Color Vision in Art and Science Oftrint from ColorVision Perspectivesfrom Different Disciplines 'O 1998Walter de Gruyter& Co.,Berlin-NewYork 1. Aging through the Eyes of Monet

John S. Werner

l.l Introduction Abstract Expressionism.In science,the physical principles pertaining to light and color laid down One of the most eventful periods for our under- by Newton in the preceding century (Newton, standingof color, both in art and in science,oc- 1704) were used to discover processesof color curred between 14 November 1840 and 5 De- coding by the eyeand brain. In short,the way both cember 1926 - the life span of Oscar Claude artists and scientiststhink about color today was Monet. In art, Monet's life encompassedthe peri- shapedfrom 1840to 1926to a degreethat may be od betweenthe Romantic pictorial tradition and unparalleledby any other period of 86 years.

Fig. 1.1:Claude Monet (1872) Impression: Soleil levant (Le Port du Havrepar la brume).[Impression: Sunrise (Portof Le HavreThrough the Mist.l Oil on canvas,48 x 63 cm. (After restoration.)Mus6e Marmottan, Paris. (Photocredit: Giraudon/Art Resource, New York.) 1. Aging through the Eyes of Monet

Fig. 1.2: Pierre Auguste Renoir (1875-76) Torse de femme au soleil. [Tbrso of a Womanin the Sun.J Oll on canvas,81 x 65 cm. Mus6e d'Orsay, Paris. 1.2 A Link betweenSunlight and Aging

Art and sciencehave at least one purposein and so too did the way he portrayedthe world. One common: to enrich the human spirit. Whatever mustadmit, of course,that changesin Monet'svi- elseone might sayabout Monet, he has certainly sion are confoundedby changesin his style of' enrichedour civilization.For when he unveiledthe painting, notwithstandingthat his statedgoal was painting shownin Figure L I at the first exhibition always to portray the subtle modulationsof light of the Soci6tdAnonyme des Artistes in 1874,he without interpretation.Monet once said: became the de facto leader of a movementthat When you go out to paint, try to forget what ob- rvouldalter the courseof Westernart history.This jectsyou havebefore you, a tree,a house,a field painting was originally called "The Port of Le or whatever.Merely think, here is a little square Havre" but is now known by its subtitle ofblue, herean oblongofpink, herea streakof I "lnrpression:Sunrise," from which a schoolof art yellow,and paint it just as it looks to you, the I u'as given its name,Impressionism. The picture exactcolor and shape,until it givesyour own captures what we now expect from an Im- naive impression of the scene before you. pressionistpainting light, atmosphere,color and (Perry,1927, p. 120) movement,all in the serviceof renderingthe feel- Monet's changingportrayal of nature throughout ingsof the moment. his life has drawn attentionto importantprocesses The Impressionistswere individualists, with dif- of visual aging but has also perpetuatedmyths ferentstyles, preferred subject matter, and aspira- aboutthe agingvisual system. The purposeof this tions.What united them, however, was a rebellious chapteris to offer a personalinterpretation of spirit againstthe ParisSalon and a desireto cap- color scienceand art in Monet'slifetime, with an ture the fleeting effects of light and color. analysisof his agingeye as it maybe derivedfrom Consider Pierre Auguste Renoir's Torso of a his art and asrelated to currentresearch on senes- Ilbntanin the Sun (Fig. 1.2)presented in the sec- cenceof humancolor vision. ond Impressionists'exhibit in 1876.Although he preferredto paint the humanform, he did so in a n,aythat captured the delicateshades and shadows E thatwere previously not recordedon canvas.But it 1.2 A Link betweenSunlight I was not just the handlingof light and color that andAging made the movementcontroversial; those pretty picturesof the Impressionistshad said"No" to the PaulC6zanne once remarked (Barnes, 1990, p. 6) classicalpictorial tradition. Great art no longer that "Monet is just an eye but my god what an hadto depictkings, popes and saints; ordinary ex- eye!" The humaneye is shown schematicallyin periencewould do. Figure1.3. Light, if it is to be seen,must first travel Someoneonce asked Renoir how it is thathe ob- throughthe variousocular media, the cornea,the tained the delicateflesh tones of his nudes for anteriorchamber filled with aqueous,the lensand shich he becamefamous. and he said in effect.I the vitreoushumor. It thenpasses through the lay- just keep painting and painting until I feel like ersof cellscomprising the retina,shown in an en- grabbing (Vollard" 1925).When pushed further largedview, where it can be absorbedby the rods about the possiblescientific basis of his tech- andcones, the receptorcells that initiatevision. niques,Renoir said that if anyof his work couldbe The clinically normal eye appearsrather stable subjectedto scientificanalysis, he would not con- overmuch of thelife span.Barring disease or trau- siderit art. Sucha reactionis not atypicalin the ma, senescentdeterioration is seldomnoticed until historyof art, but it is somewhatatypical for the mid- to late-life.At first glance,then, aging of the lmpressionists.Many of them had a deep and eye is a phenomenonof later life. Unfortunately, abidinginterest in color science.Camille Pissarro, first impressionscan be quitemisleading. A clos- for example,studied scientific literature in order er look at the visual systemshows that it is con- to perfecthis useofcolor. stantly changing throughout life (Weale, 1982; The eyesof Monet changedover his life span, Werneret al.. 1990). I . Aging through the Eyes of Monet

. Fovea/ Lens

Fig. 1.3: Schematic cross-sectionof the human eye with the retina shown in an enlarged view. The ocular media include the cornea, aqueous contained in the anterior chamber, lens and vitreous humor. The retina, shown in a magnified view, includes five principal cell types, photoreceptors (rods and cones), horizontal cells, bipolars, amacrines and ganglion cells (the axons ofwhich form the optic nerve).

One factor that is believedto contributeto age- than the usual cumulative exposureto sunlight. relatedchanges in the eyeis exposureto light itself Even as early as 1867, at age27, Monet had troub- (Werner,1991). This factormay be especiallyper- le with his vision following hours of painting in tinent to understandingMonet. Although other sunlight, and he receivedmedical advice to aban- artistshad paintedin the open,Monet was perhaps don his outdoorpainting (Stuckey, 1995). Several the first to do so on a large scale and seemingly times thereafter he reported visual disturbances under all weather and seasonalconditions. His following a day of painting in the sun. careful observationsofthe varying effectsofsun- To understandthe effects of light on the eye, it light and his insistenceon painting en plein air is necessaryto define the spectrumof optical ra- virtually guaranteedthat he would receive more diation. The visible spectrum includes wave- 1.2 A Link betweenSunlight and Aging

A C

D E r

G H ar hl

It. b- in n- ''ia";;,, al K ES Fig. l.-1: Extractedlenses of humansat various ages:(A) six months, (B) eight years,(C) 12 years,(D) 25 lt rcers. (E) 47 years,(F) 60 years,(G) 70 years,(H) 82 years,and (I) 91 years.Also shown are three types of iirtrractoLrslenses: (J) nuclearcataract, age 70; (K) cortical cataract,age 68; and (L) mixed nuclearand cor, 'e- ticul cataract,age 74 years.(From Lerman, 1980.) 1. Aging throughthe Eyes of Monet

lengthsbetween about 400 and700 nm. At 400nm distributedthroughout the eye are variousantioxi- the light normally appearsviolet in the light- dant molecules(e.g., superoxide dismutase, o-to- adaptedstate, and shorterwavelengths are called copherol, glutathione, melanin, selenium and ultraviolet(or UV) light. Becauseof absorptionin ascorbicacid) that neutralize phototoxic reactions. the stratospheric ozone layer, very little light Especiallyimportant in this respectis thepresence below 300 nm reachesthe earth'ssurface so, for of the yellowmacular pigment around the fovea(a ( practicalpurposes, the UV spectrumof sunlight depressionin the retina wherethe conephotore- I encompassesthe rangefrom approximately300 to ceptorsare most denselypacked and which pro- ( 400 nm.At the otherend of the visiblespectrum at videsour bestspatial resolution; it typicallycorre- : 700 nm, the light normally appearsred under spondsto the center of gaze) which not only ( lightadaptedconditions; longer wavelengthsare reducesthe intensityof short-wavevisible light I calledinfrared. reachingthe retina, but which also consistsof Theenergy contained within a singlequantum is carotinoidpigments that areexcellent atneutraliz- inverselyrelated to its wavelength;quanta in the ing someof the phototoxicreactions that occurin UV may containenough energy to altermolecules the eye (Kirschfel4 1982).A secondline of de- in the eyethat absorbthem, primarily by initiating fenselies in theability ofcells to replacetheir parts a cascadeofoxidative reactions that are harmful to by molecularrenewal. Visual cells continuously re- cells.This type of light damageis usuallycalled constructor replacevirtually all of their partsex- photochemicalor actinic(Werner and Spillmann, cept DNA (Young,1982). As a result,damaged 1989).These photochemicalreactions occur as constituentsof cells are replacedin a piecemeal long as we are exposedto high-energyphotons fashion.A third defenseagainst the most damaging and becausewe are exposedto them from birth, wavelengthsof light resultsfrom the tendencyof we can be assuredthat cellular deterioration,or thesewavelengths to be absorbedby the ocular senescence,begins even from the first daysoflife. media,primarily the lens, before they can reach the Experimentswith nonhuman animals verify that retina.Figure 1.4 showsthat the lensbecomes an any wavelengthof light, in sufficient intensity,may evenmore effectiveabsorber of shortwavelength damagethe eye,but the shorterthe wavelength,the light asan increasingfunction of age.One cansee moreeffective it is. Forexample, light at 325nm in how clearthe lens is in thenewborn, and that it be- the UV is about1,000-fold more effective in dam- comesdistinctly yellow in adulthoodand brown in aging the photoreceptorsand retinalpigment ep- old age.Quantitative studies with largernumbers ithelium than light at about580 nm (usuallyap- ofindividualsreveal that the density(log ofthe re- pearing yellow) in the visible region of the ciprocalof transmission)of the ocularmedia in- spectrum(Ham et al., 1982).This darnageis not creasesas a functionofage from infancythrough fundoscopicallyvisible until about48 hoursafter the end of life (Werner,1982; Weale, 1988; exposure,indicating that it is dueto photochemical Pokornyand Smith, 1997). processesand not a burn.A retinalburn seldom oc- Figure 1.4illustrates common types of cataract. curs with naturallight exposurebecause there is Nuclearcataract, which is what ClaudeMonet ul- usually insufficient energyto raisethe temperature timatelydeveloped is shownby lensJ. Cataractis of the retinaby >10'C, the approximatethreshold only an extremeof normalaging; we call the aged for thermal damage.Exposures that are insuffi- lens a cataractwhen it interfereswith functional cient to reach the thresholdfor retinal damage vision. Considerableexperimental and epidemio- may neverthelessadd to the effectsof other expo- logical evidencehas shownthat lenticularsenes- sures,accumulating over time to producecellular cenceand cataractare, in part, due to the absorp- changesassociated with normal aging (Marshall, tionofhigh-energy photons ofUV (Young,1991). 1985;Wernerl99l). ln other words,exposure to sunlightaccelerates Undernormal circumstances the eyehas several agingof the lensand is one of the significantrisk naturaldefenses to protectit from the photochem- factors for cataract. ical insult associatedwith sunlight.For example, I .3 The Trivarianceof Color Mixture: Maxwell and Helmholtz i- 1.3 The Trivarianceof Color two reflectedlights will not be resolvedas indi- F Mixture: Maxwell and vidual spotsand will, in the words of Pointillist d Helmholtz painters,"optically blend." The mixturein this ex- s. amplewill appearachromatic (gray or white).The light distributionthat reaches the eyein this exam- ,a Oncelight reachesthe retina, it canbe absorbedby ple is equivalentto that obtainedwhen the same :- threedifferent classes ofcones, the photoreceptors broad-bandlight is passedin parallelthrough each )- of color vision. The foundationfor our under- filter so that both beamsenter the eye and are su- l- standingof these processeswas laid by James perposedat the retina.The resultantmixture with- ly Clerk Maxwell and Hermannvon Helmholtz in in the eye appearsneither yellow nor blue, but :It the mid l9th Century,although Thomas Young achromatic.Such pairs of lights,that can be mixed )f ( I 802) and othersbefore him (Weale,1957) had to appearwhite, are calledcomplementary lights. 1.- speculatedearlier that normal humanvision may There is a largenumber of complementarylight in be trichromatic.Maxwell and Helmholtz under- pairs,but they are mostconveniently found using 9- stoodthe differencebetween additive and subtrac- monochromaticlights, essentiallysingle wave- ts tive light mixture,a distinctionthat would be dis- lengthsof light (e.g.,470 and570 nm). More gen- coveredsomewhat later by the Impressionists. erally,three relatively arbitrarily chosen lights are K- Subtractivemixture is familiar to most people required to match any other light distribution. :d throughplaying with paintsin childhood.As illus- Physicallydifferent lights that appearidentical are al tratedby Figure1.5, the mixture of blueand yellow called metamers.The existence of metamers tg painttypically appears green. In this example,the showsthat the appearanceof a color is not expli- of bluepigment absorbs many of thelong-wave quan- cableon thebasis ofthe physicsoflight alone,but AT ta and the yellow pigment absorbsmany of the is dueto the processesthat the light initiatesin the he short-wavequanta. What reaches the eye is primar- eyeand brain. iln ily middlewavelengths, the bandthat is reflected Maxwell's(1860) studies of additivelight mix- th by both pigments.This is analogousto passinga ture carefully documentedthe proporlionsof three ee \vhitelight with all wavelengthsthrough two suc- lights requiredto matchan achromaticstandard. E- cessivefilters, a blue and a yellow.In Figure 1.5 He describedthe results by algebraicor colormix- in the blue filter transmitsquanta primarily of short ture equations,and becauseonly three variables irs and middle wavelengthswhile the yellow filter wererequired he could illustrate the results in a tri- €- transmitsprimarily the middle and long wave- angulardiagram. He realizedthat the trivariance of n- length quanta.Light of specific wavelengthsis color mixture impliesthe existenceof threekinds gh subtractedout at eachstage and all thatreaches the of color mechanismsin the eye.In relatedexperi- i8; c'yeis that which both filters transmit,the middle ments, Helmholtz showedthat any light of the navelengths,which we usuallycall green.In sub- spectrumcan be matchedby an appropriatecombi- ct. tractivecolor mixturethe resultis alwaysa lossof nation of threeothers. From this observation,he tl- light comparedto thatwhich wouldbe transmitted too correctlyconcluded that the retinal receptors of is treflected)by either filter (pigment)alone. This daylight vision, the cones,are trivariant.Helm- ed can be appreciatedby comparingthe individual holtz'sestimates of the relativesensitivities of the ral paintreflectances or the filter transmittancesin the three cone types presentedin his Handbuchder io- top row of FigureI .5,with theresultant subtractive PhysiologischenOptik (Helmholtz, 1867) are close is- nrixtureshown on the right of the middlerow to more modern estimates(Vos and Walraven. Considernow a caseof additivecolor mixture 1971; Smith and Pokorny,1975) such as those r). s hich canbe effectedusing the same blue and yel- shownin Figure1.6 for infantsand adults. les lo* paintsor filters shownin Figure1.5. ln the Although correctabout this fundamentalpoint, isk caseof paints,a blue spotis placednext to a yel- Helmholtztook anotherstep that wentbeyond his los spotso that light from eachis reflectedto the data.To accountfor colorappearance, he proposed c1ein parallel.If the spotsare small enough, the that the responseof eachclass of receptoris di- 10 1. Aging throughthe Eyes of Monet

