogeneity and surface condition) and concentration ranges of the sample to be analysed are also necessary. Contrariwise, -PIXE allows controlling the penetration into the sample and quantitative elemental information even on trace and lighter elements (above Na) can be obtained. The simultaneous use of XRF and -PIXE gives an almost complete characterization of the elemental composition of the sample (Pappalardo et al. 2005, 114). Raman spectroscopy is a technique sensitive to vibrational states of matter, that actually constitute a fingerprint of the material. Raman spectroscopy can be used (Colthup et al. 1990, 60) to identify the molecular composition or even the different state of aggregation of the species in the sample under investigation. The advantages of Raman spectroscopy consist in the short measurement time (seconds-minutes) and in the high sensitivity to inorganic compounds. The main problem with Raman applications is that fluorescence is excited together with the molecular states and sometimes screens the measurement, making it impossible to extract data. As far as measurements on the Pontificale 492 and Bibbia Amiatina are concerned, the portable -PIXE and XRF spectrometers were entirely designed and developed at LNS/INFN laboratory and the portable micro-Raman spectrometer was a custommade instrument developed at Physics Dpt, Roma Tre University. Measurements on the ms. Piana 3.207 were performed with the Assing Lithos 3000 portable XRF spectrometer and the Renishaw In-Via Reflex Raman microscope currently owned by the Institute (Bicchieri et al. 2008, 1074). The full sets of analyses on Bibbia Amiatina and Pontificale 492, including the instruments and their specifications, are fully described in dedicated articles. (Bicchieri et al. 2001, 169; Pappalardo et al. 2008, 466). They will be only synthetically reviewed in present work, whereas a major focus will be given to the investigation campaign on the ms. Piana 3.207 and the related experimental work on pigments analysis and simulation of mosaic gold. The Bibbia Amiatina, Codex Amiatinus, ms. Laur. Amiat. 1, is the last surviving out of three copies of the Bible produced in Ceolfriths scriptorium in the area of the twin monasteries of Wearmouth Jarrow in Northumbria between the end of the 7th century and the beginning of the 8th. It is composed of 1029 parchment sheets, measures 345 mm x 540 mm and weighs around 50 kg. The manuscript is the most ancient and complete witness to the Vulgate Latin Bible and it is also a great example of English uncial writing and of illuminations in anglo-saxon style, influenced by the Mediterranean-Bizantine one. Therefore its historical interest is inestimable. The first goal of the analyses laid in the establishment of the original arrangement of the sheets present in the first booklet. Some pigments, in fact, left offsets on the adjacent sheet that were analyzed under ultraviolet light of two different wavelengths (253 nm and 365 nm). It should be underlined that the ordering of the sheets coming out from such a method did not necessarily correspond to the original one, but rather
176
i l l u m i n at i o n s : s e c r e t s , a l c h e m y a n d c o n s e r vat i o n i n t h r e e c a s e s t u d i e s
represented the sequence that the pages had likely followed for the bulk of the documents lifetime. However, scholars agreed to adopt it when re assembling the book. For a complete characterization of the precious manuscript all illuminations were analyzed coupling Raman, XRF and -PIXE and have been recognized and classified (Table 1).
Colour White Red Blue Brown Yellow Green Gray Metallic inks Composition calcite; white lead red lead; organic lakes; earths not yet identified pigment, copper-based pigments sometimes mixed with earths earths (iron and sometime copper) often mixed with orpiment or with gold orpiment, organic lakes, earths blue and yellow (orpiment) mixtures; verdigris mixed with an unidentified organic compound white (sometime calcite, sometime white lead) mixed with blue and carbon black. Silicates containing copper and iron were also used mixture with variable composition of copper, lead, silver, gold and sometime added with orpiment (arsenic) and earths (iron)
table 1 the bibbia amiatina palette
The ms. 492 Pontificale of Museo Diocesano di Salerno is dated back to a period from the end of 13th century and the incoming of the 14th. The manuscript consists of 349 parchment sheets each having dimensions 400 x 285 mm. The miniatures of excellent quality, ascribed to anonymous illuminators working in the Bologna area, were not always completed: in some cases only the preliminary design was present, in other cases the preparation was not covered by the gold-leaf. These circumstances gave the unique opportunity to investigate the different techniques used by the miniaturists and in particular the composition of the gold preparations and their comparison with the medieval recipes. Using the different penetration depth of the three applied techniques it was possible to carry out a completely non-destructive stratigraphy. Resulting palette is summarized in Table 2 .
Colour Gold Yellow White Red Blue Brown Pink Preparations for gold
Composition pure gold leaf, only rarely added with silver mosaic gold; yellow-ochre lead white red lead, cinnabar lapis-lazuri, azurite ochre, earths organic lake mixed with lead white armenian bole + gypsum; armenian bole + gypsum + white lead; gypsum + white lead
table 2 the pontificale palette.
177
The experimental results also allowed to prepare laboratory reproductions of the original illuminations, that were artificially aged and deteriorated in order to test different adhesives on them and to choose the best and less invasive preservation treatment. The Decretum Gratiani, ms. Piana 3.207 (13th-14th c.), also known as Concordantia discordantium canonum is a collection of Canon law. The code consists of 329 parchment sheets (455 mm x 280 mm) and is preserved at the Biblioteca Malatestiana in Cesena, Italy. The restoration of ms. Piana 3.207 was the topic of project Salviamo un codice 2nd edition, an initiative of Alumina Nova Charta press with the High Patronage of the President of the Italian Republic (Savoia 2008). Ms. Piana 3.207 had been preserved in a very good condition up to the 80s, when it was dramatically maimed: thirteen pages were torn off, damaging the sewing of the binding, and stolen. In the following years, the law enforcement retrieved all the missing pages but one (c.107) which is now considered lost. ICPL (now Icpal), the institutional authority appointed for library materials preservation, was responsible for the thorough scientific analysis of the manuscript and for the subsequent restoration. XRF and Raman measurements allowed us to characterize the palette (Table 3).
Colour Gold Yellow White Red Blue Brown Green Black Other colours Preparations for gold Composition pure gold leaf, only rarely with traces of copper; mosaic gold not fully reacted mosaic gold; yellow-ochre lead white red lead, cinnabar, earths indigo, azurite ochre, earths not fully reacted mosaic gold mixed with indigo carbon black (for inks, contour lines and shadows) mixture of the previously described pigments for other colours and tones white lead
table 3 the ms. piana palette.
Very interesting was the discovery that gilded miniatures had been realised either with pure gold or with mosaic gold (tin IV sulphide; Fig.1 and 2). The latter pigment, a yellow crystalline powder that closely resemble gold, was widely used in Europe in the Middle Age and in the Renaissance and in ms. Piana was applied not as goldimitation but in order to obtain particular hues. To confirm the presence of mosaic gold, a standard Raman spectrum was necessary. Due to the fact that modern preparation of stannic sulphide differs from the ancient methods and that Raman spectra from literature (Edwards et al. 2003, 2291) had been obtained from the modern compound, we decided to prepare mosaic gold following ancient recipes. Many texts were consulted, but the most detailed recipe was found in De Arte Illuminandi 1. Concerning amounts of reagents, not cited in
1. Handbook of technical recipes, written in Latin by anonymous authors in the 14th century and currently kept at the Biblioteca Nazionale di Napoli (Napoli, Italy).
178
2. Sulphur and Mercury are the alchemic pair. They are transported into the labyrinth of the transformation of matter by the external fire. In the Temple (crucible, Ed.) their marriage is celebrated and is only possible with the help of the arcane fire of salt the alchemic Priest able to transform the metals. In chemical terms, amalgam increases the reactivity of tin; mercury reduces sulphur to sulphide. In meantime elemental tin is oxidized to tin IV. Ammonium chloride decreases the reaction temperature, allowing the formation of stannic sulphide.
the recipe, we decided to use the proportions suggested by Villavecchia: Dizionario di Merceologia e di Chimica Applicata (Villavecchia 1932). The procedure foresees many steps: preparation of Hg-Sn amalgam, addition of sulphur and ammonium chloride and heating for a long time, choosing the temperature in order to obtain the desired hue. In this process we can recognize the two first steps of the alchemic Opus Magnum, the nigredo (mixing amalgam and the two others compounds the mixture becomes black) and the albedo (during heating, dense white fumes rise) 2. When no more fumes appear, heating is stopped and a gold-coloured powder is found in the crucible: this is mosaic gold. Even when observed at the microscope, the similarity between the mosaic gold prepared in the lab and the pigment used on the ms. Piana 3.207 is total and Raman spectroscopy demonstrates that the two compounds had the same structure. The effectiveness of the presented techniques in analysis of such invaluable manuscripts, the related scientific advance and the more than positive feedback from restorers who take advantage from the obtained information, confirmed once more the Institute choice of pioneering and always pursuing multidisciplinarity based on strictly non-destructive methods.
fig.1 x spectrum of mosaic gold from the original (christ robes). x ray source 96mo; collimator 0.5 mm diameter, acquisition time 600 s, x tube tension 25 kV, current 0.300 ma, esd silicon detector, resolution 160 eV at 5.9 keV
fig.2 comparison between the raman spectra of mosaic gold laboratory sample (black line, microscope image a) and of the original pigment (red line, microscope image b). excitation = 785 nm
179
Bibliography
Bicchieri, Marina, Francesco Paolo Romano, Lighea Pappalardo, Luigi Cosentino, Michele Nardone, Armida Sodo. 2001. Non-destructive Analysis of the Bibbia Amiatina by XRF, PIXE- and Raman. Qvinio: International Journal on the History and Conservation of the Book 3: 169 79. Bicchieri, Marina, Michela Monti, Giovanna Piantanida and Armida Sodo. 2008. All that is iron ink is not always an iron-gall! Journal of Raman Spectroscopy 39, no 8: 1074-78. Cesareo, Roberto, Alfredo Castellano, Giovanni Buccolieri, Stefano Quarta, Maurizio Marabelli, Paola Santopadre, Marcella Leole, Antonio Brunetti. 2004. Portable equipment for energy dispersive X-ray fluorescence analysis of Giottos frescoes in the Chapel of the Scrovegni. Nuclear Instruments and Methods in Physics Research B 213: 70306. Colthup, Norman B., Lawrence H. Daly and Stephen E. Wiberley. 1990. Introduction to Infrared and Raman spectroscopy. San Diego: Academic Press Inc. Edwards, Howell G.M., Emma L. Dixon, Ian J. Scowen, Fernando Rull Perez. 2003. Lead/ tin mirror formation from mixtures of red lead and tin sulphide. Spectrochimica Acta Part A 59: 2291 99. Malzer, Wolfgang, Oliver Hahn and Birgit Kanngieer. 2004. A fingerprint model for inhomogeneous inkpaper layer systems measured with micro-x-ray fluorescence analysis. X Ray Spectrometry 33: 229 33. Pappalardo, Lighea, Andreas G. Karydas, Despina Kotzamani ,Giuseppe Pappalardo, Francesco Paolo Romano and Charalambos Zarkadas. 2005. Complementary use of PIXEalpha and XRF portable systems for the non-destructive and in situ characterization of gemstones in museums. Nuclear Instruments and Methods in Physics Research B 239:114-21. Pappalardo, Lighea, Marina Bicchieri, Michele Nardone, Giuseppe Pappalardo, Francesco Paolo Romano, Pio Alfonso Russo, Armida Sodo. 2008. The LNS portable PIXE-alpha system for the examination of gold preparations in the Salerno 492 code miniatures. X-Ray Spectrometry 37 no 4: 466 69. Savoia, Daniela, ed. 2008. Per giusta causa. Il restauro del Decretum Gratiani (Ms. Piana 3.207) della Biblioteca Malatestiana di Cesena. Padova: Edizioni Nova Charta. Villavecchia, Vittorio. 1932. Dizionario di merceologia e di chimica applicata, vol. 4. Milano: Ulrico Hoepli.
180
Biographies
Dr. Marina Bicchieri works since 1979 in cultural heritage field and from 2000 she is the director of the chemistry department in Istituto centrale restauro e conservazione del patrimonio archivistico e librario of Rome-Italy (former ICPL). Her research activity is focused on the chemistry of library materials (cellulose, parchment, inks, pigments, charges, adhesives, sizes), the study and set-up of new restoring products and new nondestructive diagnostic methods (-Raman, XRF, PIXE-, -IR, AFM, UV-Vis-NIR, SEM) for writing supports and graphic media. She taught classes in chemistry, conservation and restoration organized by Italian and foreign Universities and Public Institutions, the European Union, and several private institutions. Michela Monti obtained her Master Degree in chemistry at University of Rome La Sapienza in 2000. From 2001 to 2002 she worked as researcher in the field of conservation of cultural heritage at the University of Udine and at the Istituto Centrale per la Patologia del Libro in Rome. From 2002 to September 2005 she worked as consultant for industrial safety, focusing on the prevention of chemical risk. Since September 2005 she works for the Chemistry Department of the Istituto Centrale il Restauro e la Conservazione del Patrimonio Archivistico e Librario (former Istituto Centrale per la Patologia del Libro), focusing on the application of X-ray fluorescence for the analysis of library materials and carrying out research projects on new methods for restoration and conservation of paper and parchment. Giovanna Piantanida, after achieving a M. Sc. in Physics (2000), worked for three years in the Italian R&D site of L.M. Ericsson Corporate. She then decided to turn back to research activity and obtained a post-doc position in the Department of Physics, at University of Rome La Sapienza, co-funded by ICPL, an institution within the Italian Ministry for Cultural Heritage and Activities. The research project was the finalization of diagnostic protocols based on physics methods, in particular Atomic Force Microscopt. Currently, she is a fellow researcher at the Laboratory of Chemistry Icpal (former ICPL) and she is pursuing PhD in analytical chemistry at University of Ljubljana. Armida Sodo is researcher at the Physics Department of University Roma Tre. After the degree in Chemistry (1998), she obtained the PhD in Chemical Physics and then got a post-doc position at ESRF (European Synchrotron Radiation Facility). Her research studies concern principally spectroscopic principally Raman investigation on Cultural Heritage; in particular degradation process analyses, material characterisation and development of instruments devoted to Cultural Heritage applications. She gives university courses and is co-author of more than 40 publications on international journals, books and conference proceedings.
181
Resumo
As iluminuras de De Rijmbijbel van Jacob van Maerlant so consideradas das mais importantes pinturas dos Primitivos Flamengos. O fol. 152v apresenta a seguinte inscrio: Doe men screef int iaer ons heren MCCCXXXII verlichte mi Michiel van der borch; i.e.: Foi no ano de 1332 que fui iluminado por Michiel van der Borch. O que permite consider-las como as mais antigas obras de arte flamenga datadas e com colofon. Os aspectos estilsticos destas iluminuras esto de acordo com o idioma que era comum tanto na Flandres como no noroeste de Frana, sugerindo que a formao de Van der Borch pode ter a decorrido. Estas pinturas, de pequena dimenso, foram analisadas por reflectografia de infravermelho (IRR), microscopia ptica, fluorescncia de raios-X dispersiva de energias (-XRF), espectroscopia por reflectncia no UV-VIS (FORS) e ainda por tcnicas de imagiologia espectral. As anlises permitiram identificar as tintas e pigmentos bem como as diversas estratgias de aplicao da cor para organizao da composio. Comparam-se os resultados obtidos com as instrues presentes em tratados tcnicos contemporneos como o Liber de Coloribus siue Pictorum e o de Peter de Saint Omer, ambos provenientes do noroeste da Frana.
palavras-chave
flandres maerlant iluminura flamenga anlises cientficas de coloribus
Abstract
The illuminations in De Rijmbijbel van Jacob van Maerlant belong to the most important Early Netherlandish paintings. On fol. 152v it has the inscription: Doe men screef int iaer ons heren MCCCXXXII verlichte mi Michiel van der borch; i.e.: It was in the year 1332 that I was illuminated by Michiel van der Borch This makes the illuminations the earliest signed and dated Netherlandish works of art. Stylistic features of these illuminations fit quite well into the idiom that was current in Flanders and Northwest France, and suggest that Van der Borch may have been trained there. These small-scale paintings were examined with infrared reflectography (IRR) and light microscopy, energy dispersive micro- x-ray fluorescence (-XRF), fiber optic reflectance spectroscopy (FORS), and some (radio)spectral imaging techniques. The analyses helped to identify paints and pigments and the different strategies that were employed to deal with colour in order to organise the composition. Terminology and formulae in contemporary technical treatises such as the Liber de Coloribus siue Pictorum, and in the De Coloribus Faciendi by Peter of St. Omer, both from North-western France are compared with results of scientific examination of Van der Borchs ajuda na configurao do texto e sugestes de Ana Catarina Sousa, Pedro Fialho Agradecimentos porilluminations.
key-words
southern flanders maerlant netherlandish illumination analyses de coloribus
early netherlandish manuscript illumination: technical aspects of illuminations in the rime bible of jacob van maerlant
a rie wa l l ert
Rijksmuseum Amsterdam, Amsterdam (The Netherlands)
The Rijmbijbel (The Hague, Rijksmuseum Meermanno-Westreenianum, MS 10 B 21) is one of the most important historical texts in Netherlandish culture. It is a conflation of the Historia Scholastica by Petrus Comestor and the Bello Judaico by Flavius Josephus. But contrary to contemporary tradition, this text written in 35.000 verses by Jacob van Maerlant, was in the vernacular rather than in Latin. Maerlant was the most important Netherlandish author of the Middle Ages. He was the first to deal in the vernacular with such diverse matters as geography, politics and the arts of government, biology and the interpretation of dreams. He was also the first to put the whole Bible in rime in the Netherlandish language, and to write a complete world history. (Van Oostrom, 1996) Therefore, the Rijmbijbel has achieved a canonical status for its significance in the development of the Netherlandish language. At the same time, the illuminations in this manuscript can be considered to be the most important Early Netherlandish paintings. Stylistic features of these illuminations fit quite well into the idiom that was current in Flanders and Northwest of France. The execution of the illuminations in the Maerlant manuscript compares very nicely with the characteristics of either the miniatures
183
in MS. Lat. liturg. d.42, or those in the Psalter MS Douce 5, both in the collection of the Bodleian Library, Oxford, in a psalter in the Kongelige Bibliothek in Copenhagen (Ms. 3384.8o), or those in the so called Ruskin Hours, presently in the J. Paul Getty Museum, Los Angeles (Ms. Ludwig IX 3; 83.ML.99). These manuscripts all share the same stylistic features, compositional organisation, colour schemes, facial types, and patterned backgrounds that are characteristic for the gothic style that was current in Flanders and Northern France in the beginning of the fourteenth century. (Lyna, 1944, Les Fastes Gothiques, 1981, 276-362) There is little documentary information on the production of those manuscript illuminations. Identified works of art of the period are scarce. If we can identify illuminations of the period and the region, we usually do not know the names of their makers. If we know the names of the illuminators, we generally do not have any certified example of their works. Only a few artists are known by name: Hennequin de Bruge, Girard de Hainault, and Jean de Saint-Omer. Only from documentary evidence do we know that Girard de Hainault worked for the dukes of Burgundy around the 1350s, and that Jean de Saint-Omer worked on one of the tombs of Jeanne d Evreux. We do not know, however, what their art works looked like. The exception being Hennequin who was, and still is, famous for his impressive design of the stunning Angers Apocalypse tapestries.(Calkins, 1979, 248-250) The drama and narrative of the figures, presented more than life size in the Angers tapestries, appears to be re-enacted, albeit in a very much smaller scale, in the miniatures of the Rime Bible. Under the full page painting in the Maerlant manuscript, of the Destruction of Jerusalem on fol. 152v is the inscription: Doe men screef int iaer ons heren MCCCXXXII verlichte mi Michiel van der borch; i.e.: It was in the year 1332 that I was illuminated by Michiel van der Borch. This makes the illuminations the earliest signed and dated Netherlandish works of art. Even though this signature and the presence of his name in the archives strongly suggest a more Northern Netherlandish origin (Utrecht), it is generally agreed that the illuminator of the Rime Bible must at least have had training in southern Flanders or northern France. Terminology and formulae in more or less contemporary technical treatises such as the Liber de Coloribus siue Pictorum (British Library, MS Sloane 1754), and the De Coloribus Faciendi by Peter of Saint-Omer (collection Jehan LeBegue, Bibliotheque Nationale, Ms 6741), both from North-western France, may therefore provide relevant information on the making of Van der Borchs illuminations (Liber de Coloribus, De Coloribus Faciendi). The Liber de Coloribus for instance, describes the manner in which red lead should be prepared and applied, and the colours that could be mixed with it: Minium molendum est cum aqua sicut uermiculum, et eiecta aqua de cornu, siccabis, et ita ponenda est glarea ovi, et sic illuminatur ex eo. Hoc minium cum nullo colore miscetur nisi cum albo et uermiculo. You grind minium with water, like vermilion; and when the water has been poured out of the vessel (horn), you dry the colour; and then add white of egg to it and use it for illuminating. This minium must not be mixed with any other colour except
184
e a r ly n e t h e r l a n d i s h m a n u s c r i p t i l l u m i n a t i o n : t e c h n i c a l a s p e c t s o f i l l u m i n a t i o n s i n t h e r i m e b i b l e o f jaco b va n m a e r l a n t
1. Observations with the microscope were done with a Zeiss stereo microscope with 8, 12, 20, 32, to 50 x magnifications. Images were recorded with a digital Leica DFC 420 C camera. We also used a AM411T-Dino-Lite Pro digital microscope with 10x ~ 50x - 200x magnifications. 2. Infrared reflectography (IRR) was done with a Hamamatsu C 2400-07 camera, equipped with a N2606 IR vidicon, a Nikon Micro-Nikkor 1:2.8/55 mm lens, and a Wratten 87c (B+W 093) filter. Digitized documentation is done with a Pinnacle PCTV framegrabber. IRR-assemblies were made with Adobe Photoshop CS2.
with white and vermilion. (Liber de Coloribus, 282-283) Analyses of red areas in the Rime Bible do indeed reveal those admixtures, and the crack pattern is indeed typical for a binding medium based on glair. 1 Similarly, the finding of a strong presence of Ca, Fe, and S in a tan-coloured pigment mixture, visible in abraded areas of the gilding could be related to a passage in the Liber de Coloribus: Ocrum si necessarium tibi fuerit in percameno, de ocro siue de gipso If you need ochre on parchment, grind it well with water, and when this has been done, add strong white of egg to it. But you must know that ochre is needed only by painters of murals, except that, when you wish to make a letter of gold, you lay it in first with ochre and gypsum., and: Gypsum is ground like ochre, but you will not have any use for it except when you wish to lay gold in books. And then you put on the gold as we have described in speaking of ochre. (Liber de Coloribus, 284-285) The finding of those elements is perfectly fitting for the use of a mixture of yellow ochre, a hydrated iron oxide (Fe2O3 . H2O), and gypsum, a calcium sulphate dihydrate (CaSO4 . 2H2O). The Liber de Coloribus is very explicit that these pigments can only be used for laying a gold ground. And indeed, those were the only instances where that combination of elements could be found. All the illuminations and small figures in the borders of the texts were begun with a fine brush dipped in a carbon black. These drawings define the outlines of the figures. No attempts to indicate shading or volume were made at this stage. The underdrawings often show uncertainties in elements that went beyond typically formulaic figures and shapes like the faces of Christ, or conventions for trees or draperies.2 Van der Borch especially seems to have had difficulties in defining a correct anatomy for the human figure. This would suggest that these, sometimes very complex illuminations were not based on, or copied from precedents in modelbooks, but invented on the parchment. (figs. 1 and 2) These drawings were made after
Fig.1 a. expulsion from paradise (detail), normal light b. infrared reflectogram assembly
Fig.2 a. creation of man (detail),normal light b. infrared reflectogram assembly
185
the text was written, the scribe writing in iron gall ink, leaving specific spaces open for the illuminator to work on. 3 This is evidenced by the fact that sometimes the illuminator made corrections in carbon black to the texts written in ink. This correction process also corresponds with contemporary recommandations: Sed si in pergamenis cum ovo distemperatos assumes, If you wish to lay black over other colours on parchment, you must not put incaustum, but know that you must take charcoal distempered with egg (De Coloribus Faciendi, 138-139 ) (fig. 3).
3. Analyses of the writing ink showed a strong presence of Fe, Ca, smaller amounts of Mn, K, Pb, and S, but also traces of Cu and Zn. This might suggest that the ink could have been made or stored in a brass vessel, or more likely, that the iron vitriol needed for the making of the ink also contained traces of zinc vitriol and other associated minerals. 4. Energy dispersive x-ray fluorescence (XRF) analyses were generally done with a Bruker Artax -XRF spectrometer, 40kV, 500A, 60 sec., Mo-anode, 0.090m capillary lens, Helium flush (1.7 L/min), over 50keV energy range. For the distribution images a dwell time of 10 seconds was allowed.
Fig.3 a. illumination with text (detail), normal light b. infrared reflectogram assembly
The outlines in carbon black provided the basis for painting. This was continued in a fairly limited palette. Larger areas appear to be blocked in with single unmixed colours. Modulation of drapery folds was accomplished by laying thin translucent washes over these colour fields. This was further emphasised in the shadow areas with emphatic strokes of carbon black and in the highlights with linear contour strokes of pure and unmixed lead white. Definition of forms is still largely dependent on the elongated contours and strokes in carbon black paint. In general, Van der Borch seems to have had a preference for the use of single, unmixed colours. The exception being the flesh tones where admixtures of vermilion in lead white frequently occur. This can be demonstrated quite nicely in the elemental distribution images of the head of Christ. 4 (fig. 4) In this image the area for the gold ground is indicated by the distribution of calcium, present in the gypsum. The (abraded) gilding is shown by the distribution of the M lines for gold. The vermilion, HgS (here indicated by the distribution of mercury M lines), is shown to be mixed throughout the flesh tone with lead white (Pb L, and Pb M). Otherwise, vermilion appears to have been very sparsely used. This practice corresponds to warnings in the treatises: Si vero minium.est in minio,: If the vermilion is very good and new, I put two parts of it, and scarcely the third part of minium. But if the minium is dusky and very old, put a half or a third part of the vermilion, and make the remainder of minium; and you must know that the older the vermilion is by nature, the darker and less useful it is; and the darker it is, less of it must be
186
5. A remarkable feature of the azurite used in this manuscript was the notable presence of barium and potassium. As in some regions barium salts are paragenetic with azurite, this may provide an indication for geographical origin of pigment.
fig.4 distribution of elements, face of Christ (detail)
added to the minium ( De Coloribus Faciendi, 142-143) The vermilion must have been very old indeed! Only a slight admixture was very occasionally found in areas that were otherwise done in red lead. Most red passages, even the blood on the face of Goliath, or dripping from the wounds of Christ, are done in red lead. Similarly, the blue areas appear blocked out only in pure azurite, a copper(II)carbonate mineral (Cu3(CO3)2(OH)2). Azurite is a difficult pigment to work with. To prepare it for painting, the mineral needed to be broken up and reduced to powder. The coarser the powder, the darker and nicer the blue. If it was ground too fine, it tended to look rather pale greenish sky-blue. The separation of the coarser, darker particles from the finer, paler ones was accomplished by a process of levigation. The mineral was ground to powder while wet. As a turbid liquid, the slurry was transferred to a bowl in which the powdered substance remained suspended by continuous stirring. Then it was allowed to settle for some time, before being poured off to a second basin. This affords for the subsequent separation of the mineral into various grades of fineness, because of the slower subsidence of the finer particles from suspension. The turbid liquid, was poured into the first of a series of bowls, and was allowed to stand for a certain time. The coarsest and heaviest particles quickly subside, leaving the finer material suspended in the water, which was drawn from above the sediment into the next bowl. The liquid is passed from bowl to bowl, remaining in each subsequently longer than in the preceding one, since the finer and lighter the particles, the more time is necessary for their deposition. In the illuminations, in particular for the patterned backgrounds, a very powerfully deep blue azurite of good quality was used. 5
187
Observation of blue passages under the microscope clearly demonstrated that the artist exploited a finer grained and thus paler blue azurite for the lighter passages. It is notable that in order to obtain a lighter blue, choosing finer grade particles was apparently preferred over admixture with white. Darker, sometimes almost black accents, in particular on the background patterns, seem to be of organic nature and may very well be woad indigo. 6 An equally strict adherence to the use of unmixed pigments in the Rime Bible was encountered in the application of copper-based greens, i.e. verdigris (Cu2(OAc)4(H2O)2), again much as recommended in the texts: Viride de Gretia in uaseoperari poteris., Put verdigris in a dish made of bronze or electrum and pour wine upon it, so that the wine may become green. Pour the liquid portion of this mixture into another bronze vessel, and pour wine once more into the first vessel from which you remove the liquid. And when this too has been saturated and removed, add wine a third time. Then you may leave off. And: Pone uiride in uino.et sic bonum erit., Put some green in wine, and rub it well with your finger; when it has settled, take the liquid part and put it in moderate sunlight, or elsewhere, to thicken. When it is fit for writing, put it in a vessel made of glass or wax, and you will be able to keep it in a good condition for a long time. If it is too dark, add a little saffron and the powder of calcined bones. If you wish to illuminate more easily, take some yolk of egg, and mix the green wine with it, and grind the green pigment with this mixture upon the stone, and temper it with the same; and thus it will be good. (Liber de Coloribus, 282-285) By treating copper acetate in this manner and making an almost saturated solution of it, the pigment dissolving into the wine converts into a very dense, syrupy, copper carboxylate substance. The pigment thereby looses its grittiness and acquires much more pleasant working properties. The illuminator obtained a splendidly deep emerald green paint that flowed easily from the brush. It was applied in relatively thick opaque layers but also in thinner, quite translucent paint films, allowing the reflective brightness of the parchment or better still, gold leaf to shimmer through. (fig. 5) The smoothness of the thin, translucent green accords beautifully well with the grainy, particulate texture of the deep blue azurite. We also found frequent passages executed in a light, opaque, pink colour, which were often glazed over with darker, deep red, sometimes purple or maroon, translucent strokes. The organic colorant for both the opaque pink as well as the translucent purple was tentatively identified as a brazilwood lake. 7 Brazilwood lakes were the most important organic reds in the Middle Ages (Wallert, 1986, 52-70, Roger, et al, 2003, 155-170). The Liber de Coloribus advised: Lignum brasilii accipitur, purpureus color, Take brazil-wood and scrape it down with a knife into a dish. And after it has all been reduced by scraping, cover it with white of egg. And when it has steeped and commenced to grow ripe, let alum be added in proportion to the quantity of the brazilwood. When the brazil has saturated it, the liquid should be drawn off and kept in another vessel; and when this has been done, add white of egg again, and when it has become saturated, draw it off If you mix white with brazil, it makes rose-
6. As no samples could be taken, a proper identification by chromatography was not possible. The optical features under the microscope and behaviour of the dark blue in infrared, were found to be consistent with indigoid colorants. 7. Fiber optic reflectance spectroscopy (FORS) was done with an Avantes AvaSpec-2048 Fiber Optic Spectrometer, based on symmetrical Czerny-Turner design with 2048 pixel CCD detector array. 200-1100nm. Mikropack HPX-2000 xenon light source.
fig.5 micrograph of green translucent glaze over reflective background (magnification 20x)
188
8. Canon EOS 400D, with 24-105mm 1:4L lens and Nikon D70 with Micro Nikkor 60mm 1:2.8D lens were used. Frequently 81A, and 85B-2B filters were used. Image processing was done with the use of digital filters and Adobe Photoshop CS2.
colour; and with azure it makes purple. (Liber de Coloribus, 284-285) In the Rime Bible, the white added to the brazil wood extract was a simple calcium carbonate white. This imparted opacity to the mixture and served as colour-stabilising substrate for the organic colorant. Images taken in ultraviolet light give a good impression about the extent of the use of this organic colorant 8. (figs. 6 and 7) It appears to have been applied quite generously in remarkably swift and open touches. The analyses of the illuminations helped to identify paints and pigments and the different strategies that were employed to deal with colour as a means to organise the composition.
Fig.6 a. tobit healing his father s eyes (detail), normal light b. distribution of organic lake in ultraviolet fluorescence
Fig.7 a. tobit setting out on his journey (detail), normal light b. distribution of organic lake in ultraviolet fluorescence
189
According to the historical technical treatises compositions should be formed on the basis of a three-step system of colour application. The points of departure are the individual fields of colour. Each of these fields of colour tends to be provided with a specifically prescribed colour for darker shades: incidis. In the next step, each field of colour is provided with its own fixed colour of a much lighter nature: matizatura. In the technical treatises of the time, such as the Liber de Coloribus siue Pictorum and De Coloribus Faciendi, this rigidly ordained system of operative forms seems to have been strictly prescribed. In the making of the Rime Bible miniatures, however, all the procedures described in the treatises seem to have been followed, except for the application of the incidis and matizatura system. There the artist allowed himself considerable liberties. Especially in his use of washes with organic colorants, consciously exploiting effects of opacity and translucence, Michiel van der Borch went beyond the rules of tradition 9. In his lively brushwork he added a remarkable sense of freedom and liveliness to the functionality of the mediaeval method.
9. In this development towards techniques that allowed greater flexibility, alternative approaches induced by the work of the great encyclopaedists may have played a role. Especially Book XIX: De Coloribus, odoribus et saporibus (pp. 848-871) of Bartolomeus Anglicus De Proprietatibus Rerum must have been influential. (Meier-Staubach, 2000, 451-469).
Bibliography
R.G. Calkins, Monuments of Medieval Art, Cornell University Press, Ithaca and London, 1979. De Coloribus Faciendi: Magistri Petri de Sancto Audemaro, De Coloribus Faciendis, in M.P. Merrifield, Original Treatises on the Arts of Painting, John Murray, London, 1849, vol. I, 116-165. Les Fastes du Gothique, le Sicle de Charles V, Galeries Nationales du Grand Palais, Paris 1981. Liber de Coloribus: D.V. Thompsom, Liber de Coloribus siue Pictorum, from Sloane MS. No. 1754, Speculum, I (1926), 280-307. F. Lyna, De Vlaamse Miniatuur van 1200 tot 1530, Amsterdam, 1933. C. Meier-Staubach, La matrialit et limmatrialit des couleurs a propos du traite De Coloribus dAvranches 235, Science antique, Science mdivale, actes du colloque international, eds L. Callebat and O. Desbordes, Olms-Weidmann, Hildesheim, 2000, 451-469. P. Roger, I. Villela-Petit and S. Vandroy, Les laques de brsil dans lenluminure mdivale: reconstituition a partir de recettes anciennes, Studies in Conservation, 48 (2003),155 170.
190
F. Van Oostrom, Maerlants Wereld, Prometheus Amsterdam, 1996. A. Wallert, Verzino and Roseta Colours in 15th Century Italian Manuscripts, Maltechnik/ Restauro, 92, (1986), 52-70
Biography
Arie Wallert holds a PhD from Groningen University. Since 1996, he has been curator for the Scientific Examination Department of Paintings of the Rijksmuseum Amsterdam, senior scientist in the Department for Conservation RMA. His principal interests are the analysis of museum objects, the characterisation of pigments, the study of art technical historical sources and the scientific examination of mediaeval manuscript illumination. Address: Department of Conservation, Rijksmuseum, Hobbemastraat 22, 1071ZC Amsterdam; a.wallert@rijksmuseum.nl.
191
Resumo
A fluorescncia de raios-X dispersiva de energias (XRF) foi utilizada na anlise de uma Bblia do sc. XIV proveniente de Npoles. Este equipamento transportvel permite uma anlise in situ dos materiais presentes nestas luxuosas iluminuras. Puderam assim ser identificados pigmentos como o vermelho de chumbo (mnio), branco de chumbo, ocre vermelho, vermelho, azuis e verde de cobre, amarelo de estanho e chumbo em combinao com pigmentos orgnicos ou folha de ouro e prata ou ainda com tinta de estanho.
palavras-chave
manuscritos iluminados fluorescncia de raios-X dispersiva de energias anlise in situ pigmentos folhas metlicas
Abstract
X-ray Fluorescence spectroscopy (XRF) has been used to examine a 14th century bible manuscript, originating from Naples. This portable equipment provides a powerful aid to analyse in a non-destructive way the materials of the rich illuminations. Mineral pigments like red lead, lead white, red ochre, vermilion, copper blue and copper green, lead tin yellow in combination with organic pigments and with gold and silver leaf or with tin paint could be identified.
key-words
illuminated manuscript x-ray fluorescence spectroscopy non destructive analysis pigments metal foils
analysis of the anjou bible

m a r i n a va n b o s
Royal Institute for Cultural Heritage (KIK/ IRPA), Laboratory Department, Jubelpark 1, 1000 Brussels (Belgium) marina.vanbos@kikirpa.be
l iev e watte e u w
K.U.Leuven, Illuminare, Center for the Study of the Illuminated Manuscript, Department of Archeology, Art History and Musicology, Blijde Inkomstraat 21, 3000 Leuven (Belgium) lieve.watteeuw@arts.kuleuven.be
Introduction to the Bible

