Leonardo's Colour and Chiaroscuro Author(S): John Shearman Source: Zeitschrift Für Kunstgeschichte, 25

Home , Andrea del Sarto, Chiaroscuro, Florentine painting, Grisaille, Masaccio, Sfumato

Leonardo's Colour and Chiaroscuro Author(s): John Shearman Source: Zeitschrift für Kunstgeschichte, 25. Bd., H. 1 (1962), pp. 13-47 Published by: Deutscher Kunstverlag GmbH Munchen Berlin Stable URL: http://www.jstor.org/stable/1481484 . Accessed: 27/02/2014 13:46

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp

. JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Deutscher Kunstverlag GmbH Munchen Berlin is collaborating with JSTOR to digitize, preserve and extend access to Zeitschrift für Kunstgeschichte.

http://www.jstor.org

This content downloaded from 128.111.215.12 on Thu, 27 Feb 2014 13:46:34 PM All use subject to JSTOR Terms and Conditions LEONARDO'S COLOUR AND CHIAROSCURO

By JohnShearman

It is unfortunatelythe case that the analysis and interpretationof colour in paintings lags far behind other aspects of formal historical criticism.The subject seems to be in some degree of dis- repute,or at the best open to suspicion,and not without reason. It is rare that observationsin this field descend fromthe general to the particular1, or fromfrank subjectivity(even quasi-mysticism) to the admittedlymore tedious but ultimatelymore rewardingobjectivity that is, for example, normally regarded as indispensablein modern studies of perspective.The following study was under- taken in the belief that colour (and its dependents,light and chiaroscuro)can just as well be sub- 2 mittedto argumentand historicalcriticism The analogy between perspectiveand colour is not casual. One initial clarificationis demanded: light, in painting,is absent, or present,or deployed and characterizedin this or that way, always as a resultof handling colour, the primaryvisual constituentof the work, in a certain fashion.This is too oftenforgotten, and lightis discussedas if it were a self-sufficientelement which arrived via the artist'sbrush. Similarly, linear space, in its absence or presence,is the productof the treatmentof the perspective of objects. It is not an accident that those artists in the Renaissance who made most discoveriesabout space also explored and definedthe possibilitiesof pictorial light; the interestin, and understandingof, each problem requires the same state of mind. This is as clear in the art of Giotto as in the words of Alberti; Leonardo is another conspicuouscase. Anotherclarification must be made. How oftenhas it been said that Leonardo was not interested in colour, but in chiaroscuroor tone? This is a statementthat is based on a modernanalytical distinct- ion, and no Renaissance text on colour can be understoodbefore the anachronismis removed. For example, in 1-504Ugo da Carpi's "chiaroscurowoodcuts" are called stampe di legno a 3 colori . There is in Leonardo's paintings and theoreticalwritings4, as in those of his contemporaries,no oppositionbetween colour on the one hand and lightand shade on the other; it is inexact to separate colour - in the customarysense of the chromaticelement of colore - fromchiaroscuro, and to say that he found the formerof secondary importancecompared with the latter. Leonardo developed both, in new directionsand for new purposes.To him theywere not separate departmentsof his art, but were in most respectsinseparable; at timesthey are complementary,at other timestheir interaction is so complex that they may be regarded,in all but the scientificcontext, as one medium. It is highlysignificant that when he talks of colour and chiaroscuroin pictorial,and not scientific,theory, the treatmentof light and shade is designatedcolore, as in Alberti. Dividesi la pittura in due parti principali; delle quali la prima e figura,cioe la linea, che destinguela figuradelli corpi e lor parti- cule; la seconda e il colore contenutoda essi termini;when this division is repeated in all essentials in a second text,la seconda e detta ombra5. The developmentof the handling of colour in Tuscan painting achieves its greatestacceleration between the earliestworks of Leonardo and the death of Andrea del Sarto. Its pace may be compared fruitfullyto those of plasticityand disegnobetween, say, Filippino and Salviati. It was Leonardo who gave the firstimpetus in each case, and in colour his contributionis measurablythe greatest.It is not my purpose to describethis contributionin all its many aspects,but ratherto demonstrateone relatively simple point and to explore its consequences. From the methodological point of view I have taken the obvious opportunity,in the second section,to checkobservations and interpretations

This content downloaded from 128.111.215.12 on Thu, 27 Feb 2014 13:46:34 PM All use subject to JSTOR Terms and Conditions against Leonardo's many notes on the subject; this expressesthe assumptionthat techniquesof analysis and termsof referenceare most relevantwhen they can be found in the literarymaterial which is closestto the work of art6.

The tonal scale of pigments and the tonal unity of colour.

One of the propertiesof the Absolute Colour of mediaeval paintingwhich most vigorously resisted the realistic tendenciesof early Renaissance art was the modelling of formsexclusively in colour. With significantexceptions, most quattrocento painting achieves the relief of form in this way. Generally, a formof a certain colour is definedby variations in the intensityor saturationof this colour; variations in intensityyield automatically a range of tonal differencesand these express lightingand relief. Each pigment,however, in its pure and fully saturated state, has its own specifictonal value; blues are inherentlydarker than yellows. If, hypothetically,we take the most familiarpigments on the Renaissance palette,fully saturated, it is possible to produce a tonal, as well as a chromatic,scale: from yellow, the lightest,down throughcinnobar (or vermilion),apple-green, turquoise, rose-red, to the darkest,lapis lazuli. If an artistworks within the conventionof colour-modellingthese tonal propertiesof pigmentsare imposed upon him and lead to certainresults. The awarenessof theseproperties in the Trecentoand Quattrocentois demonstratedby the way in whichmany artistsexploit them.An alternativeto simple saturation-modellingis the phenomenonof colour-change;this is the variation of the local-colour of a formbetween its highlightand shadow - a device much favoured for its decorative contributionby many of the Tuscan gothic artistssuch as Agnolo Gaddi or Lorenzo Monaco, and obviously sympatheticto an age whichassessed the beauty of colour quantitatively,both in the sense of variety and of brilliance. Frequently colour-changes are no more than decorative,and thereis no other logic in the selectionof these pairs, but the tonal differenceinherently present in the coupling of, say, yellow and blue, may be made to model form. Masolino, in the frescoesat Castiglione Olona, is typical of several Quattrocentoartists who consist- ently select their colour-couplesin this way, so that the tonal contrastof pigmentsalone provides an alternativeto variationsof saturation7 Anothermore importantfact follows fromthe tonal scale of pigments.If an artistin this conven- tion paints St. Peter, by traditionclothed in a yellow robe over a blue vestment,with each drapery modelled by saturation-changesand the full intensityof the pigmentused for the deepest shadows, then those two formsare bound to be plastically inconsistent.The potentialrange of tone offeredby each pigmentcannot be matched,and the modellingof the yellow draperywill be less powerfulthan that of the blue. The full meaning of this problem may be seen in Masaccio's S t. A n n e in the Uffizi (figurei); at this early stage of his career Masaccio worked without modificationwithin the technicaltradition of late gothic art. A case where the consequences are least obvious is the rela- tionshipbetween the rose-redand deep blue draperiesof the Madonna, for these are pigmentsclose to each other on the tonal scale; even so, if one compares the plasticityattained on the sleeve of the Madonna in red, with that on the knee below in blue, the inevitable disparity is apparent. More striking,however, is the disparityin potential modellingbetween formsof colours at opposite ends of the scale, between, say, the relatively strongrose-red of S. Anne's robe and the much weaker cinnobar-redof the angel at the top, and most strikingof all, if one takes the extremesof the scale, between the full saturated yellow of the highlightsof the upper left-handangel and his equally pure, but far deeper blue wing. This angel has a vestmentthat turnsfrom this full value of yellow to a full cinnobar-redshadow, whichis an example of the device of colour-change,used by Masaccio

This content downloaded from 128.111.215.12 on Thu, 27 Feb 2014 13:46:34 PM All use subject to JSTOR Terms and Conditions i..?;:"t' "I '"

?E* rlk ~ I t:1! r " 6...:.1 i;'t.f' ~??r

- ?

dl:rA;.'LF'?ii ?r" a.?4e Z~i * ~i? . * ?:~?? ?k ?.? . f ?, .i .* i. *: r 'f. .?`?: I .,? :'i -1 ;?it.' k. ?: i:?.Y 4 1

it ~EI~ ~B~L:Y' ill T:. ?n.: i:'-?? "'I?:: r :r ~ ?:'* ?,? ?k; fq':'' I ??'~r ??l~i~Si:~ : ? i.. j~3~ *5. ?r

e~yCI- ii ?, ~iC "i;P j?l~f~a: 'U13~ *

'b r-, * i

rr d2.1r: ~nnr. f r *s~ -1 i. It

~ ?~R ,I?r -~ ?I ,?r?

,.I r: I ; j 1,,*'% tt?; ii t, I ?;t?

t:'l -.?r:?ir? a ?e?:? 1? 'f - f 'La:

*? :k. *::? r I '% ~s~Aa~F':: i

I. Masaccio,Madonna and Childand St Anne,Uffizi.

This content downloaded from 128.111.215.12 on Thu, 27 Feb 2014 13:46:34 PM All use subject to JSTOR Terms and Conditions to exploitjust the differencesof tonein the pigmentsthemselves which poses the problemwe are examining. The S t. A n n e panelintroduces a secondaryaspect of theproblem-modelling forms in palerand darkervalues of the same colour.Using the techniqueof colour-modelling,the only possibilityis whatwe have here:the light blue veil of theMadonna is modelledfrom white to a relativelypale lapis for the full-shadow,whereas the dark blue robe has a highlightvalue of the same pigment alreadyseveral tones darker than the shadowof the veil, and a full shadowimmeasurably deeper again. The same differenceexists between the deeperrose-red of S. Anne'sveil and the pale red angelbelow to theright in thesame pigment. The stylisticresult of thisuse of colouris complex,but its main points may be brieflysummarized. Firstly,the colourimposes an accenton the linearqualities of the painting;the limitsof every objectare markedby a sharptransition to a new colourand to a new rangeof tonevalues: to a totallydifferent level of plasticity.Consequently the line so createdhas a specialemphasis, and a tendencyto insulateeach differently-coloured object as an autonomousfield on thepicture surface; to each individualcolour-plane, therefore, this use of colourwill introducea flattening,surface- stressingtendency. In the case of a formlike a drapedfigure, composed of elementsof morethan one colour,this polychromy will inevitablybreak up thevolume of thewhole into planes of varying plasticintensity. In theS t. An n e, forinstance, the total plasticity of thefigure of theVirgin - or of thewhole figure group - is incoherent,and appreciablyless impressive than the really powerful modellingof the formsindividually, like the foldsover the knees.Because of the tonal scale of pigments,a polychromeobject in colour-modellingamounts plastically to muchless than the sum of its parts.In theArena chapel every coloured figure is flatter,and lessof a volumetricunit, than the monochromefigures below; to carrythe argumenta stagefurther, it is also less fullyrelated to its surroundingforms. A secondresult, equally relevantto the styleas a whole,is thatthis use of colourentails the completelyfinite realization of everypart of everyform; there is no possibilityof varyingthe sharpnessof focuson surfacesright up to theircontours, because it is an attitudeto colourand formwhich excludes the notions of atmosphereor of volumesof shadowas universalelements in the painting,whereby the surfaces might become partially or whollylost to view. The thirdpoint concerns the attitude to li g h t whichis impliedby thishandling of colour.By no means,for example, can thecolour-change from yellow to red,or greento red,have beenthought of in theartist's mind as a rationalor naturalisticresult of thefall of a streamof lighton a coloured form;the same is trueof theintensification of the local colourin themore common cases of simple colour-modelling.Neither can he have considereda unityin thereaction of separatecoloured forms to a singlelight: each form makes its own reaction,and thisis conditionedin thefirst place by the intrinsicqualities of theparticular pigment in use. Masaccio,of course,even at thisearly date, was exceptionalin histime for the understanding of theaction of lightin painting.But in theS t. A n n e theimpression of lightthat exists is, so to speak,the sum of a numberof individually-litparts, and the only real differencebetween this and the lightof Cimabue,of Orcagna or even of Lorenzo Monaco,is thaton theseindividual parts there is imposeda unityof direction;all the highlights have beenorientated to one side.This stephad alreadybeen taken by Giottoand Duccio. But it is the lack of unityof responsefrom colours, more than the inconsistenciesin directionand cast shadow,that withholds the instantaneous impression of thepresence of a truepictorial light, a single, unifiedelement passing through space and conditioningthe visibility and invisibilityof objects. The A n n u n c ia t ion in the Uffizi(figure 2) formsthe beststarting-point for a discussionof Leonardo'sposition in thehistory of thisproblem 8; it is a veryremarkable position. In thispicture,

This content downloaded from 128.111.215.12 on Thu, 27 Feb 2014 13:46:34 PM All use subject to JSTOR Terms and Conditions S? r ?~rr~i?' ;?r;E~1 IC ~B~"-?, : TcF '~??.~L~'"1 ?7 Ic ?.-t.~ 1 ??: ?~' ~_~_~~~,_~ii~??? -1;*