Fig. 1.5: Top row: The number of quantaernitted frorn a hypotheticallight sourceis plotted as a fr-rnctionof wavelength.Glaphs in the middle and right show hypotheticalpaints and filters; the reflectanceaxes refer to the proportion of incidentquanta (plotted from 0.0 to 1.0) rellectedby the individual paints and the transmit- tanceaxes refer to the proportionofincident quanta(plotted frorn 0.0 to 1.0)trar.rsrnitted by the individual fil- ters. What is not reflectedby the paint or transmittedby the filter is shown on the right axes as absorption (plotted frorn 1.0 to 0.0). A blLrepaint (filter) containspigment that reflects(transmits) prirnarily short and middle wavelengthquanta, but absorbslong wavelengthquanta. A yellow paint (filter) containspigrnent that reflects (transmits)cluanta primarily at middle and long wavelengths,but absorbsshort wavelengthquanta. Middle row: Subtractivemixture using the blue and yellow paints (in a uniform mixture) or the blue and yel- 1.3 The Trivarianceof Color Mixtr,rre:Maxu,ell and Helmholtz ll

Fig. 1.6: Relativclog quantalsensitivity of thc threeclasses ol humancone photoreceptor.Sn-rooth firnctions shorv thc sensitivitiesof short-(S), middle- (M) and long-wave(L) conesin the adult(Vos. 1978),adjustcd in sensitivity accordingto the lessdense ocular media (Werner'. 1982) and absenceof rnac- ularpigrnent of infants.Squares show sensitivity of S-conesfront an infantob- tainedby Volbrcchtand Werner(1987). while white and black circlesshow sensitivitiesof an infant'sM- and L-concs,respectively, obtained by Bieberet al.(in press).

::.tlr linked to perception.Therefore, he labeled agesof l0 and85 years.All threecone types were 'r:: thrcc classesof receptorsas blue, green and foundto decreasesignificantly in sensitivityas a --'J For reasonsto be describedlater, this aspectof functionof age.A linearfunction describes the :. thcot'yis not correct and it is more accurateto datawell andthere is no statisticaljustification for ,i.cl thc receptorsaccording to their wavelength supposingthat the true function is non-linearover : rnurimal sensitivityat either short-,middle- or thisage range. In addition,the rate of changewith ,:r::-uavelengths. ageappears to be similarfor thethree cone types; I .inu psychophysical methods, Werner and approximately0. l3 log unit (260/o)per decadc. .::clc ( 1988)measured the sensitivityof the dif- Onecan think of theseresults as showing that the :J:.nt conepathways for 75 observersbetween the elderlyvisual system, at leastat this stageof pro- nof rto .i liltc'rs(in series)frorn the top row are illustrated.While thesetwo figr-rresshow the approximateappear- nit- . :,c rr ith neutraladaptation, the figure on the right showsthe physicallight distributionreaching the eye frorr fil- ' -.c ntirttrres. Iion llr)tt()nt rorv: Additive rnixtureoccurs when the light is reflectedby the two (unrnixed)paints applied in srnall and : :. thrltcannot be resolvedas discretedots by the visual system;the appearanceis achromatic.Additive ntrx- ' that .:,' .tl:o occurswhen the light is passedthrough the two filters in parallel;the appearanceis achrornatic.As nta. .:.tftrteclby the graphon the right, the light reachingthe eye is the sum ofthat reflected(transmitted) by each

)Lr- : rhcpivments (filtels) alone. l2 I . Aging through the Eyes of Monet cessing,is similarto the youngvisual systemop- 1.4 Monet'sEarly eratingat a reducedlight level. ImpressionisticStyle Severalsites in the visual pathwayare responsi- ble for age-relatedlosses in sensitivity,but the largestproportion ofthe sensitivityloss appears to While Maxwell and Helmholtzwere developing be at earlystages ofprocessing. These include in- theoriesabout the physiological basis of colormix- creasedabsorption oflight by the ocularmedia, a ing, Monet and Renoirwere in La Grenouilldreex- lossin the ability ofthe photoreceptorsto capture perimentingwith additiveand subtractivemixtures quanta(Schefrin et al., 1992), andlor an elevation on canvas.Here, many of the fundamentalsof in neuralnoise (Schefrin et al., 1995). Monet's style were developed,including painting Onealso sees a greatdeal ofindividual variation en plein air andrepresenting complex aspects of in conesensitivity within eachage. The sourcesof reflectionsand shadowon canvas.More and more, this variation are no doubt multi-faceted,but an his brushstrokesconsisted of a pure, unmixed importantone is likely to be exposureto sunlight. color, except when dark colors were formed Psychophysicalstudies (Werner et al., 1989)and throughsubtractive color mixtures. an anatomicalstudy (Marshall, 1978) suggest that ln 1921, the Neo-lmpressionistpainter Paul retinalaging, as with agingof the lens,is acceler- Signac(1921) published an historicalaccount that atedby exposureto light,especially UV andshort- characterizedImpressionism as based on these wavevisible lisht. four aspectsof technique:

Fig. 1.7: ClaudeMonet (1869)La Grenouilldre.Oil on canvas,74.6x99.7 cm. The MetropolitanMuseum of Art, Bequestof Mrs. H.O. Havemeyer,1929. The H.O. HavemeyerCollection. 1.4 Monet's Early ImpressionisticStyle l3

. Prlc'llc-composed solely of pure colors ap- contrastingtextures, and the high reflectanceof a flrr)\imatingthose of the solarspectrum; white canvascould be exploitedto capturequick- i \lir.ing on the paletteand optical mixfure; ly certainhighlights in natural scenesby having to -i. ('onrma-shapedor swept-overbrushstrokes, apply little or no paint. mg -1.Tcchnique of instinct and inspiration. Monet describedstill anotherreason for his F- rp.266) choiceof a white canvas;he said it was to estab- -\ll of thesecharacteristics can be seento somede- lish a scaleof (color)values. He studiedthe light [es rrl'r' in Monet's 1869 painting,La Grenouilldre intenselyand is said to haveattempted to under- OI rFig. I .7). standthe color of the prevailingillumination be- I Ing Signac'sfirst point is cleaq althoughthe term fore attemptingto evaluatethe colors of the land- of "pur!"' is not necessarilyused in any perceptual scape. Our sense of illumination was later )re. >i'fl:s'.He seemsto have meant only that the discussedby the Gestaltpsychologist David Katz Inrprr'ssionistsused the most saturatedpigments (1911), but it is stillpoorly understood. .nrilable.The secondpoint, mixing on the palette In describingthis periodof his life whenhe and -rndoptical mixture, refers to theuse of subtractive Renoiq among others, would paint together, .rndadditive color mixture, respectively. This point Monet said "lt was as if a veil was torn from my :: llso relatedto the fourth point, which merely eyes and I understoodwhat painting could be" :clcrs lo an unwillingnessof most Impressionists (Barnes,1990, p. 8). Soon,however, another veil to tollos strict divisionisttechniques associated would be torn from Monet'seyes when in 1870he '.irth Pointillismand Neo-Impressionism. andhis wife, Camille,moved to London.He want- \\'ith respectto Signac'sthird point, Monet's ed to escapeconscription in the Franco-Prussian .onrnra-likebrushstrokes can certainlybe identi- war as did Pissarro,whom he met there. In llc'd in many of his paintings.His brushstrokes,London, Monet and Pissarrosaw the works of that :r()\\L'\er.were quite varied (Seitz, 1956). He might lone English genius,J. M.W. Turner,almost cer- J\c dapplesofpaint to representthe ripple ofwater, tainly oneof England'smost creative painters. .'r ss irls to depictsmoke and steam. Multi-colored ;--nodicpatterns that recede in contrastand size are :.r:cdin Lu Grenouill'ireto showshadows shimmer- 1.4.1 PossibleInfluences of Turner and ing on the water.Or, as in someof his waterlily Goethe lrintings. Monetcaptures shadows reflected on the ',\rtL'r\\'ith mostly vertical lines, while the lily pads Not only Monetand Pissarro, but manyof thebud- rrL' contrastedwith thick horizontalbrushstrokes. ding Impressionistswent to Londonto see,for ex- .Ihcsc'and otherbrushstrokes were combined with a ample, Turner's The Fighting Temeraire(1838) \.rri!'tvof textures(Herbert, 1979) which were also andhis Rain,Steam and Speed(1844). Captivated quitc complex,but which generallyvaried from by light and color,Turner attempted to achievein .orrs!' in the foregroundto fine in the background, his paintingsa luminosityand brightnessthat ap- ;orrespondingto the surfacevariations on the reti- proachedthose visual experiences where light and nal ima_9e.At timescriticized as unskille4 some of color reachtheir highestlevels of complexityfor rhc'pafterns seen in his 1869La Grenouillire may thepainter - whenlight is reflectedfrom wateror it considereda preludeto abstractart. seenthrough rain, steamand fog. C'lose inspection of Monet's paintings also How did he do it? First of all, he seemedto .hos s anotherinteresting feature the weave of graspthe differencebetween additive and subtrac- rrs *hite canvasescan often be seenthrough the tive colormixture. In manyof his paintingsTurner lrrck_groundbecause they arenot completelycov- strategicallyplaced small dots of light color so crr'd by paint.The choiceof a white canvaswas thatthe additivemixture provided luster and bril- n\)taccidental. Canvas was available in a varietyof liance. This was about 40 years before the nof ;olors in the 19thCentury, but Monet alwaysused Impressionistswould exploitthis techniquemore shite (Callen, 1982).Unpainted areas provided fullv. and some 50 vears before the Neo- t4 I . Aging through the Eyes of Monet