The Anjou Bible is a unique bible manuscript originating from the beginning of the fourteenth century. The bible was ordered by Robert I of Anjou, king of Naples and was intended as a present for Andreas of Hungary who was engaged to the granddaughter of Robert of Anjou, Johanna of Naples 1. When Andreas was murdered in 1345, the Bible came in the hands of Johannas chancellor Nicoli Alifio, as a present for his diplomatic missions. The precious manuscript remained in Royal circles and in 1402 the manuscript is described in the inventory of Jean Duc de Berry (1340-1416), brother of the French King Charles V (Avril, 1969; Avril et al, 1984, 2005). At the end of the 15th century beginning of the 16th century, the manuscript ended up at the library of the Arras College in Leuven (Belgium) via the Bishop of Arras and there it stood for centuries until finally in 1970, the manuscript was deposited at the Maurits Sabbe Library of the Faculty of Theology of the Catholic University of Leuven 2. The bible contains almost the complete text of the old en new testament (fig.1). Its size
1. Born in Naples, Johanna was the daughter of Charles, Duke of Calabria (eldest son of King Robert of Naples) and Marie of Valois (a sister of King Philip VI of France). At the age of seven years (1334), she was betrothed to her six-yearold second cousin Prince Andrew of the Hungarian branch of the House of Anjou, the son of Charles I of Hungary and younger brother of Louis I. On the death in 1343 of her grandfather, Robert of Naples, his will provided that Andrew should be crowned King of Naples in his own right as well as Joans, Robert having displaced Andrews father, Charles Robert, from the Neapolitan throne. The Cardinal crowned Joan alone as Queen of Naples at Santa Chiara in Rome in August 1344. After the assassination of Andrew in 1345 (remaining unclear if under her own orders or not), Joan married three more times: with Louis of Taranto, with James IV of Majorca and Prince of Achaea and with Otto, Duke of Brunswick-Grubenhagen. 2. Nicolaus Ruterius founded the college in 1508 as a home for poor students. From then until the 18th century, it was gradually converted into a residential college, with a reception hall, rooms, a library, a chapel, two courtyards and an interior garden. After the French Revolution it was sold,
fig.1 anjou bible, full page illuminations kik/irpa brussels
193
is 420 by 280 millimeters and it contains 338 folios with two full-page miniatures and more than 160 decorated initial and marginal illuminations leaves. The text has been written by one scribe but the rich illumination is the work of at least three different hands. One of them names himself on folio 308 at the sequence of the Apocalypse: Christophorus di Orimina (fig.2). He was the best-known illuminator in the Neapolitan surroundings in the second quarter of the fourteenth century (Maere 1909, 279 and Maere 1910, 25) (Brm, 2007).
and the building passed through the hands of a succession of private owners until it was repurchased by the university of Leuven in 1921.
fig.2 anjou bible, detail of folio 308 kik/irpa brussels
fig.3 anjou bible with marginal decoration of a second illuminator kik/irpa brussels
The two full-page illuminations in the beginning are clearly made by his hand. The perspective and architectural settings are simple and straightforward and the facial expression isnt very detailed nor expressive, using the profile of King Robert and other royal members. A second illuminator added decorations when the codex was probably already in possession of chancellor Alifio since he adapted the rich decorations to the new owner. Systematically the imperfection of the edge cuttings in the parchment were camouflaged by a wide range of fantasy birds with long necks and tales, all executed in soft pastel colors (fig.3). The third hand is probably the best miniaturist and he is responsible for all the marginal work and decoration. Often classical themes and fantasy animals are added to the narrative scenes (fig.4). Despite the fame and the richness of the manuscript, the bible has almost never been reproduced nor displayed in public. The conservation state of the manuscript was bad: in the beginning of the 20th century the manuscript was re-bound but the covers were too tight making the manuscript difficult to open; the parchment was folded and pigments and gold leaves were flaking off. Therefore in May 2008, a conservation and preservation project started. The project
fig.4 anjou bible with decorations of a third illuminator kik/irpa brussels
194
a n a ly s i s o f t h e a n j o u b i b l e
3. For generous support and discussion we thank: Jan Van der Stock (K.U.Leuven, Research Center for Medieval Art), Christina Ceulemans (Koninklijk Instituut voor het Kunstpatrimonium), John Lowdon (Courtauld, London), Chris Coppens (K.U.Leuven, Tabularium,); Leo Kenis en Luc Knapen (K.U.Leuven, Library, Faculty of Theology) and Nancy Turner (Senior paper conservator, J.P. Getty Museum, Manuscript Department). 4. Macro photographs (Nikon D 70, AF microNikkor 60 mm) were documenting the techniques and the damage-assessment (Cumulus 5.5/ digital asset management software). Further infra-red photography and infrared reflectography (InfraCAM SWIR short-wave infrared video camera) was combined with highly detailed digital photographs with direct, transmitted and racking light (digital camera MAMIYA RZ 67, Digital back Light Phase I / Nikon D 70 and EOS I D-Mark II, 100 mm macro lens).
is a collaboration between the Department of Art History, K.U.Leuven, the Royal Institute for Cultural Heritage in Brussels and Museum M of Leuven and is funded by the InBev-Baillet Latour Foundation 3.
Experimental
The manuscript is now unbound and the loose folios offer a unique opportunity to examine the manuscript in detail in the laboratory. It is obvious that the study of such a precious and delicate work of art has to be done using non-destructive techniques. First an extensive series of highly detailed digital photographs with transmitted light and complementary infrared digital photographs were taken 4 (Watteeuw 2009, 168; Watteeuw et al, 2008, 310; Watteeuw and Van Bos, 2010, 147). Then analyses were performed using X-ray fluorescence spectroscopy (XRF). This XRF technique offers a multitude of advantages: it is a non destructive technique with a very small spot size of only 70 m, which is important to be able to analyze small features in a miniature, it offers the possibility to simultaneously analyze multiple chemical elements and it is a fast analysis technique: each measurement only takes 120 seconds. We used the mobile Arttax equipment (Bruker AXS Microanalysis, Germany) with a Rhodium tube and a built-in color CCD camera (fig.5). The exact location of the measuring spot can be followed (or changed) via the computer screen and also the focusing of the measuring head to the desired position is done via the computer. During the experiments an energy of 50 kV, a current of 500 A and a Mo 25m filter were used.
Results for the first full page illumination

The first full page illumination is full of self promotion for the Anjou dynasty. One can read: Rex Robertus, rex expertus in omnia scientia. Robert is sitting on a polychrome throne ornate on both sides with gilded lion heads, under a colourful canopy with on the background the gilded lilies of the Anjou dynasty. The King is surrounded by eight cardinal virtues and at the bottom we find seven vices with Diablo.
The reds
XRF analysis of the orange border (fig.6, XRF 1) results in a spectrum with a large peak for lead (fig.7a and 7b). Although XRF is an elemental analysis technique with no information on the molecular composition of the lead containing pigment, we can assume with high certainty that lead is present as red lead (Pb3O4), since we are analyzing an orange-red color. Red lead, an artificial pigment, was already known
Fig.5 arttax xrf equipment kik/irpa brussels
195
Fig.7a and 7b xrf spectrum of the orange border and arttax image of the measuring area kik/irpa brussels
Fig.6 anjou bible, full page illumination with indication of some of the xrf measuring spots kik/irpa brussels
for miniature painting and manuscript illumination in Europe since the 8 th century on (West Fitzhugh 1986, 109, Munoz Vinas, 1998, 114); A different result is obtained for the red in the baldachin (see fig.6, XRF 2). XRF measurement results in a large peak for mercury indicating without doubt vermilion (HgS) as vermilion is the only pigment containing mercury (fig.8a and 8b). Although the pigment mercuric sulfide can also be obtained from the mineral cinnabar (HgS), the use of the artificial vermilion, made by heating mercury with sulfur, was already so widespread in the 14th century that nor in Il libro dellArte from Cennino nor in De Arte Illuminandi (Thompson et al, 1933, Brunello, 1975) a recipe can be found for the preparation of this vermilion. Measurement of the red used for the arm of the throne (see fig.6, XRF 3) gave again a different result: a spectrum with different peaks indicating a mixture of pigments: mercury (indicating vermilion), lead (lead white or red lead) and than a very large peak for iron (fig.9a and 9b) . This suggests that also red earth (containing hematite Fe2O3) is present here. Associated with the large iron peak is the small peak for
196
Fig.8a and 8b xrf spectrum of the red used in the baldachin and arttax image of the measuring area kik/irpa brussels
Fig.9a and 9b xrf spectrum of the red used for the arm of the throne and arttax image of the measuring area kik/irpa brussels
titanium probably indicating the mineral ilmenite (FeTiO3) present in the naturally found red ochres (Eastaugh, 2004, 320) A bright red can be found on the border of the miniature (see fig.6, XRF 4). XRF measurement gives only a weak response with a mayor peak for calcium (fig.10). This spectrum corresponds well with the spectrum obtained for the parchment itself. As organic reds are known for manuscript illumination, it is likely that an organic red, which can not be verified by XRF measurement, is responsible for the bright red color.
1 97
Fig.10 xrf spectrum of the bright red border kik/irpa brussels
The blue
All blue areas in the full page illumination, ranging from dark blue to light blue or from architectural decoration to the clothes of the King or one of the cardinal Virtues gave all a comparable result after XRF analysis: a large peak corresponding to copper, probably indicating azurite as blue pigment (fig.11). Azurite, a basic copper carbonate (2CuCO3. Cu(OH)2) is prepared by grinding and washing of the mineral which was occurring in copper deposits in Italy, Spain and particularly Germany (azurium de Alamania).
Fig.11 xrf spectrum of a blue area kik/irpa brussels
198
The green
XRF analysis of the green socle of the throne (see fig.6, XRF 5) resulted in a spectrum with peaks for lead, copper and tin. Although the exact interpretation of this result in terms of composition nd paint layer built-up can not be given, it is likely that a mixture of pigments is used here: lead white, a copper green or copper blue and a lead tin yellow (fig.12). Different relative amounts of these pigments, as shown by different peak heights of lead and copper, give different color intensities: more copper for the darker green areas. The mixture of malachite (CuCO3.Cu(OH)2) with giallorino (an artificial made lead-tin yellow) was described by Cennino. On top of this green layer, occasionally a transparent organic layer is applied to brighten the color.
Fig.12 xrf spectrum of a green area in the socle of the throne kik/irpa brussels
The white
The white inscription on top of the illumination is based on lead white as evidenced by the large lead peak in the XRF spectrum. Lead white, the basic lead carbonate (2PbCO3.Pb(OH)2), made by reaction of vinegar to lead strips, is described in the Arte Illuminandi as the only white color suitable for illuminating practice.
Metal: paint and foil

Gold leaf is used for the scepter hold by the King or for the decoration in the baldachin. This gold leaf is either applied on a grayish ground layer or on a bright red under layer.
199
A grey ground, based on gesso, is frequently observed in the gold size used in medieval manuscripts. XRF measurement of this layer indicates only calcium which is not surprisingly as with this XRF equipment only elements higher than sodium can be detected and the sulfur associated with the calcium can not be detected. As the gold leaf can be very thin, using a red under layer gives a warmer color to the gold compared to the same gold leaf applied on a grey under layer. This red under layer is often based on Armenian Bole (clay pigmented with iron oxides), following the recipes of the Arte Illuminandi or Il Libro dellArte. XRF measurement of the bright red under layer shows however the presence of vermilion as shown by the large mercury peak in the resulting XRF spectrum. To further embellish the gilding, an orange glacis layer is occasionally applied on top of it as can be seen for the belt of the King (fig.13). However, using our XRF analysis method, no information about the composition of this layer is obtained. Silver leaf is used as well: XRF measurement of part of the decoration in the socle of the Kings throne (see fig.6, XRF 6) shows the presence of silver (fig.14a and 14b). The silver leaf could be applied using the same size as for gold leaf.
Fig.13 arttax image of the decoration on the belt of the king kik/irpa brussels
Fig.14a and 14b xrf spectrum of decoration in the socle of the throne and arttax image of the measuring area kik/irpa brussels
A surprising result was obtained when analyzing the quadrofold in the canopy or the lilies in the baldachin (see fig.6, XRF 7). Although these decorations look like gold, XRF analysis reveals the presence of tin. Since the lilies are painted on top of the blue ground, a large copper peak is present as well in the resulting XRF spectrum (fig.15a and 15b). In this spectrum also a relative small peak corresponding to mercury is present. This result could point to the use of Mosaic Gold. In the Arte
200
5. Also Theolfilus describes the use of tinpaint in his chapter: How paintings in Books are Embellished with Tin and Saffron/ If you have neither of these [i.e., gold or silver] and still want to embellish your work in some way, take pure tin, scrape it very fine, mill it, and wash it as you did the gold. Then, with the same glue, apply it on letters or other places that you wanted to ornament with gold or silver. After polishing it with a tooth, take some of the saffron with which silk is dyed, pour glair without water over it, and let it stand overnight. On the following day cover with a pencil [dipped in this medium] those places you wanted to gild; leave the rest [of the tin bare] to take the place of silver. Then, using a quill, draw fine lines with minium around the letters, foliage, scrollwork, materials of robes, and the other places to be ornamented. Theophilus 37
Illuminandi mosaic gold (stannic sulfide SnS2) is described as an imitation gold color. The preparation of this mosaic gold starts with: take one part of tin and melt it, and throw upon it one part of pure quicksilver. Mosaic gold could replace the costly true gold powder but did not show the same full brilliance of gold. Therefore it was recommended to model it up with gold powder. Analysis of the decoration with small crosses in the baldachin shows this mixture of gold and tin applied as paint (no metal leaf) on top of the lead white layer, pointing out towards the use of mosaic gold. 5
Fig.15a and 15b xrf spectrum of lilies in the baldachin and arttax image showing the measuring area kik/irpa brussels
The coat of arms

At the bottom of the page, the coat of arms of chancellor Alifio is present (fig.16). For this coat of arms, the same materials as present in the miniature were identified: Vermilion applied on a white lead layer for the red and white triangles, red lead for the orange border, silver leaf and gold leaf. A different result was obtained for the golden fess which is a gold paint shell gold and not a gold leaf. Another difference was found when analysing the blue ground of the coat of arms. Although it concerns a very intense blue colour, only a relative small amount of copper was detected during the analysis (fig.17). This result could indicate the use of the blue pigment ultramarine (Na8..10Al6Si6O24S2..4) made from lapis lazuli, a pigment that can not be identified using XRF.
Fig.16 anjou bible, detail with coat of arms kik/irpa brussels
201
Ultramarine was a very expensive pigment with well known outstanding characteristics. The coat of arms as present now is not the original coat of arms. Originally, the coat of arms of the first owner of the manuscript, Andreas of Hungary was present. But when Alifio became the new owner, all original coats of arms were removed by abrasion and over painted, as a clear possession mark. In that respect, it is maybe not surprisingly that such an expensive and precious pigment is used. This however has to be confirmed using a complementary analysis technique (like non destructive micro Raman spectroscopy). The spectrum shows also an important iron peak (red earth) which indicates the underlying original coat of arms.
Fig.17 xrf spectrum of the blue ground in the coat of arms kik/irpa brussels
Conclusion
Although having a long history, the Anjou Bible was now for the first time ever examined in detail. The combination of highly detailed photographs with transmitted light in combination with the non destructive XRF analysis provides more insight into the craftsmanship of the miniaturist. A very rich pallet of mineral and organic pigments in combination with gold, silver and tin as foil or as paint has been identified. Although the use of a non destructive technique, like X-Ray fluorescence has its limits, this technique is extremely useful when analyzing very delicate and precious works of art.
202
Bibliography
Avril, Franois, Trois manuscrits Napolitains des collections de Charles V et de Jean de Berry, Bibliothque de lcole des chartes 127 (1969): 293-328. Avril, Francois et al., Dix siecles denluminure italienne (VI e- XVI e sicles), Paris, Bibliotheque nationale, 1984 . Avril, Franois en Marie-Thrse Gousset, Manuscrits enlumins dorigine italienne, Paris, 2005. Maere R, Une Bible angevine de Naples au Sminaire de Malines, Revue de lart chrtien 59 (1909): 279-291 and 60 (1910): 25-34. Brm, Andreas, Naepolitanische Bilderbibeln des Trecento: Anjou-Buchmalerei von Robert dem Weisen bis zu Johanna I, Wiesbaden, 2007. Brunello, Franco, De arte illuminandi: e altri trattati sulla tecnica della miniatura medievale, Vicenza: Pozza, 1975. Eastaugh Nicholas et al The Pigment Compendium, a dictionary of historical Pigments, Elsevier Butterworth-Heinemann, Oxford, 2004, 320. Munoz Vinas, Salvador Original Written Sources for the History of Mediaeval Painting Techniques and Materials: A List of Published Texts, Studies in Conservation, Vol. 43, No. 2 (1998), pp. 114-124. Thompson, Daniel V jr The Craftsmans Handbook Il libro dellArte by Cennino dA. Cennini, New Haven Yale University Press, 1933. Thompson, Daniel jr and Heard Hamilton, George, De Arte Illuminandi, The technique of Manuscript Illumination, Translated from the Latin of Naples MS XII.E.27, New Haven Yale University Press, 1933. Watteeuw Lieve The conservation assessment of the Philips of Clve Book of Hours (Brussels, Royal Library, Ms IV 40), in The Quest for the Original, Symposium XVI for the Study of Underdrawing and Technology in Painting, Bruges, September 2006, Louvain la Neuve, 2009: 168-174. Watteeuw Lieve and Van Bos Marina The conservation assessment of an Illuminated Book of Hours. Understanding craftsmanship through interdisciplinary research: preliminary investigation, in Diversity in Heritage Conservation: Tradition, Innovation and Participation, ICOM CC, 15th Triennial Conference 22 - 26 September 2008, New Delhi, 2008: 310-316. Watteeuw Lieve and Van Bos Marina Illuminating with Pen and Brush. The Techniques of a Fourteenth-century Neapolitan Illluminator Explored, in The Anjou Bible, A Royal Manuscript revealed, Corpus of Illuminated Manuscripts, Vol 18, ed Lieve Watteeuw & Jan Van der Stock, Peeters 2010, 147-169. West Fitzhugh, Elisabeth Red lead and Minium, in Artists Pigments, A handbook of their History and Characteristic. Volume1, ed. Robert L. Feller Editor, 109-139, Cambridge University Press, 1986.
203
t h at i t s e e m e t h to b e t h e t h i n g i t s e l f
Resumo
Nicholas Hilliard, que se tornou famoso no sc. XVI pelos seus retratos em miniatura, perseguia o ideal de uma representao da natureza to prxima do real quanto possvel. That it seemeththe work of god and not of man. Para atingir esse objectivo, utilizou diversas tcnicas, em parte baseadas na tradio da iluminura medieval, mas tambm na sua formao como ourives. Utilizou um nmero invulgarmente elevado de pigmentos pretos, alguns obtidos a partir das sementes de frutos carbonizadas, para reproduzir o efeito do veludo preto e das sedas. Classificou tambm os pigmentos em diversos tipos (categorias? graus). Por exemplo, um tipo mdio de branco de chumbo seria o ideal para imitar o aspecto de p de uma maquilhagem, enquanto que um outro tipo, constitudo por partculas muito finas, seria o ideal para o impasto brilhante necessrio para realar um lao. Esta ltima tinta aplicada muito espessa, resultante da mistura do branco de chumbo com muita goma, apresenta um craquel caracterstico e falta de adeso. Alguns destes danos podem ter sido visveis ainda no tempo de vida do artista, sem que no entanto o tenham levado a alterar a sua tcnica. Um outro exemplo, o da aplicao de prata em p (para as armaduras e os realces nas prolas), que tambm apresenta problemas de conservao uma vez que a prata cedo comeou a escurecer (embaciar?). Resinas aplicadas muito espessas em fundos prateados, para imitar o efeito brilhante das pedras preciosas ou esmaltes, foram muitas vezes esmagadas ainda quando frescas e apresentam-se lascadas ou quebradias quando secas. Os autores reconstruram muitos destes mtodos e efeitos especiais no mbito de um projecto levado a cabo no Victoria and Albert Museum (Londres). A produo das reconstrues baseou-se tanto nos resultados obtidos por anlises cientficas de miniaturas e observao do efeito visual como na pesquisa de fontes escritas contemporneas ou anteriores. Esta abordagem revelou-se muito frutuosa, pois permitiu tanto reproduzir os magnficos efeitos visuais que tero existido nas miniaturas originais como identificar danos tpicos associados ao uso de certos materiais e tcnicas.
palavras-chave
retratos em miniatura limning hilliard tcnica pictrica alquimia
205
Abstract
Nicholas Hilliard, a 16th century English painter of portrait miniatures or limnings, was driven by the idea of depicting nature as realistically as possible. That it seemeth the work of god and not of man. To achieve his aim he used several special methods and techniques, which partly derived from medieval book illumination traditions but also from his own background as a goldsmith. He applied an unusual number of black pigments, some made from charred fruit-stones, to match the appearance of black velvet and silk cloth. He also sorted pigments into various grades. For instance a medium grade of lead white was ideal to imitate the powdery appearance of makeup and another grade, with very fine particles, for the glossy impasto paint he used to depict lace. The thick application of lead mixed with a lot of gum to achieve this effect led to typical cracks and flaking. Some damage might have already occurred during his life-time, nevertheless he continued to use the techniques, which caused them. Another example is the application of powdered silver (for armour and the highlights on pearls), which turned out to be problematic as the silver soon started to tarnish. Thickly applied resin on silver grounds to create the glossy impression of precious stones and enamel often became squashed when fresh, and chipped or flaked of when dry. The authors practically reconstructed many of these methods and effects used on limnings in the context of a research project in the Victoria and Albert Museum in London. The practical work was based on visual examination, instrumental analysis of original portrait miniatures and extensive analysis of earlier and contemporary written sources. The practical approach proved to be highly informative as it not only helped to reconstruct the amazing effects of the original appearance but also to identify typical damage inherent in the use of specific materials and techniques.
key-words
portrait miniature limning hilliard painting technique alchemy
the obsession of 16 th century miniature painters to imitate the beauty of nature

t i m e a ta l l ia n
Ethnographical Museum, Vienna and the V&A Museum, London (England)
that it seemeth to be the thing itself
a l a n d e r bys h i r e
Victoria and Albert Museum, London (England)
Introduction and sources

The information presented in this article is mainly based on source material related to the English limner Nicholas Hilliard (1547-1619) and his followers and students. Instrumental Analysis of original limnings has been performed in the conservation department of the Victoria and Albert Museum since the 80s (in particular by the late Jim Murrell and more recently by Alan Derbyshire and various students). Advances in non-destructive techniques such as digital photography, computer image manipulation, Raman Spectroscopy and XRF in the last 10 years have allowed many new opportunities to update and amend the information derived from earlier research. What makes the research on limnings so unique is an abundance of written source material. The three main sources used for this research have been the manuscripts associated with Nicholas Hilliard, Edward Norgate and John Hoskins. (see table 1). Unlike the many anonymous illuminators from previous centuries, Hilliard and his circle left us with a detailed written account of their ideas, ideals and working methods. Their knowledge originates directly from those illuminators of the Ghent Bruges School, who were called to live and work in England by Henry VIII 1. Another indispensable source of knowledge and understanding was the practical reconstruction of materials and techniques. This process was aiming to understand the reasons why certain materials have been used but from a 16th century viewpoint. Reconstruction was not always a success and in most cases a humbling experience as we were not able to achieve the same quality and effects as observed on the originals. However trial and error provided many new insights and a lot more questions.
1. Lucas Horenbout, often called Hornebolte in England, (Ghent c. 1490 to 1495 London 1544) was a Flemish artist who moved to England in the mid-1520s together with his sister Susanna and his father Gerard and worked there as Kings Painter and court miniaturist to King Henry VIII from 1525 until his death. His father Gerard Horenbout was an important Flemish manuscript illuminator (Reynolds 2006: 45).
2 07
It also suggested that some typical deterioration phenomena must have occurred quite rapidly after the making of the objects. Nevertheless limners continued to use them for some time until the 1640s. This paper aims to explore the reasons why limners, who according to their writings were aware of the risks, still continued to use certain methods and materials.
table 1
Hilliard MS Edinburgh, University Library, LaingIII174, f.1-14 Norgate MS Oxford, Bodleian Library, Tan.326 Hoskins MS (known as Gyles book) Original London, British Library, Harley 6376. Copy (Hoskins MS copy), NAL, MS copy of Harl. 6376, R.C.A. 20-995-1906. ed. Thornton and Cain 1992 ed. Muller and Murrell 1994 Extracts published in Muller and Murrell 1997:237-252 and in Murrell 1983:76-79, 80-84, 92-93
The Arte of Limning That it Seemeth to be the Thing Itself

Illumination did not stop with the invention of printed books. On the contrary, in the 16th century it was not only used for luxury books and important documents but also for portraits after life and illustrations in scholarly publications (for instance of a geographic, ethnographic, zoological or botanic nature). Hilliard and Norgate amongst others stated that it should ideally be practised by gentlemen only. First of all because it required a gentle nature and was clean to use (did not spoil expensive clothing like oil!). It is recommended for the young gentry as an innocent pastime and useful skill for a future career at court, in war and scientific studies. Further the gentry, not having to make a living (unlike the notoriously financially struggling common artist) could spend the full time required to produce excellent works of art. Also they could easily afford the best quality materials. Limning was (and is, if one aims at accurate reconstruction) indeed expensive. The pricy materials include, first of all, paint made from the metals gold and silver; pigments derived from semi-precious stones; and many exotic materials (like the binder gum Arabic or Indian lake). But also cheaper materials like earth colours and lead pigments had to be of high quality for the fine painting. Each pigment needed to be washed, ground and prepared in a particular manner. Beside all this the best quality fine parchment was required. It is often claimed to be made from the smooth skin of a (yet hairless) animal embryo (called abortive). It was so thin and fine, that it had to be mounted on a playing card (limnings are usually quite small, about 7x5cm average). Reconstruction also suggested that the skins of small mammals, in particular white rabbits, might have been a possible source or at least an alternative. It can be said that the search for a parchment support of sufficient quality has been difficult and the biggest obstacle for an accurate modern reconstruction. The 16th century gentry conveniently could
208
2. It was Hilliard who first observed that the effects of sulphurous pollution on pigments and he states that the working place should be free from the Sulphurous aire of Sea Cole (Thornton and Cain p.75). Indeed it seems to be basically atmospheric hydrogen sulfide which causes the discoloration by converting lead white (basic(II) lead carbonate) into the black compound lead(II) sulfide.
purchase materials ready prepared from professional limners, as for instance Alexander Brown (Browne 1679: 39). There is also an arcane aspect of limning. In many ways the usage of 16th century English limning is comparable to the use of photography. However it does not only aim for the naturalistic depiction after life but aspires to transmute pigment and binder into something different, for instance into a real gemstone, an enamelled jewel, a plant or even a piece of textile. Hilliard also puts a lot of emphasis on super realistic depiction, suggesting to use the material to depict itself: worketh the metals gold and silver with themselves (Thornton and Cain: 63) He even set a real diamond into one of his miniatures (coronation portrait of queen Elizabeth, 1559, Wellbeck collection). Portrait miniatures of the aging Queen Elizabeth I, in a way, also illustrate this point. They certainly did not intend to realistically depict the physical appearance of a nearly 60 year old queen. More than her wrinkles, it was essential for Hilliard to depict the essence of Her Majesty, which was the Virgin Queen, Gloriana, the English equivalent to Virgin Mary. Miniature portraits of her and even medals were treated with the highest respect, similar to icons. Any wilful damage to her picture was in fact considered to be high treason. For instance Thomas Harrison was accused of endangering Queen Elizabeths I life by placing a medal with the picture of her Majesty next to mercury sublimate, which was poisonous and had already corroded the metal. Interestingly Harrison defended himself by claiming that the damage was not intentional and that Nicholas Hilliard (!) had made the metal image and that it had dissolved again in the wooden box into quicksilver (mercury) (Auerbach pp.32-33). There is no evidence that Hilliard was an alchemist but he certainly had a chemical understanding of materials due to his experience as goldsmith. Hilliard clearly was very aware of reactive chemical processes and is probably one of the first authors describing the deleterious effect of sulphuric gases on pigments 2 and classifies in his treatise some colours as not fit to be used in limning. Most of the pigments he mentions are prone to fading in light or have the tendency to discolour when being mixed with other pigments (as vermilion HgS or orpigment As2S3) , others might change to a dull colour (the copper pigments: Verdigis, Verditer) 1. Nevertheless it has been shown by analysis that he, against better knowledge, used some of those pigments, as for instance vermilion and orpiment. Was he betrayed by his colour merchant; simply a hypocrite; or did he have good reasons not to stick to his own rules? Previous and recent Raman analysis of Hilliards palette allowed 13 pigments to be unambiguously identified: lead white, lead red, massicot, vermillion, azurite, natural ultramarine, indigo, malachite, orpiment, pararealgar, yellow ochre, red ochre, and carbon black (Derbyshire 1999, Tallian 2007).
Special effects
fig.1 detail from nicholas hilliard, selfportrait 1577, watercolour on parchment, dia. 41mm (v&a, p.155-1910)
Limnings were made to be looked upon at close distance, intimately held in the hand and to be worn in jewels. This was ideal for fine details and the use of techniques
209
taking advantage of raking light and movement. Special effects became Hilliards trademark e.g. his crisp golden calligraphy, burnished silver highlights on pearls, little sparkling life-like rubies, the bright red Indian lake crimson curtain, glistening and brilliantly blue ultramarine backgrounds, the detailed black textiles and the raised lace of glossy lead white impasto. These effects fascinated not only contemporaries but have the power to enchant us till the present day. Unfortunately most of the mentioned techniques, even though Hilliard tried his best, occasionally caused problems due to the inherent properties of the used materials. Silver tends to tarnish, Venetian turpentine resin dries very slowly and is easily squashed. Indian lake easily fades and requires for a good colour a lot of gum and sugar (a combination, which attracts dust and is prone to dry out and crack); ultramarine is of a gritty sandy consistence and sometimes (when the pigment binder ratio is not right) rubs or flakes off; ivory black tends to flake and so does the thick and heavily gummed layers of lead white. Hilliard must have been clearly in a dilemma. But interestingly he did not give up his technique. Sparkling silver looked too good in combination with gold. Indian lake was used also in reality to colour expensive bright red luxury textiles (and was one of the most costly pigments at the time). Ultramarine was a precious gemstone, making the portrait look even more like a jewel; it was adding to its material value and also provided a great contrast for gold inscriptions. Charred Ivory, even though it has its tendency to be a difficult pigment, is creating probably the most appealing and again pricy blackest black available, ideal for strong shadows or ornaments in dark costumes or the dark centre of the eye. And vermilion, condemned by Hilliard as an unfit colour, was still used by him (proved by analysis) in the face and lips as it produces a radiant red and for the very same reason this pigment was applied by the ladies as rouge and lipstick. And the raised white laces, accurately dribbled in fine lines onto the parchment like icing sugar, casting their own shadows, just looked strikingly real. It was all too tempting, exquisite and worth the risk.
Precaution and Deterioration

There are various suggestions in the treatises to prevent certain deterioration phenomena (to protect the silver, how to prepare and to use black pigments to ensure quality and prevent cracking, how to refine ultramarine and most important how to refine lead white and to prevent it from tarnishing). All these methods worked to a large extend, which is evident in Hilliards best miniatures and their excellent condition. In strong contrast to our modern reconstructions: those almost immediately showed all sorts of problems. However to a certain percentage, even Hilliard had to lose his battle against stubborn painting materials and time.
Fading Colours
The most dramatic change is probably the fading of colours of organic origin, which Hilliard probably used amongst other applications to model the features of fair ladies.
210
3. Nicholas Frayling, PhD student at the V&A/ RCA conservation course was fundamental in making Tudor miniatures and their techniques more accessible using digital media and the latest computer software. (Frayling, N. 1998) Together with Alan Derbyshire he was the first to combine source research, instrumental analysis, practical reconstruction and computer image manipulation.
Many portraits of Queen Elizabeth I appear unusually pale with white faces, faint red cheeks and red lips. This caused the notion that Hilliards portraits are particularly flat and mask-like. Due to the fashion of the time and make up, portraits of ladies were certainly much paler than those of male sitters. However they still might have had a subtle modelling. This is suggested by the portrait of Elizabeth, which was kept inside a jewelled pendent, presented to Sir Francis Drake. In comparison to those portraits of her which were more exposed to light this picture shows much stronger modelling and colour (see Figure 2). Some of the male portraits like Hilliards self-portrait (see figure 1) seem to be less faded. Hilliards treatise suggests that he tended to use (lightfast) earth colours for male sitters for a more dark or sunburned complexion.
fig.2 detail from hilliards portrait of queen elizabeth, c. 1600 (see figure5) and the portrait mounted in the drake jewel, private collection presently on display in the british galleries of the v&a. the miniature mounted in the locket shows clearly less fading
Silver
The most evident blemishes for the non-conservator are the dark dots on the pearls, caused by tarnished silver. Other treatises suggest protecting silver with garlic juice. Garlic seems to be a strange choice as it contains sulphur, which is well known to promote the tarnishing of silver. It also needs to be considered, that any (more effective) protective layers of gum or varnish could have been removed through abrasion or later cleaning attempts with solvents such as alcohol. To show the splendour of how the silver highlights really should have looked, computer manipulation is a great tool. Nick Frayling has provided stunning examples how to utilise digital media and the computer in combination with reconstruction for a better understanding of the original appearance of Tudor miniatures 3.
211
Artificial Gemstones
Less obvious, unless viewed with the microscope, is the vanished sparkle of artificial gemstones. Reconstruction experiments have suggested that some of those stones might have been squashed not too long after their making by a cover glass or careless handling (see figure 2). During reconstruction (based on the treatises) the viscous honey-like Venetian turpentine (in our case, larch resin) was mixed with pigments (Indian lake for a ruby, verdigris for an emerald and ultramarine for a sapphire). The resin was applied with a hot needle, in a warm room or near a heating source, on top of a burnished silver ground (see figure 3). Further heat was required for the stringy resin to settle from the shape of a stringy, dwarf hat into a round dome. To pick up the right amount of resin is quite difficult and to pick up too much is very tempting as the sparkling resin dome looks great at first. However, the thickly applied resin on the reconstruction was not dry even after a period of several months, and some stones were accidentally squashed by handling. Indeed there is evidence in some treatises that limners experienced the same problem. There are clear warnings to observe certain drying times 4. But even if the pigmented Venetian terpentine resin had time to dry without being squashed, the thick layer of resin tended to be brittle and crack after some time. Hilliard often depicted enamelled miniature lockets in miniatures by coating them with resin. Those in some cases chipped of completely, leaving the blank parchment or a (meanwhile tarnished) silver ground. Variations in the technique can be observed on Hilliards later miniatures. Instead of applying a tick layer of resin Hilliard tried to apply a thinner layer of coloured resin on a silver coated thick blob of lead white. But unfortunately, as mentioned, also raised layers of lead white tended to crack and chip off, leaving blank parchment.
4. The most detailed description of making a ruby can be found in a manuscript by Daniel King, DL Add. 12461 MS.
fig.3 detail of a ruby, nichlas hilliard, elizabeth i (v&a, 622-1882) [photo by nicholas frayling]. reconstruction of a ruby
Craquelure Too Gummy?