;~?l.:Y

2. Leonardoda Vinci,Annunciation, Uffizi.

immatureand inconsistentas it is in so many ways, there are already two revolutionaryprinciples of the greatestimportance: every form is modelled independentlyof colour, and every coloured object is investedwith a commonrange of tone9. In fact, the range of colour is as wide as is normal in Florentinepainting of the '70's and far wider than that of Masaccio's S t. A n n e ; the range which in Lorenzo di Credi, for example, gives rise to a furtherrange of tonal differences,from white throughyellow, apple-green,vermilion and rose-redto blue and brown, is used again here, but its effectis completelychanged. This change is the resultof a new attitudeto the relationof colour, lightand form. The modelling of form is achieved by achromaticmeans - in this case by the addition of black to the object-colour; saturation-changeand colour-changeare abandoned as ways of achievingthe tonal-changewhich representsrelief. In an early M a d o n n a by Lorenzo di Credi it is impossible to imagine the chromaticelement removed, because that alone generates the form; here, if the chromaticelement could be subtracted,every significent form would still remain. This point is clearest if we compare the angel's white vestmentwith the small area of pale blue on his collar (figure3); tonally these two have an equal range,and in the practical sense the creation of form is precisely the same: each runs throughthe same sequence of darkeningwith black. The pale blue becomes darker as the lightingdecreases, but it does not become bluer; the local colour has a fixedvalue in the blue as in the white. In one sense all colours here are affectedby light exactly similarly,and that is the sense which is vital for the continuityof the level of plasticity over each multi-colouredfigure: every colour- plane achievesor can achieve,a uniformdepth of shadow. There is one importantsense in which colours vary in theirreaction to light: the relationshipof the saturationof a given local colour to the chiaroscurodepends upon the specifictonal intensityof the pigment.Colours which are by value pale in tone - yellow for instance - are already fully saturatedin the highlight,and this continueson an even level into the shadow. This is true also of pale values of richercolours, such as the pale blue already mentioned,and the pale rose-pinkfloor tiles on the right.On the other hand colours which are inherentlydeep in tone, rose-red and blue for example, are lightened considerably in the highlightand achieve their full value only in the

This content downloaded from 128.111.215.12 on Thu, 27 Feb 2014 13:46:34 PM All use subject to JSTOR Terms and Conditions half-tonesand the full shadow. Colours of moderate depth of tone, vermilionand apple-green,are lighteneda little for the highlightbut are already at full intensityin the higher half-tones.The reasons for this apparent inconsistencyare two-fold. In the firstplace a colour plane which is requiredto be richin colour, suchas the "dark" blue robe of the Madonna (in contrastto the "pale" blue of the angel's collar and ribbons) is already of a depth of tone near to that which will be reachedby the blacknessof the shadow; if thereforethe colour were to remain constantin intensity for dark colours as for light,very little reliefwould result.It is clear that the new ideal of uniform plasticityrequires also that the level of tone of the highlightswill be approximatelyequal. One may also look at this from a rather differentview-point, and see that the achromaticmodelling implies not only the superimpositionon the local colour of a systemof darkeningwith some neutralpigment, in thiscase black,but also of lightening,if necessary,with white. The second reason follows from the first:a colour of weak intensitybecomes quickly submerged in the chiaroscuro.Consider, for example, the two cases of the Virgin's deep blue robe and its yellow lining; if the yellow followed the same sequence of lighteningin the highlights- that is, dilution with white (which would in fact make little tonal difference)the result would be, virtually, monochrome;the yellow would be entirelylost in the gatheringobscurity of the shadow. The blue, on the otherhand, and also the rose of the vestment,have a natural strengthwhich will enable them to colour the formeffectively even if they only reach full intensityin the deepest shadow: a power to retainchromatic effect into chiaroscurowhich yellow - and apple-green- have not. Leonardo's solution to this perpetual problem of the differentintensities of the palette is not entirelyrational; yet, from the aestheticpoint of view it is justified.When he varies the treatment of colour over the form- its reaction to light-changesin relation to the specificqualities of the local-colour, this resultsin some cases in a parallelism of saturation-changeto relief,represented by the monochromeelement, but it is perfectlyclear that these saturation-changesare conditioned by the chiaroscuro,and that they are neutral and inert in the complex processesof the generationof form. This early, and as yet unsophisticated,reaction to the pigment-problemmay be called by what it achieves: Tonal Unity. There are two major consequencesto notice. The first,and historicallythe more important,is the control of the inherentlydisruptive effects of polychromyupon plasticity. A figurelike the Madonna presentsnow a single, swelling, homogeneously-generatedvolume in contrast to the inevitably fragmentedeffects of colour-modellingseen in Masaccio's Madonna. Secondly, and of particularsignificance for Leonardo himself,light, colour and formare now related in a way that approximatesto, and describes,their scientific and naturalisticbehaviour. The relation of colour to light has entirelychanged'". In the relative absolutismof late Quattrocento painting light remained a functionof colour; changesin the objective nature of the surface of the form - variationsof intensityof colour, or colour-changes- created relief,and thisrelief was the indication of the lightingon the form; the incidenceof light on a given part of the formwas the resultof the objective nature of the colour at that point. Now, colour is a functionof light; it appears and disappears accordingto the lightingconditions, and its specificqualities at a given pointare governed by the fall of the light upon it, and not by the propertiesof the pigment.In other words light is perceived as an exteriorforce, outside the object and governingthe relativevisibility of theproperties of the object; the colour of the formis now, in the Albertian sense, one of its permanentqualities, ratherthan temporaryor accidental ones. In the group of early works the concept of tonal unity is always present,though there is a perceptibledevelopment. The angel of the UffiziB a p t i s m is in this respectalready an exceptional case in Florentinepainting of the '70's, yet it is tentativeand incomplete.On the other hand the

This content downloaded from 128.111.215.12 on Thu, 27 Feb 2014 13:46:34 PM All use subject to JSTOR Terms and Conditions ? ii

3. Leonardoda Vinci,Annunciation, Uffizi (detail).

Madonna in Munich (figure4) and the Portrait of Ginevra Benci, in the Liechten- steincollection (figure5), seem to show a later stage than the A n n u n c i a t i o n. In the case of the M a d o n n a there is a far less perceptibledifference in the handling of differentpigments than in the A n n u n c i a t i o n ; one is less aware that one group of colours is desaturatedin the highlight, or that another is fully intense there. The pale yellow of the robe lining, the blue of the robe, the brick-redof the vestmentor the olive-brown of the cushion - even the flesh- seem now to react in unisonto light and shade. In spite of the condition of the paintingit is clear that this is achieved by a perfectionof the methodused in the An n u n c i a t i o n. In the Uffizi An nunciation the smooth-swellingvolume of the Madonna is not only generatedin itself,but also with respectto its ambiente.The depiction of its backgroundand of the cast shadows in the same tonal range investsthe figurein a three-dimensionalatmospheric medium

This content downloaded from 128.111.215.12 on Thu, 27 Feb 2014 13:46:34 PM All use subject to JSTOR Terms and Conditions ::.::?? ?P:':':?: *: r:???: * 'i: ...;:..I;r ?::~:iii:?*W" .r r ~sa~ :J: :??:I;?Il~??aiss~ii~?i~'~LIB:~??' :i i ,1,,1,.,....,?"!""? "f' ?-~ ~.?:::".l$:!;iRptE :$:li~ifBJ:;l:Pli#:8::.:'a:?'?' ';'~' i;? ig'; .* ?;?

..* P .? ?';i: ??:!* ~;. ?sr : : Di. ..

.r6Lffil~

M t

?, :.::;

??,:..??,

??-crr,

~i~3%8s$E~1E~i .r P; ..

4. Leonardoda Vinci,Madonna, Munich.

which is another personal contributionof Leonardo's: a spatial chiaroscuro,which is a non-linear, tonal, systemof coordinates.His achievementof tonal unity bringsthe possibilityof a richdevelop- mentof this new spatial system,but a seriesof problemsstood in his way. From the propositionthat the greatestrange of tone in the modelling of a form,say a head, gives the greatestrelief, it is a logical step to the realization that this reliefis only obtained under certainconditions: in a restricted, focussed,light in a dark environment.In any other situation reflectedlight begins to play in the

This content downloaded from 128.111.215.12 on Thu, 27 Feb 2014 13:46:34 PM All use subject to JSTOR Terms and Conditions 5. Leonardoda Vinci,Ginevra Benci, Liechtenstein Collection.

shadows, and the lighterthe object-colour,the more the reflectedlights will mitigatethe depth of the shadow; fleshis naturally susceptibleof reflections.A head, therefore,placed beforea landscape raises problems; tonal unity,and also strengthof relief,will both be jeopardized if the naturalistic situationis observed. The problemwas initiallyposed, so to speak, by the UffiziA n n u n c i a t i o n ; it has an exterior setting,but with lighting variable in intensityand in kind. The variation is not according to

This content downloaded from 128.111.215.12 on Thu, 27 Feb 2014 13:46:34 PM All use subject to JSTOR Terms and Conditions naturalisticprinciples, but in the interestsof the clarityof individual objects. Probablythe expression of the subjectmatter made it difficultto investthe head of the angel with the same depth of shadow as his torso and the rightarm; two differentlighting conditions certainly exist. But the next problem lies in this: that whereas it is the head whichis given the lightingappropriate to the exteriorsetting, and the torso is given that of an interior,it is preciselythe torso that is in some sort of atmospheric unity with the dark background,while the head is sharply detached from it, and appears superimposed.

The colour-problem in Leonardo's notes.

In the whole painted oeuvre of Leonardo the subsequentstages in this problem are naturalistic adjustmentsof the aestheticprinciple of tonal unity. It was a problem which lasted for his entire career and it would be strangeif therewere no discussionof it in his writings;in fact thereis precise confirmationof the objects of tonal unityas theyhave been deduced fromthe paintings. For example, the finalconsiderations in the precedingsection are not abstract:Leonardo examined every one of them12. As is perhaps to be expected, he is undecided as to which settinghe prefers", but no-one was ever clearer on the differencesbetween the two extremesituations: Grand'errore e di quelli pittori,Ii quali spesse volte retrannouna cosa di rilevo a un lume particularenelle loro case, e poi mettonoin opera tal ritrattoa' un lume universalede l'aria in campagna, dove tale aria abbraccia et alumina tuttele parte delle vedute a un medesimomodo; e cosi costui fa l'ombre oscure,dove non po essereombra, e se pure ella ve e, ella e di tanta chiarezza, che'l e insensibile;e cosi fanno li riflessi, dove e impossibilequelli esserveduti 14 It is hardly necessary to demonstratethat Leonardo had a full understandingof light as an exterior force, and that form is revealed in the dynamic interplay of light and shade; he also distinguishedclearly between the local-colour and the accidental colour of objects, between the permanentand temporarycolouristic properties 5". Light and shade are temporaryphenomena clothing the form:ogni corpo opaco sia circundatoe superfitialmentevestito d'ombre.16 In effectthe whole theoryof tonal unityfollows logically upon these premises.In painting,white and black are not real colours, but are the modificationsof colours which indicate their lighting: nero ... bianco ... privatione e generativo... in pittura sono li principali, concio sia che la pittura sia composta d'ombre et di lumi, cioe di chiaroet scuro.17 This black/whitestructure is independentof the local colour on which it is superimposed;it is immaterialwhether an object be blue or white if the lightingconditions are the same,for the highlight and the shadow will contain the same quantityof white or black". This observationis followed by a practical recipe for achievingthis consistentmodelling of reliefby the addition of measuredquan- titiesto the blue, a recipe which is the perfectcounterpart of Cennini's colour-modellingrecipes for reliefby measuredquantities of the object-colour.In another passage the common depth of shadow is clearly stated: All colours when placed in shadow seem to be equally dark". Nowhere, in the hundreds of passages on light, shade and colour, is there any contradictionof this fundamental principle20 The question of the appearance of the object-colour through this light-shadestructure is also discussedin the writings;no definitivesolution is reached.Although some passages are contradictory in detail, the approach to the problemcorresponds to the analysis of the inconsistenciesnoticed in the paintings". The difficultylies in the problem of varying saturation, and whether the greatest brightness,intensity, or simply bellezza, occurs in the highlightsor in the half-tones;full shadow is naturally out of the question. Generally speaking he is content to accept the conclusions of the Albertian optical theory,that, since light reveals form and shade obscures it, the true colouristic

This content downloaded from 128.111.215.12 on Thu, 27 Feb 2014 13:46:34 PM All use subject to JSTOR Terms and Conditions propertieswill be most visible in the light22. The conflictarises fromthe furtherobservation, more practically linked to painting,that if the object-colourremained indeed unchangedin the highlight, light coloured objects would have a greater range of tone than dark, and greaterrelief23; in the abstract context of optical theory this is assumed to be the case 24. This difficultyis parallelled by the practical proposition that while shadows in a painting are achieved by the addition of black, highlightsare equally to be achieved by white. There is never any statementof the exact nature of the compromisewhich would solve these conflicts- probably because no stabilized solution was ever achieved - but it seems that the understandingof the differentrequirements of pigments according to their position on the tonal scale, which was deduced from the examination of the paintings,was in fact in Leonardo's mind. Two passages fromthe Trattato can only be explained in this way: (i) ... diverse colore hanno le loro bellezze in diverse parte di se medesimo,e questo ci mostrail nero haver la bellezza nell' Ombre e il bianco nel lume e l'azzurro e'verde e taneto (brown) nell'ombremezzane, e'l giallo e rosso (cinnobar?) ne'lumi,e l'oro ne reflessiet la laca (rose madder?) nelle ombre mezzani25. (2) Dove et in qual colore l'ombre perdano piu ii color naturale della cosa ombrata? II bianco, che non vede ne lume incidente,ne nisuna sorte di lume riflesso,e quello che prima perde nella sua ombra integralmenteil suo proprio natural colore, se colore si potesse dire il - bianco. Ma il nero agumentail suo colore nelle ombree o10perde nelle sue parte aluminate,e tanto piu lo perde, quanto la parte aluminata e veduta da lume di maggiorepotenzia. E il verde, e l'azuro agumenta il suo colore ne' l'ombre mezane; et il rosso e giallo acquista di colore nelle sue parte aluminate,e'l similefa il bianco; e li colori mistipartecipano della natura de colori, che compongano tal mistione,cioe, il nero misto col bianco fa berettino,il quale non Abello nell'ultimeombre, com'-'l nero semplice, et non e bello in su lumi, com'e il semplice bianco, ma la suprema sua bellezza sit infralume et ombra26. White,yellow and red (probably cinnobar or vermilion)are the colours light in themselves,which require littleor no adjustmentto make an effectivetonal contrastto theshadow; on the other hand blue, lake, green (probably the deep copper-greenwhich has now in most cases turnedto brown) and brown are the colours which are dark in themselves,and which can therefore appear at full intensityonly in the increasingshadow.