Impressionistswould carry it to its limit. Turner translation of Johann Wolfgang von Goethe's also experimentedextensively with blocks of color (1810) Zur Farbenlehre(Reynolds, 1969). To ap- placedside-by-side to studythe ways colors influ- preciatethe impact of Goetheon Turner,it is nec- encedeach other (Clark, 1960). essaryto describebriefly ideasofGoethe that have Despite having very little formal education, taken hold. Turner painstakingly worked his way through a Goethe wanted to be a painter, but lacked the

Fig. 1.8: JohannWolfgang von Goethe(1810). Bild einesMddchen in umgekehrtenFarben. [Picture of a Girl in ReversedColors.l Watercolor with penciledoutlines, 15.8 x 14.6cm. StiftungWeimarer Klassik Museen, Weimar,Germany. 1.4 Monet's Early ImpressionisticStyle l5 te's ::.:nr. \lany modern scientists dismiss his ap- i .;,^,'nlt'hre because of his polemic against ec- \.\\rr)n: history does not look kindly on attacks ave ijrrnst \L-wton. Nevertheless, there is much in (irir'rhc"Sbook that is worthwhile, beginning with the hr' :rttc'nrptto analyzesensations independently of :hc stinruli that produce those sensations.His \trtr'nr!'nt that "the theory of colors in strictness nrrr bc investigatedquite independentlyofoptics" r(ior'the. 1810, p. 163) anticipatedan oft-quoted \t.il!'nrcnt by Maxwell (1872) who asserted that "thc scienceof color must . . . be regardedas essen- Fig.1.9: Goethe's (Goethe, 1810) illustration of col- :r.rllr a nrentalscience" (p.261). Goethealso antic- oredshadows (see text). rn.rt!'d Hering's opponent-color theory (described 'rt'los ) u hen he indicatedthat there are pairs of pri- narr colors (at least four, perhapssix) and that the sultingin a shadowon eitherside. The shadowon l.rircd nrembersinteract with eachother antagonis- theright is lackingthe whitish light so,having only :rcrllr. He came to this view by careful observa- the yellow illumination,it appearsa more saturat- :rons ofafterimages, successivecontrast effects, as ed yellowthan the surround. One might expect that rr' dr'scribesin the following example: the shadowon the left, lackingthe yellowishlight I harc entered an inn toward evening, and, as a and being illuminatedonly by the whitish light, s.'ll-fhvoured girl, with a brilliantly fair com- would appearwhite, but it doesnot. Ratheqit ap- plc'xion. black hair, and a scarlet bodice, came pearsbluish, that is, tingedwith theopposite hue to inttr the room, I looked attentively at her as she that surroundingit. This is an exampleof what is :tood before me at some distancein half shadow moregenerally known as simultaneous contrast. As .'\s shc presently afterward turned away, I saw on Goethe correctly concluded from this phe- thc rr hite wall, which was now before me, a black nomenon,blue and yellow opposeeach other not lucc' surrounded with a bright light, while the just in time,as with afterimages,but alsoin space. .lr.'ss of the perfectly distinct figure appeareda Goethe describedhow perceptualprinciples hcautifirlsea-green. (Goethe, 1810, p. 83) could be exploited to good effect in painting. ( iocthc painted the negative image shown in Turneraccepted this view and usedit to his ad- f :::urc 1.8so that this "well-favouredgirl" can be vantageto exaggeratelight yellow areasby sur- :pprcciated in all her splendor through an afterim- roundingthem with dark blue areas.In a tribute .i,.:c.To produce the afterimage, carefully fixate a to Goethe,he paintedLight and Colour (Goethes ..rlrcnt point and hold the eyes steady for about Theory) the Morning after the Deluge . i \cconds; then shift your gaze to a blank field (Fig.1.10). This painting is predominantlyyellow, .:i.h as a neutralwall. It helpsto blink when look- the color that Goetheconsidered the first deriva- :rg at lhc blank field. Notice that the colors of the tive of light. It stands in marked contrast to .rticrinrasc are complementary, or approximately Turner's(1843) companionpainting (Shadeand .\,. r0 thoseof the original image. Darkness- the Eveningofthe Deluge)which is (ioc-the described analogous phenomena in the dominatedby blue, Goethe'sfirst derivativeof .nrtial domairrbased on colored shadows,a topic darkness.At this point, Turnerhad obviouslybe- :het uor.rld later engage Helmholtz and Hering come extremelyabstract, anticipating aspects of ,llcrins. 1t387)in heated debate. According to Impressionismand E,xpressionism.This work is ( ios'the'sillustration, shown in Figure I .9, a surface without precedentin WesternArt and it is safeto

Jlfl :r illuminated by a whitish light from the left and a say that Monet could not haveseen anything re- rt'llosish light from the right. Each beam will be motelysimilar to it in Europeat the time. nrnialll' obstructed by the object in the center, re- Turneralso becamequite a profoundpessimist 16 l. Aging through the Eyes of Monel

'e

t -:. ", , # {e 3::?::

'{

*, .,,r '; i +$- i#,. .j :' 'Eii&"

-5,

Fig. 1.10: JosephMallord William Turner(1843) Light and Colour (GoethebTheor.t) - the morning after the Deluge Moseswriting the Book of Genesis.Oil on canvas,78.8 x 78.8 cm. Tate Gallery,London. asexemplified by theverses attached to thetitle of Pissarrowas enthralledwith Tirrner,and al- this paintingtaken from his poem,The Fallacies of though Monet rarely commentedon other paint- Hope.He wrote: "hope, hope,where is thy market ings,he did makecomplimentary statements about now?" This aspectof Turnerwould not haveap- Rain, Steamand Speed(interestingly, an etching pealedto the Impressionists.Renoir saidhe sim- of this Turnermasterpiece by F61ixBracquemond ply wanted to paint pretty pictures.And Vincent was shown in the first Impressionistexhibit). van Gogh,more a Post-Impressionist,said that he Generally,however, Monet deniedbeing much im- hopedthat their art could "give comfort, and make pressedby Turnerand was critical ofhis exuberant life possible,in theway that Christianity once did" romanticism.In truth, Paul Signacnoted that the (Russell,1974, p. 22). Impressionists studied Turner's work and mar- 1.4 Monet's Early lmpressibnisticStyle I'7

r c'lc'dat effectsthat he producedthat they had no1 taneouscontrast. These experiments are described bccnable to achieve.Signac (1921) writes: in his 1839book (Chevreul,1839), De la Loi du They were in the first place struck by his snow Contraste Simttltant des Couleurs, a book that andice effects.They were astonished to seehow was hailed by the Impressionistsand studiedby he had succeededin producingthe sensationof Helmholtz.Chevreul's studies of contrastculmi- *hiteness of the snow while they had so far nated in a set of descriptivelaws. Figure 1.11 been unable to achievethis with their large showsone of his beautifulillustrations of how a patchesof ceruse(white lead) spreadout flat hue induces its complement in surrounding re- * ith broadbrushstrokes. They say that this mar- gions,analogously to Goethe'scolored shadows. velous result was obtained not by a uniform Many of Chevreul's ideas were tested on shite, but by a numberof touchesof diverse the canvasby Eugdne Delacroix. Like Tirrner, colors,placed side by sideand reconstituting the Delacroixoffended his contemporariesby his bold desiredeffect at a distance.(pp.2392a0) useof color,although it is doubtfulthat his paint- Turnerand Monet had much in commonbesides ings ever reahzed the luminosity of Turner's. their interestin additivecolor mixture and color Figure l.l2 showshis painting called The Lion contrasteffects. They wereboth fascinatedby the Hunt (1860-61).The carnagein Delacroix'spic- changingeffects of light at differenttimes of day. ture is madevivid enoughby his use of saturated John Ruskin (1843) cataloguedsome 60 Turner pigments,and perhaps it is doneso well becauseit paintingsaccording to the lighting conditionsas- wassaid that he never missed a feedingat theParis sociatedwith varioustimes of day,as modified by zoo (Clark, 1960).The Impressionistsstudied the rveather,atmosphere and the objectsthemselves. work of Delacroixand seemedparticularly enam- Later,Monet would illustratethe changingeffects ored by his thoughtson simultaneouscontrast, of light and atmospherein morecompelling fash- aboutwhich he is saidto haveremarked: "Give me ion throughseries of paintingsof thesame subject. themud of the streetsand I will turn it into the lus- Both artistswent to greatextremes to observena- ciousflesh of a woman"(Signac, 1921, p.238) lif ture. For example,Turner claimed to have tied you will allow me to surroundit asI please]. himselfto the deck of a ship to observea raging If it is assumedfor themoment that the "luscious storm at sea and was nearly killed. Monet de- flesh"that Delacroix had in mind waswhite, it can scribeda similarexperience. But their differences be shownby experimentthat he wascorrect. Figure rrere equally compelling;while Turner depicted 1.13presents results from an experiment(Werner naturein her fury, Monet seemsto havepreferred and Walraven,1982) in the CIE x,y chromaticity her morepastoral, but elegant,simplicity. diagram, which representsall possibleadditive color mixtures.The perimeterof the diagramrep- resentsthe loci of monochromaticlights and the 'the 1.4.2 PossibleInfluences of Chevreul light mixtures(between 400 and 700 nm) that en- and Delacroix closethe space.The coloredareas illustrate the ap- proximateappearance of theselight rnixturesin the lf Monet had beeninclined to studythe work of neutralstate ofadaptation (i.e., when the colorsare al- any scientist,it would probablyhave been Michel viewedin an unilluminatedsurround). The subject's unt- EugdneChevreul rather than Goethe.Whereas an task in the experimentwas to vary the ratio of two )out emotion-ladenanalysis ofcolor appealedto Turner lights (complementarycolors) until the mixture ring andGoethe, a lesspassionate analysis appealed to lookedwhite. The centralx showsthe light mixture .ond Chevreuland Monet. Chevreul was the Directorof that appearedwhite without anybackground or sur- bit). Dyesfor the Gobelinstapestry works, and in that rounding light. The experimentwas then repeated im- capacityhe conducteddetailed experiments on the in the presenceof various larger chromaticadapt- fant interactionsbetween threads placed side by side, ing backgrounds.The spokesconnect the neutral the c'itherwhen the colored regions were small enough white point and the adaptingbackgrounds, which nar- to mix additivelyor largeenough to createsimul- were all locatedon the perimeterof the color dia- l8 1. Aging through the Eyes of Monet