Each pigment requires a different ratio of pigment and binder. A good method to determine the right ratio was to mix the paint up in a shell. Pigment (which was stored dry in paper or boxes) and powdered gum were mixed with a drop of purified water using the ring finger. If the dried colour in the shell formed a shiny layer like oil
212
paint there was too much gum. If there was too little, the paint would powder off. If the ratio was right, the paint would smear thinly on the smooth surfaces of the shell and neither crack nor shine (and also made it easier to detect impurities as specks of dust). Some effects and some pigments required more gum than usual. Indian lake needed gum for its colour and ivory black was working better when gummed freshly just before it was used for painting. In the case of the paint used for the lace, gum was required for the shiny effect and the impasto consistency. In all these cases it was easy to use accidentally more gum than required. For this reason areas where such paints were used (red background, eye centre, black costume, lace) show an increased tendency for craquelure and for flaking.
fig.4 detail of the lace (~4mm), nicholas hilliard (v&a, p.3-1974) and a uv image of nicholas hilliards wife alice, by nicholas hilliard, dia. 59mm, (v&a, p.1541910) the uv image shows the extend of loss in the area of the ruff and also the differences in the thickness of the white paint layers
Due to their enormous historical and artistic value Hilliards limnings are treated by conservators as sparsely and non-invasively as possible. They are rarely opened and are usually not permitted to travel and further subjected to a strict lighting and display policy. The most common reason to open these early miniatures is their need for consolidation. The fine craquelure and flaking of paint might be not obvious with the naked eye but easily detected under the microscope. Flaking of the ead white, in the thickly applied paint layers of the ruff and costume, is possibly the most serious deterioration phenomenon on Hilliards work and other limnings. The difference in the thickness of the lead white paint layer in the face area (very thin) in the ruff (very thick), (and losses of paint) can be clearly seen with x-ray spectroscopy and with the help of UV examination (see figure 6).
Lead white purification

Hilliard and other sources mention lead white and ceruse as prime pigments for white paint. According to the most recent hypothesis, the difference in the colour terms might refer to different stages of preparation or different grades of quality
2 13
rather than different chemical compositions. Raman analysis on some of the most prominent miniatures by Hilliard in the V&A unambiguously detected only lead white (basic lead (II) carbonate 2PbCO3.Pb(OH)2) in all examined white areas (Derbyshire 1999). Interestingly there is only minor evidence for the blackening of lead white, which is a more common phenomenon in later portrait miniatures, in particular of the late 17th century. This is likely to be the result of careful preparation methods, which are described in great detail in the treatises. As Hoskins says: it will be no lost labour to be curious in this Colour, because it is the ground & foundation of all your other works, for if this become faulty, all is lost & cast away. The experience whereof does frequently appear in many excellent peices, being heightned with white lead unprepared are spoild & the Colour all rusty & dead (Hoskins MS Copy, NAL-R.C.A.20-995: 31). Analysing these recipes, the limners seemed to have picked out the whitest flakes from yet un-ground flake white from the outer crust of the lead coil 5. Like this they were able to avoid contamination with pure lead splinters or yellow/grey rusty flakes from the inside of the coil. These snow white flakes were ground and the hairs/dust and water-soluble salts (as lead acetate) were washed away with distilled water (till the water had no harsh taste anymore). The remaining product was dried in the sun or a chalkstone slab. To achieve the powdery appearance of lead white paint, which was required for the carnation Hilliard had to remove the fine lead white particles by washing. During this process the ground flake white pigment was stirred in distilled water, after a short while the floating fine particles were decanted in a separate shell or container. (When dry these very fine particles were rapped in paper and later mixed with gum for the glossy lace and pearls.) Than the mixture was stirred again and after a shorter period the floating particles were decanted in a different container. This middle portion was ideal for painting non-glossy areas. The bottom layer of big heavy particles was ground again and the process repeated. This simple method of particle separation was used for nearly all pigments as it is crucial to remove dust, the too fine particles and the too coarse ones.
5. Lead white was produced by suspending a coil of lead in a glass or ceramic container above vinegar in a warm carbon dioxid rich environment (recommended is a barrel with horse dung).
Mineral blues
The correct application and preparation of natural ultramarine pigment still largely remains a mystery. There is no doubt anymore (due to analysis) that Hilliard used unadulterated mineral blues (made from lapis lazuli and azurite). His blue backgrounds are surprisingly strong in colour and at the same time very fine and smooth. Limners probably purchased blue pigments made of the best quality raw material. It is unlikely that limners produced these pigments directly from the gemstones, as the preparation was a well kept secret and required probably large quantities of the raw material (Hilliard mentions ultramarine from Venice and notes its enormous costs). Again limners purified the pigment by removing very fine particles. This is of the greatest importance as particles under 5 microns appear to be colourless and would settle on the surface spoiling the blue colour underneath with a greenish or greyish cloud.
214
6. The following book is very much recommended. It shows amazing pictures from a traditional Indian miniature workshop, The materials and techniques are reminiscent to those described in the treatises: Lazaro 2005. 7. Hoskins MS Copy, NAL-R.C.A.20-995, fol.35.
The second step was again to remove coarse particles, which would hinder a smooth application and the final burnishing of the paint. This is a very narrow line. Approximate evaluation of the particle sizes of the blue pigments under high magnification showed a minimum of 5 micons, an average of 10 microns and a maximum of about 25. How the blue has been laid down can best be learned from contemporary Indian miniature painters or those miniature painters still working in the Islamic manuscript tradition 6. However, as experienced by the author there is a great difference between watching and doing it yourself. Even ignoring the lack of good quality pigment and experience, it seemed to be generally a nerve-racking and difficult business.
fig.5 nicholas hilliard, elizabeth, c. 1600, oval, 64.5x49mm (v&a, p.1-1974). a quantity of good quality azurite has been separated into various colour shades and grain sizes by the method of washing. the shell in the middle is the one hilliard would have used (the pigment is not too coarse, not too fine and has a good colour)
Black is not black

The most satisfying was the reconstruction of black pigments. According to the treatises some limners prepared black pigments themselves from the raw materials. Only in this way could they ensure the quality and authenticity of their final product. And limners were very specific what they wanted: namely cherrystone and Ivory black. Cherrystones consist of very dense almost grain-less wood. During the charring process organic materials keep their original shape a nutshell, a plum stone or a willow twig when charred would produce different pigments, of different texture and shade. Cherrystones produce a black paint with very fine regular particles, which is very nice to work with. The shade it produces is of a silvery grey/ black colour. Charring fruit stones turned out to be relatively easy. They were placed in a metal container with a lose lid and placed in an ordinary burning wood stove. Within a quarter of an hour, coloured flames escape the container and the process is then finished. Charring is a reduction process the less oxygen enters the container the better. In the case of ivory the temperatures needed to be somewhat higher. Hoskins says that the iron container needs to be of a dark red colour for about an hour. According to Hoskins MS the container was fully packed with ivory splinters and luted (sealed) with a mixture of wet clay mixed with salt 7. (The salt probably prevents sudden shrinkage of the clay. The woodstove was heated with coal to its full temperature and the container left inside till the stove cooled down. All this is described in some detail in the treatises. But every treatise seems to provide only a tiny piece of information. The combination of all sources was needed to reproduce
215
the recipe in reality is a key for understanding what is meant. After this experience there was little doubt that it was easy for the miniature painters to collect scraps of broken ivory, combs, scrapings from sculptors etc and produce their own pigment from scratch to achieve exactly the effects they wanted to. The amazing detail of Hilliards costumes shows it was well worth the effort.
fig.6 detail of the black costume. nicholas hilliard. unknown lady, 159093, oval, 59x47mm (v&a, p.9-1947). ivory black after breaking the clay seal of the crucible used for reconstruction. paint made from the reconstructed ivory black seemed to flake in a similar way as the black paint in the miniature (lower edge) also did the paint an irregular shine similar to the one on this miniature
Summary
The research into 16th century limnings is far from being finished. It rather feels that it has just been started. Further research for instance might include exploring the link with earlier manuscript illumination (i.e. Ghent Bruges School). Reconstruction provided a few new insights as well as many new questions. Which was surprising as Hilliards technique was always thought to be well understood due to the abundance of treatises based on it. However, today we know that not all of the information given is necessarily correct. Also we became more aware that the appearance of artefacts today is not necessarily the appearance intended by the artist. To study the artists intentions, the cultural circumstances in which a work of art was produced and to learn about the methods and techniques it was made with, greatly adds to the art works value and appreciation.
Acknowledgments
We greatly want to thank the large number of people investing time, sharing their knowledge and resources in helping with the research and reconstruction work on Hilliards materials and techniques: First of all the Royal College of Art and the V&A Museum, which provided an amazing platform for the research, thanks to William Lindsay, Nick Frayling and Katherine Coombs. We also want to thank Carmen Krisai-Chizzola, Mike Wheeler, Dr. Georg Kremer (Kremer pigments), David Margulies, Anita Chowdry, Dr. Spike Bucklow, Dr. Mark Clarke and the scientists: Dr. Robert Withnall, Dr. Lucia Burgio and Dr. Vincent Daniels.
216
Bibliography
Auerbach, E.. 1961. Nicholas Hilliard. London: Routledge and Kegan Paul. Browne, A.. 1679. Ars Pictora. London. Derbyshire, A. and R. Withnall. 1999. Pigment Analysis of Portrait Miniatures Using Raman Microscopy. Journal of Raman Spectroscopy 30, 185-188. Frayling, N.. 1998. An exploration of the original appearance of Nicholas Hilliards portrait miniatures using computer image manipulation. V&A Conservation Journal 28, July 1998, 4-6. Lazaro, D. P.. 2005. Materials, Methods and Symbolism in the Pichhvai painting tradition of Rajastan. Ahmedabad: Mapin Publishing. Muller, J. M. and J. Murrell (1997) Edward Norgate, Miniatura or the Arte of Limning. New Haven & London: Yale University Press. Murrell, J.. 1983. The Way Howe to Lymne Tudor Miniatures Observed. London: V&A Publications. Peacham H.. 1622 The Compleat Gentleman. London. Reynolds, Graham. 2006; The Sixteenth and Seventeenth-Century Miniatures in the Collection of Her Majesty The Queen, Royal Collection Publications Ltd. Tallian, T.. 2007. Reconstruction of Nicholas Hilliards Materials and Techniques, a thesis submitted in partial fulfilment of the requirements of the Royal College of Art fort the degree of Master of Philosophy, May 2007, The Royal College of Art / Victoria and Albert Museum, London. Thornton R. K. R. and T. G. S. Cain. 1992. The Arte of Limning by Edward Norgate. Manchester: Carcanet Press.
Biographies
Timea Tallian studied art history and fine art before she joined the conservation course at the Academy of fine Arts in Vienna (MA in 2003). Initially trained as paper conservator she developed a strong interest for ethnographical objects, portrait miniatures and the study of miniature painting techniques in general. In 2003 she embarked on a research project, at the V&A/RCA postgraduate conservation course, focusing on the practical reconstruction of Hilliards materials and techniques (Mphil 2007). After a year break, working as conservation consultant in Bhutan, she is back in London, presently working as a free lance conservator. Timea is affiliated with the Ethnographical Museum, Vienna and the V&A Museum, London. Email:- timea.tallian@network.rca.ac.uk Alan Derbyshire obtained a BSc in Physics fromU.M.I.S.T. in 1975 before going on to study paper conservation at Gateshead Technical College. He is Head of Paper, Books and Paintings Conservation at the Victoria and Albert Museum, where he has worked since 1983. For the last twenty years he has specialised in the conservation of portrait miniatures on ivory and vellum. He has written, taught and lectured extensively on the conservation of works of art on paper and on portrait miniatures. He is an accredited member of ICON. Email: aland@vam.ac.uk, Address: Paper Conservation, Victoria and Albert Museum, London SW7 2RL, Telephone:- 0207 942 2113
217
Resumo
Guido di Pietro, mais conhecido por Fra Angelico ou Frei Giovanni de Fiesole (c. 1400-1455), foi um dos mais importantes artistas do sc. XV. Exceleu tanto em pintura mural e de cavalete como nas artes da iluminura. O seu talento foi celebrado em 2007-08 no Museu de So Marco em Florena, por ocasio da exposio Fra Giovanni Angelico. Pittore miniatore o miniatore pittore, no mbito da qual se levou a cabo um estudo comparativo dos materiais e tcnicas que o artista utilizou na iluminura e na pintura sobre tbua. Este centrou-se nos materiais pigmentos e corantes utilizados por Angelico em manuscritos, em particular, numa das suas obras mais belas, que integra a coleco permanente do Museu: o Graduale n. 558. Tendo em conta que se tratam de obras em pergaminho, so de preferir mtodos de anlise que permitam uma identificao in situ. Para alm disso, tcnicas no invasivas so consideradas as mais adequadas, uma vez que difcil recolher amostras ou micro-amostras. De facto, as camadas cromticas so tipicamente muita finas quando comparadas com a pintura de cavalete. Acresce que, normalmente, as decoraes cobrem reas pequenas. Este artigo descreve o uso da fluorescncia de raios-X dispersiva de energias (XRF) em conjunto com a reflectografia de UV-VIS-NIR por fibra ptica (UV-Vis-NIR FORS), que foram seleccionadas de entre as possveis tcnicas no-invasivas e de aplicao in situ, para o estudo dos materiais utilizados no Graduale n. 558 de Beato Angelico.
palavras-chave
manuscritos iluminados fluorescncia de raios-x dispersiva de energias (xrf) reflectografia de uv-vis-nir com fibra ptica (fors) fra angelico graduale n. 558
Abstract
Guido di Pietro, better known as Beato Angelico or Fra Giovanni da Fiesole (c. 1400-1455), was one of the most important artists of the 15th century. He excelled in wall- and panel-paintings as well as manuscripts and illuminations. His talent was celebrated in 2007-08 at the San Marco Museum in Florence with a special exhibition Fra Giovanni Angelico. Pittore miniatore o miniatore pittore? in which the technique and materials used by the artist in making manuscripts and panel paintings were investigated and compared between these two different forms of art. The focus of the study in the context of the exhibition was on the materials pigments and dyes used by Angelico to produce manuscripts and, in particular, one of his most beautiful pieces located in the permanent collection of the San Marco Museum: the Graduale n. 558. Due to the fragility of parchment-based artworks, the application of in situ non-invasive analytical techniques is strongly recommended for analyzing the materials used by artists in making the manuscripts. Moreover, non-invasive analytical methodologies are usually considered to be the most suitable techniques for the investigation of manuscripts since taking samples, or even micro-samples, from such delicate art objects is generally considered unacceptable. Indeed, the manuscript paint layers are typically very thin compared to those of wall and panel paintings. Also, painted decorations usually cover relatively small areas. This report illustrates the effective use of X-ray fluorescence (XRF) and ultraviolet, visible, and near infrared fiber optic reflectance spectroscopy (UV-Vis-NIR FORS), selected from the available por ajuda na configurao do texto e sugestes de the materials used in Fialho Agradecimentos in situ non-invasive techniques, to identifyAna Catarina Sousa, Pedro making the Justino Maciel, Felix Teichner Heidi. de Sousa,Graduale n. 558 by BeatoeAngelico.
key-words
illuminated manuscripts x-ray fluorescence (xrf) fiber optic reflectance spectroscopy (fors) beato angelico graduale n. 558
non-invasive xrf and uv-vis-nir reflectance spectroscopic

analysis of materials used by beato angelico in the manuscript graduale n. 558
m . p i co l lo
IFAC-CNR, Firenze (Italy)
a . a l d rova n d i
Opificio delle Pietre Dure, Firenze (Italy)
a. migliori
Dipartimento di Fisica e INFN , Firenze (Italy)
s . g i acom e l l i
Societ Internazionale per lo Studio del Medioevo Latino, Firenze (Italy)
Introduction
In the study of artworks the application of more than one analytical technique permits one to better identify the materials and the techniques used by the artists (Clarke 2001, Ricciardi et al. 2009). These techniques are grouped in two main categories: non-invasive and invasive. Among the first group of techniques both X-ray fluorescence (XRF) and optical fiber reflectance spectroscopy (FORS) provide conservators and curators with useful information about works of art (Bacci et al. 2009, Dran et al. 2009). Their non-invasiveness enables these light and compact devices to acquire a large amount of data in situ. In some cases, it may be necessary to integrate the data obtained with some noninvasive methods with other non-invasive or micro-invasive analytical techniques. For example, FORS and XRF can also be very useful tools, in conjunction with other techniques, for locating areas for micro-sampling and for extending local data from micro-analyses to a broader scale, thus reducing the extent of micro-sampling. However, the small size of the illuminates and the presence of very fine details prohibit, in most cases, the use of the precise analytical chemical methods typically used on paintings, because of the sample size limitation. In this case, XRF and ultraviolet (UV), visible (Vis), and near infrared (NIR) FORS were selected from the available in situ non-invasive techniques and applied to iden-
m . s cu d i e r i
Museo di San Marco, Firenze (Italy)
219
n e w t r e n d s i n t h e a n a ly s i s o f m e d i e v a l m a n u s c r i p t s
tify the materials used in making the Graduale n. 558 by Beato Angelico. Guido di Pietro, better known as Beato Angelico or Fra Giovanni da Fiesole (c. 1400-1455), was one of the most important artists of the 15th century. He excelled in wall and panel paintings as well as manuscripts and illuminates. His talent was celebrated in 2007-08 at the San Marco Museum in Florence with a special exhibition Fra Giovanni Angelico. Pittore miniatore o miniatore pittore? There, the technique and materials used by the artist in making manuscripts and in making panel paintings were investigated and compared. The focus of the study, in the context of the exhibition, was on the materials pigments and dyes used by Angelico to produce manuscripts and, in particular, one of his most beautiful pieces of the permanent collection of the San Marco Museum: the Graduale n. 558.
Methodology
The artwork
The Graduale n. 558 represents one of the most famous illuminate masterpieces of Angelicos young production and was made for the church of San Domenico in Fiesole where the artist lived at the beginning of his monastic life. Fra Angelico was a versatile artist who excelled in the fresco, illuminate, and panel paintings techniques. The Graduale n. 558 was probably made in two different periods: in 1424-1425 and in 1428-1430. Consequently, the illustrations present some different stylistic characteristics, which make it possible to group the miniatures in four groups by their chromatic, stylistic, and decorative features. This choir book is constituted of 260 folios written in Testualis with text and music. These folios were bound together probably when the Graduale was added in the Leopoldo II di Lorena Collection in the 19th century. The illustrations are composed of 986 watermarked letters, 30 miniatures, and seven decorated letters.
Analytical techniques
FORS was performed using two spectroanalyzers: Zeiss model MSC501 and MCS511 NIR 1.7 operating in the 200-1700 nm range. The approximate spectral resolution was less than 3 nm in the 200-1000 nm range (MCS501) and 10 nm in the 900-1700 nm range (MCS511 NIR 1.7). An internal tungsten lamp (Zeiss model CLH500) was used for the final operative range (350-1700 nm). The sampled area was smaller than the approximate 2 mm diameter, and the acquisition time for each spectrum was less than one second. The spectrometers were calibrated using a white Spectralon 99% reflectance standard. A 0/45/45 reflectance configuration was adopted to avoid specular reflectance. The identification of the pigments was accomplished by comparing the acquired unknown spectra with spectral databases (http://fors.ifac. cnr.it). Assignments were made using both the primary and first derivative spectra. Given the thinness of the paint layers and the presence of overtone bands in the NIR associated with the vellum substrate, the assignment of peaks to pigments required
220
n o n - i n va s i v e x r f a n d u v - v i s - n i r r e f l e c ta n c e s p e c t r o s c o p i c
particular care. The NIR region was, however, found to be particularly useful for the identification of some pigments as well as gypsum. XRF analyses were performed with an Assing LITHOS 3000 portable spectrometer equipped with a molybdenum tube and a Peltier cooled Si-PIN detector with a sensitive area of 7 mm2, thickness of 0.5 mm and a berillium (Be) window 12.5 m thick. The resolution was of approximately 165 eV at 5.9 keV. For the measurements on the Graduale the X-ray tube voltage was 24 kV, the current was 300 A, and the acquisition time was 100 seconds (in case of intense X-ray counting rate the current was reduced and the acquisition time was increased). Also, a 1 mm collimator was used for an investigated area of approximately 2 mm in diameter. The distance between the investigated area and the detector was about 2.8 cm. The X-ray sending and collecting angles were approximately 45 and 50, respectively. Due to the intrinsic limitation of XRF, which does not yield results spatially resolved in depth; and, in order to avoid confusing results, the areas for analysis were chosen, whenever possible, so that no decoration was present in the corresponding area on the verso of the manuscript leaf.
Results and discussion

On the Graduale approximately 130 FORS and 50 XRF spectra distributed on about 30 pages were acquired. Pigment identification was achieved by combining FORS and XRF spectral data. From the XRF measurements the preparation of the parchment was found to be made with calcium carbonate as calcium, with a small amount of arsenic, were the only chemical elements recorded. However, the calcium element could have been related to the presence of calcinated animal bones (Brunello 1975). FORS measurements in the presently available operative range were not able to better specify the materials used to prepare the parchment (Fig. 1). FORS data, on
fig.1 fors spectra of the parchment (dashed line) and of the gypsum with iron oxide (hematite) in the preparatory layer for the application of the gold leaves (f.33v)
221
the other hand, made it possible to exclude the use of calcium sulfate bi-hydrated (gypsum) or calcium sulfate hemi-hydrated (plaster of Paris) in the ground layer (Bacci et al. 2007). This information was also confirmed by XRF because no sulfur was found in the preparation of the parchment. The gypsum was present only as a preparatory layer for the application of the gold leaves (c.33v, c.86v, c.124r) in mixture with iron oxide (hematite) based pigments (Fig. 1). In one case, XRF also found the presence of mercury, together with iron, in the preparatory layer. This was due to the presence of cinnabar, which gave a reddish tint to the layer. The gold leaves, which were extensively used in the Graduale for background and aureole areas, were made with pure gold. The XRF analysis did not show the presence of impurities such as silver, tin, lead, or copper (c.9r, c.21r, Fig. 2).
fig.2 xrf spectrum of a gold leaf, made with pure gold. from the xrf data the gold leaves did not show the presence of impurities such as silver, tin, lead, or copper (f.9r)
As found by FORS spectra on about 30 different areas of the Graduale the painted blue areas were made with ultramarine blue (lapis lazuli). Ultramarine blue was also used in mixture with a red pigment, red lake mainly, at different concentrations to depict violet-purple details (Figs. 3 and 4). Azurite was used only for the watermarked letters. This second blue pigment was positively identified by the two analytical techniques. A copper based green pigment, such as malachite, mainly, created the green zones. This pigment was used to depict both the cloths of the characters and the landscapes. Also, as illustrated by FORS measurements on 12 areas of the illuminated manuscript, it was used to make frames and decorations surrounding the painted scenes. Another green pigment, green earth, was only applied to paint complexion (flesh tones) areas in mixture with lead white, iron oxide and hydroxide (yellow and brown ochre or earth pigments) pigments, and cinnabar. In this Graduale the artist
222
fig.4 fors spectra from blue areas made with ultramarine blue (lapis lazuli). spectrum a from folium 33v; spectrum b from folium 60v; spectrum c from folium 93r
did not use mixtures of blue (azurite, indigo, lapis lazuli) and yellow (lead-tin yellow, orpiment) pigments to produce green hues/shades. In one case only, the FORS acquired spectrum resembled a mixture of ultramarine and lead-tin yellow pigments; but here the yellow pigment was painted as a glaze on the blue sky. The yellow and brown colors were obtained by using lead-tin yellow pigment for lighter areas and iron oxide and hydroxide pigments for darker and less saturated hues. The first pigment was identified by the presence of contemporary lead and tin in the XRF spectra (Fig. 5) and from its reflectance spectral shape (FORS); the second pigments showed typical FORS, in which the characteristic absorption bands
fig.3 detail of the folium 33v with reported some xrf and fors measurement
fig.5 xrf spectra of yellow and brown areas obtained by using lead-tin yellow pigment (red curve, f. 93r), iron oxide and hydroxide pigments (sienna earth, black curve, f. 21r), and orpiment (green curve, f. 60v)
223
of trivalent iron are easily detected, and XRF spectra. In the case of XRF data, the presence of iron is usually linked to manganese, which usually is related to the occurrence of Sienna and, more typically, of umber earths. In the folium 60v the dark yellow mantle of San Peter was depicted using orpiment (Fig. 6). This material was only identified in this area of the Graduale by the presence of arsenic in XRF spectrum and by its FORS spectrum, even though the reflectance spectra of most of the yellow pigments could be easily modified when mixed together or with red pigments (Fig. 5).
fig.6 detail of the the folium 60v with the dark yellow mantle of san peter depicted using orpiment (xrf measurement n. 92)
The red areas were created primarily by use of cinnabar and red lakes. In certain cases red lead (minium) was found, even if most of the time it was associated with cinnabar (Fig. 7). Cinnabar is easily detected by XRF due to the presence of mercury in the paint layer. This pigment was extensively used for the main scenes, the decorations, and the watermarked letters. The red lake, instead, was mainly used to paint the purple-violet glazes and to produce hues from pink to purple. This pigment is not seen by XRF, and its identification by means of FORS is certain only when red lakes are used as glaze on a light substrate or in a mixture with a white pigment, such as lead white. To date, to the best of the authors knowledge, there have been no systematic studies to determine how preparation, ageing, and type of lake affect the reflectance spectra. Many of the absorption features can vary depending on the preparation of the dye into a lake (Bacci et al. 2001) as well as the preparation of the paint (Bisulca et al. 2008). In some cases characteristic features are less intense
224
fig.8 fors spectra of red lead (solid line - a, f. 13v), cinnabar (dashed line b, f. 13v), and red lake (dotted line c, f. 93r)
or absent altogether, which can cause misleading results in interpretation of spectra. While it is sometimes possible to distinguish between these dyes or to broadly classify them as of animal or insect in origin, in the analysis of real paintings identification is complicated by the presence of other pigments, the nature of their application, and the effects of natural ageing. These factors can mask or alter characteristic features in UV-Vis-NIR reflectance spectra (Fig. 8). Finally, it was noteworthy to report that sometimes, such as in the folium 85v, the ring shape bronze/light brown decorations of the letter were made by using mosaic gold, a very fine powder made with tin and copper, as revealed by XRF (Fig. 9).
fig.7 detail of the folium 13v where red lead (minium) and cinnabar were found
fig.9 xrf spectrum from the ring shape bronze/light brown decorations of the letter of folium 85v made by using mosaic gold, a very fine powder made with tin and copper
225
Conclusion
The combination of high fidelity site-specific methods (FORS and XRF) proved to be a useful tool for the examination of illuminates and miniatures artists materials, as already found in easel and mural paintings. Expanding the FORS range of analysis further into the infrared, improving the XRF procedure in the detection of light chemical elements, and adding Raman spectroscopy to the list of applied techniques, would make possible more precise descriptions of the characteristics of the materials found in this kind of artworks, in particular when organic materials, such as binding media, need to be identified.
Bibliography
Bacci M., Orlando A., Picollo M., Radicati B., Lanterna G. 2000. Colour analysis of historical red lakes using non-destructive reflectance spectroscopy. Compatible Materials for the Protection of Cultural Heritage, in PACT 58, 21-35. Bacci M., Magrini D., Picollo M., Radicati B., Trumpy G., Tsukada M., Kunzelman D. 2007. Modern white pigments: their identification by means of non-invasive ultraviolet, visible and infrared fiber optic reflectance spectroscopy in Proceedings of the Modern Paints Uncovered Symposium, Tate modern, London, May 16-19, 2006. Editors T.J.S. Learner, P. Smithen, J.W. Krueger, M. Schilling. The Getty Conservation Institute, Los Angeles, 118-128. Bacci M., Boselli L., Picollo M., Radicati B. 2009. UV, VIS, NIR Fibre Optic Reflectance Spectroscopy (FORS) in Practical handbook on diagnosis of paintings on movable support, Editors D. Pinna, M. Galeotti, R. Mazzeo, European Project ARTECH, Centro Di, Firenze, 197-200. Bisulca C., Picollo M., Bacci M., Kunzelman D. 2008. UV-Vis-NIR reflectance spectroscopy of red lakes in paintings in Proceedings of the 9 th International Conference on Nondestructive investigations and microanalysis for the diagnostics and conservation of cultural and environmental heritage, Jerusalem, May 25-30, 2008. Clarke, M. 2001. The analysis of medieval European manuscripts, Reviews in Conservation 2, 3-17. Brunello F. 1975. De Arte Illuminandi, Neri Pozza Editore, Vicenza. Dran J.-C-, Laval E. 2009. X-Ray Fluorescence (XRF) in Practical handbook on diagnosis of paintings on movable support, Editors D. Pinna, M. Galeotti, R. Mazzeo, European Project ARTECH, Centro Di, Firenze, 210-213. Ricciardi P., Delaney J.K., Glinsman L., Thoury M., Facini M., de la Rie R. 2009. Use of visible and infrared reflectance and luminescence imaging spectroscopy to study illuminated manuscripts: pigment identification and visualization of underdrawings in Optical Methods for Arts and Archaeology Aug. 2009 Proc. SPIE Vol. 7391, R. Salimbeni, L. Pezzati Eds. http://fors.ifac.cnr.it. Fiber Optics Reflectance Spectra (FORS) of Pictorial Materials in the 270-1700 nm range.
226
Biographies
Marcello Picollo, geologist, is a Researcher at the Institute of Applied Physics Nello Carrara IFAC-CNR, Florence. He has been working on spectroscopic investigations of works of art since 1991. His main focus is on pigment characterization using non-invasive spectroscopic, imaging and X-ray techniques. Address: Institute of Applied Physics Nello Carrara IFACCNR, Via Madonna del Piano 10, 50019 Sesto Fiorentino (Italy). E-mail: m.picollo@ifac.cnr.it Alfredo Aldrovandi received his Master in Science in Physics at the University of Modena. He has been working from 1983 at the Restoration Laboratories of the Opificio delle Pietre Dure in Florence. His main focus is on the development and application of non-invasive diagnostic investigation techniques on artworks. Address: Opificio delle Pietre Dure, Via degli Alfani 78, 50121 Firenze, Italy. E-mail: alfredo.aldrovandi@gmail.com Alessandro Migliori received his Master in Science in Physics in June 2001 and his PhD in Science for the Conservation of Cultural Heritage in December 2004, both at University of Florence. He is an expert in techniques for compositional analysis using ion beams (Ion Beam Analysis techniques: PIXE, PIGE, RBS, NRA) and X-ray radiation (XRF), mainly in the field of Cultural Heritage. He has worked at CMAM laboratory in Madrid and Opificio delle Pietre Dure in Florence. Presently he is working as researcher at LABEC laboratory of INFN-Florence. Address: Dipartimento di Fisica e INFN (Florence), Via Sansone 1, 50019 Sesto Fiorentino, Italy. E-mail: migliori@fi.infn.it Sara Giacomelli received her Master in Humanities in 2004 with a thesis on History of Ancient Book Illumination. She received a post Master specialization in Art history in 2008 with a thesis on techniques and pigments used in miniatures. She has worked at the Museum of San Marco for the exhibitions Miniatura in Umbria del Rinascimento. Suggerimenti per un percorso artistico (2004); Fra Giovanni Angelico. Pittore miniatore o miniatore pittore? (2007-2008); LAngelico ritrovato. Studi e ricerche per la Pala di San Marco (2008-2009); Beato Angelico. Lalba del Rinascimento (2009). Her publications dealt with Umbrian and Florentine illumination artworks. Address: Societ Internazionale per lo Studio del Medioevo Latino (S.I.S.M.E.L.), Certosa del Galluzzo, Via della Certosa 1, 50124 Firenze, Italy. E-mail: saragiacomelli@gmail.com Magnolia Scudieri received her Master in Humanities specializing in Medieval and Modern Art history. She is the Director of both the Museum of San Marco and the Conservation Laboratories of Soprintendenza of the Florentine area. She is actively involved in encouraging and initiating study and research to further the knowledge on the permanent collection of the Museum of San Marco. She is responsible for creating proficient collaboration with prominent institutes dedicated to the investigation of, in particular, the artistic production of Beato Angelico and his followers. She has organized several exhibitions and published works on this topic. Address: Museo di San Marco, Piazza San Marco, 50100 Firenze, Italy. E-mail: scudieri@polomuseale.firenze.it
227
a p r o t o c o l f o r n o n - i n v a s i v e a n a ly s i s o f m i n i a t u r e p a i n t i n g s
Resumo
A caracterizao das paletas de cor utilizadas na iluminura medieval tarefa importante de um ponto de vista histrico-artstico, mas rduo numa perspectiva cientfica: difcil recolher amostras, no aconselhvel utilizar tcnicas que necessitem de contacto (i.e., IR em modo de ATR) nem levar a cabo sesses de anlise longas, devido ao stress que se poder causar aos manuscritos. Por estes motivos, necessrio utilizar tcnicas analticas que sejam no-invasivas e rpidas; para alm disso, na maioria dos casos ser necessrio trabalhar in situ, utilizando instrumentao transportvel. De entre as tcnicas possveis, a espectroscopia de Raman ser a mais informativa, dado o seu potencial de diagnstico; no entanto, requer tempos de anlise longos. A fluorescncia de raios-X dispersiva de energias (XRF) uma alternativa poderosa, mas sendo uma tcnica de anlise elementar, nalguns casos no permite chegar a concluses precisas. A espectroscopia de UV-VIS-NIR, em modo de reflectncia com fibra ptica (FORS), pode ser considerada uma tcnica preliminar promissora, mas apresenta algumas limitaes bvias. Neste trabalho proposto um protocolo de anlise para a caracterizao de iluminuras em manuscritos, de forma no-invasiva, utilizando equipamento transportvel e efectuando anlises in situ. Este protocolo permite a identificao de colorantes atravs da aplicao sucessiva de tcnicas complementares, explorando as vantagens de cada uma delas. Antes de mais, preparou-se uma paleta com os diversos pigmentos que se sabe terem sido utilizados na Idade Mdia; prepararam-se tintas base de goma arbica e clara de ovo, seguindo receitas antigas descritas em tratados medievais tais como De arte illuminandi by anonymous, Compositiones ad tingenda musiva by anonymous and Il libro dellarte by Cennino Cennini. Esta paleta constitui-se como ponto de partida para a construo de uma base de dados para a anlise espectroscpica, reproduzindo cenrios semelhantes aos encontrados nos manuscritos. assim importante enfatizar o facto dos espectros obtidos com esta paleta serem mais fiveis que os obtidos a partir do colorante em p. O protocolo comea com uma anlise global levada a cabo com FORS, sendo adquiridos espectros em todas as reas pintadas do manuscrito e comparados com a base de dados. Isto permite a identificao de cerca de 60-70% dos colorantes presentes. De seguida, procede-se a uma inspeco visual das cores com uma cmara digital acoplada a um microscpio 10-80x, por forma a obter boas imagens das reas onde se obtiveram espectros de FORS inconclusivos, i.e., misturas de pigmentos, cores degradadas, etc. Aps o que se conduzem as anlises de XRF que permite caracterizar os pigmentos metlicos (i.e., pigmentos base de ouro, prata e cobre), verificar a presena de camadas sobrepostas, identificar mordentes em corantes e lacas, e ainda identificar eventuais produtos secundrios que, por sua vez, fornecero informao til para estudos de provenincia das matrias-primas. Nesta fase cerca de 90% dos colorantes presentes estaro identificados. Finalmente a espectroscopia de Raman ser aplicada a casos que ainda levantem dvidas. Aps a aplicao destas tcnicas obtm-se muita informao, causando pouco ou nenhum stress aos manuscritos analisados.
palavras-chave
in situ fors raman xrf iluminuras
229
Abstract
The characterisation of palettes used in medieval manuscript illumination is an important task from the historical-artistic point of view, but a hard one from the scientific point of view: miniatures cannot be sampled, it is unsuitable to use techniques operating in contact (e.g. IR in ATR mode) and to perform long-lasting analytical sessions, due to the stress that can be imparted to manuscripts. For these reasons it is necessary to use analytical techniques both non-invasive and fast; moreover, in most of cases it is necessary to work in situ with portable instruments Among available techniques working in portable versions, Raman spectroscopy is the most informative, due to its diagnostic power; it requires, though, long time of analysis. XRF spectrometry is a powerful alternative but, being an elemental technique, in some cases it does not yield accurate results. UV-visible-NIR spectrophotometry in reflectance mode with fibre optics (FORS) can be promising as preliminary technique with some apparent limits. In this work a protocol of analysis is proposed for characterisation of miniature paintings on manuscripts in non-invasive way, using only portable techniques and performing in situ analysis. The protocol allows identification of colorants by successive application of complementary techniques, exploiting the advantages of each technique. First of all a palette with several pigments, dyes and lakes on parchment has been prepared with colorants that were in use in Middle Ages; paints have been prepared in gum Arabic and in egg white, according to ancient recipes described in medieval textbooks such as De arte illuminandi by anonymous, Compositiones ad tingenda musiva by anonymous and Il libro dellarte by Cennino Cennini. This palette is the base on which to build a database of spectroscopic analysis, reproducing a situation similar to the one present on manuscripts. It is important, therefore, to stress the fact that spectra collected from this palette are more reliable than spectra obtained from analysis of colorants in powder. The protocol is started with an overall investigation with FORS, collecting spectra from all painted areas of the manuscript and comparing them with the database. This allows to identify almost 60-70 % of the colorants present. Then visual inspection of the paintings is performed with a digital camera connected with a 10-80x microscope, in order to have a clear image of areas that show uncertain FORS spectra, i.e. mixtures of pigments, altered paints, etc. Then XRF spectrometry is performed in order to characterise metal pigments (i.e. gold, silver and copper pigments), to verify the presence of overlapping layers, to identify mordants for dyes and lakes and to identify contaminants in pigments yielding information useful to study the provenance of raw materials for colorants. At this stage almost 90% of colorants can be identified. Finally Raman spectroscopy is used for the most uncertain cases. After application of these techniques a wealth of information is obtained, causing little or no stress at all to the manuscripts under analysis.
key-words
non invasive fors raman xrf miniatures
a protocol for non-invasive analysis of miniature paintings