The development of Tonal Unity in Leonardo's painting.

At the conclusionof the firstsection it was postulated that an unsolved aestheticproblem arose in the A n n u n c i a t i o n. A naturalisticrespect for the exteriorlighting conditions in the head of the angel led to a less coherentrelation of formto settingthan was achieved in a more ideal, or abstract, lightingsituation assumed for the rest of the figure.The firststage in the solution of this problem was the exploitation of the second situation,and especially of the dark foil to the figure,with a greateror lesserimplication of enclosed space. The aim of theseearlier works, up to the firstMilanese period, is towards a closer and closer approximation to the complete consistencyof tone for the whole figure,flesh and polychromedrapery, and this means at once a greaterintensity of lighting and a greater restrictionof its direction; the dark foil is a necessary complementto the plastic consequencesof this development,and the naturalisticaspect is for the momentignored. In the more abstract sense there is no conflictin the use of essentiallythe same patternof backgroundfor both the Munich M a do n n a and the Liechtensteinportrait 27, for the lightingof the foregroundand of the backgroundis unrelatedin the naturalisticsense in each case, and the pictorial unity is achieved throughthe decorative functionof the colour. The darknesssurrounding the profileof the figureis essential for the atmosphericsetting of the form,and in neithercase is an enclosureimplied which would justifynaturalistically the characterof the foregroundlight. It is indeed remarkablethat this

This content downloaded from 128.111.215.12 on Thu, 27 Feb 2014 13:46:34 PM All use subject to JSTOR Terms and Conditions 6. Leonardoda Vinci,Madonna of theRocks, Louvre.

temporarysolution was so systematicallyapplied in the early works; every compositionwhich has been broughtto the stage of the chiaroscuroproblems shows this dark foil28, the functionof whichis similar to that of the ground in a relief,and if one considersthe evolution of the A d o r a tio n compositionit is clear that the half-builtchoir behind the figures29, whichprovides an apse of shadow behind them,is the answer to an aesthetic,and not an iconographical,problem. It is revealingto see

This content downloaded from 128.111.215.12 on Thu, 27 Feb 2014 13:46:34 PM All use subject to JSTOR Terms and Conditions 7. Leonardoda Vinci,Madonna of theRocks, National Gallery, London. (reproducedby Courtesyof theTrustees of theNational Gallery).

how the purpose of the interior setting of the Madon na Benois has been completely misunderstoodin Lorenzo di Credi's copy 30. The M a d o n n a B e n o i s is near a consistentand logical solutionof this problem; the dark foil is explained naturalistically,the particularizedlighting on the figureis justifiedby the setting,and the strongplasticity of the formsbelongs atmospherically to the space of the setting.

This content downloaded from 128.111.215.12 on Thu, 27 Feb 2014 13:46:34 PM All use subject to JSTOR Terms and Conditions Vasari,speaking specifically of theseearly works, is aware of the limitationsof thissolution 31 P cosa mirabileche quello ingegno,che avendo desideriodi dar sommorilievo alle cose che egli faceva,andava tantocon l'ombrescure a trovarei fondide'piz' scuri che cercava neri cheombrassino e fussinopi' scuridegli altri neri, per fareche'l chiaro, mediante quegli, fussi piz' lucido; et infine riuscivaquesto modo tantotinto, che non vi rimanendochiaro, avevon piz' formadi cose fatteper contraffareuna notte,che una finezzadel lume del dl: ma tuttoera per cercaredi dare maggiore rilievo,di trovaril finee la perfezionedell'arte. The nextdevelopment can bestbe shownby a comparisonbetween the two versionsof theM a - d o n n a o f t h e R o c k s. Thereis reasonto believethat the Louvre version was madein response to the commissionof 1483,and is thereforedocumented as beingentirely by Leonardo'shand and finishedsome time before about 149432. The documentsprove withoutany doubtat all that the NationalGallery picture is theone thatwas destinedfor the same altar as the I483 commissionbut was in an unfinishedstate when Leonardo left Milan in 1499, and was broughtto its presentstate by him duringthe secondMilanese period. The differencein the conceptionof the two paintings thereforerepresents a stylisticadvance of about fifteenyears and bridgesthe gap betweenthe Madonna Benois and the Madonna with the Yarn-Winder (firstversion), betweenthe Adoration andthe Battle of Anghiari,or betweenthe Ginevra Benci and Mona Lisa.33 Generallyspeaking, the lighting situation of theLouvre version (figure 6) is thatof theS t. Je - r o m e 34 or of the A d o r a t io n; thatis to say, thatthe figuresare sharplylit, in an exterior setting,before a darkfoil. It cannotbe too stronglyemphasized that this is n o t a paintingof a groupin a grotto;there is open sky above the group,conspiciously coloured a clear blue,and no indicationof any enclosure;the "grotto" is notreally a grottoat all, butthe rocky equivalent of the ruinedstable behind the figures in theMetropolitan Adoration studies3. Its pictorialfunction is thesame as thewall in theUffizi Adoration or therock behind S t. Je r o m e. In the Louvre versionthe individualrealization of everydetailed form, so clearlyseen in its drawing,is seenalso in thelighting; the dualismof thesituation in the A n n u n c i a t i o n is still unresolved.Only thoseflesh-forms close to thedark ground achieve a plasticityequal to thatof the drapery,even thoughpolychromy in itselfno longerpresents a tonal problem.In theforms above, morefreely enveloped in illuminatedspace, naturalistic concessions with respect to the "exterior" situationare made- as in the G i n e v r a B e n c i portrait- so thaton theone handthe inter- nal reflectionson theflesh deprive heads and handsof plasticity,on theother the persistently clear silhouetteof theshadowed side of theforms detaches them from atmospheric union with the setting35a. In the London version(figure 7) the dualismis resolved,and thismore intensely realized and intellectuallymeditated conception comes about partlythrough an apparentchange in the actual situationportrayed. The open area of sky above is drasticallyreduced, and the impressionof the enclosureof the figuresin a dark space therebycreated; it is possiblenow to read thesituation as thatof a grotto,since the continuity of thewall of rockis suggestedabove as well as at bothsides. This meansthat the lighting of thefigures may be broughtinto a rationalharmony with their context.A restricted,particularized light enters the scenethrough such openings as are shownbehind but actuallyfrom within the spectators' space to theleft. The lightis restrictednot only in thesense of its sharpnessof direction,so thatthe scattered reflections are eliminated,but also in its selective fall: it is a selectivelight seeking out thecompositionally and iconographicallysignificant forms and ignoringthe rest. In itsdynamic qualities of variabilityand selectivity,in contrastto thestatic, even, universallight of theLouvre version, it is thelight of a newera. By itsvery restriction the plasticity of everyform it touchesis augmented,and at the same timethe complementarychiaroscuro sets

This content downloaded from 128.111.215.12 on Thu, 27 Feb 2014 13:46:34 PM All use subject to JSTOR Terms and Conditions eachform in volumetricrelation one to another,and each to the grotto,more forcefully than ever beforebecause now the possibilityexists of losinga shadowedcontour. The "grotto-light"is as objectiveas M o n a L i s a's smile;the naturalistic harmony of thesituation is now complementary to theaesthetic harmony 3. It is perhapsfair to look uponthe London version therefore as theperfected solution of theearly problem;what we needto stresshere is therole of tonalunity, and whatcould thereby be archieved. To theindependently-conceived psychological states of thefigures in thefirst version there is applied a focusof emotionalresponse and physicalaction which contributes a reflectiveunity of sentiment. In thisdreamlike fusion the interweavingemotional cross-currents become superfluous: the figures do notlook at eachother, or at us. An integration,parallel in directionand equal in extent,has been appliedto theformal constituent of thepainting. This is theproduct as muchof a developingtreat-. mentof colour,and of its derivativelight, as of design.The solutionof theinitial dualism was not foundby abandoningthe progress towards unity, but rather by the subordination, first, of naturalistic to aestheticconsiderations and thenby theinvention of a situationin whichthey could coincide. This commonsolution to theaesthetic and naturalisticproblem in thelater years of theMilanese period formsthe basis of the finalsolution whereby the figurecould retainits plasticityagainst either a dark or a brightbackground, while stillpreserving its spatial,atmospheric relationship with it. Leonardo'shabitual diagonal formal constructions, implying as theydo thatthe settingrequires to be in morethan decorative harmony with the figures, make the pursuit of thisfinal solution appear inevitable.The new orientationfor thispursuit is set initiallyby the L a s t S u p p e r, whichin a verydifferent "naturalistic situation" represents a phase comparable to thatof theLondon V i r - gin of the Rocks. In the L a s t S u p p e r (figure8) therecould neverhave been a choiceof thesetting, yet it is remarkablethat even in the veryearly stages of the conception,in the Venicesketch, the need is feltfor the chiaroscuro as a supportfor the figures 37. As thepainting stands at presentit is of course impossibleto graspwith any accuracythe original atmospheric effect, but it is clearenough that the figuresof the apostlestowards the wingsare integratedby lightwith the space behindas in the London Madonna of the Roc k s. But thereare some new features;the chiaroscuro,the three-dimensionaldialogue of lightand shade,is now so emphaticallypresent as a pictorialreality thatit becomesitself expressive directly of thesubject matter; this is to be seennot onlyin thelite- rallydramatic shadow of Judas,but also in thewhole crepuscolomise-en-scene, recalling the text: And it was evening.... So, dependentalways upon thetonal unityof the colour,an entirelynew expressivemedium is givento painting,beyond the means of the Quattrocento.The secondfeature is thesetting of thefigure of Christagainst the bright background of theview throughthe doorway onto the open landscape(figure 8). This is perfectlysatisfying only by virtueof the clarityof the lighting-situation;perhaps the pointis clearerif we considerthe M o n a L is a whichis in this respectequivalent to thecentre-portion of the L a s t S u p p e r in isolation(figure 9). The M o n a L is a mustalways be visualizedwith the flanking columns more in evidencethan theynow are38; thistherefore, is a portraitin a loggia,with a landscapebeginning at an indefinable distancebehind it. The stratificationof the space thereforejustifies logically the unrelatedlighting of thefigure, restricted and without"scatter", and of thebackground, limpid and diffuse;this new "situation"should be contrastedwith that of G i n e v r a B e n ci (figure5). The figurenow stands out by virtueof thestrength of itsown modelling;it clearlydoes not belong to thesame space-and- lightunit as thebackground either in theaesthetic or thenaturalistic sense. However,the nextproblem, of the settingof a figure-groupin an open landscape,occupied his attentionin thesame years that were spent on the M o n a L i s a. This problemwas thenecessary

This content downloaded from 128.111.215.12 on Thu, 27 Feb 2014 13:46:34 PM All use subject to JSTOR Terms and Conditions cr.-. '" "' :?a;, ~?C~~ ~artv:, ~;i~~LU' jri? *'? iii' i~ ~:? i ?~ ?cl~7:~? ZY, -C~?..?'

*.... Lj

3L*? :"':S~?~?~Pa~B ~ ?:.~ ~i,. r ?1. ?e*! r'. 4".t ?.~:

?li3E *?

?r?;~ ~.I? I I ~'

- ,~-~~C-~E5~L- ~B-~2~T~~;i?..~ ~"' :.~S~S~i~i~?iL? a g ~I?;:? Inn;~~ "-*Y~~IJ~?