:4 fr**&te 2. ..'.'7;'ztZ

frq ir i'rq I

I '{' l"rg'ri

Fig. 1.11:M. E. Chevreul's(Chevreul, 1839) illustration of simultaneoushue contrast. In this figure,the col- ors that would normally be inducedby a colored disk into a neutral surround are exaggeratedby painting the surroundregions. (Colors are digitally enhanced by the authorto compensatefor fadingofthe originalplate.) gram. The data points show the stimulus that dictions assumingvon Kdes adaptationand a sub- looked white after adaptationto various colors and tractiveprocess; that is, separatesensitivity adjust- for variouscontrasts (i.e., intensity ratios oftest and ments in each class of receptor in proportion to background). The dashedcontours are model pre- their activation by the background light (Werner 1.4 Monet'sEarly ImpressionisticStyle 19

Fig. 1.12:E. Delacroix(1860/61) La chasseaux lions. [The Lion Hunt.] Oil on canvas,76.5 x 98.5cm. Potter PalmerCollection, Art Instituteof Chicago.(Photograph @ 1996,The Art Instituteof Chicago.A1l rightsre- served.) andWalraven,1982; see chapter l2). Withoutdis- Assimilationis still not well understood,but it is cussingthe experimentaldetails, one can readily known that it cannotbe explainedby light scatter appreciatethat Delacroix was correct: that any from one regionof the imageto another.When the light can appearwhite under the right conditions hatchingis fine relativeto the diametersof individ- of contrastand adaptation. ual receptors,additive light mixture occurs,but in Delacroixalso advocatedthe use ofhatchings to this casethe hatchingitself will not be visible.If the influence colors basedon a complementaryphe- elementsin the hatchingare small relative to the nomenon to contrast called assimilation or the postreceptoralelements that sum inputs fiom a Bezold SpreadingEffect (Bezold, 1874).This oc- numberof cones,assimilation may occur.When the curs when a backgroundand interlacedpattern of hatchingis courseqcontrast tlpically occurs,but de- differentcolor fail to opposeeach other as in simul- pendson a numberof factorssuch as the luminance taneouscontrast, but seemtoblend together.Figure of the patternrelative to the background(de Weert I . l4 illustratesassimilation with four differenthues and Spillmann,1995). By varying the viewing dis- that haveblack or white hatchingsuperimposed; al- tance of a suitablepattern, one can observelight though the backgroundhue within each panel is mixture, assimilationand contrast.Thus, assimila- physicallyuniform, it appearsdifferent depending tion, althoughnot due to optical mixture, could be on whetherit is interlacedwith black or white. due to a neural blending in some color channels 20 1. Aging through the Eyes of Monet

Fig. l.l3: The loci oflights that appearwhite undervarious states ofadaptation presented in the CIE x,y chro- maticity diagram. Colored areaswithin the diagram representthe approximateappearance of additive color mixtures in a neutral state of adaptation. The central x representsthe light mixture that appearedwhite in a neutral stateofadaptation for one observeqwhile the data points show the chromaticity ofthe light that ap- peared white following adaptation to the continuously presentedchrornatic backgrounds (250 Troland retinal illuminance).The backgroundstested are all on the perimeterofthe diagram and connectedto the centralx by lines. Squares,triangles and circles representilluminance contrasts(test/background intensity) of 0.2, 1.0 and 5.0, respectively.(After Wernerand Walraven,1982.) 1.4 Monct'.sEarly In.rpressionisticStyle 21 ffi

ffiw

Fig. t.lJ: Illuslrationol'assinrilation wor thc BczoldSplcading Ell'ect. The backglounclcolol is thc samc-rvith- in it l.ritttcl.but it appcarsdalkcr w,henintcrlaccd lvith blackhatching cornparecl to rvhenit is intcrlacedwrth u'hitchatching. ri'hilc othcr color channelsstill responclcliflcrently colors that clependsin its very csscl.lccon such to thc lirrnr and the backglound (Jamcsonancl rcciprocalinteractions" (pp. 123 l2zl). Hurrrch. 1989).Assimilation probably occurs as In firll agrccnrent.Monet wor-rldlater say: tnuclt irt naturcas contrast.br.rt we know tluch less For r-nc,a landscapecloes not cxist in its own aboutits phvsiologicalbasis (clc Weert. 199 l). right. sincc its appearancechangcs at every uto- It nlrl, seer-nlionr these pltcnolnena of aclapta- rrent; but its surrounclingsbring it to life thc tion. assinrilationancl contrasl that the visLralsys- ail and thc light'nvhichvary continLrally... For tem rs casily fbolccland subjcctto illusion,bLrt the nte, it is cln11,15. sttrt'ouncliugatmosphere u,hich mechanismsproducing these eflbcts are u,hat r.nakc gives objccts their tfuc valLle.(Banles. 1990. nornral color vision possiblc.As Hering (1920) p. 36) pointecloLrt: Thus, while some of thcse phenclntcnaof adapta- "The nrostimportant coltsequcltces of rcciplocal tion, assimilationand contrastdo illustrateinrpcr- interactionsare not at all thoscexpressecl in con- fcct color constaucy.they seldonr lcaclto confusion trastphenor.nena, that is, in the allegedfirlsc see- aboutan objcct'sidentity based on color.Mole gen- 'real' ing of colorsof objects.On the contraly,it erally,they supportcorrect identificationofobject is prcciselythe so-calledcorrcct seeinsofthese color by acccntuating dill-erencesbetiveen object 1 Aging through the Eyes of Monet

Fig. 1.15: Top: Claude Monet (1873) Zes coquelicots d Argenteuil. [Poppies at Argenteuil.l Oil on canvas, 50 x 65 cm. Mus6e d'Orsay, Paris. (Photo credit: Erich Lessing/Art Resource,New York.) Bottom: Claude Monet (1879) Camille Monet sur son lit de Mort. [Camille Monet on her death-bed.l Oil on canvas, 90 x 68 cm. Mus6e d'Orsav. Paris. 1.5 Monet'sYears in Arsenteuil and Vdtheuil zi and shadow,with assimilationenhancing the uni- 1.6 The Opponent Code for formity ofa single surfaceand contrastenhancing Color Appearance:Hering differencesbetween figure and ground.These ex- amplesprovide important probes to visual scien- tistsfor identifringproperties ofthe visualsystem During these years,Germany was developingits that are normally not apparentat all. The mecha- own Impressionistschool through the wonderful nisms mediating these effects are what keep us paintingsof Adolph Menzel, Max Liebermann, from being fooledmost of the time aboutthe re- GotthardtKuehl and others(Diichting and Sagner- flectanceor color ofobjectswhen the color ofthe Diichting, 1993), althoughthe most significant illuminationchanges - they makecolor constancy eventsfrom the point of view of this essaywere possible.This refersto the experiencethat the col- probablybeing carried out in the laboratory. ors of most objectsappear to be aboutthe samein Ewald Hering, Professor of Physiology in a wide varietyof lightingconditions, even though Vienna and later in Prague and Leipzig, pointed they may reflect very different spectraldistribu- out that the Helmholtzview nicely accountedfor tions to the eye. Color constancywould not be the trivarianceof color mixture, but the second possibleif the visual systemwere not ableto ad- part of the Helmholtz theory - that thereare three just its chromaticsensitivity as illumination varies. fundamentalhue sensations- is inconsistentwith our color experience.Hering (1920) proposed that thereare four elementalhues, re{ green,yellow and blue, and that all hue experienceis basedon 1.5 Monet'sYears in Argenteuil combinationsof theseelements. (For a complete and Vitheuil accountofcolor experience,Hering also proposed that thereare two fundamentalachromatic colors, Monet livedin Argenteuiland Vdtheuil from l871 black and white; seechapter 10, Shinomori et al., to 1881.During this time he createdsome of the 1997.)He was not the first to makethis proposal. most belovedmasterpieces of the Impressionist It could be pointedout that Leonardoappreciated movement.Many of thesepaintings illustrate an thiscenturies before. To which Heringreplied that: idyllic countrylifestyle with his wife Camilleand "If onewere to designatethe nomenclatureused young son Jean, such as in his painting, The ... asa four-colortheory, then ... languageitself Luncheon(1873) or the Poppiesat Argenteuil would be its author,for languagehas long since (Fig.1.15, top). This was a happytime for Monet- singledout re4 yellow, green and blue as the until 1879when his wife Camilledied. Out of this principlecolors of the multiplicity of chromatic tragedywe gain insightinto Monet'sseemingly ir- colors"(Hering, 1920, p. 48). repressibleobsession with the changingeffects of Heringfurther postulated that the hue variables are color.He wentso far asto portrayCamille's chang- organizedphysiologically in antagonisticpairs ing colorationon her deathbed (Fig. 1.15,bottom) that involveprocesses ofexcitation and inhibition, anddescribed his feelingsas follows: with the resultthat the sensationsof red andgreen You cannotknow ... the obsession,the joy, the neveroccur at the sametime andplace, nor do the torment of my days.To the point that, one day, sensationsof blue and yellow. He, therefore,re- when I was at the death-bedof a lady who had ferredto thesepaired colors as opponent.To ex- been,and still was,very dearto me, I foundmy- plain phenomenaof successiveand simultaneous self staringat the tragiccountenance. automati- contrast,Hering also proposedthat thesephysio- cally trying to identifythe sequence,the propor- logicalprocesses are organizedin an antagonistic tions of light and shadein the colorsthat death or opponentfashion across time and space. had imposed on the immobile face. Shades Since Berlin and Kay's (1969) anthropological of blue, yellow, gray, and I don't know what. survey,evidence has supportedthe view that there That'swhat I had become(Clemenceau, 1929, are a limited numberof hue namesneeded to de- p.350-351). scribeour color experienceand that they are de- 21 l. Agin-qthronsh thc Eyes of Monet

Fig. l.l6: Mean huc--nanringclata (u'hite squarcs) and avclageplcclictcd huc'- nanring(black circles) basccl on opponcnt-cauccllationlirnctions lFigulc l. l 7) fbl thrccobservcls plottccl as a firnctionof wavclcngth.Rccl-grecn is plottccl fhnr 0 to 100'fi,on thc lcll axis;bluc-ycllou'is plottcd fionr 100to 0'fi,on tltc li_qhtaxis. Thc lcttclsB. Ci.and Y clcnotcthc avcrageuavclcngth o1'thc samc obscrvcrs'unicluchues. (Alicr Wclrrcrand Wootcn, 1c)7t).) penclenton the organizationof neulophysiological ncccssaryto proch-rccan eqLrilibritur statc (e.g.. lurechanisursin the eye anclbrain (Ratlifl, 1976). ncithcr lccl nor'-qrecn). llue carrccllationfirnctions C)rrccxanrplc is sltorvnin Fiqule l.l(r. Obsclvcrs prcscnteclin FigLrrc1.17. basccl ou thcir mcthod. rvele asked to clcscribclights ol'difl-elent rvave- shou' the rcsllonscol- thc rccl-grccnancl yelloil,- length using thc terursrccl. green, yellon' anclblLre blLreopponcnt charrnclsas a lirnctiorr of wavc- (Wclner anclWooten, lc)79).Results are shown by lcngth.Rcsponses ol' singlccclls of thc rtracaquc thc rvhite sqLlarcs.Recl or gleen is plotteclfionr rnonkeyat variouslcvcls ofvisual processinglook 0 to 100% on the lcft axis.and blue or ycllon,is rcmarkablysimilal to thcselirnctions (scc chapter plottcclin the oppositeclircction. fiom 100 b 0%. -3;Zrcnncl ct al., lc)c)0;,antl there is conrpelling on the ri_shtaxis. The clatacoulcl bc plotted in this reason to lrclieve thal thesc ol]ponentllrocesses, u'u1,bcs'111sg, consistcut ri,'ith Hcrin-g's thcory. the via cortically-rcctil'icclsignals (DeVilois ancl obscrvers alurost never Llsedthc tcrrns rccl and DeValois. 1993).arc an iurportantpart ol'the neu- grccn togethcl or thc tcrrns bluc and ycllow to- ral netrvorkresponsible lbr pclccived huc. Figurc _rJcthcr.Thc lctters(8. G. Y) on thc horizontalaxes I . l6 (circlcs)courparcs hue nartringand the t.uean designatethc rvavelengthsof the liglrts that rvere ratio of recl-greento ycllou'-blucrcsponsc at each uniqLrelybluc, green or yellow lirr the averagcof wavelengthnreasulccl lry thc cancellationtask. Thc the samethrcc personsu,hose hue-nar-r-ring data are oppouent rcsponses measurccl by canccllation presented. agrec qLritc uell rr'ith color nalning. corlsistent .lanrcsonancl Hurvich (1955) clcvclopeclnreth- with the iclcathat thcscopponcnt proccsscs al'c thc ods to rreasufe psychophysicallythe ollponent neuralsubstrate of color appcarallce. proccssesof I Icring usins a hue-canccllationtask. While sorne researchers.evcn as rccetttly as The strengthol'each hLre(e.g., red) rvasquantified 20 yearsago, seemeclto fincl thc Hering vierv in- by the energy of the opponent hne (e.g. gleen) conrpatiblervith thc Maxrvcll-Hehrholtz theoly. 1.6 Thc OpponcntCodc for ColorAppearance: Heling 25

Fig. l.l7: Opponcnt-huecancellation fr"rnctious nveraged for threcobscrvers plottcdas a lirnctionof wavelength.Red-gleen is shownby whitesquares, with rcd ploltcdas positiveand gleen as negative.Blue-yellow is shou,nby black circlcs,with blucplotted as negative and yellow as positive. (After Werner and Wooten.1979.)