mau riz io ac e to
Department of Environmental and Life Sciences, University of Eastern Piedmont (Italy); Centro Interdisciplinare per lo Studio e la Conservazione dei Beni Culturali (CenISCo), Universit del Piemonte Orientale (Italy)
a n g e lo ag o st i n o
Department of General Chemistry and Organic Chemistry, University of Turin, Italy; NIS (Nanostructured Interfaces and Surfaces) (Italy)
Introduction
The characterization of palettes used in miniature painting is an important task as a great amount of information can be obtained at the disposal of art historians. It is, though, a hard task since miniatures are among the most precious and fragile artworks. Many problematic aspects must be considered: sampling from miniatures is usually not allowed by owners and institutions, so that only in situ analysis can be performed; it is not possible to use analytical techniques working at contact with the sample, such as ATR-IR; while performing analysis sessions, prolonged opening of manuscripts can cause stress to painting layers and to parchment itself, so that after few hours sessions must be closed. Due to these drawbacks, it is clear that only portable, non invasive and fast analytical techniques must be used. Another important consideration is the following: no analytical technique, when used alone, can yield all information needed to characterize palettes. Several aspects support this statement. Information must be gained either on macroscopic scale (1-3 mm spot) to identify the main components, and on microscopic scale (0.1 - 0.001 mm) to identify single components in a mixture. Mixtures can be apparent (brown red + black; pink red + white; grey white + black) or less evident (green blue + yellow; purple red + blue). Illuminators used sometimes a mixture of indigo and orpiment called vergaut or vergant, to obtain a green hue. Secondly, information must be gained either with surface techniques and with indepth techniques. Surface techniques (such as Raman spectroscopy) allow identification of colorants in the last pictorial layer, of varnishes and protective layers and of
mo n ica g u l m i n i el eo n o ra pe l l i z z i va l e n t i n a b i a n co
Department of Analytical Chemistry, University of Turin (Italy)
231
alteration compounds, while in-depth techniques (such as XRF) allow identification of underlying pictorial layers, information on preparation layers and on grounds. This can be seen from the following example. In the miniature shown in fig.1, taken from a XV century Book of hours by Antoine de Lonhy, held in the Museo Civico di Arte Antica in Turin (Italy), the Virgins robe is painted in blue. XRF analysis on the blue area shows the presence of copper, suggesting the use of azurite; Raman and FORS analysis, though, both show the presence of ultramarine blue in the surface layer, so that we can hypothesize that the author used the layering technique, that is the superimposition of different pigments (Aceto et al. 2008).
fig.1 complementary analysis on the virgins robe from a xv century miniature
In the analysis of manuscripts, among the techniques available in portable version Raman spectroscopy is with no doubt the one with the best diagnostic power (Aceto et al. 2006; Bersani et al. 2006). It requests, though, lengthy times of analysis. XRF spectrometry has a good diagnostic power (Bruni et al. 2008) but results are sometimes not conclusive, due to the fact that it is an elemental technique; moreover it cannot be used to identify organic compounds. IR spectrophotometry (Bruni et al. 1999) and X-ray Diffraction spectrometry (Duran et al. 2009), though available in portable version, are at present still difficult to be used in situ as self-consistant techniques. A good alternative to these powerful but sophisticated techniques is UV-visible diffuse reflectance spectrophotometry, whose acronym is FORS when it is used with fiber optics (Bacci et al. 1997, 28; Bacci 2000). This technique is easy to use, it requests short analysis times and it can be used in all geometrical situations.
232
Truly portable instruments are available on the market. Its major drawback is that results are hard to interpret in case of mixtures and when varnishes or patinas are present, being a surface technique. Moreover, due to the present performances of probes, it is not fit to analyze short painted areas (< 3 mm). It must be considered, though, that in miniature painting the range of colorants is usually narrow and known according to age and geographic zone, and that usually varnishes are not present. The range of information available from the cited techniques is the following: Raman and IR can yield information on compounds present, such as colorants, ligands and varnishes; UV-Visible FORS can yield information on compounds present in the surface layer; XRF can yield information on elements present in colorants as key-elements, in order to identify the colorants, and as impurities, in order to have information on the origin orf raw materials. To resume, it is clear the need to operate with more techniques, due to fact that these must be non invasive and portable, fast and with minimal impact on the object being examined, to be executed on macro and micro scale, on surface and in-depth. As said before, no analytical technique can fulfill all these requirements when used alone.
Analytical protocol
In order to address these issues, we propose the development of an analytical protocol to optimize the number and type of analysis needed in the characterization of miniature paintings. The protocol is composed by the following steps. preliminary analysis with UV-visible FORS chemometric treatment of spectral data visual inspection with digital camera XRF analysis Raman analysis As it is obvious, the application of the protocol must be preceded by a proper knowledge of bibliographic sources, in order to build spectral databases in proper conditions. This is because it is much better to compare analytical results from unknown samples with analytical standards prepared in similar conditions. A palette of colorants on parchment was therefore prepared, choosing colorants among those used by medieval artists and following recipes of medieval technical treatises such as De Arte Illuminandi (Brunello 1971a), Compositiones ad tingenda musiva also known as Manoscritto di Lucca (Caffaro 2000) and Il libro dellarte (Brunello 1971b). To simulate the painting techniques used by ancient illuminators, paints were prepared either in egg tempera and in gum Arabic. In fig.2 the resulting palette on parchment is shown.
233
fig.2 the complete palette of colorants on parchment
1st stage
In the preliminary stage, FORS analysis is executed on all painted areas of a manuscript (fig.3). Identification of colored colorants (i.e. all but black, white and grey) is performed according to their spectral features: reflectance or absorbance maxima for blue, green and purple colorants, inflexion points for yellow, red and orange colorants (fig.4).
fig.3 analysis of miniatures by fors
fig.4 spectral features in the fors spectrum
2nd stage
Multivariate analysis is performed on FORS spectra in order to identify painted areas obtained with similar pigments. Among unsupervised pattern recognition techniques, Hierarchical Cluster Analysis (HCA) or Principal Components Analysis (PCA) can be used. Upon classification of FORS spectra into different classes, identification performed on a single painted area can be extended to all items, i.e. all painted areas belonging to the same class. It must be noted that classification must be performed
234
separately on painted areas of a single hue: if all hues were classified simultaneously, the differences among hues would be stronger than the differences among pigments of the similar hue. In fig.5 a dendrogram is shown resulting from cluster analysis on blue painted areas taken from a XII century Italian manuscript: painted areas are clearly grouped into three classes, respectively made of azurite, indigo and ultramarine blue. In fig.6 a similar result is obtained by red painted areas from a XV century Italian manuscript: again, it is apparent the classification in distinct groups, in this case made of minium and cinnabar.
fig.5 dendrogram from cluster analysis on blue painted areas from xii century italian manuscript
fig.6 dendrogram from cluster analysis on red painted areas from a xv century italian manuscript
3rd stage
After performing FORS analysis, it is useful to carry out a visual inspection of painted areas in order to better understand the result of FORS analysis. This can be made through a digital camera connected with a zoom lens (in the present case a 10x-80x zoom, fig.7), in order to have a look under the microscopic scale. Visual inspection can yield useful information concerning the identification of mixtures, which can help tuning the interpretation of FORS spectra; moreover identification of altered areas and of particular features can be obtained. In the following figures some examples are shown, taken from inspection on a XII century Italian manuscript called Liber Evangeliorum, held in the Archive and Chapter Library of Vercelli (Italy).
fig.7 visual inspection on a manuscript with a digital camera and a zoom lens
235
In fig.8 a blue initial is shown. The corresponding FORS spectrum, shown in fig.9, suggests the presence of azurite but we must note that the reflectance maximum is red-shifted. An image at 80x magnification (fig.10) allows to clarify this behavior: little red particles, later identified as made of cinnabar, are present that cause the red shift in the spectrum; as a consequence we must think that the blue initial was painted with a mixture of azurite and cinnabar, possibly due to a dirty brush.
fig.8 a blue initial from a xii century italian manuscript fig.9 fors spectra of the blue initial (blue line) and of a standard of azurite (black line)
fig.10 magnified image (80x) of the blue initial
In fig.11 the miniature shows the Virgins and Saint Johns robes painted in a very weak blue hue. The image at 80x magnification (fig.12) put into evidence that residual particles of ultramarine blue are present: most probably both robes were painted in blue but the painted areas were later subjected to a phenomenon known as ultramarine sickness, in which ultramarine looses its color as a consequence of interaction with acidic agents.
fig.11 a highly degraded miniature from a xii century italian manuscript fig.12 magnified image (80x) of the blue paint on the virgins robe
236
In fig.13 it is possible, at high magnification, to see the preparation layer underlying a gold foil and to hypothesize that the preparation be of the flat gilding type, that is with glue. In fig.14 another gilding shows the presence of gold and silver foils overlapped (the so-called oro di mist). Finally, in fig.15 a translucent layer of iron-rich paint is shown, which was later identified as iron-gall ink used as pigment.
fig.13 magnified image (80x) of a gilding
fig.14 magnified image (80x) of a oro di mist gilding
fig.13 magnified image (80x) of a paint made from iron-gall ink
4th stage
After FORS analysis, identification of colorants is followed by XRF analysis (fig.16). Application of XRF is mandatory to address the following issues that cannot be fulfilled by other techniques: characterization of metal pigments such as gold, silver, tin, etc. either in foil or in powder form; characterization of uncolored colorants, i.e. white, black and grey pigments that cannot be identified by FORS; identification of overlapping layers; identification of mordants used with dyes and lakes; identification of impurities in pigments, useful to yield information on the provenance of raw materials. The last issue is exemplified in fig.17 which shows the relationship among copper and barium as determined by XRF on blue painted areas laid with azurite on a XV century Italian manuscript. It is apparent the good correlation among copper (chromophore in azurite) and barium (impurity in azurite). Same result is obtained for copper and zinc in green areas laid with malachite. These results suggest a link among the pigments and the minerals from which they were obtained: if large enough information is available, the geographical origin of the minerals used can be determined.
fig.16 analysis of miniatures by xrf spectrometry
fig.17 relationship among copper and barium in blue paints
237
5th stage
At this stage almost 80-90% of the colorants is already identified. Raman analysis (fig.18), which is together the most sophisticated and the most cumbersome of the cited techniques, can be used as far as uncertain cases are left and to confirm previous identifications. Its spatial resolution helps in fine-tuning identification. Finally, to illustrate the application of the proposed protocol, results from characterization of the palette of a manuscript are described. The manuscript is the already cited Book of hours by Antoine de Lonhy, held in the Museo Civico di Arte Antica in Turin (Italy) and dating to the XV century. The whole palette is reported in tab.1.
Hue White Blue Colorant lead white ultramarine blue on azurite azurite phtalocyanine blue Yellow gold foil shell gold mosaic gold lead-tin yellow type I Grey Black Red gold on silver carbon cinnabar minium Green Violet malachite lake (kermes?) with lead white FORS n.i. X X X n.i. n.i. n.i. X n.i. n.i. X X X n.i. XRF Pb Cu Cu, Ba Cu, Ti Au Au Sn, S Pb, Sn Ag, Au n.i. Hg, S Pb Cu, Zn Pb Raman X X X X n.i. n.i. X X n.i. X X X X n.i.
fig.18 analysis of miniatures by raman spctroscopy
table 1 palette of the book of hours by antoine de lonhy (xv century). x = colorant identified by the technique; n.i. = colorant not identified by the technique
In fig.19 a miniature taken from the Book of hours shows the presence of four different yellow pigments used by the author in order to obtain different effects. Identification was possible only by complementary application of analytical techniques, following the protocol described. The miniature in fig.20 shows a peculiar feature. Again, the application of the protocol allowed to put into evidence a later retouch in the background sky, in which the original paint in ultramarine blue was reinforced with phtalocyanine blue laid on titanium white, an intervention made later than 1935 when phtalocyanines were patented.
238
fig.19 yellow pigments identified on a xv century miniature
fig.20 original and later added blue pigments on a xv century miniature
239
Bibliography
Aceto, M., Agostino, A., Boccaleri, E., Crivello, F., Cerutti Garlanda, A. 2006. Evidence for the degradation of an alloy pigment on an ancient Italian manuscript. Journal of Raman Spectroscopy, vol. 37: 1160-1170. Aceto, M. Agostino, A., Gulmini, M., Pellizzi, E., Castronovo, S. 2008. Reconstructing the palette used in a fifteenth century manuscript illuminated by Antoine de Lonhy. Porceedings of the 37th International Symposium on Archaeometry, Siena, 12-16 Maggio 2008. Bacci, M., Picollo, M., Porcinai, S., Radicati, B. 1997, Non destructive spectrophotometry and colour measurements applied to the study of works of art. Techne, vol. 5: 28-33. Bacci, M. 2000. UV-VIS-NIR, FT-IR and FORS spectroscopies. In: E. Ciliberto, G. Spoto eds., Modern Analytical Methods in Art and Archaeology, Chemical Analysis Series, vol. 155: 321-361. New York: John Wiley & Sons Inc. Bersani, D., Lottici, P.P., Vignali, F., Zanichelli, G. 2006. A study of medieval illuminated manuscripts by means of portable Raman equipments. Journal of Raman Spectroscopy, vol. 37: 1012-1018. Brunello, F. 1971a. De arte illuminandi. Vicenza: Neri Pozza. Brunello, F. 1971b. Il libro dellarte. Vicenza: Neri Pozza. Bruni, S., Cariati, F., Casadio, F., Toniolo, L. 1999. Identification of pigments on a XV century illuminated parchment by Raman and FTIR microspectroscopies. Spectrochimica Acta Part A, vol. 55: 1371-1377. Bruni, S., Caglio, S., Guglielmi, V., Poldi, G. 2008. The joined use of n.i. spectroscopic analyses FTIR, Raman, visible reflectance spectrometry and EDXRF to study drawings and illuminated manuscripts. Applied Physics A, vol. 92: 103-108. Caffaro, A. 2000. Scrivere in oro. Salerno: Palladio. Duran, A., Perez-Rodriguez, J. L., Espejo, T., Franquelo, M. L., Castaing, J., Walter, P. 2009. Characterization of illuminated manuscripts by laboratory-made portable XRD and micro-XRD systems. Analytical and Bioanalytical Chemistry, vol. 395: 1997-2004.
240
Biographies
Maurizio Aceto graduated in Chemistry (University of Turin, Italy) and obtained a PhD in Chemical Sciences (University of Turin, Italy), His research interests concern characterisation of colorants of pictorial artworks with non-destructive and noninvasive spectroscopic techniques. Currently he is research associate at the Department of Environmental and Life Sciences, University of Eastern Piedmont, Italy; he is also a member of Centro Interdisciplinare per lo Studio e la Conservazione dei Beni Culturali (CenISCo), University of Eastern Piedmont, sede of Vercelli. Address: Department of Environmental and Life Sciences, University of Eastern Piedmont, via T. Michel, 11 15100 Alessandria, Italy. Tel.: +39 0131 360265; Fax: +39 0131 360250; E-mail: maurizio. aceto@unipmn.it. Angelo Agostino graduated in Chemistry (University of Turin, Italy) and obtained a PhD in Chemical Sciencs (University of Turin, Italy), His research interests concern application of X-ray spectroscopic techniques to the characterisation of materials of artisticarchaeological interest. Currently he is technician at the Department of General and Applied Organic Chemistry, University of Turin; he is also a member of Centro di Eccellenza NIS (Nanostructured Interfaces and Surfaces), University of Turin. Address: Department of General and Applied Organic Chemistry, University of Turin, c.so M. dAzeglio, 48 10125 Turin, Italy. Tel.: +39 011 6707585; Fax: +39 011 6707585; E-mail: angelo. agostino@unito.it. Monica Gulmini graduated in Chemistry (University of Turin, Italy) and obtained a PhD in Chemical Sciences (University of Turin, Italy). His research interests concern characterisation of glasses, ceramics and pictorial artworks with spectroscopic techniques. Currently she is research associate at the Department of Analytical Chemistry, University of Turin. Address: Department of Analytical Chemistry, University of Turin, via P. Giuria, 5 10125 Turin, Italy. Tel.: +39 011 6707618; Fax: +39 011 6707615; E-mail: monica. gulmini@unito.it. Valentina Bianco graduated in graduated in Sciences and Technologies for the Cultural Heritage (University of Turin, Italy). Address: Department of Analytical Chemistry, University of Turin, via P. Giuria, 5 10125 Turin, Italy. Eleonora Pellizzi graduated in graduated in Sciences and Technologies for the Cultural Heritage (University of Turin, Italy). Address: Centre de Recherche sur la Conservation des Collections, Musum National dHistoire Naturelle, 36 rue Saint Hilaire 75005 Paris, France.
241
Resumo
A fluorescncia de raios-X por radiao de sincrotro uma tcnica eficaz de anlise elementar, in situ e no-invasiva. O uso de novos detectores permite operar com fluxos elevados sem perca de resoluo. Apresentamos um estudo onde a SR-XRF utilizada, com um sensor CCD (thick fully-depleted), no estudo dos pigmentos utilizados numa iluminura de um manuscrito produzido no crculo de Fra Angelico, em torno a 1450. Os dados obtidos so utilizados conjuntamente com a anlise estilstica e a pesquisa em arquivos para a interpretao do processo de decorao de manuscritos no sc. XV em Florena.
palavras-chave
sincrotro fluorescncia de raios-x dispersiva de energias iluminuras fra angelico
Abstract
Synchrotron-radiation induced X-ray fluorescence is an effective technique for noninvasive, in-situ, elemental analysis. The use of new detectors enables operation at large fluxes without loss of resolution. We present a case-study application of SR-XRF with a thick, fully-depleted CCD sensor to the analysis of the pigments on the illumination of a manuscript decorated in the circle of Fra Angelico around 1450. Physical data are integrated with stylistic analysis and archive research in the interpretation of the process of manuscript decoration in XV century Florence.
key-words
synchrotron x-ray fluorescence manuscript illuminations fra angelico
the application of sr-xrf to the analysis of manuscript illumination

a case study 1
m a rco b at tag l i a
University of California at Santa Cruz and Lawrence Berkeley National Laboratory, CA (USA)
l au ra a l ido r i b at tag l i a
University of California at Santa Cruz, CA (USA)
rich a rd cel e st r e p eter den es d io nis io do e r i n g ta e s u n g k i m sa ra h z a lu sky

Lawrence Berkeley National Laboratory, Berkeley, CA (USA)
Introduction
The use of synchrotron radiation induced X-ray fluorescence (SR-XRF) is well suited for the elemental analysis of pigments in artworks since the analysis is fast and non-destructive, the beam can be made monochromatic and its energy changed to fit the analysis requirements. The small spot size probes tiny regions and disentangles individual pigments and small pitch scans can be performed to acquire detailed pigment maps. XRF 2 and PIXE 3 analyses have already provided us with fundamental data for the study of the pigments used in the decoration of Medieval and Renaissance manuscripts. Here we discuss pigment analysis by SR-XRF to clarify the methods of work and sharing of responsibilities among artists of the circle of Fra Angelico in mid-XV century Florence. This case study is based on the analysis of the illuminated opening page of a manuscript decorated by Battista di Biagio Sanguigni and another artist, around 1450. The physical data are integrated with stylistic analysis and archive research in the interpretation of manuscript illumination. This analysis introduces also a novel detector for XRF. We employ a thick, high-resistivity, front-illuminated Charge Coupled Device (CCD). CCD sensors offer several advantages over a conventional Si(Li) detectors. Their active surface is large and the high pixellisation allows us to operate with large X-ray fluxes without pile-up effects. This paper is organized as follows. First we discuss some aspects of the organisation of manuscript decoration in Florence around 1450 as it emerges from archival sources. Then we present the manuscript under study, the experimental set-up and the CCD performance. Finally we discuss the results of the analysis of the red pigment composition, in areas which appear to be due to different artists, and of the
1. In this article LAB has been responsible for the sections on art history and archive research, the other authors for the XRF analysis. 2. M Picollo et al., these proceedings and references therein. 3. M Bernasconi et al., Analyse des couleurs dans un groupe de manuscrits enlumines du XII au XV sicle avec lemploi de la technique PIXE, in Ancient and Medieval Book Materials and Techniques, Citta del Vaticano, 1993, pp. 57-101.
243
rendering of the incarnato, comparing it with that by other artists active in Florence in the same decade.
4. LB Kanter, Zanobi Strozzi miniatore and Battista di Biagio Sanguigni, Arte Cristiana, 90 (2002), pp. 321-331. 5. ASF, Conv. Soppr. 234, San Gaggio, n. 77, c. 13v (1432): note of a credit with the monastery of San Gaggio: Batista miniatore dee avere per insino a di primo di novembre 1432 per miniature duno ynnario grande notato e per miniature in uno salterio grande da coro e per miniature negli antifonari nuovi; ibid. n. 78, c. 35r (1435) Batista di miniatore de avere a di 14 ottobre (1435) lb. 33 s. 14 d. 2 per resto di sua ragione per i libri miniati al munistero ; ASF, Catasto 1433, vol. 482, c. 264r (1433): note of a credit with Bardo de Bardi per un libricciuolo gli fe; published in W Cohn, Il Beato Angelico e Battista di Biagio Sanguigni. Nuovi documenti, Rivista darte, 30 (1955), pp. 207-216, in particular pp. 213, 215-216. 6. As it is well known, it was Battista di Biagio who introduced Guido di Pietro (Fra Angelico) to the Compagnia di San Niccolo di Bari in 1417 (ASF, Compagnie Soppresse, vol 1549: Compagnia di San Niccolo di Bari in Santa Maria del Carmine, Elenco dei Confratelli c-c 17, n. 410 e a-c 18, n. 507) and lived near his house in S. Michele Visdomini. He later shared a house, located near San Domenico in Fiesole where Guido di Pietro moved after joining the Dominican order, with Zanobi Strozzi from around 1430 until 1438; Cohn 1955. 7. A Dillon Bussi, Zanobi Strozzi istoriatore e non miniatore (Indagine nel mondo della miniatura muovendo dai quattro piu importanti corali quattrocenteschi fiorentini), Rara Volumina, 13 (2006), 1, pp. 15-25. 8. A Dillon Bussi, Una glorificazione di San Pancrazio di Zanobi Strozzi, Paragone Arte, Anno 57, n. 69 (2006), pp 3-19. 9. ASF, Corporazioni Religiosi Soppresse dal Governo Francese 88, San Pancrazio, vol. 63, c. 34r, published in M Levi dAncona, Miniatura e Miniatori a Firenze dal XIV al XVI secolo, Firenze, Olschki, 1962, p 107: nel qual libro abbiavamo speso insino nel 1437 in fare fare le figure del minio della messa di san pancratio e piu altri fogliami ed adorneza di detto minio. Il quale fece
Ad faciendum storias and ad miniandum: the Sharing of Illumination Roles and the circle of Fra Angelico
Fra Angelico and the artists of his immediate circle enjoy a privileged role in the landscape of manuscript decoration in XV century Florence, as they were responsible for some of the most prestigious cycles of service books. This small group of artists included Zanobi Strozzi and Battista di Biagio Sanguigni and they all worked both as panel painters and as illuminators, at least according to the recent identification of the so-called Maestro del 1419 with Battista di Biagio 4. The profile of Battista di Biagio as illuminator, which still needs to be studied in detail, can be reconstructed from few documented commissions from the 1430s 5. The choirbooks for San Gaggio, now at San Marco, are the centerpiece for the identification of his style. Starting from these works, a large corpus of illuminations has been assembled. Battista di Biagio emerges as one of leading illuminators in Florence in the second quarter of the XV century and his relationship to Fra Angelico and the other artists of his immediate circle appears of special importance. Not only did he collaborate with Fra Angelico and Zanobi Strozzi to the decoration of manuscripts, but he also developed close personal relations to these artists 6. The analysis of documented commissions of manuscript decoration by these artists has clarified the difference of roles between istoriatore and miniatore, as it has recently been pointed out 7. Three examples are particularly pointing. The first is the 1437 commission for an illumination for the monastery of San Pancrazio 8 to Zanobi Strozzi and Filippo di Matteo Torelli 9. The second is the 1445 commission for the decoration of the antiphonaries of Santa Maria del Fiore 10 to Zanobi Strozzi, Filippo di Matteo Torelli, Battista di Niccolo da Padova and a not yet identified Bartolomeo 11. Finally, the Libro delle Ricordanze del Convento di San Marco, A relates us details of the making of the choirbooks for the Dominican foundation by Zanobi Strozzi and other artists from 1446 to 1454 12. In all these documents the sharing of roles is clearly defined. Specific artists are named as responsible for the figure in the main scene, or storia, and others for the leafy marginal decoration, fogliami e oro e altri adornamenti. These roles are generally respected: Filippo di Matteo Torelli is attributed almost exclusively the role of miniatore, while Zanobi Strozzi is documented to work on the same leaves as istoriatore. In at least one instance the exchange of leaves from one artist to the next during the decoration process is documented 13. We know from stylistics analysis that Battista di Biagio was similarly active as miniatore, responsible for the marginal decoration and the body of the letters, in the majority of the manuscripts for which Fra Angelico painted the historiated initials 14. This collaboration may date
244
t h e a p p l i c a t i o n o f s r - x r f t o t h e a n a ly s i s o f m a n u s c r i p t i l l u m i n a t i o n
Zanobi degli Strozi cioe le figure et il resto Pippo di Matteo che dacordo ebbono insino a di 17 di maggio 1437 L.20 s.-. 10. Mss. Edili 148-151 in the Biblioteca Medicea Laurenziana, see I libri del Duomo di Firenze: codici liturgici e biblioteca di Santa Maria del Fiore (secoli XI-XVI), L Fabbri and M Tacconi (eds.), Firenze 1997, pp. 79-96 (Dillon Bussi) and pp. 217-224 (Tacconi). 11. Archivio Opera del Duomo, I-1-4, c. 41t , 54 published in G Poggi, Il Duomo di Firenze, vol. II (1988), p 39, n.1630: Nobiles viri operarii antedicti omni modo etc. locaverunt Zanobio Benedicti Carocii de Strozis ad faciendum storias de uno ex antifanariis locatis ad scribendum Goro, pro pretio, et remuneratione alias fiendis. Item locaverunt magistro Batistead faciendum storias in alio antifanario predicto pro pretio alias fiendo. Item locaverunt ad miniandum unum de dictis antifanariis Filippo Mathei Torelli, miniatori, pretii (sic) alias fiendo. Item locaverunt ad miniandum alium antifanarium Bartolomeo minatori. 12. Biblioteca Medicea Laurenziana, Libro delle Ricordanze di San Marco A, San Marco 902, f 26v: Ricordo come Zenobio degli Strozzi miniature auto da me frate costantino [da Nocera] de san marcho per storie fan el primo graduale delle feste florin dodici in duo partite and following entries published in S Orlandi, Beato Angelico, Firenze, 1964, pp 116-117 and 194-195. 13. On 23 May 1447 Zanobi receives from Filippo Matteo Torelli two letters to be decorated, San Marco 902 cit., f 23r published in M Levi DAncona, Miniatura e miniatori a Firenze dal XIV al XVI secolo, Firenze, 1962, p 265. 14. LB Kanter, ad vocem Guido di Pietro, in Dizionario biografico dei miniatori italiani: secoli IX-XVI, M Bollati (ed), Milano 2004, pp. 334-335. 15. See for example and references therein: A Guidotti, Indagini su botteghe di cartolai e miniatori a Firenze nel XV secolo, in La miniature italiana tra Gotico e Rinascimento, Atti del II Congresso di Storia della Miniatura Italiana, E Sesti (ed.), Firenze 1985, pp. 473-507; A Guidotti, Nuovi documenti su Vespasiano da Bisticci la sua bottega e la sua famiglia, in Federico da Mon-
as early circa 1425 in the work for the choirbook Ms 558 for San Domenico, now in San Marco. It likely continued shortly afterwards in the decoration of Ms. Gerli 54 in Biblioteca Braidense of Milano. We recognize again his hand in the decoration of the two Psalters, Mss 530 and 531 of San Marco, which can be dated circa 1449 near the end of his life and are very close in style to the miniature analyzed here. This peculiar organisation of work, which sees two different artists working on the same initial is further confirmed by the analysis of Missal Ms 534 from San Marco where one artist, possibly Sanguigni himself, paints all the letters and their marginal decoration leaving the space for the work of the istoriatore. As the decoration of the missal was left unfinished, today we see the spaces left for the storie. Conversely, in the choirbooks for San Gaggio, the main illumination, by Sanguigni, was completed, but several foliated decorations were left unfinished. All these examples demonstrate that the decoration of the letters and their marginal extensions and those of the storie belonged to two distinct and possibly independent phases in the process of manuscript decoration, thus confirming the evidence from the archival sources. The organisation of artist workshops has been studied in details, 15 including the relations between the workshop principal and his collaborators and the role of the cartolaio in the distribution of the work of manuscript production and decoration. Since the appearance of the finished manuscript does not reveal to us chromatic inconsistencies, we are faced with the question whether different artists did work side-by-side, sharing the same pigments, or the observed chromatic uniformity is the result of a high degree of organisation of this process, which ensured the constant quality of the final product 16. While archival sources are generally scarce on such details of the process of manuscript production, in at few instances, documents confirm that single leaves were brought to the artist for being decorated. One of these is the case of a group of leaves for one of the antiphonaries being produced for Santa Maria del Fiore and brought to Filippo Maria Torelli in 1450 17. Therefore, we know about the effective organisation of manuscript production, which followed the various stages of the work, from the acquisition of the parchment to the writing, decoration and binding, dispatching the parts through the work chain. Responsible for this was in general, probably, the cartolaio but in the specific case of the choirbooks of the Duomo this role was taken by the patron itself, i.e. the Opera di Santa Maria del Fiore. It is suggestive to assume that the same form of organisation could also ensure the uniformity of the decorative work, even when storie and decorative borders where painted by different hands at different locations. The sources do not help us in clarifying this point. In fact, mentions of transactions of colours between patrons and artists, in the surviving documents for XV century Florence, are mostly limited to the gold and the azzurro, possibly due to their high cost 18, even though the quality of the colours to be employed is often mentioned in contracts 19. Now, the identification of another manuscript where Battista di Biagio shares the decoration work with an artist of the immediate Fra Angelico circle, provides us with the opportunity to better understand their working practice and confirm the stylistic evidence with physical data.
245
SR-XRF Analysis of a Miniature by Battista di Biagio Sanguigni and a Collaborator of Fra Angelico
The illumination analysed in this study is on the opening page of a manuscript containing works attributed to St. Jerome, now in a private collection, decorated around 1450 20. The stylistic analysis indicates that its decoration is the work of two distinct artists. One is responsible for the body of the initial and its foliated extensions and can be doubtlessly identified with Battista di Biagio Sanguigni, while a second artist, who expresses a style very close to that of Fra Angelico himself, paints the figure of the St. Jerome in the initial. This SR-XRF analysis was performed on beam line 5.3.1 at the Lawrence Berkeley National Laboratory Advanced Light Source (ALS). Synchrotron radiation is produced by the 1.9 GeV electron beam in one of the main ring bending magnets, which gives a continuum X-ray spectrum up to ~20 keV. Upstream from the sample a mono-chromator selects the beam energy. A collimator eliminates the white beam and defines the monochromatic beam geometry while a shutter controls the exposure time. For this study the beam energy was chosen to be 12 keV. The beam spot at the sample surface has a diameter of ~1 mm. The sample is mounted on a computer-controlled XY stage, which allows us to perform automated scans of the surface of the manuscript. X-rays are detected with a thick, front-illuminated CCD sensor developed at LBNL on high resistivity Si. The CCD has an active surface of 118 mm2 with 1024x512 pixels arrayed on a 15 m pitch and two output channels. Under the operating conditions adopted for this analysis the sensitive thickness is ~450 m, which offers high spectral sensitivity up to energies well above the beam energy. Thick, high-resistivity CCDs offer high detection efficiency and an excellent energy resolution, matching or surpassing that of conventional solid-state X-ray detectors. The single pixel noise is measured to be (4.21.0) e ENC. The energy resolution measured for reconstructed clusters from a 55Fe source is 155 eV FWHM at 5.9 keV. The CCD response is calibrated in the range 3-12 keV using thin metal foils in the same geometry used as for the manuscript analysis. Data are saved in the fits format and subsequently converted to the lcio format, while performing pixel-by-pixel pedestal subtraction and noise computation from CCD exposure taken without beam. The offline analysis is performed using custom clustering processors in the Marlin C++ analysis framework. Beam back-scattering from the sample is removed by subtracting a reference spectrum obtained with the beam aimed at a non-decorated part of the parchment. The spectrum is reweighted to correct for the effect of X-ray absorption in the kapton window and the inactive detector surface. The final spectrum analysis and peak search is performed in ROOT. Cluster shape analysis is helpful for rejecting spurious background signals, pile-up clusters and scattered photons striking the detector at large angles. This is particularly important when performing XRF on a high intensity primary beam, such as at a light source. The XRF analysis sampled a total of twenty-six points on the decorated
tefeltro: lo stato, le arti, la cultura, G Cerbeni Baiardi et al (eds.), Roma, 1986, pp. 97-111; L Indrio, Firenze nel Quattrocento: divisione e organizzzione del lavoro nelle botteghe, in Il colore dellantico, Ricerche di storia dellarte, 38 (1989), pp. 61-70; JJG Alexander, Medieval illuminators and their methods of work, New Haven 1992; Maestri e botteghe : pittura a Firenze alla fine del Quattrocento, M Gregori et al. (eds), Firenze 1992, Organizzazione, componenti e ruoli della bottega laica tra XIV e XV secolo, A Guidotti (ed.), Firenze, 2006. 16. A Dillon Bussi, Rara Volumina (2006). 17. Archivio Opera del Duomo, VIII-4-1, Giornale H, c. 54 pubblicato in G Poggi, Il Duomo di Firenze, vol. II (1988), p 47, n.1672: A Filippo di Matteo Torelli, miniature, addi 6 di luglio gli porto a chasa Martino manovale cho licenza di Pazino operai 2 asse dalbero nero entrovi charte scritte e notate del primo antifanare si fanno di nuovo perche lle minii Rimando dette cose allopera. Another case concerns the Olivetan convent of S. Bartolomeo alle Sacca in nearby Prato where leaves written in the convent scriptorium were sent to an outside workshop for being decorated, M Ciatti, Appunti e documenti per la storia della miniature a Prato nel Quattrocento, in La miniatura italiana tra Gotico e Rinascimento, Atti del II Congresso di Storia della Miniatura Italiana, E Sesti (ed.), Firenze, 1985, pp. 509-533, in particular p. 528. 18. For example, Battista di Biagio contracted a debt with the friars of Santa Maria degli Angeli for blue pigment in 1430 (ASF, Catasto 1430, vol 389, c . 354). Filippo di Matteo Torelli is paied for blue pigment in 1437 l. tre s. xi demo a Pippo miniature per oncia 1 dazuro de la magna e per oncia dazuro ultramarino , ASF, Conventi Soppressi 78, Badia vol. I, Giornale (1435-1441). 19. A Guidotti, Il mestiere del dipintore nellItalia Due-Trecentesca, in La Pittura in Italia. Il Duecento e il Trecento, E Castelnuovo (ed.), Milano, 1986, pp. 529-540, in particular p. 535. 20. L Alidori Battaglia, An unpublished miniature from the circle of Fra Angelico, The Burlington Magazine, vol. CLI, n. 1277 (2009), pp. 518-525.
246
surface of the manuscript (see Figure 1). Ten different pigments have been analyzed. In some cases, the same pigment has been sampled at different locations to study the effect of possible local in-homogeneities. Here we discuss the pigment composition on those parts which appear to be due to two different artists and the technique for the rendering of the incarnato of St. Jerome face.
fig.1 page with the initial l with st. jerome from the regula monachorum, c. 1r, tempera on parchment, 19 by 14 cm (private collection).
2 47
First, we study the composition of the red pigment at various locations on the foliated decoration and on the book hold by the St. Jerome. The spectra are shown in Figure 2. The analysis reveals Pb and a small amount of Cu in the red pigment on the leaf protruding from the letter. This is consistent with what we would expect for minium. Instead, the red pigment on the book is characterized by a large quantity of Cu with Pb, Fe with traces of Zn and Ba. This composition indicates that the artist mixed minium with another pigment, possibly an ocra rossa. We sampled three distinct spots on the leaf and two on the book. Results are consistent. Differences in the pigment composition, associated to the work of different artists in the same manuscripts have already been observed 21. The different composition of the two pigments seems to support the indication from the stylistic analysis that there are two artists working independently on the miniature.
21. M Bagnoli, Amanuensi e miniatori in un Decretum Gratiani del Walters Art Museum di Baltimora, Arte Medievale, n.s. Anno VI (2007), 2, pp. 65-74 reports subsequent interventions on the same initials; M Bernasconi et al., 1993 identifies differences in pigment composition between Florentine-style and Lucca-style artists working on BML, Conv. Soppr. 298.
fig.2 absorption- and efficiency-corrected xrf spectra for the red pigment on the book hold by the saint (black histogram) and on the marginal leaf (gray histogram).
The second part of this study attempts to elucidate the technique employed for the rendering of the incarnato. The deep shadows on the face of St. Jerome are obtained with brushstrokes in an olive colour, possibly a terra verde which follows the outline of the cheeks and the beard. The analysis of the beard shows that the pigment has Fe, Cu, Pb and traces of Ba. Instead, on the forehead there are larger brushstrokes beneath the final painted layer, revealed by infra-red photography. This technique is adopted by Fra Angelico as well as Zanobi Strozzi. The XRF analysis enables us to compare the pigment composition on the forehead, the breast and the hand. The four spectra are shown in Figure 3. The composition of the pigment used on the breast and on the hand appears to be identical. It is characterised by Fe and Pb, with traces of Cu, Zn and Ti. On the contrary, we observe a much larger Fe and
248
22. R Cambria et al, A PIXE analysis of Manuscripts Illuminated bu Francesco di Antonio del Chierico, Ancient and Medieval Book cit., pp. 103-119.
Cu content on the forehead, with traces of Mn and Cr. The rendering of deep shades on the face by underlying a Cu-based green pigment has already been observed in the PIXE analysis of ms. Plut. 66.22 of the Biblioteca Medicea Laurenziana 22. This manuscript is dated 1455 in the colophon and was decorated by Francesco dAntonio del Chierico. Other pigments differ significantly in the two manuscripts. For example, the analysis reveals only Ca, from the parchment preparation, in a pink-coloured area in the miniature of Francesco dAntonio, which suggests the use of a pigment of organic origin. On the leaf by Battista di Biagio in our manuscript the pink is characterized by Sb, Pb, Cu and Fe. These comparisons are particularly interesting since they refer to two manuscripts decorated within a few years by two of the most prominent workshops in Florence and highlight the variety of techniques and pigments adopted by these artists.
fig.3 absorption- and efficiency-corrected xrf spectra for the pigment on the forehead in the region of the deep shadow (continuous black histogram) and away from it (continuous gray histogram), on the breast of the saint (dashed black histogram) and on the hand (dashed gray histogram).
Conclusions and Perspectives