8. Leonardoda Vinci,The Last Supper,S. Mariadelle Grazie, Milan (detail).

corollary of the logical treatmentof the interiorsetting of M o n a L i s a and the L a s t S u p - p e r; it was already set by the compromiseof the late-Milanese cartoon in the Royal Academy, and this was followed by the ISoi S t. Ann e cartoon, the two versionsof the M ado n n a o f the Yarn-Winder, the Leda and the Battle of Anghiari; we do not know preciselythe stages in the evolution here, only the final result in the Louvre S t. An ne, but it is clear that one of the elementsin the final solution,the greatly enrichedpenetration of light within the complex group - or, to put it anotherway, the use of a diffusedlight without sacrificing plasticity - must have been considerablydeveloped already in response to the special problems of the battle-piece3. The other elementwhich is importantis the increasedrole of colour - in the modern sense - in this unity; and here, beforeconsidering this in detail, it would be as well to examine the changingfunctions of the purely chromaticelement up to this point, for the S t. A n n e belongs to a phase of Leonardo's developmentof which Florentinepainters were ignorantat least until Andrea del Sarto's to Paris in 1 journey 518/1940. It seemed worth while to deal with Leonardo's achievementand developmentof tonal unity at length,because of its great historicalimportance; of its two functions,the approximationto optical

This content downloaded from 128.111.215.12 on Thu, 27 Feb 2014 13:46:34 PM All use subject to JSTOR Terms and Conditions 9. Leonardoda Vinci,Mona Lisa,Louvre.

This content downloaded from 128.111.215.12 on Thu, 27 Feb 2014 13:46:34 PM All use subject to JSTOR Terms and Conditions naturalismand the creation of consistentplasticity, it was the latter whichwas most appreciated in his own day, and whichstands so directlyin the Florentinemonumental tradition; it is no exaggera- tion to say that this discovery is the essential and inevitable pre-requisitefor the classic plasticity of the High-Renaissance,by which painting achieved the corporeal homogeneityand relievo which would rival sculpture. It is impossible to imagine the aims of the early Cinquecento expressed without it. The "sculptural presence" of painted formmay well have been among Leonardo's most compellingambitions, obsessed as he was with the Paragone. The judgementof the Cinquecento is on these lines; Vasari's summaryof his historical position as a painter,at the end of the Vita, is this: Nell'arte della pittura aggiunse costui alla maniera del coloriread olio una certa oscurita,donde hanno dato i modernigran forza e rilievo alle loro figure41 and (an independetjudgement) a contemporaryof Vasari's calls him: primo inventoredelle figure grandi tolte dalle ombredelle lucerne42

Colour, chiaroscuro, and composition.

These changesof Leonardo's broughtwith themtwo entirelynew stylisticcharacteristics, whereby the actual appearance of the painting is changed; the firstconcerns the chromaticeffect of the picture,the second, the veil of positive,corporeal atmospherein which the sharpnessof the formis lost and found,whereby the tactilityof its surfacebecomes variable. The "suppression"of colour by chiaroscurois not a negative attitude to colour; the open clarity, the purity,the brightnessof Quattrocentocolour are rejected - and progressivelymore so - and colour speaks only in a dialogue with chiaroscuro,the one infused into the other, but therebyit attains a new resonanceand a new depth,and it is this moresophisticated richness which characterized Florentinepainting for a generation. There is a text whichis oftenchosen to show that Leonardo was interestedonly in the chiaroscuro, and that colour itselfmeant little or nothingto him: Qual e di piu importantia,o'che la figuraabbondi in bellezza di colori, o'in dimostrationedi gran rilevo? Sola lapittura si rende (cosa maravigliosa?) alli contemplatoridi quella per fare parere rilevato e spichato dalli muri quel, che nulla, e li colori sol fanno onore alli maestri,che li fanno, perchein loro non si causa altra maraviglia che bellezza, la quale bellezza non e virtudel pittore,ma di quello, che gli ha generati.E puo una cosa esser vestita di brutticolori e dar di se maraviglia alli suoi con- templantipel parere di rilevo43 It cannot be too stronglyemphasized that bellezza di colori does not mean "beauty of colour" in the modern sense, but the harsh brilliance of pure pigment;what Leonardo has in mind is not the choice between a picture brilliantlyand subtly coloured (for example, his own L a s t S u p p e r) and a grisaille44,but between a painting of, say, Ghirlandaio, and his own: the contrastis between the Quattrocentoand the Cinquecento. This is the changeof taste whichis reflectedin the writingof Michiel,Benedetto Varchi, Vasari and Paolo Pino 45.Leonardo in fact seldom toucheson the nebulous subject of colour harmony,but when he does, it is clear that his ideas are those which were given more succinctexpression by the greatercritical apparatus of Vasari. Restaci una secunda regola, la quale non attende d fare li colori in se di piu suprema bellezza, ch'essinaturalmente sieno, ma che la compagnia loro dia gratia l'uno a l'altro . . .46 Leonardo's personalityas a painter is partly conditioned by his "total" vision of a work, which results in the unified conception (in the mature drawings one sees the spontaneous growth of a complex design) and in the necessityof bringingthe executionof the whole to a finishsimultaneously,

This content downloaded from 128.111.215.12 on Thu, 27 Feb 2014 13:46:34 PM All use subject to JSTOR Terms and Conditions an innovation which broughtits own difficulties;the traditional fresco-technique,for example, was inadequate for his purposes.This simultaneousexecution of the whole work is seen in technicalterms in the unfinishedworks and in the evidence we have of the way he actually painted the two wall- paintings: the stories relating to the L a s t S u p p e r 47, Vasari's description of the adjustable scaffoldingfor the Battle-cartoon48, and the documentsdescribing the scaffoldingfor the Battle-piece itself49, all show how necessaryit was to be able to paint, so to speak, the whole work at once. Both for these technicalreasons, and for the psychologicalimpetus behind them, it was inevitable, one might think,that Leonardo should have broken the bonds of Quattrocento,quasi-Absolute colour. The clearly defined,mutually insulated, colour-planes imposed limitations like those of buon fresco, whichthe continuouslyfomenting creativity of his brain could not accept. We have already seen that in the early Uffizi A n n u n c ia t io n colour ceased to play the active role in the creationof formand line; it begins,on the contrary,to be broughtinto a dynamic relationshipwith light,parallel to that of formwith light. The particular technicalcharacteristics of the chiaroscuroin use up to the time of the London version of the Virgin of the Rocks altered the chromaticeffect of the painting in two ways: the reductionof the brightnessof colour (a suppression of one kind of richness),and its replacement by a homogeneous resonance in a distinctlylower key. In thissense thereis already a new unityof colour compared with the Quattro- cento; this is a sfumatounity achieved by achromaticmeans. In the earlier works the compositional functionof colour is primarilydecorative; that is to say that the colour links draw the picture-surfaceinto a unity,in the mannersensitive to the realityof the picture surface which is common in the Quattrocento. In the A n n u n c i a t i o n only the colour-linksacross the third dimension are exact, whereas those across the plane give the greater stress to the figures;no colour-value is common to both figures,so that each receives accented individuality.The colour-linkacross the depth is given, for example, by the repetitionof the silvery blue of the distantmountains in the ribbons,the collar and wing-rootsof the angel. This paler blue, which occurs all over the left-handhalf of the painting,in the flowersbelow and in the sky above as well as in the distantlandscape and the angel, is the same blue whichoccurs more intenselyon the robe of the virgin; similarlythe glowing ruby-redof the angel's robe is repeated more softlyon the bed-coveron the right. In the Liechtenstein Portrait and the Munich Madonna the decorative functionof the colour answers the essentiallydifferent requirements of the format;the consciousness of the depth axis is correspondinglygreater, and the necessityof integratingthe surface-patternby repetitionsacross the plane could be ignored. In the Ma don na, the olive-browncushion in the bottom leftcorner repeats the colours of the middle-distancelandscape, while the blue of the robe (softerthan in the An n un c i a tio n) repeats that of the distantmountains. In the Po r tr a i t, the orange-brownof the middle-distancetrees is precisely the same as the highlighton the right shoulder of her gown, and the blue of the lacing on the frontis equal to the strongestblue in the landscape, in the distanttrees and the churchspires. The colour-compositionof the Paris versionof the V i r g i n o f t h e R o c k s is a development from the Florentineworks only in the sense of a greatersubtlety and maturity,a new richnessof orchestration;the principle is a very finebalance of the still sharply individualized polychromyof the draperiesof Virgin and angel. The blue of her robe, for example, by far the most intensecolour in the picture and of an extraordinaryrichness, is given progressivelyweaker echoes in the distant landscape-vistaand the quadrant of sky above, then in the flowersjust to the rightof her head and in the pale irises beneath the Giovannino, and finally in the cool grey ledges of the foreground parapet of rocks and the feathersof the angel's wings. The warmest patches of rock, ruddy-brown,

This content downloaded from 128.111.215.12 on Thu, 27 Feb 2014 13:46:34 PM All use subject to JSTOR Terms and Conditions below Christ on the leftand along the very top, seem like reflectionsfrom the glowing brilliance of the Angel's robe, and the complementarygreen of the lining thrownover his leftshoulder is softly echoed in the landscape vista, diagonally opposite. In spite of the disfiguringvarnish, and what appears to me to be a considerablequantity of overpaint, all of whichadds an artificialdegree of unityto the colour-planes,the "separateness"of colour at this stage - a featurewhich furtheraccentuates the individualityof the formsto which they are fixed - is still apparent; what is also apparent is an intentionto use colour as an "accent" on the subject matter,but this has not yet found a new means of expression.The new concept of crescendo or focus emergesfirst in the La s t S u p p e r 50. It is essentialto see the L a s t S u p p e r in the correctnatural lighting,that is in the late after- noon or early eveningwhen the lightcomes stronglyfrom above and fromthe left,and leaves almost a thirdof the compositionin half-shadowon this side. Withoutthis lightingthe balance is disturbed. Leonardo has compensatedthis shadow by a greaterstrength of colour on the leftthan on the right, the colour and by strongercontrasts of colour51and modelling within the shadow; in comparison, is much softeron the right,with many of the draperiesapproaching a silvery-grey.This increase of plasticityand colour in the more weakly-lit part is a device which he seems to have learnt from Masaccio 52;it has anotherfunction, apart fromthe balance in the generalimpression on thespectator, and this is that while the directionand the temporal significanceof the actual light has been most subtly re-deployedin the painting,it has been, so to speak, extended to the part of the wall that is actually in shadow. When the paintingis seen in the natural light,therefore, the balance in the chromaticeffect from one side to another is perfect,but there is no articulated pattern in its distributionas there is in Ghirlandaio's L a s t S u p p e r: the colour-planes are not unified by repetition.In the present state of the painting53it is clearly impossibleto say whetherthere was ever a colour-chaincoordinat- ing the colours - perhaps therewas; but it is even now possible to see how colours have lost their particularity.No colour-value is exactly repeated anywhereelse, yet thereis an elementof homogene- ity, of spontaneousharmony, which runs like a ground-bassbeneath the cantilena of the individual colours, and is one featurewhich pulls the whole long compositiontogether. The other is the subordination of the whole to a dominant centre; the figureof Christ, in the colour composition,is isolated by the simplicityof the colour-shapes- a red diamond and a blue equilateral triangle- and also because these two colours, apparently endlesslyechoed and re-echoedin the other figures, are the strongestin the whole painting54. Moreover, if we are rightin reconstructingthe rich green of James' tunic as being about equal in intensityto the blue and rose of John'svestment and robe55, there was originally a crescendo towards the centre, an organic movement in the colour which continuouslyfocuses, even now, the spectators'attention on the quiet figureof Christ56. This "colour-focus", distinctfrom the old method of a particularized accent, is the expressive equivalent of that kernel of formal activitywhich Leonardo had introducedinto the plastic pattern the as early as the Benois Madonna and the Adoration, as a stresson the essence of for subject matter; it may be seen even more plainly, without so many necessary qualifications condition,in the London versionof the Vi r g i n o f th e R oc k s. The change in colour between the two versionsis very striking;in the firstit is gay, lyrical,finite and wide in range: in the second, brooding, austere, elusive and restricted,virtually, to two im- pressivechords of deep blue and golden-peachon the Madonna. To see this second picturein a light approximatingto a churchinterior is to realize two things;the firstis the futilityof the discussions as to whetherSt. Johnor the Christ Child is the subject of the painting.The augmentedintensity of the blue in theseconditions leaves no doubt whatever that it is, in Leonardo's own words,uno quadro