.lohanuesvon Krics (1882), Prol'essorof Phys- The data on age-relatedchanges in color vision g. iology in Freiburg.put fbrth a zone tlteory befbre presentcdso far havcdcalt only with the first zone ns thc turn of the centuryin which color vision bcgins (i.e..photoreceptors), but there rriglit also be age- )d. l ith the activity of thlee classesof cone pl'rotore- rclatcd changesin the later stagesof color pro- icptol whicl'ris re-mappcclollto tlte ner.rral-oppo- cessir.rg.Schcfiin anclWerner ( 1990)measured the :rcntpfocesses of Hcring. A more troclernversion balar.rccpoints of thc opponcnt mcchanisrr.rs(that uc ,,l such a zone moclcl bascd on equations of is, the spcctraluniqr.rc hr-rcs) in 50 observcrsrar.rg- ok .lru.r.rcsonand Hurvich ( 1968)is presentedin FigLrre ing in agc f}om l3 to 74 years.Of particularinter- lcr i. I 8. It Lrtilizesthc ideathat feceptorsactivate ueLl- est are the loci of unique blue and unique yellow, :ul processesin cithcr an excitatoryol inhibitoly bccausethey can be describedr-rnder r.nany condi- ES. :irshion.and differentcontbinations ofall threere- tions by a linear moclclancl will, tlielefble, be un- ncl .r'lltols producethe two opponent-clrlornaticpro- afl-ecteclby age-relatedlosscs in light intensityas- 't t- .asses.The direct connectionsbetween l'eceptors sociateclwith lcnticulal' senescellce.Thc rcsults lre .irrtlopponent r-nechanisn-rsshown in Figure l.l8 show that there is no significant change in the i0l1 .rfc a consiclerableoversinrplification to illustlate wavelengthof these unique hues over this wide rch :iLnctionalrelations. Anatomical and pliysiological age rallge. That these resnlts reflect a ulole genef- lie .tudies have der-nonstratcdulu-nerous interactiorrs al patternof stabilityin color perceptionacross the '..'t\\'een ()I l cells at various levels betweenthe pho- lif'e span was confirmed in a color-naming expeli- tnt :!)r!'ceptorsand higher-levelcolor rrechanisnrs.It ment using r.norenaturalistic, broad-band surfaces, lhe .hould alsobe r-rotedthat morc complex non-linear color chips fi'or.nthc Unifonn Color Samplcs of I(r(lels are required to describesome aspectsof the Optical Society of America (Schefrir.rar.rd lls r,rlor perception,even though tl-relinear eqr-rations Welner. 1993). Not only did yor"rngand old use the .i:cclin Figure 1.18plovide an excellentfirst-order' samewords to describehues, but they did not dif- rry. .tcscliptionof cancellationfunctions. fer significantly in the proportions of different 26 1. Aging throughthe Eyesof Monel

Fig. l.l8: A modernversion of von Kries's1882 zone theory (Kries, 1882), based on the equationsof Jarnesonand Hurvich (1968)linking the activity of thethree classes ofcone photoreceptors(S, M andL) to red-green(left) and yellow-bh"re(right) opponent mechanisrnsas a function of wavelengthfrom 400 to 700 rlm. The shapesof the opponerrt-responseflnc- tions dependon the sign (+ or -, correspondingto neuralexcitation or inhibition)ofthe signalsfrorn the conesand their neuralweights. The parameters(spec- ified in tcrmsof quantaat thc cornca)and equations nsed to generatethe firnctionsale slrown at the bot- torr.

Fig. l.l9: Heterochromaticbrightness sensitivity change per decadeis plotted asa function ofwavelengthat the cornea(black circles)and at the retina(white squares).The horizontal[ine at zerodenotes no age-relatedchange. The thick- er horizontal line at +0.05 showsthe mean increasein brightnessssensitivity per decadebetween 420 and 560 nrn. (After Kraft and Werner,1994.) 1.7 \tonet s Responseto Pointillismand Dirisionism 27 hues neededto describethe stimuli. Similarly, The techniqueis, in principle, not different from when askedto describethe percentageof chro- that usedby Turner,although in practiceit was an matic and achromatic components of spectral extraordinaryleap from Turnerbecause the points lights (kaft and Werner, 1997), and when asked were applied by Seuratin a much more systemat- to find a mixture of unique blue and yellow that ic and consistentmanner. Seurathad studied all appearedwhite (Werner and Schefrin, 1993), the color scienceavailable to him includingwork youngand old did not differ significantly. by Chevreuland Helmholtz.His masterpiecein- Theseresults demonstrate a surprisingdegree of cludes dots of varying sizes to achieve the stabilityin color perceptionacross the life span, Pointillist goal of increasingthe luminosity of eventhough age-related changes in the lens alter paintingsby placingsmall dots of purecolor side- the spectraldistribution of light reachingthe reti- by-sideto producean optical mixture in the eye or na.This would seemto be possibleonly if the vi- to achievestrong hue contrastswith largerdots. sual systemcompensates for or adaptsto those Pissarro,van Goghand Matissetried their hand changesin retinal illumination that occur with at Pointillism but werenot satisfiedand soonaban- lenticular senescence.Clear evidence for this doned the technique.Their disillusionmentwas compensationwas found in a study by Kraft and due, in part, to the fact that the appearanceof a Werner(1994) of the brightnessof monochromat- Pointillist painting dependsso critically on the ic lightsfor 50 observersranging in agefrom I 9 to viewingdistance. At somedistances, the paintings 85 years.Each monchromaticlight (420-700 nm; do not havethe brillianthues intended, butjust the 16 wavelengths)was matchedin brightnessto a opposite- the hues appeardrab and desaturated. white standard.In the sameindividuals, the densi- EvenSeurat's ( 1887-88) Les Poseuses, with itstiny ty of the ocular media was estimatedso that dots, is a disappointmentin this regard(Ratliff, brightnesssensitivity could be speciJiedat the 1992).Interestingly, the mechanismthat underlies retina, as shown in Figure 1.19 (white squares). the successof the Pointillisttechnique as an art On average,brightness sensitivity at middle and form also underliesits limitations.The problem long-wavelengthsdoes not changesignificantly, arisesbecause the signalsfrom conephotorecep- but at short wavelengthsit actually increaseswith tors are integratedin the visualpathways to form age.This increaseoccurs only at the retina,and receptivefields, areasof the retinathat activatea undernatural conditions it is not enoughto elimi- particularcell by excitationor inhibition,depend- natesome brightness loss (at the cornea)for violet ing on where the light falls within the receptive lights in the elderly.However, across 6 decadesit field (Wieseland Hubel, 1966),as illustratedby impliesa doublingof brightnesssensitivity, which Figure1.21. The three-dimensionalprofile on the goesa long way towardrestoring constancy that left illustratesthe response increase when stimulat- would otherwisebe disruptedby the yellowingof ed by greenin thereceptive field centerand the in- thelens. hibition by red in its surround.The other cell shownhas a blue-yellow,antagonistic center-sur- round organization.These receptive field profiles are spatialfilters for color processing,modeled 1.7 Monet'sResponse to hereas the differenceof two Gaussianfunctions, Pointillism and Divisionism one representingthe distributionof excitationand a secondhaving lower amplitude but broaderarea, In the eighth Impressionistexhibit in 1886, representingthe distributionof inhibition.Many Georges Seurat showed how the Neo- individualcells can be modeledin this way, al- Impressionistswould take the next stepin apply- thoughtheir responses often do not follow theper- ing the color scienceof the time to painting.His ceptualred-green or blue-yellowaxes (Lennie and (1884)I SundayAfternoon on the Island of the D'Zmura. 1988).Nevertheless, the combinedac- GrandJatte (Fig. 1.20)is one of the most cele- tivity of manysuch cells having overlapping recep- bratedexamples of the techniqueof Pointillism. tive fields appearsto result in a networkthat forms 28 l. Aging throughthe Eyes of Monet

Fig. 1.20: GeorgesSeurat (1884) Un Dimancheupris-nidi ii lle de La Grande-Jute.[A SundavAliernoonon the Island o.f'theGrand Jatte.l Oll on canvas,207.5 x 308 crn. Helen Bitch Bartlett Memorial Collection,Art Instituteof Chicago.(Photograph O 1996,The Art Instituteof Chicago.All lights reserved.) a mosaicconsistent with Figure1.2 l. Suchrecep- thedots, at a particularviewing distance, are small tive fieldsprovide the kind of mechanismrequired enoughto approachthe sizeofindividual conere- to explain the spatial-chromaticopponency de- ceptors,additive color mixturewould be expected scribedby Hering(1920). Ofcourse, each individ- and the region would appearachromatic. If the ual receptivefield will producea responseonly to dots were somewhatlargeq but small enoughso stimulationby an edge,but if the entirereceptive that the yellow and blue fell within the excitatory field is uniformlyilluminated by a particularcolor and inhibitory regionsof the receptivefield. re- it will not respond.These receptive fields help to spectively,they would likely canceleach other's explainhue contrastat borders,but not the induc- effects,rather than producecontrast. The result tion of hue acrosslarge areas of the visual field would be an achromaticcolor. Notice that in this suchas observedwith Goethe'scolored shadows. case,the effect is similarto additivecolor mixture. How an entirearea is filled-in perceptuallywith a but it is reallya neuralmixture that depends on the uniformhue is not clear,although a numberofneu- size of the areas over which information is rophysiologicalhypotheses involving propagation summedin the visual pathways.Whether small acrosscortical regions beyond the areaofthe clas- dabsof paintproduce cancellation, contrast or as- sical receptive field have been described similationdepends upon the "fit" betweenthe size (Spillmannand Werner, 1996). of the dots imagedon the retina (and hencethe Considerthe consequencesof this kind of neu- viewing distance)and the size of the receptive ral organizationfor how we perceivea Pointillist fields. Receptivefields are known to increasein painting. Supposean artist placessmall dots of size with retinal eccentricity,although there is a paint, say a blue and yellow dot, side-by-side.If rangeofreceptive field sizesrepresenting each re- 1.7 Monet's Responseto Pointillism and Divisionism 29