This case study offered us with the opportunity to address some issues in the practice of manuscript decoration in mid-fifteenth century Florence. We attempted to combine the results of SR-XRF analysis with stylistics analysis and archival research. Even within the limitations imposed by data collected from a single manuscript, the picture which emerges is very consistent and quite encouraging for the continuation of such studies. Manuscript decoration was a well-organized and structured process
249
with different artists contributing as miniatori and istoriator. This seems to be the case not only in the decoration of major cycles, but also on minor commissions consisting of a single illuminated initial per manuscript. Results of physical elemental analysis confirm the evidence from stylistic analysis and documents. The results of this case study highlight the need to tightly integrate future campaigns of elemental analyses with well-defined art history problems which can be answered by the knowledge of pigment composition. At the same time it is apparent that more data on pigment analysis should be made available to support the work of art historians in the understanding of manuscript decoration practice and techniques across chronological and geographical boundaries. The effect of the intense X-ray beam from a light source on the pigment and the parchment is an issue which is often raised by conservationists and curators and deserves further studies. Proton irradiation is known to induce the so-called dark spot phenomenon which appears to be related to the generation of colour centers 23, but there is only limited understanding on the mechanisms responsible for pigment damage and on their reversibility 24. Contrary to the case of protons, in SR-XRF, the X-ray energy is often below threshold for displacement damage. We carried out some observations on the pigments on a XV century bas-de-page after exposure to a 15 keV X-ray beam at the ALS. Observations were performed immediately after the irradiation, after several months and after more than a year. No effect was detectable. However, it would be important to further study the issue. Finally the use of advanced pixellated sensors, as the high-resistivity CCDs adopted for this case study, offer new opportunities for high resolution spectroscopy with large photon fluxes, which will benefit the study of manuscript pigments both in terms of the exposure time and sensitivity to trace elements.
23. J Absil et al., Study of color centers induced by PIXE irradiation, Nuclear Instruments and Methods in Physics Research, B198 (2002), pp. 90-97. 24. E. Enguita et al., Damage induced by proton irradiation in carbonate based natural painting pigments, Nuclear Instruments and Methods in Physics Research, B219-220 (2004), pp. 53-56.
250
Resumo
A fragilidade e pequeno tamanho das iluminuras torna difcil a aplicao de mtodos qumicos que requeiram pequenas amostras como comum fazer em quadros. Em alternativa, tcnicas com boa resoluo espacial e passveis de aplicao in situ, como a fluorescncia de raios-X dispersiva de energias (XRF) e microscopia Raman tm sido utilizadas para a identificao de pigmentos. Ainda que o nmero de tcnicas no-invasivas esteja a crescer, so todas de aplicao pontual e como tal no so utilizadas para investigar a obra de arte por inteiro. Recentemente descobrimos que combinando os dados obtidos atravs de tcnicas de imagiografia multiespectrais, baseadas em reflectncia de VIS-NIR e luminescncia, com mtodos pontuais como a espectroscopia de UV-VIS por reflectncia com fibra ptica (FORS) e XFR, possvel identificar e mapear os pigmentos principais utilizados em iluminuras bem como visualizar o desenho preparatrio e eventuais alteraes na composio. Descrevem-se exemplos do procedimento experimental e dos resultados obtidos em diversos flios iluminados.
palavras-chave
imagiografia multiespectral fluorescncia de raios-x dispersiva de energias fors iluminuras
Abstract
The fragility and small dimensions of illuminated manuscripts prohibit the use of chemical methods that rely on small samples being taken from the artwork as is typically done for paintings. Instead site-specific in situ techniques such as X-ray fluorescence (XRF) and Raman spectroscopy are commonly used to identify pigments. While the number of non-invasive analytical techniques is growing, all are site-specific and thus cannot be used to survey the entire work of art. Recently we have found that by combining results from multispectral visible/infrared reflectance and luminescence imaging spectroscopy with those from site-specific methods such as fiber optics reflectance spectroscopy (FORS) and XRF, it is possible to successfully identify and map the primary pigments on medieval illuminated leaves, as well as visualize the underlying preparatory sketches and compositional changes. Examples of the experimental workflow and results obtained by using it to study several illuminated leaves are presented.
key-words
multispectral imaging x-ray fluorescence fors manuscript illuminations
combining visible and infrared imaging spectroscopy with site specific, in-situ techniques for material identification and mapping
pao l a riccia r d i , j ohn k. d e l a n e y
Scientific Research Dept., National Gallery of Art, Washington, D.C. (USA) paola.ricciardi@gmail.com j-delaney@nga.gov
Introduction
The work being carried out at the National Gallery of Art, Washington DC (NGA), is aimed at adapting the imaging spectroscopy methods, originally developed for remote sensing, for use in conservation science, and in particular to explore the resulting synergy when these results are combined with in situ chemical methods. Of particular interest is the study of illuminated manuscripts, which have a simpler palette than many paintings and often cannot be sampled. Among the many questions often addressed in the analysis of illuminated manuscripts, two are most relevant for conservation science. Namely, the identification of the materials used, particularly the colorants, and elucidating the construction techniques used such as layering and the use of preparatory sketches. Site-specific analytical techniques, such as X-ray fluorescence (XRF) and Raman spectroscopy,
253
n e w t r e n d s i n t h e a n a ly s i s o f m e d i e v a l m s
are usually employed to identify the pigments and possibly the binders used for the illumination. While such methods are analytically powerful, they cannot be used to survey the entire surface of a miniature. With both techniques, spectra are usually acquired on visually identified sites thought to be representative of the pigments and mixtures used for the illumination. Such visual inspection may not always ensure an adequate representation of the pigment diversity present. On the other hand visual imaging techniques such as microscopic observation, infrared reflectography (IRR), and UV-excited fluorescence, are employed to visualize the preparatory sketches and understand the layer structure of the miniature. Radiometric calibration of images acquired in numerous visible/infrared spectral bands can be used to produce quantitative results, thus transforming the imaging process into imaging spectroscopy, which can provide chemical information about materials. This is done by generating reflectance and luminescence spectra at each pixel and thus over the whole imaged art object. In recent years, research carried out mostly on paintings has proved that imaging spectroscopy techniques can be effectively used for material identification and mapping on works of art (Casini et al. 1999, Delaney et al. 2005, Delaney et al. 2009a, Delaney et al. 2010). However such imaging spectroscopy methods have required light levels too high for works of art on paper or parchment. Here we present some results regarding the use of a high sensitivity multispectral visible/infrared camera system (400 to 2500 nm, 15 spectral bands), which operates at illumination levels of approximately 150 lux. We have used this camera system to acquire reflectance and luminescence images of several 14 th century Italian illuminated leaves in the NGA collection, notably, one attributed to the workshop of Pacino di Buonaguida, representing Christ in majesty with twelve apostles, and another attributed to Niccol da Bologna, representing the Birth of John the Baptist. Combining the imaging spectroscopy results with data from low light fiber optics reflectance spectroscopy (FORS) and X-ray fluorescence, we have found that the primary pigments on these medieval illuminated leaves can be identified and mapped, and preparatory sketches and changes can be visualized (Delaney et al. 2009c, Ricciardi et al. 2009).
Summary of experimental techniques

Imaging spectroscopy involves the acquisition of spatially co-registered images in many spectral bands in order to allow construction of a spectrum at each point in the spectral image set (Fig. 1). In the case of reflectance imaging spectroscopy the spectral features collected relate to the electronic transitions and some vibrational modes of the artists materials. By grouping similar spectra maps can be made, and the average spectra from each map can be compared to data-bases to help with the identification. Luminescence imaging spectroscopy exploits the observation that a subset of artists pigments, mostly organics dyes, is luminescent when excited with UV/blue light. This is useful to help in the identification of pigments which have
254
n e w t r e n d s i n t h e s t u d y o f m e d i e va l i l l u m i n at e d m a n u s c r i p t s
fig.1 (top) schematic representation of the working principle of reflectance imaging spectroscopy, with the reconstruction of the reflectance spectra for a blue and a red area from the multi-spectral images of christ in majesty with twelve apostles by the workshop of pacino di buonaguida (visible image detail at top left); (bottom) high-resolution reflectance spectra of the same areas collected with the fiber optics reflectance spectrometer. the gray gaussians and rectangles represent the wavelength ranges analyzed by the imaging cameras
255
reflectance spectra that lack unique features, as may be the case for organic dyes, and it is for this reason that it is useful for the analysis of illuminated manuscripts. As noted before, the use of imaging spectroscopy for the study of light-sensitive materials such as illuminated manuscripts can be problematic; to date in fact, many of the multi-spectral imaging (MSI) camera systems having high spatial fidelity require high light levels. The study of works of art on paper and parchment requiring low illumination (about 150 lux, comparable to a few times gallery light levels) is currently carried out at NGA with an optimized low-noise, high-sensitivity 4 mega pixel Si-CCD imaging camera. This system is used to collect 12 narrow-band reflectance and luminescence images in the visible to near infrared range (400 to 950 nm, 40 nm FWHM). While multispectral imaging spectroscopy in this spectral range (visible to near infrared) is a powerful tool to spectrally separate and cluster similar materials it is not robust enough to identify directly most artists materials. Thus site-specific methods such as XRF and FORS can be used to aid in the identification. Infrared reflectograms (IRR) in the 750 to 950 nm and 1000 to 2500 nm range can be acquired using the Si-CCD camera (see above) and a highly sensitive InSb camera equipped with infrared filters, allowing the observation of preparatory sketches and compositional changes at the same low light level used for MSI. IRR is also useful to distinguish between the materials used for underdrawings; most iron-based inks are in fact totally transparent in the 1000-2500 nm range, while carbon-based inks and most metalpoints remain visible throughout this range. Fiber Optics Reflectance Spectroscopy (FORS) is used to collect higher resolution reflectance spectra in selected areas in the 350 to 2500 nm range, allowing access to the short-wave infrared region, which yields important information for the identification of certain materials such as azurite (easily recognizable by two absorption bands at about 2285 and 2352 nm), lead white, and gypsum. Derivative spectra can be used to distinguish for example between red lead and vermillion, thanks to the difference in the position of the transition edge (~570 vs. ~595 nm, respectively). The interpretation of spectra acquired on mixtures is, however, not always straightforward. The presence of numerous parchment-related absorption bands in the infrared range also makes it harder to separate which spectral features actually relate to the pigment. In both cases, additional data treatment is required. X-ray fluorescence (XRF) analyses are carried out using a Bruker ARTAX Pro instrument equipped with a Rh tube and 75 m capillary optics. A helium flush is used to be able to detect light elements (> Na). Whenever possible, areas for analysis are chosen so that no decoration is present in the corresponding area on the verso of the manuscript leaf, in order to avoid confusing results due to the intrinsic limitation of XRF which does not yield results spatially resolved in depth. It is worth noting that XRF analyses of works of art on parchment are difficult due to the low density of this material, which creates a large amount of inelastic scattering in the spectrum compared with the small amounts of material actually being analyzed in the manuscript. XRF data are interpreted using traditional methods, i.e. the combination of elements identified in each spectrum is compared with possible pigment/mixture
256
compositions. This leads to several possible assignments and thus not always to a unique answer. Comparing these results with those from the spectral methods can help finalize the assignment to the most likely candidate.
Methodology
First, the multispectral reflectance image cube is constructed from 12 reflectance images (400 to 950 nm). Second, the luminescence image cube (500 to 900 nm) is constructed from images acquired while the manuscript is illuminated with blue light (380 to 450 nm) to excite the fluorescence. Next, infrared reflectance images are collected in three spectral bands: 1100 to 1400 nm, 1500 to 1800 nm and 2000 to 2400 nm, and added to the multispectral reflectance cube. The reflectance and luminescence images are processed to correct for pixel and illumination non-uniformity and calibrated to reflectance or emittance using standards. These calibrated images are then spatially registered, i.e. aligned together to remove lens and filter related image shifts. Spatial maps of the primary spectral elements can be made directly by probing the low-resolution reflectance image cube with a reference spectrum. The reference spectrum can come from a spectral library, from a user selected site in the image cube, or as result of principle component based analysis with software designed to find the primary spectral components. The cube is probed with the spectral angle mapper algorithm which finds pixels having similar spectra to the reference spectrum. These maps are produced using all 12 spectral bands as opposed to creating false color maps from just three bands, ensuring more meaningful results. In a similar manner the luminescence spectral cube can be probed and maps of areas having the same emission spectra found. These maps then can be used to guide selection of sites where more definitive in situ analysis can be performed, such as XRF, Raman spectroscopy, or fiber optics reflectance and luminescence spectroscopy. An example of the experimental workflow used during the study of the Christ in majesty with twelve apostles by the workshop of Pacino di Buonaguida, is illustrated in Fig. 2. It should be noted that imaging in the 400-1600 nm range can be easily performed with relatively inexpensive equipment, quite commonly found in Conservation Departments.
Case Studies
In the case of the illuminated manuscript leaves analyzed at NGA, the described analytical methodology yielded distribution maps for several pigments typically used for illuminations in the Middle Ages (Brunello 1975), i.e. regions of azurite, ultramarine, vermillion, and brown earth, along with some indication of pigment layering in the blue and red areas. In two separate cases (Delaney et al. 2009b, Ricciardi et al. 2009), the use of imaging spectroscopy allowed, among other things, to map selected blue
257
fig.2 experimental workflow used to collect and analyze the reflectance images followed by site-specific analysis, as carried out during the study of christ in majesty with twelve apostles by the workshop of pacino di buonaguida (rosenwald collection, accession number 1952.8.277, image courtesy of the national gallery of art, washington dc)
areas not identified by other techniques, where a layer of ultramarine was painted over azurite, a practice which was common in 14th century Italy (Bomford et al. 1989). XRF and FORS analyses of selected sites, chosen following the indications provided by the MSI maps, specifically showed the presence of lead white and mosaic gold (tin sulphide). The presence of organic dyes was identified and mapped through the analysis of the luminescence image cubes. The reflectance and luminescence datacubes therefore allowed mapping the pigment distribution and layering on the illuminated leaves in a totally non-invasive and relatively fast way, not requiring a time-consuming detailed observation under the microscope. They also helped guide the choice of the sites to be analyzed by XRF and FORS for more accurate pigment identification. During the analysis of Birth of John the Baptist by Niccol da Bologna (Ricciardi et al. 2009), infrared reflectograms revealed a few preliminary sketches, either executed with a carbon-based ink on a fine brush, or with a metalpoint. The near infrared images also showed that the skin tones were loosely painted and do not always match the finely detailed preparatory lines.
258
Additional remarks
In order to produce results which are relevant from the point of view of art historical issues, scientific analyses are usually applied to many, if not all, the illuminated manuscript leaves contained in one manuscript. The final goal is a comprehensive characterization of the manuscript, obtained by the identification and mapping of the primary pigments, and by the observation of preparatory sketches. The availability of a relatively quick procedure would allow surveying a large number of leaves, increasing the possibility to obtain significant results and providing a chance for extensive comparisons. The analysis of each miniature, including data acquisition and treatment, following the methodology presented in the previous sections, at the moment requires two days for the MSI and IRR, at least one day each for XRF and FORS. Significant, if preliminary, results on a single illuminated manuscript leaf can thus be obtained in about a week. This is therefore a quick and effective methodology compared to the time (and effort) which would be required to obtain pigment identification and mapping at a comparable level using only site-specific methods, such as Raman spectroscopy and XRF, which would have to be performed on hundreds of spots. In the future starting with the MSI-derived maps improved results could be obtained by combining XRF, FORS, and Raman spectroscopy into a single setup, which would allow acquiring different kinds of information on the same sites, yielding a complete characterization of the materials in situ (e.g. identification of the binding media and organic pigments).
Conclusions
The combination of reflectance and luminescence imaging spectroscopy, FORS, and XRF has proved useful for the identification and mapping of the primary pigments on medieval illuminated manuscript leaves. Imaging spectroscopy, even in only a few bands, has shown its utility for pigment mapping; the imaging was made quantitative thanks to calibration and this allowed reconstructing reflectance and emission spectra. This approach of combining high fidelity site-specific methods (FORS and XRF) with the mapping capability of multispectral reflectance and luminescence imaging spectroscopy appears thus to be a useful tool, providing improved in situ mapping and identification of pigments on illuminated manuscripts in a relatively short time. This methodology can be further refined, for example by improving the correlation between the luminescence and reflectance images in order to improve the characterization of organic dyes and yellow pigments. Finally, the identification of the organic binding media could be attained by adding Raman spectroscopy or mid-IR spectroscopy to the list of analytical techniques.
259
Acknowledgments
The authors gratefully acknowledge financial support from the Andrew W. Mellon Foundation and the Samuel H. Kress Foundation. They also wish to thank their colleagues Michelle Facini, Lisha D. Glinsman, Doug Lachance, Mathieu Thoury and Ren de la Rie for their respective roles during the experimental part of this work.
Bibliography
Bomford, David, Jill Dunkerton, Dillian Gordon, Ashok Roy. 1989. Art in the Making Italian painting before 1400, London: National Gallery Company Ltd. Brunello, Franco. 1975. De Arte Illuminandi e altri trattati sulla tecnica della miniatura medievale. Vicenza: Neri Pozza Editore. Casini, Andrea, Franco Lotti, Marcello Picollo, Lorenzo Stefani, and Ezio Buzzegoli. 1999. Image spectroscopy mapping technique for non-invasive analysis of paintings. Studies in Conservation 44: 39-48. Delaney, John K., Elizabeth Walmsley, Barbara H. Berrie, Colin F. Fletcher. 2005. Multispectral imaging of paintings in the infrared to detect and map blue pigments. In Scientific examination of art modern techniques in conservation and analysis, 120-136. Washington D.C.: The National Academies Press. Delaney, John K., Mathieu Thoury, Michael Palmer, Jason G. Zeibel, Roy T. Littleton, E. Ren de la Rie. 2009a. Visible and infrared imaging spectroscopy of paintings: pigment mapping and improved infrared reflectography. In O3A Optics for art, architecture and archaeology II, Proceedings of SPIE Vol. 7391, ed. Luca Pezzati and Renzo Salimbeni, 739103. d.o.i. 10.1117/12.827493. Delaney, John K., Michelle Facini, Lisha D. Glinsman, Mathieu Thoury, and Paola Ricciardi. 2009b. Application of imaging spectroscopy to the study of illuminated manuscripts. Poster presented at the 37th AIC annual meeting, May 19-22, in Los Angeles, USA (and manuscript in preparation). Delaney, John K., Jason G. Zeibel, Mathieu Thoury, Roy Littleton, Michael Palmer, Kathryn M. Morales, E. Rene de la Rie, Ann Hoenigswald. 2010. Visible and Infrared Imaging Spectroscopy of Picassos Harlequin Musician: Mapping and Identification of Artist Materials in situ. Applied Spectroscopy, 64(6): 584-594. Ricciardi, Paola, John K. Delaney, Lisha D. Glinsman, Mathieu Thoury, Michelle Facini, and E. Ren de la Rie. 2009. Use of visible and infrared reflectance and luminescence imaging spectroscopy to study illuminated manuscripts: pigment identification and
260
visualization of underdrawings. In O3A Optics for art, architecture and archaeology II, Proceedings of SPIE Vol. 7391, ed. Luca Pezzati and Renzo Salimbeni, 739106. d.o.i. 10.1117/12.827415.
Biographies
Paola Ricciardi has been at the National Gallery of Art, Washington DC, since 2008 as the Samuel H. Kress fellow in imaging science. Her research focuses on the application of in situ analytical methods to the study of light-sensitive works of art. She received her undergraduate degree in Physics from the University of Rome La Sapienza in 2003 and her PhD in Conservation Science from the University of Florence in 2008 with a thesis on the application of Raman spectroscopy to the analysis of ancient pottery, porcelain, and glass. Address: DCL, National Gallery of Art, 2000B South Club Drive, Landover, Md. 20785; paola.ricciardi@gmail.com John K. Delaney is the senior imaging scientist at the National Gallery of Art, Washington DC, where his research focuses on the development of in situ imaging methods for art conservation and the understanding of the optical properties of varnishes. He is also a research professor in the School of Engineering and Applied Science at George Washington University in DC. He received his PhD from the Rockefeller University and completed post-doctoral studies at the University of Arizona and the Johns Hopkins University School of Medicine. Address: as for Ricciardi; j-delaney@nga.gov
261
recenses noir. histoire dune couleur
noir. histoire dune couleur

ditions du seuil, 2008
ma ria a del a id e m i r a n da rita ca rva l h o
Instituto de Estudos Medievais e Instituto de Histria da Arte, Faculdade de Cincias Sociais e Humanas, Universidade Nova de Lisboa, Lisboa (Portugal)
michel pastoureau
Noir. Histoire d'une couleur (2008) marca um etapa importante nas publicaes mais recentes sobre a cor. No seguimento de Bleu. Histoire d'une couleur (2002), obra decisiva para a introduo desta temtica na Histria e da Histria da Arte, Michel Pastoureau brinda-nos com mais um volume em que a cor apresentada do ponto de vista do historiador, mas cujo impacto recai sobre variadssimos ramos de conhecimento. Muito embora se trate de um estudo monogrfico sobre o preto, o autor mantm-se fiel aos princpios enunciados nas suas obras pioneiras, das quais no podemos deixar de citar Couleurs, images, symboles (1989), alertando-nos para a necessidade de o estudar num contexto social, cronolgico e simblico, no esquecendo a sua dimenso cientfica, material e tcnica. Confessa-nos igualmente que esta obra se insere num desejo mais vasto de construir uma histria das cores nas sociedades europeias, da Antiguidade romana at ao sc. XVIII. Com efeito, longe de estudar o preto isoladamente, o autor aborda-o enquanto parte integrante de um sistema de cores. A este respeito, afirma (2008, p. 12): Une couleur ne vient pas seule; elle ne prend son sens, elle ne fonctionne pleinement du point de vue social, artistique et symbolique que pour autant quelle est associe ou oppose une ou plusieurs couleurs. Pastoureau chama a ateno para a importncia da conscincia das dificuldades inerentes ao estudo da cor, nomeadamente as alteraes que esta sofre atravs dos tempos e a especificidades dos meios atravs dos quais se transmite. Descodificar o documento, ponderar a procura de significados, no caindo na armadilha de aceitar passivamente a forma de nomear as cores, ser um trabalho rduo mas profcuo para o historiador. Neste sentido, Pastoureau (2008, p.15) d-nos um exemplo bem claro: Croire, par exemple, qu'une porte noir prenant place dans une miniature du XIIIe sicle ou dans un tableau du XVII e reprsente une porte vritable qui a rellement t noir, est la fois naf et anachronique. Cest une erreur de mthode. Dans toute
isbn 9782020490870 39
264
1. Um mordente liga-se simultaneamente fibra txtil e ao corante, conferindo assim resistncia cor quando da lavagem do tecido. O io ferro poderia ser introduzido atravs de compostos como o sulfato de ferro, tambm utilizado para obter tintas de escrever ferroglicas.
image, une porte noire est d'abord noir parce qu'elle s'oppose une autre porte, ou une fentre, voire une autre objet, qui est blanc, rouge ou d'un autre noir. Dotada de grande clareza, esta obra evidencia uma notvel erudio, expressa na riqueza das notas e vasta bibliografia. Revela um profundo conhecimento histrico, particularmente da histria social do vesturio e da pintura, que so sem dvida os grandes eixos nos quais o preto toma lugar na ordem das cores. O carcter transdisciplinar deste estudo est patente na relao que estabelece com grande agilidade, por exemplo, no perodo medieval, entre o bestirio, a herldica e o preto. Tambm relevante a ligao que vai tecendo, ao longo do texto, entre a construo material do preto e a sua carga simblica e social. No caso do vesturio, as dificuldades na obteno de um preto brilhante, homogneo e duradouro deviam-se aos materiais a partir dos quais era produzido cortia, razes de rvores ou partes de plantas ricas em taninos que, mordentadas com o io ferro 1, produziam tons acastanhados ou acinzentados, ou seja, um mau preto. Seria preciso que letrados, burgueses e prncipes adoptassem o preto como cor do vesturio para que se divulgasse o preto obtido base de noz de galha, produto dispendioso, j que o de melhor qualidade era importado da Europa Oriental ou do Norte de frica. No entanto, o mau preto tambm podia ser apangio de virtude e austeridade, caso do hbito dos monges cluniacenses e mendicantes. No que respeita estrutura da obra, o autor optou por uma organizao cronolgico-temtica (ditada pelo prprio percurso do negro no mundo ocidental), desenvolvendo de forma coerente e estimulante aquilo a que se propunha, em quatro perodos. Mergulhamos na perspectiva histrica de longa durao, concepo sempre presente nas suas obras. Na Antiguidade e at ao ano mil, o preto assumia-se como uma cor positiva porque ligada a terra frtil. Inseria-se num sistema tridico em combinao como o branco e o vermelho. Contudo, a sensibilidade judaico-crist aliou, desde cedo, o negro s trevas (a luz condio para a vida), morte e progressivamente ao inferno. esta caracterizao do negro que marcar o perodo seguinte. Com efeito, a cristianizao do mundo ocidental e a oposio ao paganismo e ao mundo islmico, leva a que o negro adquira todas as conotaes malficas. Aplica-se ao demnio, e a todas as formas animais que lhe esto associadas. Esta atitude face ao preto no uniforme. Devemos ter em conta, por exemplo, que o hbito negro dos cluniacenses em oposio ao hbito branco dos cistercienses, antecipa em muito o carcter moral do preto associado pobreza e austeridade que ir dominar o vesturio laico e religioso protestante. atravs da herldica e dos seus cdigos que Michel Pastoureau nos apresenta uma nova ordem das cores, em que o preto se desvincula do seu sentido negativo, fazendo-se representar no leo das armas do conde da Flandres e na guia do imperador do Sacro-imprio. Para a mudana de atitude face ao negro, contribuem ainda, por um lado, o clima austero decorrente da Peste Negra, e por outro, as leis sumpturias que, vedando burguesia cores como o azul e o vermelho, permitiram a escolha do negro para o vesturio de uma classe social abastada. Assim, a partir da 2. metade do sc. XIV os negros
265
produzidos nas tinturarias vo adquirindo uma qualidade superior, j que mercadores e legistas o exigiam. No sc. XV, prncipes como Filipe o Bom vestem-se j de negro. Com a Reforma Protestante o preto, agora uma cor valorizada, contudo excludo da ordem das cores. Uma nova sensibilidade propicia o que Pastoureau designa por cromofobia (sendo as cores associadas seduo, ao luxo e ao pecado), o que ir generalizar o uso do preto, afirmando-o como negao da cor, ou como uma cor parte. A descoberta da imprensa vem inaugurar um mundo a preto e branco, reforando a oposio entre cores e preto. De facto, esta cor, nos scs. XVI e XVII assume um papel histrico na cultura ocidental atravs da sua utilizao sistemtica no texto impresso e na gravura. Valoriza-se nesta poca o papel intelectual do livro, sendo por esta razo desprestigiadas as imagens coloridas to caractersticas do livro iluminado. Toda esta atitude face ao negro confirmada pela descoberta do espectro das cores de Newton que o exclui da ordem das cores ao demonstrar que a luz branca uma mistura de todas as cores (violeta, indigo, azul, verde, amarelo, laranja e vermelho). Na sequncia desta descoberta, o sculo das luzes assiste a um recuo dos negros. O vesturio apresenta cores alegres e luminosas, de tons pastel, principalmente na gama dos azuis, rosas amarelos e cinzentos. O preto perde terreno, vindo apenas a revalorizar-se no final de sculo, no contexto de um exotismo literrio. A difuso e a criao de uma nova paleta de cores vivas justificam esse recuo do negro, que s vai ser valorizado no sc. XIX com o romantismo, que trs consigo o culto da melancolia, noite, da morte e mesmo do macabro. Se valorizado nos meios intelectuais, particularmente na literatura, o negro recebe igualmente a conotao negativa provinda dos sombrios meios operrios da II revoluo industrial, com os quais os outros grupos sociais no se queriam identificar. No entanto, o preto torna-se omnipresente na vida quotidiana, mesmo em ambientes financeiros, numa atitude tica em parte herdada pelos protestantes. A fotografia e o cinema contribuem em muito para o j mencionado mundo em preto e branco, mas so os pintores e designers, que explorando tons e texturas, produzem pretos excepcionais e expressivos, promovendo-o como uma das cores emblemticas da modernidade. O significado desta cor no se esgota no universo das artes: socialmente o preto est presente nas bandeiras das ideologias contemporneas, a cor utilizada pelos totalitarismos mas tambm pelos movimentos de libertao, no deixando de estar igualmente associado a supersties e mesmo por vezes a um bestirio do diabo. Tal como nos primeiros tempos, os mltiplos significados do preto (ou dos pretos) s podem ser entendidos nos seus contextos sociais. No se poder ficar indiferente apresentao grfica do presente volume, cujo design honra o estudo e a cor que apresenta. Tal como num manuscrito medieval, a construo do volume cuidada, criativa e funcional, quer na encadernao escolhida quer na ligao entre o texto e as excelentes imagens. No estaremos face a um objecto produto do design contemporneo, que nos far tomar uma posio afirmativa face interrogao: O preto uma cor?
266
r e c e n s e s m e d i a e va l pa i n t e r s m at e r i a l s a n d t e c h n i q u e s
mediaeval painters materials and techniques: the montpellier liber diversarum arcium
archetype, london, 2011
ste fan os k rousta l l i s
stefanos.kroustallis@gmail.com
mark clarke
isbn 9781804982647 50
Mark Clarke en su libro Mediaeval Painters Materials and Techniques: The Montpellier Liber diversarum arcium ofrece una edicin crtica (con el estudio codicolgico del manuscrito) del tratado medieval de tecnologa artstica Liber diversarum arcium, junto con la primera traduccin al ingls del texto en latn. El autor acompaa la traduccin de las prescripciones con extensos comentarios histricos, tcnicos y artsticos y, adems, realiza un estudio comparativo de estas recetas con otras similares de otros tratados. Los tratados medievales de tecnologa artstica se han convertido en las ltimas dcadas en una importante herramienta de trabajo para varios campos de estudio, como la historia del arte y de la tecnologa o la conservacin, al describir los materiales y tcnicas empleados por los artistas y artesanos en el pasado. Hoy en da se conocen ms de 400 tratados medievales de tecnologa artstica (desde textos extensos hasta recetas singulares), cuya recopilacin ya llev a cabo el autor en su obra (The art of all colours. Medieval recipe books for painters and illuminators, 2001, Archetype). En el presente trabajo el autor se centra en el estudio de uno de estos tratados (o recetarios) el Liber diversarum arcium que forma parte del manuscrito Ms H 277, folios 81v-101v, de la Bibliothque Interuniversitaire, section medecine, de Montpellier (Francia). El tratado es un texto extenso de alrededor de 580 prescripciones tcnicas agrupadas en cuatro secciones (o libros) y dedicadas, principalmente, a las tcnicas pictricas (preparacin y mezcla de pigmentos y aglutinantes, barnices, soportes, dibujo, tmpera, pintura al leo, pintura mural, pintura sobre vidrio y ce-
2 67
rmica, as como otras tcnicas decorativas auxiliares). El Liber diversarum arcium es annimo, data aproximadamente de 1430 y fue escrito en latn, en Venecia o sus alrededores. Mark Clarke en su estudio argumenta, de manera convincente, sobre un ncleo original del tratado que fue escrito cerca de 1300 y refleja prcticas artsticas del norte de Europa. A lo largo del siglo XIV, a este tipo de prescripciones, corregidas y aumentadas, fueron aadidas prcticas de influencia italiana, bajo la mano de un compilador y revisor experto en las tcnicas descritas. En la introduccin del libro Mediaeval Painters Materials and Techniques el autor explica de manera breve y concisa la historia del tratado Liber diversarum arcium, su uso como fuente de tecnologa artstica, su importancia dentro del esquema de la transmisin de los conocimientos tcnicos y su relacin con el resto de los tratados medievales. El resto del libro se organiza en dos captulos, con varias secciones cada uno. En el primer captulo se contextualiza el Liber diversarum arcium dentro de la tratadstica medieval de tecnologa artstica y se resaltan las similitudes textuales con algunos de ellos. En la siguiente seccin se estudia la manera en la que el tratado fue compuesto y compilado, as como la manera en la que sus recetas tcnicas fueron reelaboradas, reincorporando material nuevo y actualizado. En el segundo captulo se exponen las principales caractersticas de la tcnica pictrica coetnea al tratado de Montpellier y se resaltan las analogas tecnolgicas entre ellas (propias del siglo XIV). Adems se hace hincapi en la importancia del Liber diversarum arcium en la historia de la pintura al leo y se argumenta sobre las razones que hacen el tratado tan importante e interesante para el estudio de las tcnicas pictricas medievales. Sigue la traduccin al ingls del texto, jerarquizando las recetas para facilitar las referencias y las llamadas cuando es necesario. A continuacin, el autor ofrece extensos y detallados datos y comentarios tcnicos e histricos sobre los procedimientos descritos que, aparte de su valor documental, sirven como argumentos a las tesis propuestas por el autor anteriormente. Un vocabulario ingls-latn de trminos tcnicos sirve como puente entre la traduccin y la nueva transcripcin del texto (fue transcrito por primera vez por Libri, 1849, en Catalogue gnral des manuscrits des bibliothques publiques des departaments, v. I). El libro se complementa con varios apndices donde se renen el estudio codicolgico del manuscrito, prescripciones y referencias a otros tratados de tecnologa artstica que facilitan la comprensin del texto y, sobre todo, una muy interesante lista de referencias modernas sobre el Liber diversarum arcium y la manera en la que fue considerado y utilizado por la historiografa especializada. Tal como seala Mark Clarke la publicacin del libro Mediaeval Painters Materials and Techniques se debe al hecho de que el Liber diversarum arcium constituye el tratado medieval de tecnologa artstica ms interesante y exhaustivo que sintetiza el estado del arte y la tcnica de la pintura en la Europa del siglo XIV. Adems, la estructura y el contenido del tratado muestra un claro afn prctico, estrechamente relacionado con el aprendizaje de la tcnica de la pintura, propio de un manual de taller. Las prescripciones tcnicas del tratado y su organizacin en la preparacin de pigmentos, dibujo y tcnicas pictricas avalan, precisamente, la propuesta del autor y
268
lo sitan al mismo nivel de inters que el clebre Il libro dellarte de Cennino Cennini. El Liber diversarum arcium permite sacar conclusiones ms generalizadoras sobre la historia y la tcnica de la pintura en la Europa de los siglos XIV-XV. Mencionar slo dos de los casos que Mark Clarke ha sabido resaltar en su estudio: en primer lugar, el tratamiento que se da a la pintura al leo, no como una novedad sino como un conocimiento tecnolgico subyacente que estaba a punto de eclosionar estticamente, al darse cuenta los artistas de las posibilidades que ofreca tal tcnica. En segundo lugar, toda la informacin tcnica que ofrece para los juegos de luces y sombras y la imitacin de texturas, como las famosas telas tornasoladas que aparecen con mucha frecuencia en la pintura de la poca. Otro elemento a destacar es la metodologa de trabajo en la traduccin y transcripcin del texto, que incluyen elementos como la incorporacin en la traduccin de los trminos en latn para los materiales cuando el autor considera que es necesario para facilitar la comprensin y evitar equvocos. De este modo se consigue evitar o mitigar el problema en la traduccin de trminos tcnicos del latn vulgar y en la identificacin de un nombre con una sustancia concreta. Otro punto que el autor ha sabido gestionar muy bien es la transcripcin del texto latn, adaptando los principios de la crtica textual a la realidad de un texto tecnolgico, sin pretensiones literarias y al que no se puede atribuir a un nico arquetipo: el texto de las prescripciones tcnicas se ha cotejado con similares en otros manuscritos y se han sealado los pasajes adicionales o divergentes. Finalmente, me gustara destacar el hecho de incorporar en la publicacin el estudio codicolgico del manuscrito entero y no simples referencias a los folios del tratado concreto, un hecho que permite sacar conclusiones interesantes sobre la transmisin de la tecnologa tcnica e indagar sobre sus autores, su finalidad y su pblico. En definitiva, la publicacin del Mediaeval Painters Materials and Techniques: The Montpellier Liber diversarum arcium de Mark Clarke explora al mximo los posibles enfoques (histricos, artsticos, tcnicos, estticos, lingsticos) de un tratado de tecnologa artstica y estoy seguro que pronto se convertir en una referencia obligatoria a la hora de tratar estos textos como una fuente para la historia del arte y disciplinas afines.
269
va r i a d e s i n t e r ac t i o n s e n t r e s c r i p to r i a p o rt u g a i s au x i i e s i c l e
des interactions entre scriptoria portugais au xii e sicle

rmy co rdo n n i e r
artuslemerle@gmail.com
Introduction
Le corpus des manuscrits de lAviarium de Hugues de Fouilloy est riche de 128 manuscrits dont une soixantaine sont illustrs de tout ou partie du programme iconographique 1. Ces illustrations ont t conues par lauteur lui-mme si lon en juge par le contenu de la ddicace et du prologue du trait 2. Le fait que lauteur du texte soit galement le concepteur des illustrations confre ces dernires un statut particulier. Hugues explique en effet que, dans la mesure o il destine son ouvrage un lectorat peu cultiv, les images ont pour fonction de faciliter la transmission de linformation ceux qui auraient des difficults saisir la subtilit du texte. Cest probablement cette valeur didactique qui a valu au Trait des oiseaux de connatre une telle diffusion travers toute lEurope mdivale 3, y compris au Portugal o ont sont conserves trois prestigieuses copies dont au moins deux furent ralises dans une abbaye lusitanienne 4: Lisboa, Arquivo Nacional Torre do Tombo, ms. Lorvo 5 (anc. 90), Portugal, Saint-Mamede de Lorvo (OSB puis OCist. vers 1200), 1184 (Era 1222 colophon) 5. Porto, Biblioteca Pblica Municipal do Porto, ms. 43, Portugal (Santa-Cruz de Coimbra ?), db. du XIIIe s., Santa-Cruz de Coimbra (Aug.) 6. Lisboa, Biblioteca Nacional de Portugal, ms. Alc 238, vers 1200-1210, SantaMaria dAlcobaa (OCist.) 7.
1. Le texte a t dit et traduit en anglais par Willene B. CLARK, The Medieval Book of Birds: Hugh of Fouilloys Aviarium, Binghamton, Center for Medieval and Early Renaissance Studies, State University of New York (Medieval and Renaissance Texts and Studies, 80), 1992. Ce livre saccompagne dune tude introductive, dun catalogue des manuscrits illustrs recenss par lauteur, et dune liste des autres exemplaires connus lpoque. Signalons galement ldition partir des trois manuscrits concerns par cette tude et accompagne dune traduction en portugais par Maria Isabel REBELO GONCALVES, Livro das aves Hugo de Folieto: Traduo do latim et intraduo por Maria Isabel Rebelo Goncalves, Lisbone, Colibri, 1999. Pour la liste des manuscrits voir Baudouin VAN DEN ABEELE, Trente et un nouveaux manuscrits de lAviarium: regards sur la diffusion de luvre dHugues de Fouilloy, Scriptorium, 57-2, 2003, p. 253-271 (ici p. 264-267). Il faut y ajouter lexemplaire de Sevilla, Biblioteca Colombina, 7-2-21 (XIIIe s.), cat. Y. F. SAEZ GUILLEM, Catalogo de manoscritos de la Biblioteca Colombina de Sevilla, Seville, 202, notice 390, p. 467-468 qui ma t signal par B. Van den Abeele que je remercie chaleureusement; et lexemplaire de Bruxelles,
Ces trois manuscrits, tous illustrs du cycle iconographique complet 8, comptent parmi les plus beaux exemplaires du corpus. Quant celui qui provient de Lorvo, il est le plus ancien Aviarium dat ce jour. Sil y a de fortes chances pour que ces trois manuscrits aient t copis au Portugal, cette question nest pas encore entirement
272
KBR, ms II 2313 (XVe s.) que jai dcouvert lors de mes recherches doctorales, voir Rmy CORDONNIER, Un 128 e exemplaire de lAviarium de Hugues de Fouilloy: Bruxelles, KBR, ms. II 2313, Revista Signum, 2010, 11-1, p. 358-411. 2. Charles DE CLERCQ, Le rle de limage dans un manuscrit mdival (Bodleian, Lyell 71), Gutenberg-Jahrbuch, 37, 1962, p. 2330 et du mme: Hugues de Fouilloy, imagier de ses propres uvres?, Revue du Nord, 177, 1963, p. 31-43. 3. Voir la carte donne par B. VAN DEN ABEELE, 2003 (art. cit. notre note 1), p. 269. 4. Il existe galement une version partielle de lAviarium en ancien portugais ralise au XIVe sicle et actuellement conserve Brasilia: Livro das aves. Reproduo fac-similar do manuscrito do sc. XIV; introduo; leitura crtica; notas e glossrio. Edio preparada por Jacira Andrade MOTA, Rosa Virgnia MATOS, Vera Lcia SAMPAIO e N. ROSSI, s. l., Instituto nacional do livro, Ministrio da educao e cultura (Dicionrio da lngua portugusa. Textos e vocabulrios, 4), 1965. Voir aussi: Serafim DA SILVA NETO, Textos medievais portugueses e seus problemas. Rio de Janeiro, MEC / Casa de Rui Barbosa, 1956, p. 40-45 ; Pedro de AZEVEDO, Uma verso portuguesa de historia natural das aves do sec. XIV, Revista lusitana, 25:1-4, 1925, p. 128-147; Maria Adelaide VALLE CINTRA, Bibliografia de textos medievais portugueses, Lisboa, Publicaes do Centro de estudos filologicos, 1960, p. 68-69 et Rosa Virgnia MATTOS E SILVA, Amrico Venncio LOPES MACHADO FILHO, Fontes para o conhecimento da lngua portuguesa de trezentos: Os mais antigos manuscritos portugueses existentes no Brasil, Srie Estudos Medievais, 2: Fontes, 2009, p. 189-202 5URL: http://www.fclar.unesp.br/poslinpor/gtmedieval/interno.php?secao=fontes). 5. Mrio MARTINS, O Livro das aves, Broteria, 77-5, 1963, p. 413-416, brve description codicologique du mansucrit; Firmino CRESPO et Frade FERNANDO, Anotaes e comentrios sobre o Livro das aves, Geographica, III-9, 1967, p. 20-39, qui dcrivent le manuscrit
clarifie. Les chercheurs qui se sont pench sur la question auparavant nont pas put donner de rponse dfinitive 9. Je ne prtends pas non plus pouvoir le faire ce stade de mes recherches. Mais dans le cadre dun rcent sjour au Portugal 10, jai eu loccasion de travailler directement sur ces manuscrits et dy faire des observations qui me permettent de proposer de nouvelles hypothses de recherche.
Le trait des oiseaux