This content downloaded from 128.111.215.12 on Thu, 27 Feb 2014 13:46:34 PM All use subject to JSTOR Terms and Conditions de una nostra dona 57; the blue has an overwhelmingvitality in the whole composition.The second point is that the robe when seen in this way is as plastically modelled as the head or the yellow lining58; this is not normallyobvious. In this case one can be reasonably sure that all the colour in the painting is linked togetherin a cycle; blue turnsto yellow throughgreen, and the yellow returnsto blue throughbrown, warm-grey and cool-grey.The two positive, full-strength,colours are concentratedon the figureof the Virgin, so that She seems to contain the quintessenceof the whole chromaticrange, and She becomes the true focus of an effectembracing every part of the painting. The flexibilityof the colour that is developed at thisstage is dependenton thefundamental break with traditionmade earlier,when colour was released fromits functionof creatingform. The independentdevelopment of formand colour then made possible is in this particular case turned towards the expressionof the iconographic content. Colour therefore,in the fully-developedstyle of tonal unity, ceases to be a static element,and becomes fluid,surging, dynamic; the dynamic characterof this medium,with its new expressiveand compositionalpotential, recalls the similarchange in the characterof lightbetween the two versions, and distinguishesthe High Renaissance artistfrom the Quattrocentoone. This intensificationof the qualities immanentalready in theLouvre version,which seems reasonably to be the resultof the unusual situation of the artistbeing required to recreatea work of the same subject for the same altar, profitingfrom the experienceof seeing the firstin position,- this inten- sificationis the answer to a specificproblem, and it is a little difficultto assess on this basis the general stylisticcharacter of the lost works of the period which is spanned by the execution of the London version.The only paintingof this period which does survive is the M o n a L i s a, and the featuresthese two have in commonmay be summarizedto provide a rough idea of the legacy which Leonardo leftto Cinquecento Florentinepainting. It is, of course, notoriouslydifficult to see any colour at all in the M o n a L is a; one's first impressionis only of a greenishhaze, dimly seen towards the top. One readily perceives,however, what a really essentialpart colour plays in the whole, how it is vitally presentin every particle, if one visualizes its appearance if it were, or rather if it could be, executed in grisaille. First, of the functionsof this partially-hiddencolour, we mustnot discountthe naturalisticevocation of vitality, whichform alone cannot give, and whichmay still be appreciated afterlong scrutinyin a reasonable light59. But two furthercharacteristics must have appeared revolutionaryto his contemporaries;the firstis the perfectionof the plastic consistencythrough tonal unity (it is quite uncompromising),and the second is the way in which all the colour seems to be infused,as a scarcely tangibleelement, into tone. The yellow sleeve is a dull glow compared with the ceramicbrilliance of Ghirlandaio or Credi, but it is no less yellow. Leonardo now stands,with Titian, on the far side of a dividing line in the historyof colour in painting. In Cinquecento painting one does not measure colour, so to speak, in termsof pigment- its refinementor degree of saturation- but in termsof its organic interaction with the tone of the whole picture.Colour is summonedto visibilityout of shadow by the action of light; it appears, certainly,as a propertyof a certain form,but that propertyis only a particular localized nuance of the continuous,unified colour-material of whichthe pictureis constructed.There is, however, a corollary of this,and that is that as the colour becomes submergedin chiaroscuro,so surfacesdisappear also; the two are plainly affectedin parallel. Finally, the last stage in Leonardo's developmentseen in the S t. An ne (fig. Io) will serve to show that, even to its creator,the colour-systemof the period that Florence knew was not a stable thing,but a transitionto anothersystem 60. One of the relative limitationsin the styleof the London Virgin of the Rocks and of the Mona Lisa is that each colour plane is in itself homogeneousand "fixed" in value, with the inevitable result that a group of colour-planes remains

This content downloaded from 128.111.215.12 on Thu, 27 Feb 2014 13:46:34 PM All use subject to JSTOR Terms and Conditions individualized as far as their colour is concerned, however much this may be tempered by the envelopingchiaroscuro; it is a limited stage of developmentwhich Hetzer still found in Titian's A s s u n t a and he described,in a memorablephrase, its separatenessof colour: fast wie bei einem Glasgemeilde1. The strikingcharacteristic of the S t. An n e, in comparison with the other works we have considered,is the much greatersoftness in all its elements;much of this is beyond analysis and lies deep in the matiere,but one may discern quite clearly the softeningalmost to extinctionof the differencesof colour-planes62. All the draperieshave a virtually uniformgrey, only faintlytinted, for theirhighlights; the real colour palpitates uncertainlyin the half-tonesand is lost again in the shadow. The shadow itselfis now changed; it is no longer dense, and opaque, for on the one hand light penetratesfreely and it is alive with reflections,and on the other it is not black, but - so far as one can see - a chromaticgrey. This latent colour in all the shadow is entirelynew, and it has thisresult: the already delicatelymuted palette - blue-grey,rose-grey, red-brown and lilac-brown- achieves values that are never really stabilized on any form; each "colour-plane" - the expression is hardly adequate any more - is subtlyshifting in value all over, and each colour seemspotentially presentin the next. This developmentis extremelydifficult to definein its preciseextent; thereis the dangerof claimingfor Leonardo, in thisinterpenetration of colours,the achievementof laterpainters. The approach is fundamentallydifferent in method from the Venetian broken-colourwhich had already appeared63. It is not so mucha question of the division of colour, a featureto be expressed in the brushwork,but an instabilityin its appearance in the three-dimensionallight-and-shadow system64; its value as a positive factor in the colour-harmonyis not to be denied, but it seems also that it is part of the solution of the problems presentedby the chiaroscuro-and-spacerelationship whichbecame criticalafter 50oo. It is noticeable in front of the original that the constructionof the figure-groupbefore the landscape's change of tone fromdark to light is achieved without the break in the unity of lighting and atmospherewhich has to be justifiedby the loggia in the M o n a L i s a. The diaphonous softnessof all surfaces and silhouettesis partly responsiblefor this; there is also a considerable quantitative change in the proportion of areas of light and shade, in the directionof more light, compared with the London Vir gin of the Roc ks or the S t. Anne cartoon in the Royal Academy. Sympatheticto this is the raisingof the tone of the shadows, whichis the automatic result of replacing black by neutralized colour. Furthermore,if colour has re-entered,in a new way compatiblewith optical rationalism,into the constructionof form,then it followsthat the foreground figuresare constructedin the same way as the distance against which they are set. If one could see the Mo n a L is a with its filterof varnish removed, there would almost certainly appear the actual change in the quality of the paint between figureand backgroundthat is seen in the London Virgin of the Roc ks (where, of course, the two are not contiguouson the surface,so that the contrastis of small importance); in the S t. An n e, so far as one may judge througha less dense layer of varnish,the technicalmeans in the whole painting is uniform.This material unity of all colours as of all forms is strikinglyparalleled in the developments in technique in the late drawings65; entirely new techniqueswere, in thesecases also, inventedby Leonardo forthe expression of a world of art in harmonywith his vision of the continuityand homogeneityin the world of nature. Conte x t.

The development that has been traced in Leonardo's use of colour is not properly assessed in isolation,or simplyin contrastto an early panel by Masaccio. There are some interestingprecedents

This content downloaded from 128.111.215.12 on Thu, 27 Feb 2014 13:46:34 PM All use subject to JSTOR Terms and Conditions I'~~.~-~'~'IPE~L~""""as"a~B~?~LB~B~-a#l~ si~'? jl. ~l"i??''' ? ? ~ ~: ~.t~~.jSL~.h'??;:~7t? ?1 i' %;.7 -? a.-S~L~! Pi.~ E~lx ;???. ?il- *:? :: ::7~,*i t,?::j~: 3~?7.:3.:~.?1;: ~ J. I;;i ~35~0~: ?:?d:l? ill? i 'if;:L: ?? l-..??lr? ;J~?nl;,??:~i~TP;::':Jin' i: -.???~:e s:~ i*Jir~~??: ;j .~['[~:*li~:?:~'::L;i:.?s.?. ?i ?r c?? ~?.r? ~-7~ "~ :???'?:, '?'rc??:? ?:?"' ?o? ?.??:?:t?... :?? .?*.? "?.:.4-: :~ iiIjkj ;:*~ti:l-i~.:m dd tt ?; ?~-~f~f~ i' ?*!~.I? i~tl~f~'~i ?r5 ?, Y~:* t~d~~P. :r: it~~ff:L?-?' ?!,7T:I': yr?' ;; .~? :'~??~':""? ?, ''~':??i r. "'?'~i;?~:?r~ *;I ~t..: ::?: ?-??:? ~~ !:~? r. ~ ?i~ 12f: i 'IS - rC~rp~lm~l-~~:?k~.P~ `~i~qpFP~j~i~El P -~P ' ??rl? :i *k~.;. i ":. ?"!~r~E~ C a~Eil"~?;~~i"~ilC~Fx~~s-*~ ;~~lii ?'"?;;:....- ,.... :?~~JrYLlk I IE .It. I1J~BL~'~Yr~b~"llE~Ia~B~l-- ;.:? ?? .:? ~F~P'r:e:?~ ?~ t?. ?$~t~P:-;-~C~_: L~:"- -51 'r'i ' ~U ~r?~I~L rll y

*:i:~CillE ~E .~ ,?

~?. Ij;Z~~ ?~ ?:. :] :.;:-?~?-~????'b?? :?:*j,.*I I~if~:'l

~Q--~Sp~rq *.i 'r ? :?9'~" .C.:" '~.*?! ~?? :i -~??.: .r? Y-?: ?. iP

;s h?:

?::?

io. Leonardoda Vinci,St Anne,Louvre.

This content downloaded from 128.111.215.12 on Thu, 27 Feb 2014 13:46:34 PM All use subject to JSTOR Terms and Conditions for his discoveriesand a parallel (but not identical) developmenttook place independentlyin Venice. In conclusion,therefore, a briefsketch of this situation,and also of the immediatereflexes in Floren- tine painting,is necessary.In doing this I am not attemptingto sketchany sort of developmentin Florentinecolour, but only to round out more fully the characterof Leonardo's achievementwhich, like any otherhistorical phenomenon, gains meaningfrom a considerationof its context. It is unlikelythat such a profoundstylistic change should be unaccompaniedby comment,or by a sympatheticrevolution in taste, and some documentationon both counts may make this enquiry seem less like an analytical exerciseunrelated to historicalevents. The S t. A nn e in the Uffizirepresents only Masaccio's starting-pointin his handlingof colour. It is probable that his last work known to us is the Cathedra Petri (figure II) in the Brancacci chapel, and here the differenceis remarkable;the impetusof Masaccio's developmentwas so great that the contrastsexist even betweenthis and his earlierfrescoes in the same cycle. The point is simple: not only is there no substantial differencein the plastic intensityof blue and yellow draperieson the figureof St. Peter, but the same is true of flesh-formsand, most remarkably,of the white-habitedkneeling figures before him; muchof Leonardo's spatial correlationof volumes by tone is also prefiguredhere, and of course it is in these frescoes(for the same set of reasons) that appears the firstinstantaneously-felt pictorial light of the Renaissance. However, the extent of Masaccio's adumbrationof Leonardo's early use of colour seems to be somewhat exaggeratedby the condition of the frescoes;overpainted they certainly are (how muchis not clear) and the superimposeddirt and smoke probably give a false impressionof an atmosphericchiaroscuro, and obscure one important fact: in those small areas that are free from this distortion66 it seems that colour-modellingis still predominantlyin use. Probably Masaccio modified this only by reinforcingthe tonally-weaker pigmentswith some achromaticaddition 67. It is importantthat in Masaccio's case no technicalrevo- lution accompanied this change,for this probably inhibitedfurther development in the Quattrocento, and surelymade it relativelyeasy for this demonstrationto be quicklyforgotten 68 It is possible that therewere even some precendentsin frescofor Masaccio. In Giotto'slatter chapels in S. Croce thereseems to be an appreciably greatercontrol over the propertiesof pigmentsthan was visible in the Arena chapel69.And, on reflection,it is strikinghow unifiedin tone is Cavallini's L a s t Judgement in S. Cecilia 70 AfterMasaccio, it seems that only Filippo Lippi attemptedto follow him, and that not for long; but Lippi's early colour-style may be significantsince it was firstof all a translation of the C a t h e d r a P e t r i experimentinto panel-painting(a step which,so far as I know, was not taken by Masaccio, and which does in some ways pose the greatertechnical problem) and secondly because the earlierstyle of Lippi as a whole is an importantpart of Leonardo's background71 It is possible that Masaccio's dramatic inventionis more than the solution to a practical aesthetic problem and was promptedby theoreticaloptical conclusions drawn by Brunelleschi.Certainly in Alberti's della Pittura the logical counterpartof his perspectivetheory is a light-shadowsystem which takes for granted the situation created practically by Masaccio, and this systemis entirely based on abstract, optical considerations.Alberti's theory of colore7" bears some relation to the exactly contemporarypractice of Filippo Lippi, but I doubt if the relationis one of cause and effect; the perspectivetheory could fairly easily be convertedto workshop practice,whereas the theoryof colore (in spite of the protestationsparliamo come pictore) is so totally unrelated to practical problemsthat it is not surprisingthat it found no interpreteruntil, possibly,Piero della Francesca, and Leonardo. It is no longer necessaryto demonstratethat Leonardo had digested della Pittura7 , and much of what he says on this subject is a criticismof it; but in his person was presentedthe

This content downloaded from 128.111.215.12 on Thu, 27 Feb 2014 13:46:34 PM All use subject to JSTOR Terms and Conditions 1I. Masaccio,St Peterenthroned, Brancacci chapel, Florence.

unique opportunityfor a full, understandingreconciliation of earlier theoryand practice, and this criticalsynthesis is his point of departurein the Uffizi A n n u n c i a t i o n. In Venetian painting from the 1470's onwards, in Bellini and Antonello, a development took place which is in some respectssimilar to Leonardo's; this appears to be promptedby Flemish art, where some of the simpler characteristicsof tonal unity appeared at least as early as the mature works of the Master of Flemalle. While the Venetian development seems to be independent of Leonardo, it is more difficultto assess the possibilityof Flemishart, especially in its technicalsolution to the problem,being to some extent a common source of inspiration7. While this possibilityis not to be excluded, it seems more likely that Leonardo draws the ultimateconclusions from propositions