+ f

Fig. 1.21: Receptivefield profiles for red-greenand blue-yellow cells.These cells havean antagonisticcen- ter-surroundspatial organization such that one color in the centerproduces excitation and the opponentcolor I in the surroundproduces inhibition. 1

gion of the visual field. The Neo-lmpressioniststhe problem;rather, it only definesa new set of' I did not know aboutreceptive fields, but certainly distancesat which one has hue cancellationvs. knew about the perceptualphenomena that they contrast. Perhaps anticipating this rejoinder, ; produce.At just the right distance,the receptive Signacreferred to Rembrandt: fields will be activatedto producethe additive 'A painting is not to be sniffed,"said Rem- 0 mixturesand contrasts intended by the artist.How brandt. When listening to a symphony,one doesone know what that distance is? In notingthat doesnot sit in the midst of the brass,but in the the brillianceof Pointillistpaintings depended on placewhere the soundsof the differentinstru- s the viewing distance,Pissarro suggested the gen- ments blend into the harmony desiredby the It eral rule that a Pointillistpainting be viewedat a composer.Afterwards one can enjoy dissecting s distance that is three times the diagonal. Of the score,note by note,and so studythe manner course,this adviceonly makessense if the sizeof of orchestration.Likewise, when viewing a di- thedots has a fixed relationto the sizeofthe paint- videdpainting, one should first standfar enough ing as a whole(Weale,1971) - whichwas appar- awayto obtainthe impressionas a whole, and ll entlynot alwaysthe case. then comecloser in orderto studythe interplay i- PaulSignac emphasized that the besttechnique of the colored elements,supposing that these e practiced by Neo-lmpressionistswas "Divi- technicaldetails are ofinterest. (Signac, 1921, e sionism"not "Pointillism,"by whichhe meantthat p.264). e the paint shouldbe appliedwith small distinctive In the meantime,Monet tried other approaches n strokes,not tiny points.Signac wrote: "The Neo- that usedelements of Divisionism,but without a a Impressionistdoes not paintwith dots,he divides" rigid applicationof the technique.One example is (1921, p. 207).lt is difficult to seehow this solves shownin Figure1.22, Bend in theEpte River near 30 l. Aging through the Eyes of Monet

a:.:f,:- tf k,. t,r3

'1't.-s.r-o":,1Jt tlrd;

.*l; 1.i t, -.". :tr.7 T{ --l

rl: .l

,til ':i: ;r*. fi r $q

:1fr.W,

Fig, 1.22: ClaudeMonet ( I 888) Un Tournantde I'Epte. [Bend in the Epte River near Giverny.JOil on canvas, 74 x 93 cm. PhiladelohiaMuseum of Art: The William L. Elkins Collection.

Giverny(1888). The foliage of the treesfollows centsite which he capturedin numerouspaintings. very much the style of Seurat,but it is set apart Gloriousthough his gardenswere, Monet also il- from the water and sky which maintain his lustratedthe subtleundulations of light and color comma-likestrokes. The resultis just as luminous in more mundanespots such as in the hay stacks as the Pointillist and Divisionist attempts,but behind his house.There he painted a seriesof seemsaltogether more spontaneous and natural. canvasescapturing different conditionsof light, atmosphereand weather.His approachwas me- thodical;rising early in the morning even in the depthof winter,he caughtthe first glimpseof sun- 1.8 Hay Stackand Cathedral riseat his chosenlocation, typically rested at mid- Series day andthen returned to catchthe settingsun. Figure1.23 shows how splendidlyMonet captures By 1890,at the age of 50, Monet had reacheda the light in the hay stacks.On the top, the moming high standing in the art world and had found fi- light falls upon the snow,and the yellow hay stacks nancialsecurity. He rebuilt an old farmhousein are surroundedwith the blue colored shadowsde- Givernyand employed six gardenersto indulgehis scribedby Goethe.On the bottom,Monet showsa love of horticultureand flowers.It wasa masnifi- greenishshadow induced by the reddishcolor of the

l I Hay Stackand CathedralSeries

d.o,. :. *::'.-. r.1- ",i

.. : r:i;; +:',r,-i,-'rl.l

t*a, - ' - '- ;".' --a*i--. ,'.'-. i#rit 32 1. Aging through the Eyes of Monet

Fig. l.24z Left: Claude Monet (1894) Le portail, brouillard matinal. [Rouen Cathedral: The Portal (Early Morning).1 Oil on canvas,100 x 65 cm. Museum Folkwang,Essen. Right: Claude Monet (1894) Le portail (soleil). [Rouen Cathedral: The Portal (Early Afternoon).] Oil on can- vas, 100 x 65 cm. Board of Trustees's,National Gallery of Art, Washington. hay stack in the late afternoon sun. We now regard Monetpainted several other series, including the this simultaneous contrast effect, exaggeratedby Gare St. Lazare,the Seinenear Giverny and Monet on the canvas,as due to the reciprocalneural- ous scenesfrom his gardens.In the cold winter opponentresponses across the visualfield. 1892.the 52-vear-oldMonet renteda larqe Monet's hay stack series includes more than directly acrossfrom the Cathedralof Notre Da 30 canvases,from differentvantage points and dis- in the nearbvcitv of Rouen,For the first time, tances,and in different lighting. In order to depict would oaint an outdoor scene from indoors fugitive effects, he worked on as many as seven Unlike many cathedrals,there was little space canvasessimultaneously, apparently dashing from afford an unobstructedvantage point so we see oneto anotheras the light would change.The can- cathedralfacade cropped (Fig. 7.24), due appar vaseswould then be takento the studioand placed ently to the restrictedview through the window side-by-sidefor retouching.The serieswas intend- Monet was apparently frustrated, not so m ed to be displayedtogether. That the subjectmat- from the view, but from what he was trying to ter is monotonousand uninspiringto many people complish.No otherpaintings occupied so much is besidethe point. Monet said: "I wastrying to do his time. the impossible... to paint light itself" (Myers, He left Rouenafter the winter andreturned again 1990,p. 92). the next year,but still failed to completehis project 1.9 Monet Returnsto London JJ

until he returnedhome and paintedfrom memory. morethan 100canvases (Tucker, 1995). It is inter- ln all, he managed30 canvaseswhich he placed estingto comparehts Boats on theThames in 187I side-by-sidein his studioto finish. They all are with his Houses of Parliament painted 30 years signed 1894,the year of their completion.Once later (Fig. 1.25).The atmospherein the latter is again,the contentof the seriesis not particularly richeq as we seethe fiery sun barelypenetrating criticalto Monet(cf., Pissarro,1990). Each canvas the denseclouds. lndee4 the pageantofcolors in representsa moment in time associatedwith differ- many canvases from this series anticipated ent light and atmosphere.In Figure 1.24,the paint- Fauvism.It is almostas though Monet is nowchal- ing on the left representsearly morning, and the lengingTurner on his own turf, asTurner had once sun can be seengently breaking through the mist done with the great French landscapist,Claude nearthe spireswhile the light is occludednearer Lorrain.Turner not only copiedsome of Lorrain's theground. On theright, it is afternoon,and the en- paintings, but insisted in his bequest to the tire facadebasks in the sunlight.The shadowsare National Gallery that his copies hang next to createdwith sculptedmounds of paint,whereas the Lorrain'soriginals for comparison,a requestthat brightareas include patches where the canvas is not is still honored.Some critics pointed out that now completelycovered. Monet'spaintings should be hungnext to Turner's, If Monet had wantedto representthe physical calling them both greatImpressionists worthy of situationaccurately he wouldhave had to makethe comparison(Gage, | 972). afternoon scenereflect severalthousand times more light to the eye comparedto the canvasde- pictingthe morningscene. He couldnot havedone that,but he didn't needto becauseour visualsys- 1.10 Water Lilies and Cataracts tem is remarkablyinsensitive to ambientlight level over a large range.How, then, doeshis painting Monet'sfinal motif, which occupiedhim for well conveythe obviousimpression that the afternoon over25 years,was based on his gardenat Giverny, sceneis much brighterthan the morning scene particularlyhis waterlilies and his Japanese-style when there is relativelylittle differencebetween bridge.The bridgeon the top in Figure 1.26was thetwo paintingsin theiraverage light reflectance? paintedin 1899and the one on the bottomabout Monet seemsto havediscovered another funda- 20 yearslater. What was the basis for theenormous mentalcharacteristic of the visual system.As the difference?In the interveningyears, Monet's cat- overalllight levelincreases, so doesthe perceived aractshad matured. One sees not only a shiftin col- contrast- yellowsand bluesbecome more yellow ors from bluesand greensto yellowsand browns, or blue,blacks become blacker and whites become but alsoless distinct forms on thebottom painting, whiter.The lossof sensitivityto absolutelight level no doubtdue to thescattering oflight causedby his may resultlargely from sensitivityadjustments of cataracts.This is an opticaleffect that neuralpro- the conepathways, but the changesin contrastre- cessespreviously described cannot compensate. quirean explanationat a neurallevel in the oppo- Althoughthe onsetof his visualloss was gradu- nentpathways (Jameson and Hurvich, l975). al, he seemsnot to haveremarked about it until about1 908, at age68. Fouryears lateq a Parisdoc- tor confirmedthe diagnosisof bilateralcataracts made by Monet's country physician. As his 1.9 Monet Returnsto London cataractsbecame worse, he found it impossibleto paintin bright light or to depictscenes with bright In the fall of 1899, Monet traveled back to backgrounds again,to be sure,due to the scat- London.From the SavoyHotel and St. Thomas's teringof light andthe concomitantdegradation of Hospitalhe would paint WaterlooBridge and the the retinalimage. Housesof Parliament.He madeseveral trips over Despite his poor vision, Monet pursued his the next five years,during which he cornpleted dreamof manv vearsto createvast canvasesthat 34 1. Aging thlough the Eyes of Monet

t- ,+I

Fig. 1.25:Top: ClaudeMonet (1871)Boteuur clans le Bussincle Lottth'es. [Bocrts on the Tltanres,Lontlon.J oil on canvas.17 x72 cnl. privatecoilection. Monte carlo. Bottom: Claude Monet (1904) Lonttres,le Puilentent,Troute clesoleil tluns le ltttuil- lard. [Hou.sesof Puiliuntenl(Ru.r's o/ Sunin the Rtg).] oil on canvas,g l x 92 cm. Mus6e d'Orsay,Paris. l.l0 WaterLilies andCataracts J)