Le De avibus fut rdig au milieu du XIIe sicle par Hugues de Fouilloy, prieur dune communaut de chanoines rguliers augustiniens, installe entre Amiens et Corbie 11. Ce trait est ddi un certain Rainier, ancien chevalier devenu frre convers, qui le Prieur de Saint-Laurent se propose denseigner les principes de la vie rgulire 12. Les deux principales sources doctrinales du trait sont dailleurs la Rgle bndictine qui donne sa structure thmatique au trait, et le Pastoral de Grgoire le Grand auquel Hugues largement puis. LAviarium est organis de faon binaire afin de reflter les deux aspects de la vie religieuse au Moyen ge: la vie contemplative et la vie active la vie contemplative tant principalement consacre la lectio divina, et la vie active la pratique des bonnes actions et de la prdication. Le choix du symbolisme des oiseaux est motiv dabord par le fait quHugues sadresse des religieux, traditionnellement reprsents par des oiseaux dans la symbolique chrtienne 13. Dans la mesure ou il crit pour des lecteurs qui ne sont pas forms es lettres, Hugues a pris soin dadapter son propos en lenrichissant de figures peintes destines en faciliter la comprhension ceux qui ne sauraient pas se contenter du texte. Son Trait des oiseaux est donc illustr de miniatures qui forment des compositions complexes relies de faon troite pour relayer le message de lauteur auprs des illiterati 14. Le cycle iconographique complet de lAviarium comprend trente miniatures, toutes sont annotes. La miniature de la ddicace, le diagramme de la colombe, le diagramme des trois colombes, le diagramme de lautour, la miniature du palmier, la miniature de la tourterelle et la miniature du cdre et des passereaux, illustrent la premires partie du trait. Ce sont les compositions les plus originales. Les vingttrois autres miniatures illustrent la seconde partie du Trait des oiseaux. Elles sont beaucoup plus simples car elles ne font que reprsenter loiseau en question dans le chapitre quelles accompagnent. Mais elles sont aussi accompagnes dune inscription place soit en incipit du chapitre, soit dans la bordure du cadre de lillustration sil y en a un, soit simplement autour de la miniature. Ces compositions picturales sinscrivent dans la tradition de lexgse visuelle 15. Les diagrammes et les miniatures de la premire partie du trait expliquent la signification allgorique des oiseaux concerns et les figures idalises de la seconde partie ont pour fonction de mettre en vidence la dimension signifiante des autres oiseaux dcrit par Hugues. Limportance du rle des images dans lconomie du sens de lAviarium explique le grand nombre dexemplaires illustrs qui nous sont parvenus, mme si la tradition nest pas exempte derreurs de copies ou de rinterprtations 16.
273
Spcificits iconographiques des exemplaires portugais

Les diagrammes et les miniatures de la premire partie
Les trois exemplaires portugais sont, dans lensemble, assez fidles la principale tradition iconographique du trait 17. Ils prsentent nanmoins quelques variations intressantes signaler pour ltude de la tradition portugaise du Trait des oiseaux. Ainsi, dans laviaire de Lorvo, un arbaltrier et un archer sont reprsents sur la page en regard de celle du diagramme de la colombe (f. 5v.). Ils ne font pas partie du cycle originel du trait. Peut-tre est-ce une vocation du Psaume 90, 3-5: Cest lui-mme qui marrachera au filet de loiseleur et dune parole meurtrire. Il te mettra lombre sous ses paules, et sous ses ailes tu espreras. Sa vrit tenvironnera de son bouclier, et tu nauras pas craindre dune terreur nocturne, dune flche volant le jour. Lassimilation de ces personnages des figures ngatives voir dmoniaques est corrobore par la mutilation apotropaque des visages dont ils ont fait lobjet 18. La thmatique du Psaume 90 est par ailleurs trs proche de celle des premiers chapitres de lAviarium et du diagramme de la colombe dont les inscriptions voquent la recherche de scurit et lespoir du repos ternel. Comme le suggre W. Clark 19, il se peut aussi quil sagisse simplement de figures destines combler lespace vide du feuillet cr par le rejet du diagramme de la colombe sur la pleine page en regard. Lenlumineur sest peut-tre inspir darchers reprsents dans dautres manuscrits du scriptorium dAlcobaa (Clark donne lexemple de Porto, BM, ms 31, f. 207v.) 20. Les miniatures du cycle traditionnel de la premire partie du trait prsentent quelques modifications plus ou moins importantes par rapport au cycle originel. Les termes clericus et miles qui identifient les oiseaux de la miniature du prologue nont pas t reports. Dans le diagramme de la colombe il manque les inscriptions pennas vitutum et volabo desiderio dans les coinons suprieurs, et les mots timor, desiderio, spes et amor autour de loiseau dans le mdaillon central. Dans le diagramme des trois colombes, la colombe du christ est dissocie des deux autres mais les trois oiseaux restent visibles en mme temps car ils sont figurs sur deux pages en vis-vis. Le rubricateur a oubli le mot nigra entre anima et formosa dans linscription de la bordure du mdaillon de la colombe de No. Le diagramme de lautour, sens reprsenter une rose des vents, est ici remplac par un mdaillon contenant la figure de loiseau. Les inscriptions habituellement places sur les bras de la croix servent de rubrique aux chapitres 12, 13 et 14 21. La miniature du palmier est partiellement inscrite dans un mdaillon dans le cadre duquel est reporte la rubrique du chapitre 24 dj inscrite de part et dautre du tronc. La miniature de la tourterelle perd son apparence cruciforme. Le mdaillon central avec loiseau est ici encadr par les rubriques inscrites normalement dans le stipex de la croix. Les inscriptions du patibulus servent quant elles de rubriques aux chapitres 25 et 27. La miniature du cdre et des passereaux est un arbre stylis. Des rinceaux senroulent autour de son tronc. Son feuillage dessine une mandorle losange fond rouge, au centre de laquelle est figur un jeune Christ imberbe au nimbe crucifre 22, revtu dune robe et envelopp
et surtout ses enluminures; voir aussi la notice de Maria Adelaide MIRANDA, dans A iluminura em Portugal, Identidade e Influncias, Lisboa, BN, 1999, p. 190. 6. Johana LENCART, notice dans A. A. NASCIMIENTO, J. F. MEIRINHOS, Catlogo dos cdices da Livraria de Mao Do Mosterio de Santa Cruz de Coimbra na Biblioteca Pblica Municipal do Porto, Porto, Biblioteca Pblica Municipal do Porto, 1997, p. 199-203. 7. Maria Adelaide MIRANDA, notice dans A iluminura em Portugal, Identidade e Influncias, Lisboa, BN, 1999, p. 184. 8. Lexemplaire de Santa-Cruz est hlas lacunaire du premier feuillet, qui devait comprendre la miniature du prologue et le diagramme de la colombe. 9. A Iluminura em Portugal: Identidade e Influncias, catlogo de exposio, orientao cientica Adelaide Miranda, Lisboa, Ministrio da Cultura/Biblioteca Nacional, 1999, p.184191; W. CLARK, 1992 (op. cit. notre note 1), p. 41-50, cat. n. 23, 24, 47; Antonio CRUZ, O Livro das Aves. Um Cdice Ignorado Idntico ao de Lorvo, Revista de Cincias Historicas, 1, 1986, p. 161-174; Maria Adelaide MIRANDA, A Iluminura Romnica em Santa Cruz de Coimbra e Santa Maria de Alcobaa. Lisboa. Faculdade de Cincias Sociais e Humanas, 1986 p.181-184; Mrio S. J. MARTINS, O Livro das Aves, Broteria, 77-5, 1963, p. 413-416. 10. Jai plaisir remercier cette occasion lquipe du projet FFCT/FCT/UNL Color in medieval illuminated manuscripts: between beauty and meaning et tout particulirement Adlade Miranda, pour son chaleureux accueil. 11. Sur Hugues de Fouilloy, en plus de ce quen dit W. CLARK, 1992 (op. cit. notre note 1), p. 5-10, voir notamment: Jean BRIAL, Hugues de Fouilloi, prieur de Saint-Laurent de Heilli, dans Histoire Littraire de la France, XIII, Paris, Victor Palme, 1879, p. 492-507; Henri PELTIER, Hugues de Fouilloy, chanoine rgulier, prieur de Saint-Laurent-au-Bois, Revue du Moyen ge latin, 2, 1946, p. 25-44; Ivan GOBRY, Hugues de Fouilloy, Dictionnaire de spi-
2 74
ritualit, Paris, t. 7, 1969, col. 880-886; Cosimo Damiano FONSECA, Hugues de Fouilloy entre lordo antiquus et lordo novus, Cahiers de civilisation mdivale, 16, 1973, p. 303-312. 12. Jacques BERLIOZ et Rmy CORDONNIER, Le convers et les oiseaux. Monde animal, morale et milieu monastique: le De avibus dHugues de Fouilloy (XII e sicle), dans Lhomme-animal, histoire dun face face, Strasbourg, Adam Biro/Muses de Strasbourg, 2004, p. 72-81. 13. Voir Rmy CORDONNIER, La plume dans lAviarium dHugues de Fouilloy: snfiance(s) dune proprit aviaire, dans F. POMEL (dir.), La corne et la plume dans la littrature mdivale, Rennes, Presses Universitaires de Rennes, 2010, p. 167-202. 14. Willene B. CLARK, The Illustrated Medieval Aviary and Laybrotherhood, Gesta, 21, 1982, p. 63-74; Rmy CORDONNIER, Haec pertica est regula. Texte, image et mise en page dans lAviarium dHugues de Fouilloy, dans B. VAN DEN ABEELE (d.), Bestiaires mdivaux. Nouvelles perspectives sur les manuscrits et les traditions textuelles, Louvain-la-Neuve, Institut dtudes mdivales, 2005, p. 71-110. 15. Sur lexgse visuelle voir notamment: Anna C. ESMEIJER, Divina quaternitas, a preliminary study in the method and application of visual exegesis, Assen, Van Gorcum, 1978; Patrice SICARD, Diagrammes mdivaux et exgse visuelle: le Libellus de formatione arche de Hugues de Saint-Victor, Paris-Turnhout, Brepols (Bibliotheca Victorina 4), 1993; E. MAGNANI et D. RUSSO, Histoire de lart et anthropologie, 3. Exgse textuelle, exgse visuelle, Bulletin du centre dtudes mdivales dAuxerre [En ligne], Histoire de lart & Anthropologie. Sminaires, mis en ligne le 22 octobre 2009 (URL: http://cem.revues.org/index11323.html); Natasha F. H. OHEAR, Contrasting Images of the Book of Revelation in Late Medieval and Early Modern Art: A Case Study in Visual Exegesis, Oxford, Oxford University Press, 2010. 16. Rmy CORDONNIER, Enluminure et spiritualit: le sens du signe et ses difficults, Histoire et images mdivales, Thmatique 15: les arts de la couleur, fvrier 2009, p. 24-29.
dans un pallium. Il bnit de la main droite et de la gauche maintient un livre pos sur sa cuisse. Il est assis sur un arc de cercle tandis quun autre arc lui sert de reposepieds. Six mdaillons fond bleu sont dessins dans le trac du feuillage de part et dautre du Christ et un septime au sommet de la mandorle qui se termine par une palmette fleurdelise. Chaque mdaillon comprend un oiseau. Laviaire de Santa-Cruz est enlumin dans un style trs lgant et color. Le premier feuillet a t hlas dcoup (longlet est encore clairement visible), il devait comprendre nen pas douter le prologue et la ddicace ainsi que la miniature du prologue et le diagramme de la colombe qui sont donc perdu ce jour. Comme dans lexemplaire de Lorvo, la mise en page du diagramme des trois colombes isole la colombe du Christ de celles de No et David. Mais cette fois, lune est sur le recto du feuillet et les deux autres sur le verso et de ce fait lunit visuelle de lensemble est entirement perdue. Le rubricateur a fait le mme oubli que celui de Lorvo: le nigra entre anima et formosa. Il avait en plus inscrit nigra au lieu de nivea dans la bordure du mdaillon de la colombe du Christ, mais cette erreur a t rectifie par le correcteur au Moyen ge. Les autres miniatures de la premire partie du trait sont, iconographiquement et du point de vue de la composition, identiques celles de lexemplaire de Lorvo. Les inscriptions du digramme de lautour et de la miniature de la tourterelle sont, l aussi, devenues les rubriques des chapitres 12, 13 et 14 sur les vents et 25 et 27 sur le palmier. Le rubricateur, qui ntait dcidment pas brillant, a fait une troisime erreur. Dans linscription du mdaillon de la tourterelle il a inscrit in nido pour in nidulo, corrig par le relecteur monastique. Les modifications les plus importantes apparaissent finalement dans la miniature du cdre et des passereaux. Le Christ y est dsormais barbu et moustachu. Il tient un sceptre crucifre au lieu dun livre dans la main gauche. Il nest plus assis sur un arc et ses pieds reposent dsormais sur une extension des rinceaux du tronc. La miniature est aussi plus colore. Lenlumineur a peint en vert lintrieur du tronc, la bordure centrale du losange et la hampe du sceptre, des rehauts de vert servent modeler les plis du pallium. La robe du christ, certaines parties du corps des oiseaux ainsi que les bras de la croix du nimbe et la bordure extrieure du feuillage sont en jaune ple. Les dcorations de la ceinture du Christ et certaines parties des rinceaux du tronc sont en rouge. Le nimbe du Christ est bleu. Enfin, laviaire dAlcobaa est dun style beaucoup plus fruste que ceux de Lorvo et Santa-Cruz et a vraisemblablement t illustr par le scribe lui-mme ou en tout cas par un peintre qui ntait pas artiste de formation. La miniature du prologue y est nanmoins complte. Dans le diagramme de la colombe, les inscriptions pennas vitutum et volabo desiderio dans les coinons suprieurs sont absentes lexception du mot penne inscrit dans lcoinon suprieur gauche. Il ny a pas dinscriptions non plus dans le mdaillon central. Le diagramme des trois colombes reoit la mme mise en page qu Santa-Cruz et reproduit le problme du recto-verso. Le diagramme de lautour est l encore transform en un simple encadrement situ non plus au centre de la page mais aux deux tiers de la colonne de texte de droite. Le rubricateur a plac les inscriptions de la croix en tte des chapitres 12, 13, 14, 15. La miniature du pal-
275
mier est au milieu de la colonne de texte de gauche. Et sa rubrique est aussi porte deux fois: entre les branches du feuillage et sous le tronc en tte du chapitre 24. La miniature de la tourterelle est transforme en un simple mdaillon plac au milieu de la colonne de texte de droite, en regard de la miniature du palmier. Les inscriptions du patibulus sont places dans lentrecolonne et les inscriptions du stipex servent de rubriques aux chapitres 25 et 27. Comme dans laviaire de Lorvo, le cdre est stylis pour devenir une mandorle losange. Outre le style, on note les diffrences suivantes: disparition des arcs sur lesquels le Christ tait assis, ce dernier tient un sceptre crucifre au lieu dun livre dans la main gauche, et on ne compte que six mdaillons au lieu de sept. En revanche, lenlumineur a figur les plaies des pieds du Christ qui reposent sur un escabeau. Le tronc et le feuillage sont teints en vert.
17. Voir R. CORDONNIER, 2005 (art. cit. notre note 14), p. 76-79 et du mme, Lillustration du De avibus de Hugues de Fouilloy: symbolisme animal et mthodes denseignement au Moyen ge, thse en 4 vol., non dite, Universit Charles de Gaulle, Lille, 2007, vol. 1 p. 121-174. 18. Sur le thme du diable oiseleur voir: B. G. KOONCE, Satan the Fowler, Mediaeval Studies, 21, 1959, p. 176-184. 19. W. B. CLARK, 1992 (op. cit. notre note 1), p. 45. 20. Il a galement t suggr que ces personnages pourraient voquer le miles du prologue, voir A. MIRANDA, A . LEMOS, C. MIGUEL, M. J. MELO, On Wings of Blue: The history, material and technique of the Book of Birds in Portuguese scriptoria, dans L. U. ALFONSO (ed.), The Materials of the Image/As Matrias da Imagem, Lisbonne, Universidade de Lisboa, 2010, p. 171-184 (ici p. 175). Toutefois lagressivit de ces archers rend cette interprtation difficilement recevable dans la mesure o le miles du prologue est prsent comme quelquun stant converti la vie religieuse. 21. Daprs la numrotation de ldition de CLARK, 1992 (op. cit. notre note 1). 22. Comme le signale M. A. MIRANDA, 1999 (art. cit. notre note 5), p. 190, on retrouve ce Christ jeune la croix dans lexemplaire de Bordeaux, BM, ms. 995, f. 67v. Elle mentionne aussi lexemplaire de Valenciennes, BM, ms. 101, f. 176, mais si le Christ y est jeune et glabre, en revanche il tient une sphaera mundi la place de la croix. 23. Cette lacune nest pas mentionne par W. B. CLARK, 1992 (op. cit. notre note 1), p. 286287, mais elle est signale par M. I. REBELO GONALVES, 1999 (op. cit. notre note 1), p. 50. 24. W. B. CLARK, 1992 (op. cit. notre note 1), p. 42.
Les portraits doiseaux de la seconde partie

Les portraits doiseaux qui illustrent la seconde partie de laviaire de Hugues ne prsentent pas de diffrences notables dans les exemplaires portugais par rapport la principale tradition du corpus. Il faut mentionner cependant labsence de la figure du hibou dans le manuscrit de Lorvo en raison dun arrachement de la moiti suprieure de la page o se trouvait la miniature 23. Dans les manuscrits de Lorvo et de Santa-Cruz tous les oiseaux sont figurs dans un mdaillon dont la bordure comprend chaque fois la sentence rime qui rsume le trope principal du chapitre. Les deux manuscrits prsentent par ailleurs une certaine parent formelle. Dans lensemble les figures prsentent les caractristiques du groupe dHeiligenkreuz dtermines par W. Clark 24. On retrouve notamment dans ces deux exemplaires lun des petits du plican mystique dessin lhorizontale sur la droite de la composition. Il y a aussi cette manire particulire de figurer lautruche avec son croupion emplum de touffes presque pileuses. Dans le manuscrit de Santa-Cruz lautruche sloigne de lchassier de Lorvo pour sapparenter un rapace palmipde. Lenlumineur de Santa-Cruz la en outre affuble de deux petites oreilles qui caractrisent habituellement les rapaces nocturnes. Dans laviaire de Santa-Cruz lenlumineur a fait dborder les plumes de la queue de la grue sur linscription du mdaillon et a report en bleu dans la marge adjacente le mot que sa peinture avait occult: viuent. Lhirondelle de Santa-Cruz a laile droite releve, ce qui nest pas le cas dans lexemplaire de Lorvo. On peut aussi relever la position particulire de la cigogne dans ces deux manuscrits, o elle est reprsente en train de craqueter, cest-dire la tte renverse en arrire sur le dos et claquant du bec. Dans lexemplaire de Santa-Cruz elle lve en plus la patte gauche. Le hron de Lorvo tourne la tte vers larrire alors que celui de Santa-Cruz regarde devant lui, mais lun et lautre prsentent la mme ondulation du cou. Les caladres sont figurs comme des palmipdes. Dans lexemplaire dAlcobaa seul le corbeau, le coq, le vautour et lhirondelle sont dans un mdaillon. Pour les figures du corbeau et du vautour, le mdaillon a visiblement t rajout aprs dans la mesure o il passe par-dessus le dessin de loiseau. Le coq et lhirondelle posent un autre problme car leur style est nettement diffrent du reste du manuscrit, mais je reviendrai sur ce point plus bas.
276
25. W. B. CLARK, 1992 (op. cit. notre note 1), p. 41-51. 26. Le premier comprend laviaire dHeiligenkreuz plus les exemplaires de Zwettl, StiftsB, ms. 253, v. 1200, Bruxelles, KBR, ms. 8536-43 (Flandres ou France du Nord?, v. 1250, et Oxford, BL, ms. Lyell 71 (Lombardie?, v. 1300). Le second runit nos trois manuscrits portugais plus les exemplaires de Troyes, BM, ms. 711 (fin de XIIe s.), Rome, B. Cansanatense, ms. 444 (v. 1220-1230), et Paris, BNF, ms. 2495 (XIII e s.). Le troisime sous-groupe runit les manuscrits de New Haven, Yale UL, Beinecke ms. 189 (v. 1200), Bordeaux, BM, ms. 995 (2e moit. XIIIe) et Cambrai, BM, ms. 259 (rgion parisienne, v. 1230-1235). 27. B. VAN DEN ABEELE, (ar t. cit. notre note 1), cat. 11, p. 258. Sur ce ms. voir Martine SAINTE-MARIE, Note sur un trait des oiseaux conserv parmi les manuscrits de la Socit archologique de Montpellier, Mmoires de la Socit archologique de Montpellier, 21, 1993, p. 393-401. 28. B. VAN DEN ABEELE, (art. cit. notre note 1), p. 264. 29. M. A. MIRANDA, 1999 (art. cit. notre note 5), p. 190 donne DEN[TUR]. 30. Comme le souligne W. B. CLARK, 1992 (op. cit. notre note 1), p. 286-287, lordre des cahiers a donc vraisemblablement t invers lors dune reliure antrieure et les cahiers contenant le texte dIsidore de Sville sur la cration de lhomme et de la femme devait initialement se trouver avant lAviarium et les extraits de lExameron sur les animaux. Cette hypothse est corrobore par la concidence entre le trou situ cot du genou du Crateur dans la miniature de la cration dve et une tache brune que lon observe peu prs au milieu de la page dincipit de lAviarium. Lensemble constituait une suite logique dcrivant la signification des cratures vivantes en commenant par lhomme et en poursuivant par les oiseaux et les mammifres. 31. Saint Mamde de Lorvo est pass sous plusieurs obdiences depuis sa fondation. Ini-
Relations entre les manuscrits