This content downloaded from 128.111.215.12 on Thu, 27 Feb 2014 13:46:34 PM All use subject to JSTOR Terms and Conditions indigenousto Tuscan painting,and certainlyhis contributioncrystallizes the aims and ideals of the local monumentaltradition. Many of the characteristicsof Leonardo's colour reappear in the work of Perugino, and this is historicallyimportant since Perugino's influenceduring Leonardo's long absence in Milan in the 1480's and '90o's provided to some extent a colouristic substitutefor the latter which effects,for example, the earliest works of Fra Bartolomeo and Albertinelli75.But Perugino's understandingof the potentialof thisnew use of colour was limited; it went no furtherthan thecreation of plastically- coherentsimple forms,and exploits none of the volumetricand spatial possibilities,nor the compositional possibilitiesconsequent upon the liberationof colour fromthe provision of form,and shows no awareness of the rational or naturalisticproblems involved. Chronologically assessed,it is clear in any case that Peruginofollows Leonardo's firstdemonstrations in the 1470's. The firstFlorentine to explore and develop all the potentialitiesof Leonardo's colour-style(up to whose intellectual of that was remarkable.Fra Bartolo- 15o8) was Fra Bartolomeo, penetration style meo's own development cannot be discussed here, but an altarpiece such as the Accademia Marriage of St. Catherine (1512) can illustratewhat is necessary.The tonal unity of colours is of course complete and is the basis for the homogeneousrilievo throughoutthe work, and at the same timefor the instantaneousimpression of a powerfullight. The presenceof this light,and of its complementaryshadow, gives the coordinates of the spatial design with practically no assistancefrom linear constructions.Darkness is the condition of space without light, and the light penetratesselectively and draws out form into visible volume. Fra Bartolomeo at this stage makes full use of the new capacity of the handling of colour to vary the realization, or the tactility,of form,so that the sculpturally-plasticforeground figures contrast with those deep in the half-lightof the niche, as insubstantialas shadows. In the earlier quasi-Absolute colour-systema colour-composition automatically created a tone-composition,and the two were indivisible: the tonal composition was the product of colour and line (figure i). After the Leonardesque evolution tonal compositionsbecame the product of light, form,and space, so that colour, now independent,was free to surge dynamicallyover the picturesurface, as in the L a s t S u p p e r. In Fra Bartolomeo's altarpiece again the colour-compositionis independentof the tonal one, and plays principally the complementaryrole of articulating,and thereforere-asserting, the surface. The style of tonal unity in this direct Leonardesque form characterizesFlorentine painting in its most vital currentsat the end of the firstdecade of the Cinquecento and for most of the second; it characterizes,for example, not only the School of S. Marco (Fra Bartolomeo, Albertinelli,Sogliani and their followers) but also'the rapidly risingand ultimatelytriumphant rival, the School of the Annunziata, circulatinground the figureof Andrea del Sarto. The earliest works of Andrea, of who Pontormo,Rosso and Puligo fall withinthis definition7. It was, on the otherhand, theseartists progressivelyformed the most critical reaction to it, and the point of departurefor theirincreasing criticismwas a work which was both influencedby, and a protestagainst, Leonardo's colour-style. Michelangelo's D o n i T o n d o (figure12) provided for Florentinesan alternativeapproach to the problem which was very differentin character,chromatically more intense,and more directly based upon the Florentinetradition of colour-modelling. Leonardo's new mode of vision embracedphenomena which were strictlypictorial: light,corporeal space, and chiaroscuro,which brought a subjective flexibilityof form,so that its limits and even its substancemight be momentarilymade less tangibleor lost altogether.Michelangelo, by his whole training,saw formin termsof solid matihre,and was not interestedin phenomenaoutside the enve- lope of surface. Chiaroscuro would have been anathema to him at this stage, and the workshop practiceof his only masterin painting,Ghirlandaio, was already on these lines (throughindeed with

This content downloaded from 128.111.215.12 on Thu, 27 Feb 2014 13:46:34 PM All use subject to JSTOR Terms and Conditions F-.--"17:':j .?I~~C~: *' *..

'? S.: ? .~: ? ..?~: . . , ... . 1g

" " X dbi:.:4 ,,' ~ ~ .

?" .,

S ".. .i " ?

12. Michelangelo,The Doni Holy Family,Uffizi.

a ratherdifferent purpose). The derivationof the chromaticscale of the tondo fromGhirlandaio has oftenbeen remarkedupon and is precise; but the derivation is selective and suggeststhat Ghirlan- daio's role is that of a channelfor the Massacciesque. In the D o n i T o n d o pure colour creates form; if the chromaticelement were removed, all form would cease to exist. Tonal changes, and so modelling, are exclusively provided either by colour-changesor by saturation-changes.It is in characterwith the ideals of a sculptor that form should be interpretedas modulation of substance,in this case chromaticsubstance (a quality of the material), rather than as a visual impressionof the form,modified by exterior phenomena. The colour retains a degree of Absolutismwhich is at variance with the optical rationalismof Alberti and Leonardo. A corollary of the sculptural attitude to form is the will for complete clarity,and perfectlysharp and consistentrealization, of the surface; so one finds,here, the uniformlyprecise focus on all formsand a particularemphasis on theirlimiting outlines.

This content downloaded from 128.111.215.12 on Thu, 27 Feb 2014 13:46:34 PM All use subject to JSTOR Terms and Conditions The colour-changesin the D o n i T o n d o follow one Florentine tradition in exploiting the natural tonal differencesof pigments.Joseph's robe, lying over his knees, is the most striking example: a fully-saturatedyellow in the highlightsturns smoothly to a pure and luminouscoral-red which provides all but the very deepest points of shadow; here, only, the red is reinforcedwith a little rose-madderand brown. It is easy to findsimilar passages in Ghirlandaio and Filippino, but it is questionable whetherthey ever reach such vivid intensity.A second case, however, is in the draperybehind the nudes in the leftbackground: this turnsfrom green to pure rose in the shadows77. Both in the case of colour-change,and in the more frequenttransition in one chromaticvalue from a white highlightto a densely-saturatedshadow (Joseph's slate-blue vestment,the Virgin's lapis- lazuli robe and rose vestment,and also the flesh-tones),the forcefulnessof the plasticity depends upon the strengthof the colour; only the practical problemsof grinding,purifying, and media set a limit to thisplasticity, and this goes a long way towards explainingthe really exceptionalbrilliance of the colour in thispainting 78. All this is clearly very differentfrom Leonardo's practice; yet the two artists have one vital featurein commonin theiruse of colour, which I believe is an influenceof the older upon the youn- ger, however differenttheir methods. This featureis the common and unified("synchronized") range of tone of all colour planes. An indication of the importanceMichelangelo attached to this is the reinforcementof the red shadow of Joseph's robe, which provides an extremeshadow of a depth commensuratewith the blues and the rose-red,but which the yellow pigment would not reach. The functionof this tonal unity of the colours of the figure-groupis also the same as in Leonardo: to provide a uniformdevelopment of plasticitythroughout. The subsequent development of early Cinquocento colour in Florence oscillates between the techniquesof Leonardo and of Michelangelo.The situationis very confusing,particularly as Michel- angelo's later colour seemsto have moved significantlycloser to Leonardo's v,.But in general,Andrea del Sarto, Rosso and Pontormo all seem to have reacted against the restraintsthat Leonardo placed upon the purelychromatic attributes of painting,and each of them,from about 15 20 onwards,varied at will the proportionsof colour-modellingor of chiaroscuroin theirworks. However, no artistwho was trained in the Tonal Unity of High Renaissance painting could affordto forgetits formal and expressivepotential, and the many adjustmentsthat follow Leonardo do not contradict,but on the contrarydevelop further,his fundamentalaesthetic principles. In 1522 a group of artistsdiscussing the frescoesin the Carmine admired Masaccio's maniera si moderna.. nel coloritoso. It is probable that - at least in part - they were commentingupon the advanced tonal unity of his colour, as Vasari did when he remarked:dipinse le cose sue con buona unione e morbidezza, accompagnando con le incarnazioni delle teste e degl'ignudi i colori dei panni ... come fa il vivo e naturale ... '. This remark is a compressionof his advice in his della Pittura: Ne si debbono vestire gl'ignudi di colori tanto carichi di corpo, che dividano le carni da'panni, quando detti panni attraversassinodetti ignudi; ma i colori dei lumi di detti panni siano chiari simili alle carni, o gialletti,o rossigni... (etc.), purche tragghinoallo scuro, e che unitamente si accompagnino nel girare delle figurecon le lor ombre: in quel medesimomodo, che noi veggiamo nel vivo in his letterto Tintorettoof on the A o an r - ...82. Aretino, February 1545, p o 11 d M a s y a s, had already appreciated the same point:... lo intendeil vostro spiritointendente il dove si distendono i colori chiari e gli oscuri. Per la quale intelligenziale figureignude e vestite mostrano se medesimenei rilievi lor propri 83. Vasari on anotheroccasion explains that ... colori. . carichidi corpo, siccome usavano di fare gid alcuni pittori make the figurestoo obviously painted with pigments and prevent them being di rilievo e naturali84. More acutely still, he observesthat sharp contrastsof pigmentsmake a painting

This content downloaded from 128.111.215.12 on Thu, 27 Feb 2014 13:46:34 PM All use subject to JSTOR Terms and Conditions piz' prestoun tappeto colorito,o un paro di carte da giocare,che carne unita, o panni morbidio altre cose piumose,delicate, e dolci85. Carpets and playing-cardsdo indeed preserve,naturally, the qualities of Absolute Colour 86 In the Cinquecento,beauty of colour is no longer the quantitativebellezza. Dolce makes this point clear: Ne creda alcuno, che la forza del colorito consista nella scelta de'bei colori; come belle lache, bei azzurri, bei verdi, e simili; perciochequesti colori sono belli parimentesenza che e'si mettanoin opera; ma nel sapergli maneggiareconvenevolmente .... Altri . . non sanno imitar la diversita delle tinte de panni, ma pongono solamentei colori pieni, come essi stanno, in guisa che nelle opere loro non si ha a lodare altro che i colori87. Pino, characteristically,had already made the same point more pungently: ... non pero intendo vagghezza l'azzurro oltra marino da sessanta scudi l'ontia, 6 la bella lacha, per ch'i colori sono ancho belli nelle scatole da se stessi88. Before Pino, even, Aretinohad made the same point in 537: . . i miniatoritengono del disegno dei mastri de le finestredi vetro, e il far loro non e altro che una vaghezza di oltramarini,di verdi azurri, di lacche di grana e d'ori macinati, studiandosi in una fragola, in una chiocciola e simili novelluzze. Ma l'opra vostra e tutta disegno e tutta rilievo . . ogni cosa e dolce, sfumata,come fusse a olio89. This point of view is already seen in the passage of Leonardo's quoted above (p. xx): li colori sol fanno onore alli maestriche li fanno. This reversalof taste,which from the literaryevidence is as true of Venice as of Florence,has two facets: that true beauty does not lie in material richness,and that the proper businessof painting is the pursuitof light and shade, or rilievo. It seems probable that both ideas may reflectthe influence of classical aesthetics.The firstproposition is to be found already in Alberti'sdella Pittura in a form that presupposesan antique source90. Pliny's descriptionof Apelles's paintingmust have been particularly importantsince it is the only extended passage whereby the style of an antique painting could be visualized in colouristicterms: the final process was a glaze to prevent the brilliance of colour offendingthe sight,and from a distance subtly to give austeritatemto over-vivid colour91. Paolo Giovio's appreciationof Raphael's colour is expressedin thesePlinian terms92. Anotherremark by Pliny may be part of the foundation of the second proposition: Quod inter haec (lumen) et umbras esset appellarunt tonon.. (The relation of light and shade they called the strengthof pain- ting93.) This may explain Alberti's summaryof colore as ricevere il lume and Dolce's la principal parte del colorito (e) il contendimento,che fa il lume con l'ombra94 Justas Leonardo's effectiverevolution in colour was prefiguredin Masaccio, so the general change of taste in the Cinquecento was foreshadowedin the Quattrocento,most obviously in Alberti. When Albertisays, in the introductionto the della Pittura,that in Masaccio's paintinghe found a standard recreatedwhich was not exceeded by that of antiquity,he could only have assessed the latter on the literarytradition. I believe that what he meant,at least in part, was the rediscoveryof austeritatem (Manetti's puro, sanza ornato) and of rilievo or ricevereil lume. It is undoubtedlytrue that Leonardo expressed these two propositionswith even greatereffect. It is a possibility,then, that Leonardo's colour-stylewas not only the cosummationof a traditionwithin the monumentalart of Tuscany, and of the inheritenceof optical theory,but also of an ideal derived from the literatureof anti- quity95 Leonardo's greatestcontribution to Florentineart was energy,both formaland psychological.His developmentof tonal unity of colour goes far beyond its simplestresult: a coherenceof plasticity in the painted form which may rival that of sculpture; it gives energy to light and colour, and throughthe dynamicvariability of tactility,or plastic force,it contributesto the energyof forms.