Fig. 1.26: Top: Claude Monet (1899) Le bussittaux nvnphias. fweter- lillies antl Japanese Bridge.J Oil on canvas, 91 x90 cm. The Art Museurn,Princeton University. From the Collection of Williarn Church Osborn,Trustee ofPrinceton University(l9l4 I951), Presidentofthe MetropolitanMuseum of Art (1941-1947);given by his family.(Photo credit: Clerl Fiori.) Bottom: Claude Monet (1918-22) Nymphias: Le pont Japanais. [Japanese Bridge.] Oil on canvas, 89x ll6 cm. The Minneapolis Instituteof Arts. 36 L Aging through the Eyes of Monet would surroundthe entire interior of a room, de- little blue in his paintingsat this time,presumably picting water and plants in a mannerthat revealed becausehis densecataracts would have transmit- the elusivebrilliance of nature.He pursuedthe ted so little short-wavelight that blue would have projectwith vigor and evenbuilt a largenew stu- beenindistinguishable from black.The compensa- dio to accommodatehis large-scalecanvases. tion processesdescribed above may havereached Monet called this project a bit of a folly and the physiological limit so that by the summer of referred to the paintings as his "Grandes 1922 he found it necessaryto slop painting. D6corations."At this point he refused cataract Clemenceaufinally convincedMonet to go ahead surgery, fearing it would make his vision even with cataractsurgery in his right eye that year. worse.Yet he could no longer discriminatebe- Within six monthsof cataractextraction, Monet tweenmany of his paints,relying instead on read- developeda secondarycataract, an opacificationof ing the labelsfor selectinghis colorsand thenre- the posteriorcapsule ofthe operatedeye. This is a membering their precise arrangement on his commoncomplication of cataraclsurgery; and al- palette.He realizedthat manyof his pictureswere thoughit did not surprisehis physician,the opaci- quite dark; and on severaloccasions, after com- ty was a traumaticdevelopment for Monet. In July paring thesecanvases with his earlierworks, he 1923 the cloudy membrane was extracted in slashedand destroyedthem. Monet's home in Giverny. Monet was prescribed During these years, Monet received frequent glasses,but they causedhim to experiencedouble visits from his long-time friend, Georges visionand optical distortion. He discoveredthat his Clemenceau,a remarkableman who hadascended vision improvedif he coveredone eye, usually the to therank of Premierof France.By thebeginning left. Now Monet complainedthat throughhis left of 1917,the end of the first World War was in eye,with a remaining cataract,everything was too sight, and Monet had agreedto donatetwo large yellow, while throughthe eye with the cataractre- panelsofhis "GrandesD6corations" to Franceto moved(aphakic eye), he experiencedeverything as celebratethe armistice. Clemenceau.however. too blue. Figure 1.27 showshis paintingsof his convincedhim to donatenot two panelsbut all 12 House Seenfrom the Rose Gardenwhich are be- thatMonet had planned" with the stipulationthat a lievedto be paintedwith only oneeye or the other building would be constructedto housethem in (Lanthony, 1993). The differencebetween the the mannerthat Monet had envisioned.The num- views throughthe diflerent eyesis striking. ber waslater increased to 19 panelsin orderto ac- Onemight wonderwhy the differencein yellow commodatethe architect'splan to housethem in filters in Monet'seyes should so strikinglyaffect the Orangeriedes Tuileries in Paris.There was, of his choiceof colors.After all, in the aphakiceye, course, some question in Monet's mind about Monet's retina should receivemore short-wave- whetherhe would havethe energyand the eyesight lengthlight reflectednot only from the scenehe is neededto completethis enormousundertaking. trying to depictbut equallyfrom theblue paints on Monethad once said that he wishedhe hadbeen his palette.The net effectwould thereforeseem to born blind and then suddenlymade to seeso that requirethe samematch between the sceneand the he couldpaint his impressionswithout the biasof canvaswith or without a denseyellow cataract. prior experience.Soon, a version of his wish From this point of view, the yellow filter should would be granted.By 1922,at the age of 82, he haveno effecton Monet'spaintings because it lies had becomeessentially blind in the right eyeand in front ofboth the originalscene and his palette. had only a little usefulvision in the left eye,ac- This analysiswould be correctifour visualsys- cording to his medical records.Determined to tem were capableof respondingto each wave- Fit continue,he saidthat "I will paint almostblind as lengthoflight separatelyand ifthere wereenough IH Beethovencomposed completely deaf" (Stuckey, pigments to match each wavelength. However, Mi 1995,p. 251). To improvehis vision for a few neither of these two requirementsis met. As Bo hoursat a time, he useda prescribedmydriatic to Clemenceauput it to Monet, "The steelof your Se, dilatehis pupils (Ravin, 1985).Monet usedvery eyesightbreaks the crust of appearances,and you Mr Water Lilies and Cataracts

Fig. l.27: Top: Claude Monet (1925) La maison vue du jardin altx roses. [House Seenfrom the Rose Garden.J Oil on canvas, 8l x 92 cm. Mus6e Marmottan,Paris. Bottom: Claude Monet (1925) La maison vue dujardin aux roses. [House Seen from the Rose Garden.l Oil on canvas, 89x100 cm. Mus6e Marmottan, Paris. 38 l. Aging thlough the Eyes of Monet penetratethc'ir-rner substance of tliings in order to glassesthat Monet desclibcclas qr"ritesatisfactory. clecomposcit into projectofsof light rvhichvou re- No doubt this outcome\\'as dLre to chronraticadap- colxposeu'ith the brush so that vor-rnlry reestab- tation in his visual systeuro\cr this periocl. lish subtly upou ollr retinasthe ef-fectol'scrrsatior.rs In .Tuly 1925. tliree ycars afier his original in their Iullest intensitv" (Clemenccau.1930. cataractcxtraction. Monet clcclaredthat his color pp. 18 l9). Expressecllcss poetically. the painter r isiou uas conrpletelylestorccl and his t.t.tooduas cloesnot rnakea physicalrnatch betnecn the orig- cbLrllier.rt.Non ll-5vears olcl. he resutueclIiis n'ork inal scene and the canvasbut a visr,ral(approxi- and cornpletc-dnot thc lc) plomised canvasesfbr mately mclanreric)nratch. These nratchcsdepencl his "GlanclcsDccolatious" but 22. ort tlte sltrrPcsof tltc lcccplor lctiott:lteclrt Monet never sari' thesc canvAsesclisplayed in (Fig. 1.6).anti the latterarc nrodifiecl (ns specifiecl the' roor.n that he hacl cnr isionecl.He' died on at the conreu) by age-rclatcdcharrgcs in the lens -5 Decenrber1926. Ncxt to his bed nas a book (WLight. l92li 29). In short. thc scncscentor' opened to Bauclclaire'spocnr. "The Slrangcr" cataractouslcns moclificsthe receptorscnsitivities (Vidal, 1956).I1 -soes. "Tcll nrc"enignlatical rran. irr tn'o u'ays: First. rt ultct'.sthe .shultt'.sof thc le-- whom do you love best.vour lather'.your It-tother. ceptor action spectra(spccificcl at thc colnca) by vour sister.or vour brothcr'.)"To uhich tltc rlan leducing scnsitivity nrorc at short \\ravclengths rcplied. "I lovc the cloucls." It ri'as fitting that than at nricldlcand long rvavelengths.Sccond. it Clemenccaurvas by his siclcattcl lvas thc one to raducc.tthc rclutit'c ltcigltt.so1'the action spectra. close his evcs.Monet oncc saiclto C'lcnrcncc-au. aftecting S-cones nrorc than M- ol L-cones. "l)ut your hanclin n-rineand lct Lrshelp onc anoth- I)r'occsscssuclt as color nratching\\hich dcpencl cr to scc thrrt-usbctter" (Tuckcr. lt)95. p. 225).T\itl ()n llrc shitl)csof tltc pltiltolcccPlorlre lrort :pL'cllil rrrorrthsalicl closin-qMonct's cyes. ('lcrtrcnccau cannot bc conrperrsatcclncLrrally bcciruse eacli hclpcclthc rvolld to scc bcttcr through Monct by conc typc otrcvsthc principlcof'unirariarrcc (i.c.. opcnin-uthc Oranscric clcs Tuilcrics n'ith his ttll tb:orltatl quantaproclucc thc sarlrcrcsl-ronsc in "Grandcs Dccolations." a photorcccptor).r.r,hilc rclativc scnsitir,ities or' hcightsol'thc curvcs can bc conrpcrrsaledto a largeextcrtl. Thc nct cll'cclis thatlcnticulal scncs- cenccallccts colol rnatchcsn'rolc than color ap- l.1l Summarv peamncc(Wcrner. I 9c)(r). Followingcataract sllrltcry. paticnts ollcn reporta Thc lifc spanol'thc In'rplcssionistC'lauclc Monct, reslrrgenccof blue expclicnccclthrouqh thc opcrat- llt40 to 192(r.cnconrpassccl sonre ol'thc nrost in'r- eclcyc. as rvoulclbc cxpcctcclclue to thc rcmoval ol portant clevclopnrentsin hori,colol is nou uncler- thcir dcrrscycllon lcns.but thcy aclaptrather qLrick- stoodin art unclsciencc. ln his paintings.Monet ly ovel thc courseol- lvccks or months.Monet clicl rnadcthc cphcmcralclflcts of light anclcolor his not aclaptso cluickly.but soureof his dilf iculty mly ccntralsr.rt-r.jcct nlatter. IIc is likcly to havclrecn in- havebccn lvith his irlpaticncc.u'hich lcd his physi- flLrenceclby.l. W v. Gocthc via.l. M.W. Tut'tteraucl cian lrcqucntly to changc the prcsclibcclcolorcd M. E. Chci'r'culi'ia E. Dclacroix.nho cxploitecl lenses that r'"'ereintenclccl to eqllatc color appear- neu icicasabout aclclitivc color nrixturcancl simul- ance in thc tr,voeyes (Ravin. l9tl5). This proccss taneolls contrast. Paintings fl'orr this periocl comurcncctlrvhen a Parisophthalnrolo-uist, Jacqucs provide usclirl illustlatiorrsof thc discovcriesancl Man'its.1-r'cscribed a pair of glasscsthat rrreletintccl tltcoriesol Nltrrtet'seortte rttPtrtatics in sciL'rlcc.itt- vellot ish-qrc'en.This c-IinrinatedMonct's plevit'lrs clucling.1. ('. Marrrell" II. r. Heln.rholtz.[-.. Hcling conrplaintthat he seesnothing br-rt lrlr"rc. Up to that arrcl.1. r'. Krics. althor"rshhc *as probably not di- point.hc rclieclon his lcli cye lbr painting.but nou, rcctly inflLrenccclby thenr.Their scicntific theories su'itchcclto his right cyc. Monet ancl Dr'. Mawas are col'ncrstolresftx cLrrrcntthinkin-q abor.rt colot' tlied glasscsrvith varioustints during the next tn'cr vision and providea uselirl ll'amervorkfirr analvz- vears.In the end.thcy scttledon a pair of untinterl ing age-relatcclchanges in color perccl.rtion. 1.1I Summary 39

Monet'svision changedduring his life, perhaps press).M- and L-conesin early infancy: Il. Action duein partto acceleratedaging caused by sunlight spectraat 8-weeksofage. Vision Res. to which he was often exposedby virtue of his Callen, A. (1982). Techniquesof the Impressionists. paintingen plein air. Light, especiallyUV light, (London: QED Publishing). Chevreul, M. E. (1839). De la Loi du Contraste may acceleratethe normal age-relatedchanges in Simultan6 des Couleurs. Paris: Pitois-Levrault. the lens and photoreceptors.The cone pathways [trans. by F. Birren (1967) as The Principles of lose sensitivityon a continuousbasis from early Harmony and Contrast of Colors and their adulthoodto old age.When expressed in termsof Applications to the Arts.] (New York: Van Nostrand sensitivityat the cornea,S-, M- and L-conesap- Reinhold Company,Inc.). pearto losesensitivity at approximatelythe same Clark, K. (1960).Looking at Pictures.(London: John ratewith age.This is somewhatsurprising because Murray). senescentchanges in the lens producea selective Clemenceau,G. (1929).Claude Monet. Reprintedin: loss in the amountof short wavelengthlight that Monet: A Retrospective,C.F. Stuckey,ed. (New pp. can reach the retina, a reduction commonly York: Spadern,1985), 350-366. Clemenceau,G. (1930). Claude Monet: The Water thoughtto reducesensitivity of the visual system Lilies. (GardenCiry New York: Doubleday,Doran to blue hues.Monet's reaction to his own senes- & Company) centlens. culminating in a cataract,has been taken de Weert,C. M. M. (1991).Assimilation versus con- to support this view Recent studies,however, trast.In: Frorn Pigmentsto Perception,A. Valberg show that the visual system adaptsto normal and B. B. Lee, eds. (New York: Plenum), lenticularsenescence and activelyrebalances the pp. 305-31 1. sensitivityof color mechanismsto supportcon- de Weert, C. M.M. and Spillmann, L. (1995). stancyof color perceptionacross the life span. Assimilation: Asymmetry betweenbrightness and darkness?Vision Res. 35, 1413-1419. DeValois,R. L. andDeValois, K. K. (1993).A multi- stagecolor rnodel.Vision Res.JJ, 1053-1065. Acknorvledgements DLichting,H. and Sagner-Diichting,K. (1993). Die Malerei des deutschen Impressionismus.(Koln: I am -rrratefulto the Alexandervon Humboldt DuMont Buchverlag). Foundation(Bonn, Germany) for a SeniorScien- Gage, J. (1972). Turner: Rain, Steam and Speed. tist An'ard. Studiesconducted in my laboratory (London: Allen Lane the PenguinPress). were made possiblethrough the supportof the Goethe,J. W. v. (1810). Zur Farbenlehre.(Ttbingen: NationalInstitute on Aging (grantAG04058). The J.G. Cotta. Diederich,Jena). (Translation arranged helpfulcomments of Michelle L. Bieber,Philippe and edited by R. Matthaei as: Coethe's Color Theory.New York: Van Nostrand"1971.) Lanthony.Jonathan O. Roberts,Brooke E. Schefrin, Ham, W. T., Mueller, H. A., Ruffolo, J. J., Guerry,D., ElizabethJ. Smith,Vivianne Smith, Lothar Spill- and Guerry,R.K. ( 1982).Action spectrumfor reti- mann, Floyd Ratliff and Michael Wertheimerare nal injury from near-ultravioletradiation in the gratefullyacknowledged. aphakicmonkey. Am. J. Ophthalmol.9-1, 299-306. Helmholtz, H. v. (1867). Handbuch der Physio- logischen Optik. (Hamburg: Voss). [third edition translated as: Helmholtz's Treatise on Phys- References iological Optics; J. P.C. Southall, Ed. Rochester, NewYork: Optical SocietyofAmerica, 1924.] Barnes,R. (1990).Monet by Monet.(New York: Herbert, R. (1979). Method and meaning in Monet. AlfredA. Knopf.1. AIA 67,90-108. Berlin,B. and Kay, P. (1969).Basic Color Terms, Hering,E. (1887).Ueber die Theoriedes simultanen Their Universalityand Evolution. (Berkeley: Contrastesvon Helmholtz. Pf. A. 40, 172-191. Universityof CaliforniaPress). Hering, E. (1920). Outlines of a Theory of the Light Bezold, W. v. ( 1874) Die Farbenlehre.(Braun- Sense. L.M. Hurvich and D. Jameson, trans. schweig:Westermann). (Harvard U. Press,Cambridge, Mass., 1964;origi- Bieber,M., Knoblauch,K., and Werner,J. S. (in nally publishedby Springer,Berlin). 40 l Aging through the Eyes of Monet