tat de la question
Selon le classement propos par Clark 25, les trois manuscrits portugais appartiennent au groupe des aviaires illustrs dit dHeiligenkreuz, du nom de lAbbaye autrichienne o est actuellement conserv le manuscrit type du groupe (Heiligenkreuz, StiftsB, ms. 226, France ?, fin du XII e sicle). Les quatorze manuscrits quil comprend seraient issus dun modle dit modle A, qui tait soit loriginal soit trs proche de ce dernier. Du point de vue textuel, laviaire dHeiligenkreuz est lexemplaire qui possde le plus de points communs avec le reste du corpus. Cest le groupe le plus nombreux et qui comprend les plus anciens tmoins de lAviarium, ce qui en fait le meilleur reprsentant potentiel du trait originel. La majorit de ces manuscrits sont dorigine cistercienne, et ils semblent rattachs de prs ou de loin au scriptorium de labbaye de Morimond. Clark distingue trois ensembles au sein de ce corpus 26, mais les diffrences textuelles qui les distinguent sont minimes. Plus rcemment B. Van den Abeele a ajout un manuscrit au groupe, celui de Montpellier, Socit archologique, ms. 8 (fin XIIe - db. XIIIe s.) 27, dont il souligne la parent des illustrations avec les exemplaires portugais en prcisant: il pourrait donc sagir dun chanon entre les manuscrits franais et ceux du Portugal 28. Les aviaria de Lorvo et Santa-Cruz sont trs proches stylistiquement et les chercheurs qui sy sont intress sont tous daccord pour situer leur production au Portugal. Pour le premier la chose est vidente grce aux deux colophons prsent dans le manuscrit: ad honorem Dei et sancti Mametis in Monasterio laurbanense est scribtus [sic] liber iste. In diebus Johannis abbatis FINITO LIBRO DONA DEN[IQUE] 29 LARGIORA magistro. Era M.C.C.XXII (f. 67) et: Scriptus est liber este ad laudem et honorem Dei omnipotentis et sancti Mametis laurbanensis monasterii temporum regis Alfonsi, in diebus Johannis abbatis. Era M.CC.XXI (f. 90v.) 30. Pas de doutes donc sur son lieu de production ni sur la datation qui le situe dans la priode bndictine de labbaye 31. La date est donne selon lre espagnole, dont le point de dpart correspond lapplication de la loi romaine en Espagne en 38 av. J.-C., ce qui donne 1184 pour le premier colophon et 1183 pour le second 32. Or Lorvo, ne devint une abbaye cistercienne de femmes quen 1206, sous linfluence de Dona Teresa, fille du roi Sanche I, et confirme dans ses statuts en 1213. Le scribe a t identifi comme tant Egeas, celui du fameux Beatus de Lorvo 33. Les enlumineurs de lAviarium sont galement les deux mmes qui se sont charg dillustrer le Beatus. Le plus dou des deux tant cependant moins prsent dans lAviarium que dans le Beatus 34. Bien qutant le plus ancien manuscrit dat du corpus, lexemplaire de Lorvo a t copi au moins dix ans aprs la mort dHugues de Fouilloy (survenue probablement vers 1173) et plus de vingt ans aprs la rdaction probable du trait que je situe entre 1130 et 1160. Certes au XIIe sicle les voyages prenaient plus de temps que de nos jours, mais en vingt ans un texte avait bien le temps dtre largement diffus y compris jusquau lointain Portugal pour reprendre les mots de Charles de Clercq 35. Par ailleurs, les modifications apportes au digramme de lautour et la miniature de la
277
tourterelle montrent quand mme une nette divergence par rapport au cycle originel. La suppression de la composition cruciforme de ces deux miniatures les ampute dune grande part de leur signification exgtique. De mme, la sparation de la colombe du Christ de celles de No et David dans le diagramme des trois colombes rend impossible la lecture ascensionnelle de ce dernier et en supprime la dimension anagogique. Contrairement ce que suggrait Ch. de Clercq, lexemplaire de Lorvo tmoigne plutt de ce que vingt ans aprs la rdaction du trait, son cycle iconographique avait dj subit dimportantes modifications et dperditions de sens. En outre, les miniatures ne sont pas les seuls lments du trait avoir subit des modifications. En effet, le texte a aussi t augment par linterpolation de deux passages de lExameron dAmbroise de Milan (V, 21, 74) entre les chapitres 15-16 et 16-17 36. Il est en outre suivit sans transition dune dizaine dautres passages du mme texte, auquel le scribe sest content dajouter une srie de rubriques indiquant de quel animal traite le passage adjacent:
[fin de lAviarium] De infantulo sit, ut oculo tuae mentis eius perfidiam deprehendas, [f. 145v.] et prior cursum verborum eius impedias, in (sic)pudentiam eius et acumen disputationis obtundas. Quod si te ille praevenerit, vocem tibi aufert: et si obmutueris, solve amictum tuum, ut sermonem resolvas. Et si in te insurrexerit lupus, petram cape, et fugit. Petra tua Christus est. Si ad Christum confugias, fugit lupus, nec terrere te poterit. Hanc petram quaesivit Petrus, cum titubaret in fluctibus, et invenit; quia dexteram amplexus est Christi. De leone pardo quod alia odorem refugiat Quid dicam allio homines delectari, et illud ad escam sumere, quod et leopardus fugit ? Denique sicubi allium aliquis confricandum putaverit, leopardus inde exsilit, nec resistit. Cuius venenata fera odorem non potest sustinere, id tu pro cibo sumis, et visceribus infundis internis ? Sed medicatur interdum doloribus. Sumatur pro medicamento, non pro cibo: sumatur [f. 146] ab aegrotantibus, non ab epulantibus. Medicamentum Ieiunium salubre est coporis quaeris, et ieiunium fugis; quasi maius aliud remedium reperire possis. serpens Ieiuni De sputo ieuni hominis moritur hominis sputum si serpens gustaverit,
tialement cr sous une version remanie de la regula communis de saint Fructueux. Labbaye devint progressivement bndictine entre 1064 et 1109, sous linfluence de labb rformateur Eusbe. Voir J. M. GOMES DA SILVA ROCHA, Limage dans le Beatus de Lorvo, thse en 4 vol., non dite, consulte sur place, Universit Libre de Bruxelles, 2008, vol. 1 p. 71-72. (Je remercie chaleureusement Alain Dierkens et Jacqueline Leclercq-Marx pour mavoir signal et prt un exemplaire de cette tude). 32. W. B. CLARK, 1992 (op. cit. notre note 1), n. 2 p. 45. 33. Voir W. NEUSS, O Comentrio do Apocalipse do Lorvao e Suas Iluminuras, Coimbra, 1929; Anne DE EGRY, Um estudo de O Apocalipse do Lorvao e sua relao com as ilustraes medievais do Apocalipse, Lisbonne, 1972. 34. W. B. CLARK, 1992 (op. cit. notre note 1), p. 45. 35. Charles DE CLERCQ, La nature et le sens du De avibus dHugues de Fouilloy, dans A. ZIMMERMANN et R. HOFFMANN (ds.), Methoden in Wissenschaft und Kunst des Mittelalters, Berlin, de Gruyter (Miscellanea Mediaevalia, 7), 1970, p. 279-302, ici p. 302. W. B. CLARK, 1992 (op. cit. notre note 1), n. 1 p. 46, a mal interprt les propos de Charles de Clercq et lui fait dire que laviaire de Lorvo aurait t copi daprs le premier exemplaire. Or de Clercq ne fait quaffirmer lexistence du cycle iconographique dans lexemplaire originel partir du constat de la transmission relativement stable de ce cycle au sein du corpus, y compris dans des exemplaires produits loin de la rgion dorigine du trait. 36. W. B. CLARK, 1992 (op. cit. notre note 1), p. 286 transcrits par M. I. REBELO GONALVES, 1999 (op. cit. notre note 1), p. 76 et 78. 37. Ambroise de Milan, Exameron, VI, 26-29 (PL 14, c. 123-273 - Paris, 1845). Jai rendu le texte comme il apparait sur le manuscrit sauf pour lorganisation sur deux colonnes, choisie ici pour gagner de la place, dans le manuscrit le texte est longues lignes. M. MARTINS, 1963 (art. cit. notre note 5), p. 413; F. CRESPO et F.
Vix infantulo coeperunt dentes prorumpere, et iam novit sua arma tentare. De catulo Nondum catulo dentes, et tamquam habeat, ore proprio se quaerit ulcisci. De cervo Nondum cervo cornua, et tamen fronte praeludit, atque ea quae nondum expertus sit, tela minitatur. [f. 145] Lupus si prior hominem viderit De lupus, vocem eripit, et despicit eum tamquam vocis ablatae victor. Idem si se praevisum senserit, deponit ferociam, non potest currere. De Leo gallum et maxi leo fortitudine Me album veretur. De medicamine capre Capra vulnerata dictamnum petit, et de vulnere excludit sagittas. De remediis Norunt et bestiae remedia sua. bestiarum Leo aeger simiam quaerit, ut devoret; quo possit sanari. Leopardus capreae agrestis sanguinem bibit, et vim languoris evitat. Omnis fera aegra canis hausto curatur sanguine. Ursus aeger formicas devorat. Cervus oleae ramusculos mandit. Ergo ferae norunt ea [ex]petere quae sibi prosint ; tu ignoras, o homo, remedia tua! Tu nescis quomodo virtutem eripias adversario, ut te tamquam praeventus lupus effugere non pos-
278
FRADE, 1967 (art. cit. notre note 5), p. 30 et M. I. REBELO GONALVES, 1999 (art. cit. notre note 1), p. 33 donnent la liste des rubriques ajoutes dans le manuscrit. W. B. CLARK, 1992 (op. cit. notre note 1), n. 2 p. 45, identifiait ces passages des extraits la seconde famille du Bestiaire latin tout en soulignant par ailleurs que cette famille du Bestiaire, dorigine anglaise, nest connue par dautres sources sur le continent qu partir du milieu du XIIIe sicle. 38. A. G. DA ROCHA MADAHIL, Os cdices de Sant Cruz de Coimbra, Boletim de Biblioteca da Universidade de Coimbra, 8, 1927, p. 386-391; A. CRUZ, Santa Cruz de Coimbra na cultura Portuguesa da idade mdia, Porto, Marnus, 1964, p. 130-136; Maria Adelaide MIRANDA, A iluminura romnica em Santa Cruz de Coimbra e Santa Maria de Alcobaa, Lisboa, 1996, p. 409-442. 39. W. B. CLARK, 1992 (op. cit. notre note 1), p. 46; J. LENCART, 1997 (art. cit. notre note 6), p. 199, le situe dans la premire moiti du XIII e s. et (p. 200) mentionne une note au f. 117 avec la date Era 1312, soit 1274, ce qui nous donne un terminus ante quem. Le codex contient en outre un catalogue des manuscrits de labbaye qui dbute avec la mention dun don fait en 1207 (Era 1245) par Pierre Vincent, chanoine de Saint-Vincent de Lisbonne. 40. M. A. MIRANDA, 1999 (art. cit. notre note 5), p. 135-139. 41. W. B. CLARK, 1992 (op. cit. notre note 1), p. 46-47. 42. En grande partie reprises par M. A. MIRANDA, 1999 (op. cit. notre note 5), dans ses notices sur les aviaires de Lorvo et Alcobaa, p. 184 et 190. 43. W. B. CLARK, 1992 (op. cit. notre note 1), p. 47. 44. Hugues de Fouilloy, De avibus. Trait des oiseaux (extraits). Fac-simil du manuscrit 177 de la Mdiathque de lAgglomration troyenne, extraits trad. en fr. par Rmy Cordonnier, Paris, 2004. Ce manuscrit est mentionn en 1472 dans le catalogue de labbaye compos par Pierre de Virey, voir Franoise BIBOLET,
moritur. Vides quanta vis ieiunii sit; ut et sputo suo homo terrenum serpentem interficiat, et merito spiritalem. Quantam Dominus etiam minusculis infudit prudentia[m]! ris Turtur nido suo, ne pul- De nido turtulos suos incurset lupus, squillae folia
superiacit. Novit enim quod huiusmodi folia lupi fugere consuerint. Novit vulpecula quomodo posteritatem foveat suam: et tu ignoras, tu negligis quomodo adversum lupos nequitiae spiritalis posteritatem vitae huius habeas tutiorem? 37
Le scriptorium de Santa-Cruz tait bien actif la fin du XIIe et au dbut du XIIIe sicle 38. Il est donc trs probable que laviaire y ait t produit. Toutefois, bien que leur style soit diffrents, laviaire de Santa-Cruz, nous lavons dit, est iconographiquement trs proche de lexemplaire de Lorvo. Clark a relev de nombreux points de contact entre liconographie des deux exemplaires, mais situe la production de laviaire de Santa-Cruz aprs celui de Lorvo 39. Selon Clark, son style plus nerveux loigne lenlumineur de Santa-Cruz de la production traditionnelle locale et le place sous linfluence de lart roman franais. Le monastre Sainte-Marie dAlcobaa, fond en 1153, est la 53e fille de Clairvaux. Sa construction ne dbute effectivement quen 1178, mais son scriptorium est dj pleinement en activit la fin du XIIe sicle. En tmoigne un ensemble de manuscrits de grande qualit produits cette poque et qui se dmarquent nettement de la production de Lorvo 40. Pourtant, W. Clark, a relev une caractristique stylistique commune certains oiseaux des manuscrits de Lorvo et dAlcobaa, ce qui lui fait dire que lexemplaire dAlcobaa aurait pu avoir t copi Lorvo. Il sagit du motif en forme de soleil radiant (sunburst) que lon trouve sur les genoux du Christ de Lorvo et sur les paules de plusieurs oiseaux du mme manuscrit ainsi que sur les figures du coq et de lhirondelle dAlcobaa. Paralllement, Clark a relev la prsence dune initiale au dbut du premier chapitre de laviaire dAlcobaa, donc le style est trs proche, mais pas identique, de celui dune Bible produite Santa-Cruz (Porto, BP, ms. 31, SC 2) 41. Au vu des ressemblances et des divergences entre les trois exemplaires portugais, W. Clark avait mis des hypothses de localisation sans pouvoir se prononcer avec certitude quant leurs interrelations ni sur le lieu de production de laviaire dAlcobaa 42. Pour elle, si les trois manuscrits ont t raliss dans la rgion de Coimbra, le scriptorium le plus susceptible den avoir possd le modle commun serait Santa-Cruz dans la mesure o il sagit dune maison augustinienne, soit lordre auquel appartenait Hugues de Fouilloy 43. Mais lenqute se complique avec linclusion ncessaire dans le corpus dun quatrime tmoin illustr. Il sagit de lexemplaire actuellement conserv Troyes (Mdiathque municipale, ms. 177), dat lui aussi de la fin du XIIe sicle ou du dbut du XIIIe sicle, et qui a appartenu lAbbaye cistercienne de Clairvaux 44. Les miniatures des aviaires dAlcobaa et de Clairvaux sont virtuellement identiques lexception, l encore, des miniatures du coq et de lhirondelle. Leurs textes sont trs proches galement, bien que celui de Clairvaux soit moins fautif que celui dAlcobaa 45. Clark considre donc comme possible que laviaire dAlcobaa ait t copi
279
daprs celui de Clairvaux, mais les diffrences entre les deux manuscrits ne lui permettaient pas de laffirmer. Linverse est galement envisag par lauteur, qui argue du caractre plus portugais que claravalien de lexemplaire de Clairvaux. Il aurait pu alors avoir t la copie et non le modle. Les deux manuscrits cisterciens peuvent aussi, selon Clark, avoir t raliss Alcobaa daprs un mme modle et lun des deux offert ensuite Clairvaux 46. W. Clark conclue en proposant sans laffirmer une origine portugaise pour les aviaire de Lorvo, Santa-Cruz, Alcobaa et Clairvaux. Tous les quatre auraient alors t copis indpendamment mais partir dun mme modle aujourdhui perdu. Elle ajoute pour finir que nombre de variantes textuelles et certaines des caractristiques iconographiques que lon trouve dans ces manuscrits se retrouvent galement dans deux autres exemplaires illustrs : Laviaire dAvignon (Paris, BNF, ms. lat. 2495, db. XIIIe s., dont lorigine est pour le moment indtermine 47) et celui de Rome (Bibliothque Casanatense, ms. 444, v. 1230-1240 48, ayant peut-tre appartenu labbaye cistercienne de Montier-en-Argonne 49) 50.
Portraits doiseaux illustrant le De Avibus dHugues de Fouilloy (Manuscrit de Clairvaux, Troyes 177), dans Benot CHAUVIN (d.), Mlanges la mmoire du Pre Anselme Dimier, prsents par Benot Chauvin, [II (Histoire cistercienne), vol. IV (abbayes)], Beernem, B Chauvin, 1984, p. 409-447, ici p. 409. 45. W. B. CLARK, 1992 (op. cit. notre note 1), p. 47-48. 46. F. BIBOLET, 1984 (art. cit., notre note 42), p. 411, ne partage pas son avis et considre que le style de laviaire correspond la production claravalienne de la fin du XIIe sicle. 47. W. B. CLARK, 1992 (op. cit. notre note 1), cat. 43, p. 301-302; R. CORDONNIER, 2009 (art. cit. notre note 11), p. 27-29 et du mme Le coq et les cloches dans liconographie, dans Fabienne POMEL (d.), Cloches et Horloges: le Temps au Moyen Age, Rennes, Presses universitaires de Rennes (sous presses): On sait simplement quil fut offert aux Clestins dAvignon au XVe sicle, par le cardinal Jean Allarmet de Brogny (1342-1426) vque dOstie, puis archevque dArles. Ce dernier ayant termin sa formation clricale Avignon, il est probable quil ait fait don du manuscrit aux Clestins de la ville ce moment l. Mais on ne sait pas comment il est entr en possession du livre auparavant. Cest une compilation des traits dHugues contenant outre lAviarium: Le clotre de lme, La mdecine de lme, Les noces spirituelles et charnelles et une Visio cuiusdam monachi, galement attribue Hugues de Fouilloy. 48. Ada DI MORICCA CAPUTI (dir.), Catalogo dei Manoscritti della Biblioteca Casanatense, Roma, Libreria dello Stato, Istituto Poligrafico dello Stato, 1949-1978, vol 5, p. 73-74 ; W. B. CLARK, 1992 (op. cit. notre note 1), cat. 49 p. 306. Vincenzo DE GREGORIO, La Biblioteca Casanatense di Roma, Napoli, Edizioni Scientifiche Italiane (Pubblicazioni dellUniversit degli Studi di Salerno. Sezione di studi Filologici, Letterari e Artistici, 23), 1993. 49. Anne-Marie TURCAN-VERKERK, Les manuscrits de la Charit, Cheminon et Montieren-Argonne. Collections cisterciennes et voies
Nouvelles dcouvertes
De nouvelles donnes codicologiques peuvent dsormais tre ajoutes au dossier, et devraient permettre de faire avancer la problmatique de linterrelation entre les manuscrits concerns. Lorsque jai consult le manuscrit dAlcobaa, jai en effet pu faire des observations intressantes sur les miniatures du coq et de lhirondelle qui posaient problme en raison de leur nette diffrence stylistique avec les autres illustrations du manuscrit. En y regardant de plus prs, jai pu constater que la portion de parchemin qui a reu les deux miniatures tait plus sale et plus veloute que le reste de la page, comme sil avait t gratt. En outre, lun des cercles du mdaillon du coq est nettement estomp comme si lui aussi avait subit une tentative deffacement. Finalement, linspection de ces deux miniatures laide dune lampe de Wood a confirm mon hypothse en rvlant les traces lgres mais bien visibles dun dessin sous-jacent diffrent de celui des miniatures actuelles. Le coq et lhirondelle de laviaire dAlcobaa sont donc des repeints. Pour lhirondelle, le dessin de la tte de lancien oiseau est encore visible droite de laile de la figure actuelle et il correspond au dessin de la miniature de Clairvaux. Pour le coq, le cercle intrieur du mdaillon qui a t gratt dans le manuscrit dAlcobaa correspond celui prsent dans la miniature du coq de Clairvaux. On distingue aussi la trace dune ancienne rubrique gratte sous le mdaillon du coq dAlcobaa, qui pourrait correspondre celle qui encadre le mme mdaillon dans laviaire de Clairvaux. Le dessin sous-jacent de loiseau est en revanche trop effac pour que lon puisse objectivement le comparer celui de Clairvaux. Il me semble donc que lon peut dsormais considrer les aviaires de Clairvaux et dAlcobaa comme des exemplaires jumeaux. Les deux miniatures qui empchaient cette affirmation jusque l ont t ajoutes a posteriori daprs un des deux autres exemplaires portugais, ce qui explique que lon y retrouve certains motifs typiques de ces scriptoria comme les soleils radiants sur les paules. Cette hypothse est par ailleurs corrobore par les nombreuses corrections textuelles qui ont t faite
280
de transmission des textes (IXe - XIXe sicles), Paris, CNRS ditions, 2000, p. 250. 50. W. B. CLARK, 1992 (op. cit. notre note 1), p. 49. 51. W. B. CLARK, 1992 (op. cit. notre note 1), p. 47. 52. Ces observations ont pu tre faites laide dun clairage spciale la lumire rasante (selon un angle de 30 15 degrs) qui rvle nettement les irrgularits du manuscrit et dvoile les trous de pochoir ainsi que les traces de poinon pour la rglure et pour certains tracs de dessins prparatoires. Voir notamment larticle dIns CORREIA, Highlight manuscripts third dimension access, document and display micrometric details. Some examples shown at ms. ANTT, Lorvo 5 De avibus, dans ce mme ouvrage, p. 284. 53. Cette mthode a t observe dans un autre aviaire par W. B. CLARK, 1992 (op. cit. notre note 1), p. 57, 93 voir aussi de la mme: The Aviary-bestiary at the Houghton Library Harvard, dans Willene B. CLARK et Meradith T. MCMUNN (ds.), Beast and birds of the Middle Ages, the bestiary and its legacy, Philadelphie, 1990, p. 26-43; Dorothy MINER, More about Medieval Pouncing, dans Helmut LEHMANN-HAUPT (d.), Homage to a Bookman: Essays on Manuscripts, Books, and Printing Written for Hans P. Kraus on his 60 th Birthday, Berlin, Gebr. Mann, 1967, p. 87-107; Mojmir FRINTA, Punched Decoration on Late Medieval and Miniature Painting, 2 vol., Prague, Maxdorf, 1998-2000.
dans lexemplaire dAlcobaa. Ces corrections sont clairement reprables car lencre dcriture qui a servit les inscrire est beaucoup plus sombre, presque noire, que lencre brune du manuscrit. Or W. Clark, qui mentionne ces corrections, souligne que cette encre noire est typique dAlcobaa et Santa-Cruz 51. On peut se demander pourquoi lon a souhait reprendre seulement deux miniatures du cycle. Pour le coq il semble que la rponse se trouve dans les sentences rimes et rubriques qui accompagnent lesdites miniatures. En effet, dans laviaire de Clairvaux, le rubricateur a inscrit la sentence rime du coq autour du mdaillon: gallus alis se percutiens est doctoris aliis exemplum praebens, et dans la bordure de celui-ci il a reformul une des ides principales du texte courant: Intelligentia galli prudentia magistri. Or dans le manuscrit dAlcobaa, seule la sentence rime est inscrite dans la bordure du mdaillon. Il ny a plus de rubrique autour du mdaillon, il a du y en avoir une mais elle a t visiblement gratte. Pour faciliter son travail le correcteur du manuscrit dAlcobaa a probablement gratt entirement lancienne miniature, son mdaillon et les rubriques pour les remplacer par la composition actuelle. On ne peut hlas pas en dire autant de la miniature de lhirondelle car la rubrique qui encadre loiseau est correcte et les traces de grattage sur le parchemin ne concernent que la figure de loiseau. Il nest pas non plus possible pour le moment de dterminer lequel des aviaires de Lorvo et de Santa-Cruz fut lexemplaire de rfrence du correcteur dAlcobaa, si tant est que ce soit lun de ces deux l. Toutefois, une autre dcouverte codicologique rcente tendrait dsigner lexemplaire de Lorvo. Les recherches menes par lquipe de conservateurs et de chercheurs en charge de la restauration du manuscrit 52 ont en effet montr que le contour de trois miniatures de ce manuscrit a t piquet pour servir de pochoir (le merle, le phnix et le paon) une technique largement utilise au Moyen ge pour faciliter la copie des enluminures 53. Il me semble en outre que laviaire de Lorvo a peut-tre servit de modle celui de Santa-Cruz. Sur certaines des miniatures de ce derniers on distingue en effet un trac sous-jacent lger mais qui est encore visible a et l, notamment entre les pattes de lautour (f. 92) ou la queue de la cigogne (f. 104). Or ce trac ressemble beaucoup celui de la miniature idoine dans lexemplaire de Lorvo. Nous avons donc l une poigne de nouveaux indices qui permettent, sinon dapporter des rponses dfinitives, du moins de rorienter les recherches et de proposer de nouvelles hypothses concernant les interactions entre ces quatre manuscrits. Laviaire de Lorvo a t utilis comme modle, mais est-ce pour celui de Santa Cruz ou pour un tout autre manuscrit ? Les aviaires de Clairvaux et dAlcobaa taient bel et bien des manuscrits jumeaux lorigine, mais ont-ils t copis lun sur lautre ou partir dun modle commun, et tait-ce au Portugal ou en France ? Les aviaires dAlcobaa et de Santa Cruz ont fait lobjet de corrections textuelles et iconographiques, mais est-ce partir de lexemplaire de Lorvo ou dun autre manuscrit ?
281
Conclusion
Si le manuscrit de Lorvo a servit de modle pour celui de Santa-Cruz, lenlumineur de Santa-Cruz sest rappropri le programme iconographique du Trait des oiseaux et en a adapt lexpression son propre style. Il se dtache des influences mozarabes encore fortement prsentes dans lexemplaire de Lorvo 54, pour se tourner vers le nouveau courant stylistique qui fleurit en Europe de louest autour de 1200, marquant le passage du roman au gothique dans lenluminure. Un style qui se caractrise notamment par abandon progressif du formalisme plastique des artistes romans 55 par un assouplissement de la ligne et une multiplication de plis serrs dans les draps. Cela se remarque notamment dans les vtements du Christ de la miniature du cdre, qui abandonne par ailleurs le type lusitanien jeune et imberbe pour un Christ barbu plus franais. Dans les figures animales cette nouvelle expression se traduit par une plus grande attention aux dtails, des figures plus animes, plus mouvementes, plus vivantes 56. On observe galement un abandon progressif de la mise en couleur par aplats couvrants au profit dune technique plus impressionniste, constitue de petites touches multicolores destines rendre le dtail du plumage. Enfin les techniques de dessin voluent. Le fameux motif en soleil radiant utilis Lorvo pour dessiner les paules des oiseaux est nettement moins prsent dans les enluminures de Santa-Cruz dont lartiste semble prfrer un motif de palmette pour dessiner larticulation des ailes. Toutefois, je suis tent de voir deux mains dans la ralisation des miniatures de laviaire de Santa-Cruz en raison du contraste qualitatif que lon observe dans certaines miniatures. Ainsi le diagramme des trois colombes et les miniatures de lautour, du palmier, du cdre et des passereaux, du plican, du nictocorax, du corbeau, du coq, de la grue, de lhirondelle, de la cigogne, du hibou, du geai, de loie, du hron, du phnix, me semblent moins abouties et plus maladroites dans leur dessin et leur mise en couleur que les autres miniatures du manuscrit. Je les attribuerais donc un enlumineur plus proche de lancienne tradition romane locale. Alors que les miniatures de la tourterelle, de lautruche, du vautour, du milan, du merle, du caladre, de la perdrix, de la caille, de la huppe, du cygne, du paon et de laigle, montre une plus grande suret et une plus grande finesse du trac ainsi quun sens plus aige de la couleur 57. Toutefois je dois admettre que la distinction nest pas vidente et que pour certains oiseaux tels que lautour, lautruche ou la caille lattribution lune ou lautre main est loin dtre vidente. Les connections entre les scriptoria de Lorvo et de Santa-Cruz peuvent aussi en partie sexpliquer par leur histoire. Outre une grande proximit gographique, en 1109 labbaye de Lorvo a t confie aux autorits ecclsiastiques de Coimbra par Henri de Bourgogne (1066-1112 comte de Portugal partir de 1093). Elle retrouvera nanmoins son indpendance dix ans plus tard. Il est vident que cette priode daffiliation temporaire a renforc les liens entre les deux maisons et probablement occasionn des changes importants entre les scriptoria. De son cot, Santa-Cruz est un produit conjoint de la Rforme Grgorienne et de la nouvelle monarchie. Sa
54. M. MARTINS, 1963 (art. cit. notre note 5), p. 413; M. A. MIRANDA, 1999 (art. cit. notre note 5), p. 190. 55. Franois AVRIL, Lenluminure lpoque gothique 1200 - 1420, Paris, Famot, 1979, rd. Bibliothque de lImage, 1995, p. 8. 56. M. A. MIRANDA et al., 2010 (art. cit., notre note 20), p. 174. 57. A ce sujet on consultera M. A. MIRANDA et al., 2010 (art. cit., notre note 20).
282
58. J. M. GOMES DA SILVA ROCHA, 2008 (op. cit. notre note 28), vol. 1 p. 73-74. 59. Jos MATTOSO, Portugal Medieval. Novas Interpretaes, Lisboa, INCM, 1992, p. 111. 60. Maria Adelaide MIRANDA, A iluminura Romnica em Portugal, dans A iluminura em Portugal, Identidade e Influncias, Lisboa, BN, 1999, p. 139.
fondation en 1131 concide avec le dplacement de la cour royale Coimbra. La ville devient alors le centre du pouvoir de la nouvelle monarchie et permet Alphonse I daffirmer son autonomie face au pouvoir des puissants lacs du nord 58 et de renforcer sa position en se rapprochant de la zone de conflit avec les musulmans. Dans ce contexte particulier, le choix des chanoines augustiniens tait stratgique en raison de leur rputation drudition, de leur vaste diffusion dans toute lEurope et de la mixit de leurs vux qui leur permet de bnficier de laura de pit propre aux religieux rguliers tout en cultivant le contact avec les populations laques. Ils taient alors bien placs pour diffuser la rforme et appuyer le pouvoir royal nouvellement instaur 59. Toutefois, venant juste de sinstaller dans la rgion, mme sils avaient apport des manuscrits avec eux, les chanoines de Coimbra ont certainement eu besoin denrichir leur fonds initial. La prsence dune grande, riche et ancienne abbaye bndictine non loin de l, faisait de Lorvo le lieu tout dsign o trouver des manuscrits susceptibles daider les religieux de Santa-Cruz de combler les lacunes de leurs armaria. Les diffrences stylistiques sexpliquant alors par linfluence franaise (languedocienne ou aquitaine 60) des chanoines de Coimbra qui ont ralis leur exemplaire et certainement contribu au dveloppement du gothique lusitanien dans lenluminure. En dpit des ces nouvelles dcouvertes, plusieurs questions majeures restent en suspend pour le moment. Elles concernent essentiellement le doublon Clairvaux-Alcobaa, pour lequel je ne suis toujours pas en mesure de trancher au sujet de leur lieu de production. Ces manuscrits proviennent-ils tous-deux du scriptorium de Clairvaux, de celui dAlcobaa ou chacun de leur scriptorium dorigine o ils auraient alors t copis daprs un modle commun? Jespre quune tude comparative plus pousse de ces deux tmoins et notamment de leurs textes permettra de faire avancer la question. Il faudra galement se pencher sur le cas des aviaires dAvignon, de Rome et de Montpellier, afin de dterminer comment ils sintgrent au corpus, et tenter daffiner la situation des aviaires portugais au sein du groupe dHeiligenkreuz.
283
va r i a h i g h l i g h t i n g m a n u s c r i p t s t h i r d d i m e n s i o n
highlighting manuscripts third dimension
access, documentation and display of micrometric details

ins co rreia
DGARQ National Archives of Torre do Tombo and IEM, Faculdade de Cincias Sociais e Humanas UNL, 1069-061 Lisboa (Portugal) ines.ineia@gmail.com
Introduction
Medieval scribes complained of the difficulty and physical effort of the long and hard labour of illuminating or copying manuscripts with the dim light coming through the windows of the monasteries 1. We can make some interesting reflections on the characteristics of that light. It would reach the working table at an extreme angle 2, similar to a raking light, modelled by the orientation of the room and time of day. The position of the scribe and the angle of incidence of the light source are not clearly attested by written sources, but may be found in iconography 3. It is not rare to find the scribe or copyist monk depicted as seated in profile at their working table, sometimes framed by the arches of the cloister 4. When illuminated in the same way, with light from the side as it would have been in a Scriptorium, i.e. more parallel to the parchment (grazing or raking light), a manuscript can expose an extraordinary dimension texture. If this seems redundant, imagine the difficulty in following lines ruled with a punctorium (blind ruling) without such raking light (Figure 1). Such light could also help when painting, where overlapping strokes can be better controlled because of similar lighting. In modern reading rooms the lighting is generally more diffuse and homogeneous, and in consequence the surface texture of a manuscript is not perceived in the same way. Therefore there is increasingly a danger that the evidence of the manufacturing processes and later uses of a medieval manuscript that are detectable within its third dimension (subtle surface deformations, subtle surface textures, such as blind ruling, stylus markings, pricking) will be lost (at worst) or not appreciated (at best). Conservation treatments can remove or reduce this evidence through flattening. Furthermore, increasing reliance on digitised im-
1. Glenisson, J. (dir.). 1988. Le Livre au Moyen ge. Paris: CNRS, 40-41. 2. We are not considering the cases in which the light enters through the roof. 3. Metzeger, B. 1968. When did scribes begin to use writing desks? in Historical and literary studies, Pagan, Jewish and Christian (New Testament Tools and Studies, 8). 2 d. Leyde: BNF, 123-132; III-XIX. 4. Azevedo Santos, M.J. 2001. As condies tcnicas e materiais da cpia de manuscritos na Idade Mdia in Catlogo da Exposio Santa Cruz de Coimbra: A cultura portuguesa aberta Europa na Idade Mdia. Porto: BPMP.
284
fig.1 illuminating the folio with cold light at a sharp angle, highlights the ruling marks of a blind tool, lorvo 5, f. 12
ages of manuscripts by scholars means that these clues can be overlooked (because these features are rarely visible under the types of illumination used for photography and scanning). Within this framework, we propose the revaluation of digital objects created from the transfer programs of medieval manuscripts, in which the gradual increase in resolution does not compensate, yet, for the bi-dimensional perception of the folio. We propose to improve the access-preservation of these manuscripts based on a recovering of optical effects by variable and raking-light as a result of manipulating the orientation and intensity of a proper light source, during direct physical access. We propose the use of raking-light, not as an optional or secondary procedure, but as an essential and routine measure, which simultaneously improve our perception of important details in the manuscript. The illumination of objects from a light source at an oblique angle or almost parallel to the surface reveals significant information on the codex production, and can thus help improve the understanding of the relations between monastic institutions, theological and doctrinal influences, authors receptivity, migration of models or illuminators and several other contextual matters. Some of these relations are revealed in, on or beyond the surviving material of a unique cultural heritage.
From a low tech method to high tech possibility
figs.2 crow in f. 33, left, and, right, detail captured with raking-light (optic fibre). the effect of a deliberate mutilation is recognized in the marks left by a tool tip
5. Sequence of images obtained with a different type, intensity and orientation of light in order to obtain specific information. 6. Clark, W. 1992.The Medieval Book of Birds. Hugh of Fouilloys Aviary. Bringhamtom: Medieval & Renaissance texts & studies. p 41; Clark, W. 1982. The Illustrated Medieval Aviary and the Lay-Brotherhood. Gesta. 21(1): 63-74.
In this context, we deal with a so-called low tech method based on a specific lighting plan, (LP) 5 applied to particular manuscripts. Direct access to the manuscript itself would be justified to assess three-dimensional surface details that are not recorded on digital format. The method proved to be useful for the study of an important medieval manuscript, De Avibus from Lorvo Monastery (Lorvo 5), both in terms of assessing its conservation status, and helping the interpretation of marks left by copy processes, habits from handling or owner censorship (Figure 3b). Dated, by colophon to 1184, it is known (Clark, 1992) as the earliest surviving dated copy of Hugh of Fouilloy text 6. In this fully illustrated manuscript the depiction of the twenty-two
285
birds is very natural; the illuminator combines in his drawings the classical tradition of the representation as nature and as allegorical symbol 7, (Table 1). Its importance in medieval art justifies its presence in numerous thematic exhibitions 8. Five years have elapsed since a conservation treatment, which was carried out because of the risk of structural collapse. Intervention included disassembly of the structure, stabilization and restoration of parchment support and binding. Interdisciplinary projects had been developed since then, revealing much about this Portuguese manuscripts production 9 and suggesting that there is more to be revealed. Simultaneously, the dismantling of the codex, freeing the folios of the structure, has also facilitated the scanning of high resolution (300dpi), which is currently available on-line 10 (Figure 6). We will focus on lighting of type B, which is, as mentioned before, a low tech easy to manipulate source of light. At very sharp angles to the surface of the manuscript, details such as surface defects or distortions can be easily revealed, mapping not only physical changes but also dimensional phenomena like retracting, curling or slight undulations, Table 1-B. At a closer look, we can observe depth of thick paint layers, tool incisions and even subtle scribal corrections 11. Table 1
Manuscript Lorvo 5, observed through the microscope with optic fibre lighting, rakinglight and a standard homogeneous source
7. Miranda, M.A., Lemos A., Claro A., Miguel C. and Melo M.J. 2010. On Wings of Blue: The history, materials and techniques of the Book of Birds in Portuguese scriptoria. in Afonso, L.U. (ed.) The Materials of the Image. As Matrias da Imagem, Lisboa: Campo da Comunicao. 8. Clark, W., Ib idem, p.40. For further information about Portuguese Book of Birds, see Rebelo Gonalves, M.I. 1999. Livro das aves / Hugo de Folieto. 1st edition. Lisboa: Edies Colibri. 9. Projects: 1) An interdisciplinary approach to the study of color in Portuguese manuscript illuminations, POCTI/EAT/33782/2000; 2) The identity of Portuguese medieval manuscript illumination in the European context, PTDC/ EAT/65445/2006; 3) Colour in medieval illuminated manuscripts: between beauty and meaning, PTDC/EAT-EAT/104930/2008. 10. http://digitarq.dgarq.gov.pt, Reference code: PT/TT/MSML/B/5. 11. Michael Douma, Visible & Beyond, from http://www.webexhibits.org/pigments/, public service of the Institute for Dynamic Educational Advancement (IDEA).
A Stereoscopic Light
B Raking Light
C Standard Light
Macro for lapis lazuli, f. 16, Hawk.
Parchment distor tion produced by over tension of sewing, f. 58v-46. System of optic fibre with rheostat, which reveals information about macroscopic texture and planar distortion 1.
General view, displaying text, data on colour, opacity and gloss, f. 54. Standard lighting from scanning technology, optical recording is converted into digital image with variable resolution.
This blue pigment area will be characterized at the molecular level by microRaman and micro FTIR.
1. Michael Douma, Visible & Beyond, from http://www.webexhibits.org/pigments/, public service of the Institute for Dynamic Educational Advancement (IDEA).
286
12. Ins Correia, What Image physical disturbance can tell? (in progress). 13. Articulated optic fibre was used with rheostat supply. 14. Michael Arnott and Ian Bevan, Pouncing. The Aberdeen Bestiary. http://www.abdn.ac.uk/ bestiary/codicology.hti#pounce. 12/04/2003. 15. IDAP, Improved Damage Assessment of Parchment. Assessment, data collection and sharing of knowledge 2007, European Commission, Directorate-General for Research. Final report, 2008.
Figures 3 and 4, show three examples of LP application to Lorvo 5, which became the basis of a current research in image damage and its relation with transmission and censorship 12. Images indicated A were digitized using standard lighting from analogue capture, which was not enough to reveal micrometric details of texture. It also shows some interference coming from verso side writing, which is increased by the high intensity of the lighting. On the other hand, when the manuscript was observed under filtered natural light (35 lx), eyes easily adapted to this low levels of light intensity and, at the same time, more was revealed on the texture surface with an additional raking light source 13. The first example, Figure 3b and Figure 4b, shows the possibility of surface assessment by reducing the translucent character of parchment, revealing even subtle creases. If light direction is changed during examination we can understand the causes of mechanical damage such as shrinkage or local distortion. This example is also significant because it enabled detection of the fine holes along the bird contour, just as in the Aberdeen Bestiary, where such holes can be found on several of the images, and which have been largely studied as pouncing, one of most popular copying techniques. 14 To better document the impact of the pricking tool, the verso side was documented (Figure 3b) as well as the next folio, f.62, where no holes were found. During the last conservation treatment (2005), while backlighting was used to help parchment infill, all the images with this kind of contour prick were documented on the conservation report, but it was never contextualised as an artistic or historic issue. The studies on this manuscript did not
fig.3: a), c) details acquired with a standard homogeneous light source (>800 lx) for the Phoenix and the Partridge, respectively; b), d) texture revealed by raking-light at 300 (50 lx)
287
Fig.4: a) detail acquired with a standard homogeneous light source (>800 lx) for the Ostrich (f. 40); c), d) texture revealed by raking-light at 300, with 35 and 50 lx, respectively; b) stilus incisions, in f. 61v, revealed by raking-light at 150, 50 lx
proceed to a systematic texture examination. Looking at the next examples it is easy to agree on the importance of the details unveiled. The Ostrich (Figure 4) seen under raking light at 200 reveals two distinct, previously unnoticed, types of physical damage: a deep incision along the contour made with some sharp tool, which is clearly seen under 50 lux and a deliberate scratching on the image. A blind trace, not so abrasive, is documented for the Partridge, f.62v, but this time, out of bird contour, suggesting two different drawing options. The indications detected on the surface of Lorvo 5 thus can increase our knowledge on its reception and transmission. Medieval manuscripts are full of significant texture maybe due to the light angle, with which they were written and illuminated. Raking lighting is shown to be a shortcut to bring this past to light.
Conclusion - Texture conservation and display

Parchment is particularly sensitive to relative humidity and temperature changes. Even a small variation induces changes in its dimensions such as curling and waving. Conservation procedures, such as cleaning, even with a small amount of moisture, may increase the stiffness, accelerate chemical reactions and produce irreversible shrinkage, even at room temperature. Also, flattening and pressure with heavy load on moist parchment may cause considerable change of colour and transparency. 15 These treatments have a direct impact on appearance: decreasing surface texture they may decrease or remove the evidence related to the manufacturing processes and
288
later uses of the codex. Reducing the level of information should always be avoided, but the risk is higher if information at the micro level is not even acknowledged. This evidence may be disclosed with raking-light, applied with a low light intensity to avoid photodegradation. We propose that this lighting may bring us closer to what might have been the medieval conditions of its manufacture and use. It is concluded that, for an accurate recording of medieval manuscripts, namely to assess the conservation status, raking-light examination is an indispensable tool. As far as we know, digital scanning is not carried out systematically under raking light; this could be easily achieved, in order to preserve and highlight manuscripts third dimension. Selected manuscripts, with relevant micrometric details noticed by conservators or researchers, should be also digitized with raking light. One of those manuscripts should be, undoubtedly, The Book of Birds from Lorvo.
Acknowledgements
The National Science Foundation, FCT-MCTES, is gratefully acknowledged for the PhD grant SFRH/BD/44192/2008 and for funding the project PTDC/EAT-EAT/104930/2008. The author would like to thank the Director of Torre do Tombo, Silvestre Lacerda, and to Mark Clarke, Maria Joo Melo and Adelaide Miranda for helpful suggestions; the author is also grateful to Mark Clarke for his generous assistance in the text editing.
Bibliography
Azevedo Santos, M.J. 1998. A Ars Scribendi: Textos e Imagens. Coimbra: Universidade de Coimbra (Humanitas). Vol L. Batori, A. 2003. Innovation in Preserving and Conserving Book Heritage. Rome: Instituto Centrale di Patologia del Libro. Brownrigg, L. 1995. Making the Medieval Book: Techniques of Production. Oxford: The Red Gull Press. Driscoll, M.J. and Ragnheiour, M. (editors) 2008. Care and Conservation of Manuscripts 11: Proceedings of the Eleventh International Seminar Held at the University of Copenhagen 24th-25th April 2008 (Paperback). Conpenhagen: Museum Tusculanum Press, University of Copenhagen. Gilissen, L. 1977. Prolgomnes la codicologie Recherches sur la construction des cahiers et la mise en page des manuscrits mdivaux. Gand: ditions Scientifiques StoryScientia S.P.R.L. Miranda, M.A. 2001. A produo do livro: do monge ao arteso. A iluminura e o iluminador no contexto de produo do cdice. Coimbra: Cmara Municipal: INATEL: ADDAC. Rebelo Gonalves, M.I. 1999. Livro das Aves. Lisboa: Edies Colibri.
289
v a r i a v o lt a d e u m v e r m e l h o
apresentao de edio d o livro de como se fazem as cores , sob o olhar da cincia e tecnologia
ma ria j oo m e lo e c ata r i n a m i g u e l
Departamento de Conservao e Restauro and Requimte, Faculdade de Cincias e Tecnologia, FCT, Universidade Nova de Lisboa, Lisboa (Portugal)
volta de um vermelho
ma ria a del a id e m i r a n da
Instituto de Estudos Medievais and Instituto de Histria da Arte, Faculdade de Cincias Sociais e Humanas, FCSH, Universidade Nova de Lisboa, Lisboa (Portugal)
O processo de preparao da sntese de um vermelho medieval que a seguir se apresenta o resultado de muitos anos de investigao em torno de um dos mais interessantes tratados tcnicos medievais, O livro de como se fazem as cores 1. Para este contriburam, com inmeras tentativas de reproduo da receita, os muitos alunos de Histria e Tcnicas de Produo Artstica do Mestrado em Conservao e Restauro da Universidade Nova de Lisboa, e mais recentemente, Catarina Miguel no mbito do seu projecto de doutoramento, Le vert et le rouge. A presente edio distancia-se em muitos aspectos de outras excelentes deste tratado, pois deseja actualizar prticas e materiais vindos de um passado longnquo. O que significa reproduzir, com sucesso, a receita descrita pelo praticante medieval. Se numa receita culinria, o sucesso medido aferindo se o produto final esteticamente apelativo e saboroso, no nosso caso, o sucesso medido pela beleza (e durabilidade) da cor obtida. O desafio intelectual foi grande e podemos afirmar que este um trabalho em aberto e novas descobertas na histria e arqueologia do passado nacional permitiro maior rigor na reconstruo destes pigmentos e corantes medievais. Assim, optamos por uma edio on-line 2, de modo a permitir uma rpida partilha e discusso com outros investigadores e curiosos; o que, esperamos, levar a uma actualizao eficiente. Centramo-nos neste nmero especial no caso do vermelho 3, talvez a reproduo mais complexa de todo o tratado, para que o leitor participe na histria do vermelho dos vermelhos, o mais utilizado na iluminura medieval, HgS.
1. A edio crtica pioneira foi a de Blondheim, a que neste momento se junta a mais recente tese de Devon Strolovich. A datao deste tratado tem intrigado estudiosos tendo sido reavaliada recentemente por um grupo de investigadores portugueses, em que se destaca a participao de Lus Afonso, Ivo Cruz e Antnio Joo Cruz. Nas palavras de Lus Afonso, O livro de como se fazem as cores das tintas, translated into English as The book on how to make the colours of the paints, is a Portuguese technical text written during the Late Middle Ages dealing with the preparation of artists materials. It is composed by forty-five chapters of different lengths, most of them concerning materials and techniques to be used in the illumination of manuscripts. (...) The text on colours is written in Portuguese language but using Hebraic script, making it an example of the practice of Iberian aljamia writing, that is the use of Arabic or Hebrew script to write a text in an Iberian romance language.
290
Para mais informao consultar a bibliografia abaixo elencada. Blondheim, S.1928. An old Portuguese work on manuscript illumination. Jewish Quarterly Review. 19: 97-135. Strolovitch, D. 2005. Old Portuguese in Hebrew Script: convention, contact, and convivncia. PhD Dissertation. Cornell University. 116-184; Cruz, A.J. and Afonso, L.U. 2008. On the Date and Contents of a Portuguese Medieval Technical Book on Illumination: O livro de como se fazem as cores. The Medieval History Journal. 11: 1-28. Strolovitch, D.L. 2010. Old Portuguese in Hebrew script: beyond O livro de como se fazem as cores, in -Afonso, L.U. (ed.) The Materials of the Image. As Matrias da Imagem, Lisboa: Campo da Comunicao, 29-43. Afonso, L.U. 2010. New developments in the study of O livro de como se fazen as cores das tintas, in L. U. Afonso (ed.) The Materials of The Image: As Matrias da Imagem, Srie Monogrfica Alberto Benveniste 3. Volume. Lisboa: Ctedra de Estudos Sefarditas Alberto Benveniste da Universidade de Lisboa, 3-27. Castro, I. 2010. Notas sobre a lngua do Livro de como se fazen as cores (ms. Parma1959) in Afonso, L.U. (ed.) The Materials of the Image. As Matrias da Imagem, Lisboa: Campo da Comunicao, p.87. 2. www.dcr.fct.unl.pt/LivComoFazemCores, e a seleccionar arquivo digital. 3. O vermelho um sulfureto de mercrio de frmula qumica, HgS.
Chapter15.Tomakevermilion
Figure 1. Main steps in the reproduction of the process of making vermilion.
"To make vermilion, take five pounds of quicksilver, that is mercury, and place it in a bottle or large glazed bowl, and take a pound of very fine virgin sulphur. And pour the powdered sulphur over the quicksilver little by little until it is well incorporated, always stirring it with a dog's foot that has its hair and wool, until the fire turns to ashes. [10r.] And once the fire has thus died down, place it in two new pots that are made like bottles, broad below and narrow above. And seal them, leaving only a small hole through which the vapor will escape. And place the pots on the fire on their holders and cover them well with clay, and place a bowl over the holes. And when you see the smoke coming out red and not malodorous, place a thin spit in it. And if anything sticks to the spit, remove the pots from the fire and let it cool. And once it is cool break the pots and you will find the vermilion made. With these measures you ill made as much vermilion as you wish: for a tera of mercury take five pounds of sulphur, and for five pounds of mercury take one pound of sulphur. And regulate the first in such as way that it does not [10v.] burn, and keep the fire moderate, neither quick nor slow. On this note, if by chance the vermilion burns, break the pots and grind it and incorporate it and mix it with another measure of mercury and sulphur, and place them in other pots and proceed as described. And pay attention to the vapours that escape, thus you will never ruin anything." [1]
The proportion present in the treatise was followed: 1.047g of mercury and 0.1674g of sulphur were weighed and ground in an agate mortar with a pestle. To improve the incorporation of both compounds, sulphur was slowly added and mixed with mercury, until all were bound and a silver-black-greyish compound was formed. What is observed may be described as in the treatise "until the fire turns to ashes" [1]. This step takes time and patience. Black mercury sulphide, was then transferred into the base of the clay crucible, which was covered and sealed with fresh clay. Experiments placing the pot directly into the fire were carried out. When needed, small amounts of water were dropped over the necessary areas to avoid flames or lowering the temperature. For maintaining the embers, combustion air was introduced with the aid of a wooden air blower. In a successful experiment, after two and a half hours, the pot was taken from the fire and cooled to room temperature. Afterwards, the olha was opened and vermillion was found inside its base.
Rationalization / Chemical reactions