This content downloaded from 128.111.215.12 on Thu, 27 Feb 2014 13:46:34 PM All use subject to JSTOR Terms and Conditions NOTES

1 It is a pleasure to recall two distinguishedexceptions, the more so since I feel incalculably but profoundly indebted to them: H. Siebenhiiner,Ober den Kolorismus der Friihrenaissance,Leipzig, 1935, and T. Hetzer, Tizian, Geschichteseiner Farbe, Frankfurt,1948 (2nd edition). 2 This article is a condensed and re-arrangedversion of one chapter from my thesis,"Developments in the Use of Colour in Tuscan Painting of the Early SixteenthCentury", delivered at London Universityin 1957; the substanceof it was given as a Public Lecture at the Courtauld Institutein the Summerof 1958. During the initial work, and in many subsequent conversationson this subject, I have been fortunatein, and most gratefulfor, the guidance and inspirationof JohannesWilde. 3 From the inscriptionon a portraitin the Castello dei Pio, Carpi, reproducedin L. Servolini,La Xilographia a chiaroscuroitaliana ..., Lecco, 1930. 4 References to the various publications of Leonardo's writings have been abbreviated as follows. - The Trattato, or Libro della Pittura (Codex Vaticanus 1270), as "Trat."; the numbersrefer to the paragraphs in the edition by H. Ludwig, Lionardo da Vinci, Das Buch von der Malerei, Vienna 1882. "Mc. C.", followed by a numberrefers to the page in E. McCurdy, The Notebooks of Leonardo da Vinci, London 1938 (2 vols.). "Richter", followed by a numberrefers to J. P. Richter,The LiteraryWorks of Leonardo da Vinci, Oxford

1939 (2nd edition). 5 Trat. i i i and 133. In one of the Paragone texts (Trat. 3Pa) he says, speaking of sculpture,Questa anchora non e imitatricede' colori, per ii quali ii pittore si affatica a trovare, che le ombre sieno compagne de' lumi ..., which is very close to Alberti's broadest definitionof colore: ricevereil lume. Alberti,defining the three parts of painting, states: Ultimo, piu distinto determiniamocolori et qualita' delle superficie,quali ripresentandoli,che ogni differenzanasce da lumi, proprio possiamo chiamarlo receptione di lumi. (L. B. Alberti, della Pittura, ed. L. Mallk, Florence, I95o, p. 82.) 6 This was the method used with strikingresults by Siebenhiiner,Loc. Cit., for the analysis of earlier quattrocento painting. Another line of enquiry, which for practical reasons I have had to omit, is the changes in the use of colour broughtabout by changesof technique. 7 One example: the St Ambrose (M. Salmi, Masaccio, Paris, 1934, pl. CLXXIX) wears a robe with turquoise highlightsand deep wine-red shadows, and a lining turningfrom yellow to red. The principle is applied to architectureas well as to figures.Colour-changes are frequentlyinterpreted as shot-silk,which is very seldom correctin Renaissance painting. 8 A new analysis of the originsof this painting is given by G. Passavant, "Beobachtungenam Verkiindigungs- Institutesin bild aus Monte Oliveto", Mitteilungendes Kunsthistorischen Florenz, February 1960, p. 7iff: that the compositionwas by Domenico Ghirlandaio, and largely overpainted by Leonardo. This hypothesis seems to me convincing.The colour remains in all essential parts Leonardo's, and is unimaginable in Ghir- landaio. 9 With the exceptions,especially in the fleshin some parts, mentionedbelow, p. 22 and n. 3 5a. ro I hope is clear that in this analysis "colour" means the chromaticconstituent of the wider complex which Leonardo would have called colore, which would include, of course, the neutral chiaroscuro.From his point of view (in the synthesis,as opposed to the analysis) these elementsremain indivisible. 11 cf. della Pittura, ed. Mallk, p. 56. 12 (I) For the differenttypes of lighting,cf. for example Ms. B. N. 2038, 29r. (McC. II, 265). (II) The different degrees of contrast produced: Trat. 120, 86, 712 (B. N. 2038, 33v.) and E 32v. (McC. II, 346). (III) The necessityof using a light appropriate to the environment:Trat. 94, 498, 414 (Ms. B. N. 2038, 33r., McC. II, 272-3), G 33v. (McC. II, 352) etc. (IV) No reflectionsin a dark environment:Trat. 157. 13 cf., for example, E 32v., (McC. II, 346) and Trat. 712. No opinion seems to be datable throughthe Mss. 14 Trat. IIo. S5 cf. Trat. 454 and 886 (Ms. G 24r., McC. II, 297). i6 Ms. C. A. 25oa (Richter,I, iii). The whole passage, a Proemio to the projected books on light and shade, is particularly relevant to this discussion. 17 Trat. 213 (cf. also 207). i8 Trat. 433, 514 and especially 703, where colour-changesare excluded and black alone is to provide the vera ombra de colori. cf. also Trat. 706: Del colore de' l'ombre, e quanto si scurano. Si come tutti Ii colori si tingono nell'oscurita'delle tenebredella notte, cosi d' l'ombra di qualonche colore finiscein esse tenebre. adonque tu, pittore,non osservare,che nelle ultime tue oscuritds'abbia a conoscere li colori..

This content downloaded from 128.111.215.12 on Thu, 27 Feb 2014 13:46:34 PM All use subject to JSTOR Terms and Conditions '9 Ms. E I7v. (McC. I, 258; see also Trat. 201). 20 Two apparent exceptions: (I) Richter I, 265 "A shadow is always affectedby the colour of the surface on which it is cast", is a mistranslationof W. 19 076a: L'ombra participa sempre del color del suo obbietto. Obbietto, in the optical theory of Alberti and Leonardo, means an object placed opposite the surface in question; this is thereforeone of many texts on reflectedcolour in shadows. (II) Trat. 592 (Ms. E 30v., McC. 1I, p. 343): Qualita dellonbre. Infralle equali alleviatione diluce tal proportione sia da osschurita asschuritadelle generateonbre Qual sara daosschuritaasschurita delli choloridove tali onbre sichongiunghano. (From the facsimile in C. Ravaisson-Mollien, Les Manuscrits de Leonardo de Vinci, C, E, K, de la Biblio- theque de l'Institut,Paris 1888.) Generate onbre are cast shadows, not body-shadows, and this text refers to their apparent change in value according to the relative darkness or lightnessof adjacent surfaces,an abstract optical problem which need not concern us here. 21 See above, p. 17. 22 B. E. g., Trat. 207, and Ms. N. 2038, 33r. (McC. II, 273-4). In the same Ms., however (3Iv., McC. II, 272), it is stated that the part which is between the light and the shadow is pi colorita. (Both these texts were included in the Trattato, 21o and 419 respectively.)Ms. B. N. 2038 is datable c. 1492. The firstconclusion is repeated in the much later Arundel Ms. (B. M. 263, I69r., McC. II, 379) and in numerousother places. 23 Ms. B. N. 2038, 26a (McC. II, 258) and Trat. 492. 24 e. g. Trat. 713b. 25 Trat. 206. 26 Trat. 692; for the use of the word bello here, see n. 44. 27 At firstsight the settingof this head against the dark backgroundis deceptively like the Angel's silhouette in the Annunciation (cf. Figs. 3 and 5). But in this case the light falls full on the face from the right, with the purpose - one imagines for personal reasons - of reducing shadow to a minimum;the dark foil does not delineate the profileof the head, as in the angel, because the head is at no point actually in contact with it. On the contrary,the enclosing hair is related tonally to the dark foil. 28 For example, the most advanced stages of the Madon na with the Cat (A. E. Popham, The Drawings of Leonardo da Vinci, London, 1947, Nos. 9A, which shows a window behind, and 9B, which simply has a dark background). The dark foil is designated campo by Leonardo: see Trat. 120. 29 Which may also be, as it is sometimesinterpreted, the Virgin's Cloth of Honour. 30 The two are reproduced togetherin E. Hildebrandt, Leonardo da Vinci, der Kiinstler und sein Werk, Ber- lin 1927, PP. 244-5. 31 Vasari, Vite.., 1568 edition, ed. Milanesi, IV p. 26; the referencecomes between commentson the M a - donna of Clement VII, and the Adoration. This is not a suitable 32 place for yet another discussion of these documents.For the point of view expressed here, see G. Castelfranco in Raccolta Vinciana XVIII, my commentsin the BurlingtonMagazine, Novem- ber and the I96I, p. 475, ensuing correspondencebetween Cecil Gould and myselfin the BurlingtonMaga- I zine, January 1962. If, as believe with Castelfranco and others,the documentsestablish that two paintings were made in response to the contract of 1483, then the first(the Louvre version) must be la dicta nostra dona facta a olio per lo dicto florentinomentioned in the Supplica of the early '90o's. 33 For the provenance of the London version see M. Davies, in the Catalogue of the Earlier Italian Schools, National Gallery London, i95i, p 211 ff. In spite of the unusually positive documentationthat the London version is Leonardo's work, this is denied in many quarters. My acceptance is based partly on the documents and partly on a visual estimateof its quality; moreover,it differsin style,compared with the Louvre in a that is consistentwith Leonardo's version, way development between 1483 and I5O8 (as I hope I can demontratepartly here) and I do not see how these stylisticadvances can logically be attributedto Leonar- do's followers.See also note 58. 34 This comparison is suggestedby the small sketchat Windsor (Popham, Op. Cit., 16o) which seems to come between the two. 35 Popham I59. 35aIt will be noticed that the unity of tone is frequentlybroken in the flesh;two possible explanations for this appear in the notes. Leonardo quotes the fleshof young people as a case in which heavy shadows should be avoided, apparently as an exception to the regular depth of shadow (Trat. 419, B. N. 20o38,31 v., McC. II, 272). According to his optical theory,reflections will be most effectiveon the pale surface of flesh (Trat. 160 and 162-3) so it follows that only in a very restrictedlight will the degree of relief be unaffected, and equal to that of darker drapery. His painting becomes progressivelymore logical in this sense. Alter-

This content downloaded from 128.111.215.12 on Thu, 27 Feb 2014 13:46:34 PM All use subject to JSTOR Terms and Conditions natively, another observation may apply in the An nunciation and in the Louvre Madon na of t h e R o c k s: no reflectedlights will appear in shadows, even in the open air, near corpi ombrosi,come .. prati di varie altezze d'herbe (Trat. I57). In the A n n u n c i at io n, according to Passavant's analysis (above, n. 8), the Angel's head was less repainted by Leonardo than, for example, his left hand; it is questionable how much,in that case, this was consciouslyan application of the principles mentionedhere. 36 The design of the original a n c o n a is far fromclear, but it seems probable that it provided for the central panel a settingwhich was, in a sense, cavern-like.The M a d o n n a was enclosed by shutters(in the Natio- nal Gallery); moreover I believe that the absidem .. que vadit ante imaginem beatissimevirginis marie ad modum incastri (from the payment to the carpenter del Maino, 7th August 1482) was a baldacchino (Du Gage, Absides, in Glossarium Mediae et Infimae Latinitatis), like the vaulted canopies fairly frequently found in North Italian elaborately-carved ancone (such as Foppa's polyptych in S. Maria in Castello, Savona) and oftenplaced over the image of the Virgin. A baldacchino is mentionedas part of the a n c o n a in question in the inventoryof 1781. I wonder if Leonardo did not exploit such a situation in the second version having seen the firstin position. 37 Popham 162. This drawing has been described as a fake by A. M. Brizio in Raccolta Vinciana XVIII; I believe it is either an authentic,but routine,drawing by Leonardo, or a pupil's copy of one (see the Bur- November lingtonMagazine, I96I, p. 475). 38 Bodmer,Klassiker der Kunst, Leonardo, 1931, p. 362/3. 39 There are a few referencesin the notes to this three-dimensionalaspect of chiaroscuro,of which the most notable are Ms. G 19r. (McC. II, 283-4) and Trat. 440, which is virtually a text for the chiaroscuroof the Battle of Anghiari. 40 In the version that we know; at the same late date Leonardo drew the extremeconclusions fromthe dark setting in the Louvre B a p t i s t, in which it appears to me that the shadowed profile is in some places physicallynon-existent on the panel. 41 Vasari, ed. Milanesi, IV. p 5o. 42 Sabba da Castiglione, 1554; compare Paolo Giovio, Leonardi Vincii Vita (soon after 1527): Plasticem ante alia penicillo praeponebat, veluti archetypumad planas imagines exprimendas,and the anonymous sonnet appended by Vasari to the 1550o edition of the Vite: Perspicuas picturae umbras,oleoque colores, Illius ante alios docta manus posuit. 43 Trat. 123. 44 On the contrary,cf. one of the Paragone texts (Trat. 45): Se il bronzo rimano nero e bruno, questa pittura e piena di vari e vaghi colori. For other examples of bellezza in this sense, cf. Trat. 191 (in which stained glass is said to give maximum bellezza) and Ms. E i8 a (Richter I, 286). Naturally Leonardo still uses bellezza as beauty (for example Trat. 36). Michelangelo also uses the word in the firstsense; at the commen- cementof the Sistine Ceiling he writes to Frate Jacopo of the Gesuati, in Florence: m'e di bisognio di cierta quantitd d'azzuri begli .. and particularlyinsists che sieno begli (Milanesi, Lettere,p. 379); this was quoted should be translated or by Tolnay, Michelangelo,II, Princeton 1945, P. i85, but I feel sure that begli "pure" "brilliant", ratherthan "beautiful". 45 See below, p. 41. 46 Trat. 19o a. 47 cf. Bandello's eyewitnessreport reprintedin Leonardo da Vinci, La Vita di Giorgio Vasari, ed. G. Poggi, Florence 1919, p. 2 1-2. 48 Vasari, ed. Milanesi, IV, p 43. 49 G. Gaye, Carteggio Inedito.., Florence 1840, II, p. 89. Leonardo's invention defied descriptionin normal bureaucraticterms: ... 4 ruoteper fare il carro a Lionardo da Vincio overo ponte. to have been considered cf. Carlo 5o The Last S up p e r seems always unusually "colourful"; Torre, II Ritratto di Milano, 1674, p. 164 (.. veggonsi ancora vivi sembianti, figure in iscorsi sforzosi, colori risplendente. .). sr On the leRft,green against red, yellow against blue, etc., on the right,soR blues, pink-greys,lilac etc. 52 Compare the two ends of The Raising of the King's Son in the Brancacci Chapel, the right third of which is normally in shadow (and would have been from the original window). Leonardo would have rememberedthis fresco without its portions by Filippino, so that the contrastwould have been at its most obvious. 53 For estimatesof the original colours remaining,compare F. Wittgens,Il restaurodel Cenacolo di Leonardo,