Jameson,D. and Hurvich, L. M. (1955). Some quan- Perry,L. C. (1927).Reminiscences of ClaudeMonet titative aspects of an opponent-colors theory: from l8E9 to 1909.Am. Mag. Art 18, (March, I. Chromatic responsesand spectral safuration. J. No. 3), I 19-126. Opt. Soc. Am.45, 546-552. Pissarro, J. ( 1990). Monet3 Cathedral: Rouen Jameson, D. and Hurvich, L. M. (1968). Opponent- 1892-1 894. (ltior lbrk Alfred A. Koopf). response functions related to measured cone pig- Pokorny.J- ad Smiil !: (1997). How much light ments. J. Opt. Soc. Am. 58, 429430. reachesilE mim? ln Colour Vision Deficiencies Jameson,D. and Hurvich, L. M. (1975). From con- XIII, C. R- Carooiur ed- (Dordrecht:Kluwer), trast to assimilation: In art and in the eye. Leonardo pp.49l-51l. 8, 125-131. Ratliff. F. (19?6f-On the psychophysiologicalbases Jameson,D. and Hurvich, L. M. (1989). Essaycon- of universalcola tcflns. Pro.Am. Phil. Soc.124 cerning color constancy. A. Rev. Psychol. 44 3 I t-330. 1 ">'' Ratliff, F. ll992r. Frrl Signacand Color in Neo- KaIz,D. (1911).The World of Colour.fEnglish trans- Impressimisu. ()icrr York: The Rockefeller lationby R. B. Macleod andC. W. Fox,published Universityh,ess|. in 1935)1,(London: Kegan Paul, Trench, Trubner Ravin,J. G. (1965)-Voct's cataracts.J. Am. Med. & Co). Ass.25J. 394-399- Kirschfeld"K. (1982).Carotenoid pigments. Proc. R. Reynolds.G. (l96el Ttnner.(London: Thames and Soc.London B. 216,7l-85. Hudson). Kraft, J.M. and Werner,J. S. (1994).Spectral effi Ruskin. J. (l&{3}. Modern Painters.Vol. L Of ciencyacross the life span:Flicker photometry and Generalniritcs ad of Truth. (London:Smith, brightnessmatching. J. Opt. Soc. Am. A. .11, Elder& Co.l" t2t3-r221. Russell,J. ( 1974I Thc Mcaningsof ModernArt. Vol. Kraft, J.M. and Werner,J. S. (1997). Age-related l. The Sccm Rcrolrin- (NewYork:The Museum changesin saturationof non-spectrallights. In: of ModcrnAnl John Dalton's Color Vision Legacy, C.M. SchefrinB. E. ead$trncr. J-S. ( l 990).Loci of spec- Dickinson, I. J. Murray and D. Carden, eds. tral unigrr hn eoqghout the life span.J. Opt. (London:Taylor & Francis),pp. 553-560. Soc.Am- A. 7.305-31I Kries, J.v. (1882). Die Gesichtsempfindungenund Schefrin,B- E. od \fnrncr. J. S. (1993).Age-related ihreAnalyse. (Leipzig: Veith & Co.). changesin thc cobr rppcaranceof broadbandsur- Lanthony,P. (1993). La cataracteet la peinturede faces.Color Rcs ad Appl. /8. 380-389. ClaudeMonet. Points De VueNo. 29. 1215. Schefrin. B. E. Sliml K.. and Werner.J. S. Lennie,P andD'Zmura, M. (1988).Mechanisms of (1995).Cootnlrrirc of mural pathwaysto age- color vision. CRC Critical Rev. Neurobiol. 3, related losscs io cLmic discrimination. J. Opt. 333-340. Soc.Am. A- ,L 1233-1241. Lerman, S. (1980). Radiant Energy and the Eye. Schefrin,B. E.. frytrs. I S.. Plach,M., Utlaut, N., (lrlewYork: Macmillan Publishing Co.). and Switkes,E. (1992).Sites of age-relatedsensi- Marshall,J. (1978). Ageing changes in humancones. tivity loss in a sborr-rare cone pathway.J. Opt. In: XXIII Concilium Ophthalmologicum,Kyoto, Soc.Am. A. 9. 355-363. K. Shimizu and J.A. Oosterhuis,eds. (Elsevier Seitz, W. (1956)- Mocr and abstract painting. North-Holland,Amsterdam), pp. 375-37 8. Reprintedin:MqEA Rcaospective,C. F.Stuckey, Marshall, J. (1985). Radiationand the aging eye. ed.(NewYort: Spedcr&1985), pp. 367174. OphthalmicPhysiol. OpL 5, 241-263. Shinomori,K., Schcfrin. B. E., and Werner,J. S. Maxwell, J.C. (1860).On the theory of compound (1997).Spectral nslnnisrns of spatiallyinduced coloursand the relations ofthe coloursofthe spect- blackness:data ad $otitative model.J. Opt. Soc. rum.Philos. Trans. R. Soc.London 150.57-84. Am. A. 14.372-387- Maxwell,J. C. (1872).On colorvision. Proc. R. Inst. Signac,P. (1921). D'E€Soe Delacroixau N6o- G.8.6.26017t. Impressionnisnr.(Paris: H. Floury). [trans.and Myers,B. (1990).Methods of the Masters:Monet. ed. by W. Silverrmn in F. Ratlitr (1992) Paul (London:Park Lane). Signac and Color in N+lmpressionism. New Newton, l. (1704). Opticks: Or, a Treatiseof the York: The Rockcfeller University Press, Reflexions,Refractions, Inflexions and Colours of pp.193-2851 Light. (London:S. Smithand B. Walford). Smith,V C. andPokorny. J. ( 1975).Spectral sensitiv- 1.1I Summary 4l

ity of the foveal cone photopigments between 400 Osborne and J. Chader, eds. (Oxford: Pergamon and 500 nm. Vision Res.1l 16l-171. Press),pp. 621-645. Spillmann. L. and Werner,J. S. (1996). Long-range Werneq J. S. and Schefrin, B. E. (1993). Loci of interactions in visual perception. TINS 19, achromaticpoints acrossthe life span.J. Opt. Soc. 428434. Am. A. 10. 1509-1516. Stuckey, C. F. (1995). Claude Monet: 1840-1926. Werner,J. S. and Spillmann,L. (1989).UV-absorbing (Chicago:The Art Instituteof Chicago). intraocular lenses: Safety, efficacy, and conse- Tucker, P.H. (1995). Claude Monet: Life and Art. quences for the cataract patient. Graefe's Arch. (New Halen: Yale UniversityPress). Clin. Exp. Ophthalmol.2 2 7, 248-256. Vidal, H. (1956). RememberingClaude Monet. Werner,J. S. and Steele,V G. (1988). Sensitivityof Reprinted in: Monet: A Retrospective, C. F. human foveal color mechanismsthroughout the Stuckey. ed. (New York: Spadem,1985), fife span.J. Opt. Soc.Am. A. 5,2122-2130. pp.349 350. Werner,J. S. and Walraven,J. (1982).The effectsof' Volbrecht.V J. and Werner,J. S. (1987). Isolation of chromatic adaptation on the achromatic locus: short-ualelength-sensitivecone photoreceptorsin Role ofluminance, contrast,and backgroundcolor. 4-6-ueek-old human infants. Vision Ptes. 27, Vision Res.22, 929-943. 469478. Werner, J. S. and Wooten, B. R. (1979). Opponent- Vollard A. ( 1925). Renoir, an intimate record. chromaticmechanisms: Relation to photopigments Reprintedin: Renoir:A Retrospective,N. Wadley, and hue naming.J. Opt. Soc. Am. 69, 422434. ed. (Nes \brk: Hugh Lauter Levin Associates, Werner, J. S., Peterzell, D. H., and Scheetz,A. J. 1987).pp. 30,+-308. (1990). Light, vision, and aging. Optom. Vision Vos, J. J. ( I 978). Colorimetric and photometricprop- Sci.67,214-229. ertiesofa 2'fundamental observer.Color Res.and Werner,J. S., Steele,V G., and Pfoff, D. S. (1989). Appl.3.125-128. Lossof humanphotoreceptor sensitivity associated Vos, J. J. and \\alraven, P.L. (1971). On the deriva- with chronic exposure to ultraviolet radiation. tion of the foleal receptorprimaries. Vision Res. Ophthalmology96, | 552-1558. 1/, 799 8r8. Wiesel, T. N. and Hubel, D. H. (1966). Spatial and Weale,R. A. ( 1957).Trichromatic ideas in the seven- chromatic interactions in the lateral geniculate teenth and eighteenth centuries. Nature 179, body of the rhesusmonkey. J. Physiol.,Lond. 29, 648-65l. I I 16-1156. Weale,R. A. (1971).The deathof Pointillism.The Wright, W. D. (1928-29). A re-determinationof the Listenert{ \larch). 273-275. trichromatic coefficients of the spectral colors. Weale, R.A. (1982). A Biography of the Eye. Trans.Opt. Soc.-i0, 14l-164. (London:H.K. Lewis& Co. Ltd.). Young, R. W. (1982). The Bowman lecture, 1982. Weale,R. A. (1988).Age and transmittanceof the Biological renewal. Applications to the eye. T. human cnstalline Iens. J. Physiol., Lond. -19J, Ophth.Soc. 102,42-75. s77-87. Young, R. W. (1991). Age-Related Cataract.(New Werner,J. S. ( 1982).Development of scotopicsensi- York: Oxford University Press). tivity and the absorptionspectrum of the human Young,T. (1802). On the theory oflight and colors. ocular media.J. Opt. Soc. Am. 72, 247-258. Philos.Trans. R. Soc. London 92, 1248. Werner,J. S. ( I 991). The damagingeflects of light on Zrenner, E., Abramov, I., Akita, M., Cowey, A., the eyeand implicationsfor understandingchanges Livingstone,M., andValberg, A. (1990).Color per- in vision acrossthe life span. In: The Changing ception: Retinal, geniculate,and cortical mecha- Visual Sr,stem: Maturation and Aging in the nisms. In: Visual Perception: The Neuro- CentralNenous System,P Bagnoli andW. Hodos, physiologicalFoundations, L. Spillmann and J. S. eds.(Nerv \brk: Plenum), pp. 295 309. Werneq eds. (San Diego: Academic Press), Werneq J. S. (1996). Visual problems of the retina pp.163-203. during ageing. In: Prog. Retinal Res., N. N.