In the first step, mercury and sulphur are ground to produce metacinnabar, a silver-black compound with a cubic crystal structure, which is the kinetic product of this reaction [2, 3, 4], being thermodynamically stable only for high temperatures, above about 370C. The thermodynamic stable form at room temperature is the hexagonal mercury sulphide (vermilion). For more details please see [2-11].
Reproduction
The olhas (clay container) were designed, handmade using white or red clays, and fired as necessary to obtain a ceramic pot. In this reproduction, we used a white crucible made of two parts: a base where metacinnabar is heated and a cover.
291
Figure 2. Reaction scheme for vermilion synthesis.
In a second step, metacinnabar is heated at 350370C, and will rearrange into the hexagonal form that corresponds to the red product, vermilion. It is important to stress that, for our experimental conditions, the production of red mercury sulphide vermilion is a solid-state reaction and not a sublimation process [12, 13]. When sublimation occurs, for temperatures higher than 580C, a black product, not yet characterized, is formed. This product can be a mixture of meta and hipercinnabar [3].
temperature control over reaction yield, sand baths were used and temperature was measured over time in the sand. For our experimental conditions, starting with about 0.5 g of black metacinnabar, we found that introducing the pot in the sand bath heat at 285C for 2h30 and afterwards rising the temperature (heating rate of 15C/min) until circa 350C for 2h30 produced the best results.
Missing / Obscure indications

Pot design: There is no precise information about the shape of the olha. Strolovitch in his translation refers to it as a vessel like bottles, broad below and narrow above", [1]. Heating temperature and time: the only information given is place the pots on the fire and "keep the fire moderate, neither quick nor slow", [1]. Usually, embers temperature fall between 620670C. We do not know if the pots should be placed directly on the fire neither for how long (a couple of hours? all day?). Where was this fire made? In a special apparatus or just on the ground? How was it maintained and controlled? We infer, from the lack of information, that the experiment time was to be expected more in the 1-2h range than in the 3h-6h. Heating time and red smoke: The information present in the treatise says and when you see the red smoke coming out red and not malodorous, place a spit in it. If anything sticks to the spit, remove the pots from the fire and let them cool [1]. However, no red smoke was observed in any of the many experiments conducted. The only smoke observed was black, probably the result of metacinnabar sublimation.
Key aspects
Reaction stoichiometry: one mole of sulphur reacts with one mole of mercury. Mixing mercury with sulphur to produce black mercury sulphide may be achieved by a thorough grinding, heating or using amalgam. In our reproductions we tested both thorough grinding and amalgam formed by heating sulphur: both worked in the same way, although in the text grinding is described and with very specific details. Temperature at which black metacinnabar is converted into red vermilion is "the crucial" parameter. To transform the black mercury sulphide form (-HgS) into the red form ( -HgS), it is necessary to avoid temperatures above about 400-450C. On the other hand, the higher the temperature the more efficient will be the solid state reaction that enables the conversion of the cubic black form into the red one. To test the influence of
Figure 3. Examples of vermilion in Portuguese medieval illuminations. From left to right: Santa Cruz 20, f.137v: pure vermilion paints for the red and a mixture with an organic dye for the dark red; Alcobaa 419, f.91v: pure vermilion paint; De Avibus (Lorvo 5), f.25: vermilion is found in mixture with the extender calcium carbonate; Lorvo Apocalypse, f.210 with red lead.
Comments
heating temperature: The reference of red smoke as a signal for the complete transformation of metacinnabar into cinnabar, prompt us to considerer that a sublimation process could be present. In fact this smoke colour was never observed. Only a black smoke was seen due to the overheating of the bottle (olha). heating time: Although there is no specific information about the heating time, it is clear that this reaction takes hours and not days. pot design: The olhas base thickness is essential on the inset temperature control, namely on allowing a controlled heating rate and on maintaining constant temperature inside the olha, undoubtedly the two most determinant parameters of this recipe. the dog's foot: "always stirring it with a dog's foot that has its hair and wool until the fire turns to ashes" is possibly one of the most beautiful instructions found in a medieval treatise. Those who have tried to mix sulphur with mercury know how "fugitivas" the small drops of mercury may turn to be. The use of a fluffy surface could help in capturing Hg, facilitating the grinding and reaction of big amounts of Hg and S, as those described in the text. Also, the description "until the fire turns to ashes" not only conveys a beautiful image but it also depicts accurately what is observed; indeed, during the grinding, the yellow sulphur and the bright mercury are transformed into a greyish colour that may be described as "ash colour".
Vermilion in Portuguese medieval illuminations

Vermilion red is an important colour in Portuguese Medieval manuscripts; it was used both to paint the rubricae and in the illuminations, displaying a very good conservation condition. Vermilion as a proteinaceous paint was applied as a pure pigment or mixed with red lead or/and calcium carbonate and white lead. The later compounds were added as extenders as they did not affect the final colour. In the Lorvo collection, we found pure vermilion in the lettering and mixtures of vermilion with variable percentages of red lead (from 5% up to 40% wt) or other additives in big size illuminations. To produce dark reds, it was mixed with an organic dye, as found in Santa Cruz 20.
Works cited
[1] - Strolovitch, D. 2005. Old Portuguese in Hebrew Script: convention, contact, and convivncia, PhD Dissertation, Cornell University, 116-184. [2] - Dickson, F.W. and Tunnel, G. 1959. The stability relations of cinnabar and metacinnabar. American Mineralogist. 44: 471-487. [3] - Pattrick, R.A.D., Bell, A.M.T. and Vaughan, D.J. 2010. Structural evolution of aqueous mercury sulphide precipitates: energy-dispersive X-ray diffraction studies. Mineralogical Magazine. 74: 8596. [4] - Sharma, R.C. and Chang, Y.A. 1993. The HgS (Mercury-Sulfur) System. Journal of Phase Equilibria. 14:100-109.
292
[5] - Melo, M.J. and Miguel, C. 2010. The making of vermilion in medieval Europe - Historically accurate reconstructions from The book on how to make colours, in Kroustallis, S. and Del Egido, M. (eds.), Fatto DArchimia: history and identification of artificial pigments, Madrid: IPCE. [6] - Miguel, C; Miranda, A.; Lopes, J.A.; Melo, M.J. and Clarke, M. 2011. A Study in Scarlet vermilion red and colour paint formulations in medieval illumination, in Preprints of the XVI ICOM-CC Triennial Conference, Lisbon. [7] - Miguel, C., Clarke, M., Melo, M.J., Miranda, A. and Oliveira, M. The book on how one makes colours of all shades in order to illuminate books Revisited, in Technology and Interpretation Reflecting the Artist's Process. London: Archetype. [8] - Charnock, J.M., Moyes, L.N., Pattrick, R.A.D., Mosselmans, J.F.W., Vaughan, D.J., and Livens, F.R. 2003. The structural evolution of mercury sulfide precipitate: an XAS and XRD study. American Mineralogist. 88: 1197-1203. [9] - Munir, Z.A., Kashkooli, I.Y. and Street, G.B. 1973. Sublimation of IIB-VIA compounds. V. Relative thermal stability and heat of transformation of blackmercury sulfide (metacinnabar). High Temperature Science. 5: 8-15. [10] - Potter II, R.W., and Barnes, H. L. 1978. Phase relations in the binary Hg-S. American Mineralogist. 63: 1143-1152. [11] - Rodic, D., Spasojevic, V., Bajorek, A. and Onnerud, P. 1996. Similarity of structure properties of Hg1-xMnxS and Cd1-xMnxS (structure properties of Hg-MnS and Cd-MnS). Journal of Magnetism and Magnetic Materials. 152: 159-164. [12] - Gettens, R., Chase, M., and Feller, R. L. May 1972. Vermillion and Cinnabar. Studies in Conservation. Vol. 17, 2: 45-69. [13] - Rinse, J. 1928. The vapour pressure, dissociation, and transition point of mercury sulphide. Recueil des Travaux Chimiques des Pays-Bas. 47: 28:32.
Moreira de S, A. 1960. Revista da Faculdade de Letras, 4: 210-223. Cennini, C. 1960. The Craftsman's Handbook The Italian Il libro dellarte, translated by D.V. Thomson. New York: Dover Publications. Theophilus. 1979. On Divers Arts. Smith, C.S. and Hawthorne, J.G. (eds). New York: Dover Publications. Plinio. 1985. Histoire Naturelle, Livre XXXIII. Paris: Belles Lettres. Merrifield, M.P. 1999. Medieval and Renaissance Treatises on the Art of Painting: original texts with English translations. London: Dover Publications.
Other Garcia-Moreno, R. and Thomas, N. 2008. Cinnabar or vermilion?, in Art Technology- Sources and Methods. London: Archetype.141-143. Bruquetas, R. 2010. El bermelln de Almadn: de Plinio a Goya, in Kroustallis, S. and Del Egido, M. (eds.), Fatto DArchimia: history and identification of artificial pigments, Madrid: IPCE. Afonso, L.U. 2010. New developments in the study of O livro de como se fazem as cores das tintas, in Afonso, L.U. (ed.) The Materials of the Image. As Matrias da Imagem. Lisboa: Campo da Comunicao. 3-27
Appendix Vermilion characterization. Produced following "The book on how to make colours"
Colour Table 1. Colour coordinates, Lab*, for vermilion paint reconstructions using two different binders (arabic gum and parchment glue) applied over filter paper and parchment. Support Filter paper Parchment Binder Parchment glue Arabic gum Parchment glue Arabic gum L 49.72 49.64 46.64 46.56 a* 36.97 38.30 36.18 37.24 b* 19.96 22.68 26.30 27.77
Spectroscopic characterization
Further reading
Written Sources Blondheim, S. 1928. An old Portuguese work on manuscript illumination. Jewish Quarterly Review. 19: 97-135.
a.u.
XRD diffractogram acquired with a Philips XPert diffractometer using monochromatised CuK radiation.
15 25 35 45 55 65
2 teta ()
Raman spectrum acquired with a Labram 300 Jobin Yvon spectrometer with exc= 632.8nm; characteristic bands @ 253, 285 -1 and 343 cm
EDXRF spectrum acquired through ArtTAX spectrometer, with a molybdenum anode.
293
va r i a i n t e r a c t i v e t e c h n o l o g y to e x p l o r e m e d i e va l i l l u m i n at i o n s
interactive technology to explore medieval illuminations

a ndr rica rd o, n u n o cor r e i a , ta rq u n i o mota
Centro de Informtica e Tecnologias de Informao, Faculdade de Cincias e Tecnologia, Universidade Nova de Lisboa
rita ca rva l h o
Instituto de Estudos Medievais, Faculdade de Cincias e Sociais e Humanas, Universidade Nova de Lisboa
Introduction
The main goal of the installation is to allow wide dissemination and awareness about color in Portuguese medieval illuminations, the theme of the project within which these components are being developed 1. This theme, as well as the historical context surrounding it, was already the object of some dissemination to non-scientific audiences, namely in several workshops where illuminations and color paints were produced using techniques similar to the original ones. The installation builds on this previous work. The design and development process was carried out by a multidisciplinary team from areas including computer engineering, design, illustration, art history, history, chemistry and conservation sciences, with several iterations on the content and technology. It also works as a trial for interaction experiments regarding innovative computational interfaces and how they can be developed and used in a cultural heritage setting. The installation has three components that can work independently or as a full installation. The following sections describe each of these components and how they contribute to the overall experience.
Virtual Scriptorium
This component aims to provide an individual experience of producing an illumination, using a digital platform. It works as an extension of a real workshop on medieval
294
fig.1 e 1a image creation. workshop of medieval illumination
fig.2 paint production. workshop of medieval illumination
fig.3 image creation. virtual scriptorium
fig.4 paint production. virtual scriptorium
illumination which takes place regularly at the University, attracting participants with several ages and backgrounds (fig. 1). Their challenge is to reproduce an image from the Book of Birds or Apocalypse of Lorvo, both romanesque manuscripts from the Monastery of Lorvo. The session includes the production of paints, based on medieval recipes (fig. 2). Similarly to the workshop, this digital component aims to support the construction of an illumination from the Book of Birds. Using a tablet PC and a pen instead of parchment and brushes, the user is guided through several steps in order to construct an illumination (fig. 3). Details such as the brush thickness, or the running down of the ink/paint in the pen brings a certain manual feeling to the application. As in the real workshop, users can produce their own paints (fig. 4), through actions like grinding carefully a lapis-lazuli stone in order to have a proper pigment (as in the color blue), or mixing it with the binder. The application contains also information concerning each illumination from the Book of Birds as well as the corresponding texts. The goal is not to create an experience equivalent to a real workshop (differences, such the ones regarding touch and smells are obvious), but to provide a new approach to the production of an illumination, with its own specificities. Being easy to set up, this application can be placed in cultural sites with broad audiences such as museums, public libraries and archives, helping to divulge the ancient art of illuminations through a playful and engaging experience.
295
fig.5 initial diagram. interactive panel
fig.6 color diagram. interactive panel
Interactive Panel
The interactive panel provides an overview of color from a scientific, social, artistic and historical point of view. Special relevance is also given to the illuminated codex production process and to its historical context. Users may explore several items such as colors, scriptorium, codex or Romanesque Portugal, visible in a playful initial panel (inspired by medieval genealogical diagrams) (fig. 5). Some of those items will now be described. Besides a brief scientific explanation of color, the item colors (fig. 6) explores the way paints of different colors were made in the Middle Ages. It also shows their applicability at that time and their possible social meaning(s). The item codex also plays an important part, showing the enormous importance of the book at that time (in spite of its limited access), and displaying how the page layout of an illuminated manuscript was carefully designed regarding its different functions. It will display virtual versions of the codices Book of Birds and Apocalypse of Lorvo that the user may flip as in a real codex. Since the monks were the agents of codex production at that time, the item monastery provides an important overview of the site where the workshops or scriptoria were placed, showing also a glimpse of a monks life at that time. In scriptorium the user can explore the instruments and materials used by the copyist or by the illuminator during the process of production of the codex, and also acknowledge the different tasks involved. Romanesque Portugal (fig. 7) is an item in which the user can find concise information about the context of the Iberian Peninsula between the second half of the 12th century and the first of the 13rd century, namely the multicultural Portuguese society, its organization, and the art that was produced.
296
fig.7 e 7a Romanesque Portugal. interactive panel
2 97
The panel contains also a gallery of existing illuminations that were previously done with the tablet. All these items contain images largely based on illuminations analyzed within the research project. This panel has the advantage of allowing access by multiple users at the same time and of establishing a connection with the tablet, relating practical and theoretical knowledge.
Augmented Book
The augmented book is the most innovative component of the installation in terms of interface device. In opposition to the other two components, it provides a physical experience of the book while the digital content augments this experience. All of the electronics are hidden within the covers of the book, to make the book as similar to a real codex as possible (fig. 8). While manipulating a real codex (a reproduction of a Portuguese medieval one) containing sensors that detect its orientation, a model of that same book is displayed, accompanied with information relating to its binding, that is, to the actions and elements that turn a gathering of folios into a book as a functional object. For instance, as the user turns the book, facing its spine, visual and written information will be provided about the way quires were sewed. Information concerning the parchment and some relevant elements of the folios are also accessible through actions like opening the book or turning pages. The book can detect when specific pages are open, enabling the display of virtual content related to that page. Within each page, it is also possible to navigate through the content performing swipe gestures with the hand in front of the book. All of
fig.8 e 8a testing prototype for the augmented book
298
these actions can be performed whether the book is standing on the table or is being held by the user.
Conclusions and future work

Prototype versions for the three components were developed, and testing with good results was achieved, specially the Virtual Scriptorium component. The development process is iterative and the different members of the team contribute at each iteration in the content, technology and desired user experience. Current and future work includes content development to add additional contextual information in the Interactive Panel and information about the book as an object in the Augmented Book component. Better integration of the three components regarding the software is also a relevant issue for future work. The ultimate goal is to deploy the complete installation in a museum, library or similar institution and for this preliminary contacts have been carried out. Extended user tests will be conducted to assess the experience that is proposed.
299
va r i a m a st e r t h e s i s
studies in medieval manuscript illuminations

master and phd thesis from faculty of social sciences and humanities and faculty of sciences and technology, new university of lisbon (unl)
Master thesis (or graduate)

CLARO, Ana Caracterizao dos materiais e estudo de conservao de um manuscrito iluminado, Apocalipse do Lorvo (1189). Caparica: FCT/UNL, 2004. MOURA, Laura Caracterizao dos materiais e estudo de conservao da folha de rosto do Foral manuelino de Vila Flor, (1512). Caparica: FCT/UNL, 2004. SOUSA, Lus Iconografia Musical Medieval. Lisboa: FCSH/UNL, 2004. LEMOS, Ana Um Novo olhar sobre o livro de Horas de D. Duarte. Lisboa: FCSH/UNL, 2009. CUSTDIO, Delmira A Luz da grisalha. Arte, Liturgia e Histria no Livro de Horas dito de D. Leonor IL 165 da BNP. Lisboa: FCSH/UNL, 2011.
Master thesis ongoing

ARAJO, Ana Rita Os Livros de Horas (sc. XV) na coleco do Palcio Nacional de Mafra: estudo e conservao. Caparica: FCT/UNL, ongoing
304
va r i a p h d t h e s i s a n d p o s d o c p ro j e c t s
PhD thesis
CLARO, Ana An interdisciplinary approach to the study of colour in Portuguese manuscript illuminations. Caparica: FCT/UNL, 2009. SOUSA, Lus Speculum Musicae. Iconografia musical na arte do final da Idade Mdia em Portugal. Lisboa: FCSH/UNL, 2011.
PhD thesis ongoing

CASTRO, Rita The Book of Birds in Portuguese scriptorium: preservation and assess. Caparica: FCT/UNL, ongoing. CORREIA, Ins Estudo Arqueolgico dos Cdices Iluminados do Fundo Laurbanense (Scs. XII - XIII) A recepo dos manuscritos no reflexo de intervenes passadas, presentes e futuras. Lisboa: FCSH/UNL, ongoing. CUSTDIO, Delmira Relaes artsticas entre Portugal e a Flandres atravs dos Livros de Horas. Lisboa: FCSH/UNL, ongoing. LEMOS, Ana Catlogo dos livros de horas iluminados de origem francesa nas coleces pblicas portuguesas (primeira metade do sculo XV). Lisboa: FCSH/UNL, ongoing. MIGUEL, Catarina Le vert et le rouge: A study on the materials, techniques and meaning of the green and red colours in medieval Portuguese illuminations. Caparica: FCT/UNL, ongoing.
Pos-Doc projects ongoing

BILLOTA, Maria Alessandra Juridical manuscripts production and illumination in Portugal between 14th and 15th centuries and theirs connections with manuscripts production and illumination in the French Midi (specially Toulouse, Avignon and Montpellier) and in the North-Mediterranean regions (Italy and Catalua). Lisboa: IEM, FCSH/UNL, ongoing. MIGULEZ CAVERO, Alicia Anlisis del lenguaje gestual en el Apocalipsis de Lorvo y su comparacin con otros beatos y obras artsticas romnicas. Lisboa: IEM, FCSH/UNL, ongoing. CORDONNIER, Rmy The bestiary of Portuguese medieval art and though. From the expression of a multicultural aesthetic to the elaboration of a specific symbolism (12th-13th c.). Lisboa: IEM, FCSH/UNL, submitted.
305
va r i a n ot c i a s
notcia
Colour In Medieval Illuminated Manuscripts: Between Beauty and Meaning
In this project, funded by FCT-MCTES, we explore the issues related with the symbolic and social meaning of colour in medieval Portuguese illuminations, produced during the twelfth and first quarter of the thirteenth century in Alcobaa, Lorvo and St Cruz monasteries. Colour use and production in Portuguese medieval illuminations was a consequence of the technology available as well as of cultural and artistic options; by defining the specificities of its use and production we contribute to fingerprint the influences of the three different cultures that coexisted in Portugal at that time, Arab, Jewish and Christian. We approach this subject from an art history and molecular sciences point of view, aiming to characterize the monastery scriptoria and their evolution during the twelfth and first quarter of the thirteen centuries. We started by quantifying the dominant colours and its combinations, in national manuscript collections; namely, in the manuscript collections of Alcobaa, Arouca, Lorvo and St. Cruz. We will afterwards proceed to the comparison with other international collections. Quantification is performed by computer image analysis of colour areas. As degradation processes affect our perception of colour, examination at the molecular level is essential
in order to avoid misinterpretations of the colour meaning and distribution. The binding media, the invisible component of a paint colour, may also have a fundamental influence on colour perception as they play a key role in the colour changes over time. Particular attention is devoted to their complete characterization by using non-conventional techniques recently applied in the field of cultural heritage, such as ELISA- antigen-antibody assay. We also explore new ways of sharing our results with the general public, with the main focus on children, contributing to divulge the ancient art of illuminations through the exploration of modern computer interaction technologies. We intend to achieve an engaging, intuitive and easy to use interactive system, where physical objects similar to the ones used in real life, when making the illuminations, will be used. This installation will simulate the illuminations creation process in the medieval period, addressing several aspects from the materials origin and production methods to the painting process. It will also show users the historical and social context of that epoch and reveal the meanings of the used colours and images depicted. This interactive installation can be easily deployed at cultural sites, enriching an exhibition, or at institutions where the books are preserved. Finally, a book, describing the main findings and breakthroughs of our research is being prepared. The tasks and objectives proposed within this project will promote innovation that will contribute to a better access and conservation of medieval illuminations .
300
The team is constituted by 19 members, including five PhD students and five research fellows. The art history research is coordinated by Adelaide Miranda, the molecular characterization by Maria Joo Melo, the image analysis by Joo
Lopes, and the multimedia interfaces by Nuno Correia. Consultant Mark Clarke contributes with his expertise to the overall project as well as Augusto Aires do Nascimento. Fruitful partnerships are maintained with the BNP, for the
Alcobaa nucleus), BPMP, for the Santa Cruz collection and DGARQ-ANTT, for the Lorvo manuscripts).
Maria Joo Melo

DCR e Requimte, FCT-UNL
team, from left to right: andr ricardo, eduardo dias, teresa romo, tarqunio mota, rmy cordonnier, nuno correia, ana matias, conceio casanova, teresa serra, mark clarke (consultant), maria joo melo, adelaide miranda, rita carvalho, ins correia, catarina duarte, catarina miguel, joo lopes and rita castro. team members not present in this photo: ana lemos, mafalda sagarrua and solange muralha. from citi and di/fct-unl; ibet-unl; requimte and dcr/fct-unl; iem and dha/fcsh-unl; iceta-ff-up; iict and dcr/fct-unl; vicarte-fct-unl
301
O projecto IMAGO
O projecto Imago 1, uma base de dados sobre iconografia medieval (abrangendo dois campos de estudo: o da iluminura e o da escultura), visa colmatar a inexistncia, em Portugal, de um importante utenslio de pesquisa para os estudiosos da imagem, semelhana dos existentes nos restantes pases da Europa. Foi financiado pela FCT-MCTES, atravs do projecto POCTI/EAT/45922/2002. Sediada na pgina da Faculdade de Cincias Sociais e Humanas da Universidade Nova de Lisboa e integrada no Instituto de Estudos Medievais e Instituto de Histria da Arte, a base de dados sobre iconografia medieval conta j com algu-
mas fichas no campo da iluminura e da escultura, elaboradas por um grupo de investigadores, um bolseiro do projecto (Joana Rama/escultura) e dois colaboradores (Ana Lemos e Lus Sousa/iluminura) sob a orientao dos Professores Doutores Jos Custdio Vieira da Silva e Maria Adelaide Miranda, coordenadores do projecto. Permitindo aos investigadores o acesso a dados fundamentais no campo da investigao, tem por objectivo contribuir para a construo de um centro de iconografia medieval portugus.
1. ver artigo de Joana Rama na revista n. 7, 2009 do IHA, p. 284.
Para sua consulta ir at http://imago. fcsh.unl.pt
Ana Lemos
Instituto de Estudos Medievais
302
The awakening of the Manueline Foral charters

The awakening of the Manueline Foral charters: science and technology insights into the masterpiece is a three year project (2011-2014) financed by Fundao para a Cincia e a Tecnologia carried on by a multidisciplinary team including experts from Centro de Fsica Atmica da Universidade de Lisboa (CFAUL), Instituto dos Museus e Conservao laboratrio Jos de Figueiredo (IMC) and Arquivo Nacional da Torre do Tombo (ANTT). A Foral charter was a royal document in Portugal and its former empire, whose purpose was to regulate the juridical relation between a master, such as the king, and a collective entity, such as a village. In the beginning of the 16th century, King D. Manuel I of Portugal promoted a large restructuration of the historical written memory of the realm. Within this restructure, Foral charters produced since the 12th century were copied into renewed codices in the contemporary language and adjusted to the present community rules. This was one of the most ambitious works carried out by D. Manuel for whom the esthetic of the charters was as important as the message they carried in. In this context they were written on parchment in gothic style characters and illuminated with precious ornaments such as the coat of arms and the armillary sphere used to express the authority of the realm. These Foral charters became known as Manueline foral charters and represent
the most important written heritage of the Portuguese Art in the Renaissance. With this project, a systematic analytical (physical and chemical), paleographic and codicological study of Manueline foral charters will be performed by applying methodologies that will enable the identification of materials (pigments, dyes, fillers, binders, inks, parchment) and methods used by the scribes and artisans. The major outbreak is the thorough analytical study of around fifty charters by non-destructive techniques either in situ (energy dispersive X-ray fluorescence analysis, diffuse reflectance, optical microscopy) or by micro-sampling (X-ray diffraction, Raman and Fourier transform infrared analysis). This way, the already fragile pieces, wont have to leave the controlled environment of the archive where they are kept nowadays. Transcription and high resolution digital color photographic documentation of all the foral charters will be made, to enable online access, and preventing further
damages to this heritage. Whenever necessary, the foral charters, whose state of deterioration would not recommend their manipulation, will be restored beforehand. The project will invite other research teams working on related subjects to participate in workshops where results on the Manueline foral charters will be also divulged. Information regarding the project will be available in http:// manuelin.cii.fc.ul.pt/ForaisManuelinos. Coordination team: Ana Isabel Seruya (CFAUL) and Maria Lusa Carvalho (CFAUL) Other team members: Marta Manso (CFAUL), Sofia Pessanha (CFAUL), Agns Le Gac (DCR/FCT-UNL and CFAUL), Mrio Costa (CFAUL), Isabel Cabao (CFAUL), Marcel Besnard (CFAUL), Antnio Candeias (IMC), Filipa Roldo, Jos Manuel Garcia (CML), Teresa Figueiredo (ANTT), Snia Domingos (ANTT), Ariana Webb (ANTT).
Maria Lusa Carvalho

Centro de Fsica Atmica (UL)
fig. charter of sintra, 1514
303
va r i a s e m i n r i o e e x p o s i o
os livros de horas do palcio nacional de mafra e a cultura artstica do sculo xv

Seminrio e exposio
No mbito da tese de doutoramento em histria da arte intitulada Catlogo dos livros de horas iluminados de origem francesa nas coleces pblicas portuguesas (primeira metade do sculo xv). Anlise estilstica e iconogrfica, desenvolvemos o estudo de um conjunto de 8 manuscritos conservados na biblioteca do Palcio Nacional de Mafra. Resulta este seminrio das questes surgidas da problematizao sobre os manuscritos em termos artsticos, culturais e sociais bem como da ponderao de cada um deles enquanto obra de arte singular. Pretende-se, assim, ao juntar alguns dos especialistas nas diferentes reas do saber, apreender a cultura artstica portuguesa do sculo xv de forma a tentar perceber, do conjunto de manuscritos deste perodo existentes em Portugal, os que foram adquiridos na poca quatrocentista e quinhentista, por razes devocionais e de gosto, e os que resultam de uma aquisio posterior por coleccionadores portugueses. Importa esclarecer o que at data se tem vindo a dizer sobre o assunto, isto , o facto de no nosso pas se ter dado preferncia a livros de horas oriundos da escola flamenga, e determinar, se possvel, a data em que os manuscritos chegaram at ns. Desenvolvendo, por outro lado, o estudo das relaes com os manuscritos iluminados da poca, numa tentativa de determinar a sua datao mais rigorosa, bem como a caracterizao estilstica da arte dos iluminadores, visa este seminrio um debate sobre toda a problemtica envolvente. E porque o estudo deste ncleo de manuscritos no seria possvel sem a colaborao de alguns dos especialistas da rea, gostaria de aqui deixar uma palavra de reconhecimento aos que no podero estar presentes, Roger Wieck (Pierpont Morgan Library, Nova Iorque), Lieve Watteeuw (Centro Illuminare, Universidade de Louvaina, Bruxelas), Michel Pastoureau (EHESS, Paris) e Patricia Stirnemann (IRHT/CNRS, Paris).
organizao
instituto de estudos medievais palcio nacional de mafra
comisso institucional
mrio pereira (pnm) teresa amaral (pnm)
comisso cientfica
aires augusto do nascimento (fl-ul) maria adelaide miranda (iem/fcsh-unl) maria joo melo (dcr/fct-unl)
curadoria da exposio
ana lemos (iem/fcsh-unl)
comisso executiva
ana lemos (iem/fcsh-unl) ricardo naito (iem/fcsh-unl) rita carvalho (iem/fcsh-unl) teresa amaral (pnm)
informaes
frequncia no seminrio mediante inscrio limite de inscritos: 50 participantes perodo de inscrio: de 1 a 24 de novembro
propina do seminrio
estudantes | 10 euros outros | 20 euros
informaes adicionais / inscries

instituto de estudos medievais faculdade de cincias sociais e humanas av. de berna 26-c, 1069-061 - lisboa 217908300 coloquios.projectos@fcsh.unl.pt http://iem.fcsh.unl.pt
Ana Lemos
IEM/FCSH-UNL
306
va r i a s e m i n r i o e e x p o s i o
25 Nnovembro
Seminrio Os livros de horas do Palcio Nacional de Mafra e a cultura artstica do sculo xv
10h00 Abertura Mrio Pereira (Palcio Nacional de Mafra), Maria Adelaide Miranda (IEM/FCSHUNL) e Ana Lemos (IEM/FCSH-UNL) 10h30 Problemtica dos livros de horas: As particularidades da coleco de Mafra Aires Augusto do Nascimento (Academia das Cincias e CEC/FLUL) e Ana Lemos (IEM/FCSH-UNL) 11h00 - Pausa 11h20 O uso do livro em conventos e mosteiros femininos portugueses visto atravs das marcas de posse Fernanda Maria Guedes de Campos (BNP) 11h40 A biblioteca do Palcio Nacional de Mafra - os catlogos de 1755 e 1819 Teresa Amaral (Palcio Nacional de Mafra) 12h00 Os livros de horas e a devoo dos leigos na baixa idade mdia: novas questes Maria de Lurdes Rosa (IEM/FCSH-UNL) 12h20 A iluminura quatrocentista e a noo de espao pictrico segundo Pierre Francastel. Uma problemtica histrica revisitada Carlo Moura (IHA/FCSH-UNL) 12h40 - pausa para almoo (livre) 14h00 La dvotion aux saints dans les livres dheures: lexemple de Paris au XVme sicle Claudia Rabel (Institut de Recherche et dHistoire des Texts - CNRS) 14h20 A pintura mural portuguesa do sculo XV Lus Urbano Afonso (FLUL) 14h40 A iluminura hebraica portuguesa do sculo XV Lus Urbano Afonso (FLUL) e Dbora Matos (FLUL) 15h00 A herldica portuguesa do sculo XV: propostas para o seu estudo Miguel Metelo de Seixas (CHAM-IEM/ FCSH-UNL) e Joo Bernardo Galvo-Telles (CLEGH-ULL) 15h20 A reencadernao de manuscritos caractersticas e significados Ins Correia (IEM/FCSH-UNL) 15h40 Willem Vrelant/Juan de Carron: elaborao de um percurso atravs da comparao de modelos para o livro de horas dito de D. Leonor da BNP Delmira Custdio (IEM/FCSH-UNL) 16h00 - pausa 16h20 Os livros de horas iluminados do Palcio Nacional de Mafra Ana Lemos (IEM/FCSH-UNL) 16h40 O que nos dizem os materiais da cor sobre os livros de horas do PNM? Maria Joo Melo (DCR/FCT-UNL), Ana Rita Arajo (DCR/FCT-UNL) e Ana Lemos (IEM/FCSH-UNL) 17h00 Debate e concluses
26 novembro
Exposio Os livros de horas iluminados do Palcio Nacional de Mafra Inaugurao pelas 16h00 (Entrada livre)
3 07

RHA - Medieval Colours

Hochgeladen von

Dokumentinformationen

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

RHA - Medieval Colours

Hochgeladen von

Copyright:

Verfügbare Formate

Medieval Colours

between beauty and meaning

Entrevista com Michel Pastoureau

Les couleurs de lenluminure: recettes de Michelino da Besozzo et dAntoine de Compigne

propos dune notice sur le vermillon

Binding media in medieval manuscript illumination: a source research

Colours versus colorants in art history: evaluating lost manuscript yellows

The colour of medieval Portuguese illumination: an interdisciplinary approach

Illuminations: secrets, alchemy and conservation in three case studies

Analysis of the Anjou Bible

A protocol for non-invasive analysis of miniature paintings

The application of SR-XRF to the analysis of manuscript illumination a case study

Michel Pastoureau. Noir. Histoire dune couleur. ditions du Seuil, 2008

VARIA Des interactions entre scriptoria Portugais au XIIe sicle

Interactive technology to explore medieval illuminations

NOTCIAS Colour in Medieval Illuminated Manuscripts: Between Beauty and Meaning

The awakening of the Manueline Foral charters

Studies in medieval manuscript illuminations: Master and PhD thesis

Maria Adelaide Miranda, FCSH-UNL

Maria Joo Melo, FCT-UNL

Mark Clarke, UvA

1. with my researches in the history of colors, I seem to please myself greatly.

* Biografia adaptada da Wikipedia

conduzida por MARIA ADELAIDE MIRANDA ANA LEMOS

1. From Material to Medium

colour and meaning

2. Colours in medieval book illumination: Colour systems and attempts at interpretation

colour and meaning

4. Colours in the Codex Aureus of Echternach

colour and meaning

colour and meaning

colour and meaning

colour and meaning

colour and meaning

colour and meaning

1. Munich, Bayerische Staatsbibliothek Clm. 4453, towards 997.

colour and meaning

colour and meaning

colour and meaning

colour and meaning

colour and meaning

colour and meaning

fig.3 saint louis psalter, paris, bnf, ms.10525, fl. 9r.

fig.4 isabella psalter, cambridge, fitzwilliam museum, ms.300, fl. 1v

colour and meaning

byza n t i n e re ci pe s a n d b ook illum ination

byzantine recipes and book illumination

The Postbyzantine tradition

byza n t i n e re ci pe s a n d b ook illum ination

Byzantine texts on manuscript production

byza n t i n e re ci pe s a n d b ook illum ination

7. Oxford, Bodleian Library, ms. Canon. gr. 39.

Recipes and workshop practice

Chrysography and gilding

byza n t i n e re ci pe s a n d b ook illum ination

10. Fuchs, Oltrogge (forthcoming).

byza n t i n e re ci pe s a n d b ook illum ination

byza n t i n e re ci pe s a n d b ook illum ination

t he st ra s b ourg fami ly t e x t s : ori gi n al i t y and survival

the strasbourg family texts: originality and survival

The Strasbourg Manuscript and its technical tradition

The Strasbourg family database

t he st ra s b ourg fami ly t e x t s : ori gi n al i t y and survival

DB_Manuscripts Key manuscript Key edition1