This content downloaded from 128.111.215.12 on Thu, 27 Feb 2014 13:46:34 PM All use subject to JSTOR Terms and Conditions Atti del convegno di studi Vinciani, Florence 1953, p. 39, and Arturo Bovi, La visione del colore e della luce nella Cena, Raccolta Vinciana XVII, 1954, p. 315. 54 Bovi judged that the blue is still somewhat overpainted,whereas Wittgensinferred that this had been removed; my personal impression,from the ground, is that there is a good deal of old and new paint remaining. 55 cf. Bovi: Frammentidi coloresi scopronodello sfacellodella veste.. e risplendonodel verdechiarissimo originarioche si riflettenei tonidelle argentee stoviglie sul tavolo. 56 The settingaround the figuresis predominantlyneutral in colour, but the whole palette, with special emphasis on the reds and greens,is repeated in the wreaths in the lunettes above. The distant view in the centre (which is said to be free from overpaint), seems also to pick up the paler blues and greens of the figuresin the normal way. 57 From the undated S u p p li c a of the early 'go's; Leonardo's authorshipmay be deduced from its phrase- ology. In this case he was referring,I believe, to the Louvre version,but it is clear fromthe documentsthat there was no change of subject between the two versions, since both were made in response to the same contract,without amendments. It is already true of the Louvre version that, when it is seen in a light approximatingto that for which it was intended,the increased relative intensityof the blue makes the Madonna the true subject of the picture, but the intensificationof this effect,and the sacrificeof all inessentialdecorative colour in the second version, is yet anotherexample of Leonardo's deeper appreciation of the painting'sdestination. The explanation of the changed appearance of the paintings in lower illuminationis to be found in what is known as the "Purkinje effect"(C. Ladd-Franklin, Colour and Colour Theories,London 1932, p. 58); the importanceof this for the study of paintings seems not to have been appreciated. The decrease of illumina- tion on a complete colour-scale results in a shiftof emphasis, (or of relative intensity)from the red end of the spectrumtowards the blue. In paintings, therefore,the distortionsof modern gallery-lightingmay affectthe symbolismof colour, the relative plasticityof colours, or a finely-balancedcolour-composition. 58 The painting of this lining to the robe is worth special attention.In Leonardo's notes it is observed that a coarse texture(like the blue robe) will have no "lustre"; the extremehighlight of a smooth silk-texturewill be entirely lustre, and the highlightproper will be introduced, along with the colour, around this. The colour-reflectionswill appear to intensifythe local colour in half-shadowswhere they are reflectedfrom one part of a formto anotherpart of the same (Trat. 223-4, etc.). In the London version,the fluctuatingvalue of the colour in the silk lining is remarkablysensitive in its exact analysis of the source, nature, and colour of the light falling on its sharp folds. It is a measure, as true as the notes, of the completelythree-dimen- sional perception of light at this stage. It seems to me that details like these support the attributionto Leo- nardo of this version; not only is the practice intimatelyrelated to the novel analysis in the notes, so that we seem to see one mind at work, but also no work by any pupil that I have ever seen displays a comparable intellectualgrasp of suchproblems. 59 See the evocative descriptionin Sir Kenneth Clark's Leonardo da Vinci.., Cambridge 1952, p. i16. In fact the impressionof vitality given by the M o n a L i s a (and, one must imagine, the L e d a) became an importantfactor in the style of Andrea del Sarto and of his followers. 6o In the case of the L eda and of the Madonna of the Yarn-winder, both compositionsin an exterior setting,we may infer in the lighting a respect for these conditions partly from the copies, and partly from a considerationof the converse of the argumentabout M o n a L i s a. Exactly how this was done cannot be determined,but that it was done in the second Florentineperiod is historicallyimportant. 6r Hetzer, Op. Cit., p. 133. Leonardo's departure fromthis point was not, of course, so complete or so important as Titian's 62 The condition of this painting, in some ways unusually satisfactory,requires a little mental reconstruction, not only in those unimportantparts where the degree of finishis uncertain.The most importantchange is in the Virgin's robe, which has lost its tonal qualities owing to the phenomenonof "Ultramarine sickness".It may be reconstructedas tonally unified,also in the shadow, with the other draperies, on the basis of the Uffizi copy; the tone of the highlightshas changed little, if at all. There is considerable overpaint on this drapery,especially on the hip, and this attemptsto rectifythe loss of the shadows. 63 Perhaps the most importantearly example is Bellini's S. Zaccaria altarpiece, though it had already appeared in some parts of the Pesaro C o r o n a t io n, and in Carpaccio's painting of the '90's. Fra Bartolomeo broughtback this device fromVenice, among other Venetian influences,and it is to be found spasmodically- used in altarpieces by him and by Albertinellifrom I509; the Florentinewho firstexploited its potentialities was Andrea del Sarto. Michelangeloalso uses it in the second half of the Sistine ceiling.

This content downloaded from 128.111.215.12 on Thu, 27 Feb 2014 13:46:34 PM All use subject to JSTOR Terms and Conditions 64 For example St Anne's leftsleeve: grey,variable over a short range fromyellowish to bluish; its exact value is never established, but it makes an effectivecontinuity with the lamb and with the hills in the left background. 65 e. g. the studies for the Trivulzio Monument,or the W i n d s o r A 11 e g o r y (Popham 102, 123). 66 For example, the small area revealed by removal of the baroque altar-frame,reproduced in M. Salmi, Masaccio, Cappella Brancacci, Milan, (U. N. E. S. C. 0.), 1949, pl. XVI. 67 The increasingrationalism of Masaccio's light is seen in the progressiverejection of colour-changein modelling; in the earlier frescoesyellow is oftenshaded with red, whereas in the later ones it is shaded towards grey-brown. 68 How completelyit was forgottencan be seen in the Brancacci Chapel itself,comparing Masaccio's frescoes with Filippino's of the 1480's. It is impossible for me to mention here all the many exceptions to colour- modelling that exist even within Florentinepainting of the Quattrocento. None of them,to my knowledge, anticipate Leonardo in any significantway, with the possible exception of the technique of the Pollaiuoli which,while not being directed towards tonal unity,may have been importantas a precedent for the infu- sion of neutral pigmentinto colour. There may well be a technicallink between such works as the Me r - canz ia Virtu es and the Angel in the S. Sal vi Baptism (Uffizi). 69 Again, this is seen partly in the rejectionof colour-modellingbetween the Arena and Bardi Chapels. 70 I owe this commentto JohnWhite. 71 Compare, for example, the Angel in the S. S a lvi B a p t is m with the leftkneeling Saint in the Barba- dori Altarpiece; similarly,Fra Bartolomeo's debt to this source is demonstratedby the relation of his mature drapery-style(for example in the 1509 S. Marco Altarpiece) to Lippi's in the S. Lorenzo An nu n - c ia t io n. One of the Angels in the Barbadori Altar is exactly repeated in an early Ridolfo Ghirlandaio, an A d o r a t io n once in Berlin (S. Freedberg,Painting of the High Renaissance in Rome and Florence, Cambridge 1961, II, fig.278). 72 The most thoroughanalysis is still Siebenhiiner,Op. Cit. 73 See, most recently,V. P. Zoubov, Leon-Battista Alberti et Leonard de Vinci, Raccolta Vinciana XVIII, 1961. 74 Leonardo's interestin Northern technique is shown by his note on Jean Perrial, Ms. C. A. 247r. (Richter 1379; for the date - either1494-5 or 1499 - see G. Calvi in Raccolta Vinciana, III, p. 99). 75 E. g. the A n n un ci ati on in the Duomo at Volterra, 1947, by Fra Bartolomeo,or the Holy Fam i ly tondo in the Pitti by Albertinelliof about the same date. 76 For Andrea, the Noli me tangere of c. I5io in the Uffizi; for Pontormo, the Madonna from Poggio Imperiale (now in the Uffizi,Mostra Pontormo, 1956, No. 14), c. 1515; for Rosso, the M a d on n a for the in the Staedel Institute,Frankfurt, c. 1515 (K. Kusenberg,Le Rosso, Paris 1931, P1. XII); Puligo, M a g d a 1 e n in the JohnsonCollection, Philadelphia (also c. 1515). 77 This example is particularlyclose to the upper-rightangel in Masaccio's S t A n n e. I have not mentioned in the text the colour-changeon the lining of the Virgin's robe, which is an exemplary though rare case of the descriptionof shot-silk. 78 Michelangelo was insistenton the perfectionof his materials in painting as well as in sculpture; cf. the lettermentioned above, n. 44. 79 The direction seen in the later parts of the Sistine Ceiling was probably continued in the L e d a of 1530o; to judge from the probable influenceof the latter upon Pontormo's N o li m e T a n g e r e, this was the case. For an assessmentof the pictorial qualities of the L e d a see J. Wilde, Notes on the Genesis of Memorial for Fritz London Michelangelo's"Leda", Essays Saxl, 1957, p. 279- 8so Vita of Perino del Vaga, Vasari ed. Milanesi, V, p. 604. 8i Vasari, II, p. 288. 82 Vasari, della Pittura,Cap. IV, ed. Milanesi, I, p. 18o. 83 Aretino,Lettere sull'Arte, ed. F. Pertile and E. Camesasca, Milan 1957, II, p. 52. 84 Vasari, della Pittura,Cap. IV, I, p. 179. 8s Vasari, Loc. Cit. 86 The same is inevitably the case with tapestry; Armenini (II, 2 and 7) condemns the colours of tapestries as confusingand jarring. cf. already Giulio Romano in his letter to the Steccata, 1541 (F. Hartt, Giulio Romano, New Haven 1958, p. 249). 87 L. Dolce, Dialogo della Pittura ed. P. Barocchi,Bari 1960, p. 184-5. (x557), 88 Paolo Pino, Dialogo di Pittura,Venice 1548, p. 18. 89 Aretino,ed. cit., I, p. 45.

This content downloaded from 128.111.215.12 on Thu, 27 Feb 2014 13:46:34 PM All use subject to JSTOR Terms and Conditions 90 Alberti,della Pittura,ed. Mall6, p. 76. 9' K. Jex-Blake and E. Sellers, The Elder Pliny's Chapters on the History of Art, London I896, p. I32. If it seems improbable that a text in Pliny can affectRenaissance style,this at least was not the view held in the Quattrocento. Facio (De Viris Illustribus,before 1457) claimed that Jan van Eyck learnt the proprietates dei colori from reading the ancient authors, especially Pliny: see J. Schlosser Magnino, La Letteratura Artistica,2nd ed. Florence and Vienna 1956, p. i i o. 92 V. Golzio, Raffaello nei documenti..., Citta del Vaticano, 1936, p. 192. 93 Pliny XXX, 5o; Jex-Blake and Sellers, p. 96. The word tonos cannot really be translated: "tone" is certainlyinadequate, and "relative strength"may be nearer. 94 Dolce, ed. cit., p. 183- 95 Bandello, as an eye-witness,records a conversation between Leonardo and the Cardinal of Gurk, on the possibility of modern paintings rivalling those of antiquity che tanto da i buoni scrittorisono celebrate (Poggi, Op. Cit., p. 22). Leonardo was compared with Zeuxis and Parrhasius as early as 1497; see Luca Pacioli, de Divina Proportione,ed. C. Winterberg,Vienna 1896, p. 41.

This content downloaded from 128.111.215.12 on Thu, 27 Feb 2014 13:46:34 PM All use subject to JSTOR Terms and Conditions