Charting Color from the Eyeof the Beholder

A century ago, artist Albert Henry Munsell quantified colors based on how theyappear to people; specializations of his system are still in wide scientific use

Edward R. Landa and Mark D. Fairchild

Say the phrase "school bus" to anyAmerican, and the image sure to

spring to mind is a vehicle of a partic-ular color—a standard yellow that isone of the most easily recognized andevocative product colors in the UnitedStates. But how is it that all school busesacross the country are exactly the sameyellow color?

It is neither practical nor feasible forall school buses to be painted with thesame batch of paint. Instead, specificcolor standards exist that assure that allpaint manufactured for school buses iswithin a color tolerance—and that thecolor stays within that tolerance whenthe paint is applied to the bus.

Where do such standards comefrom? Are they simply calculationsmade by a physicist or chemist wield-

EdwardR. Landa is a research hydmlogisi withthe U.S. Geological Siin'ey in ResUm. Virginia. Hereceived his B.S. in geolog\from Ihe City- CollegeofNew Yorkandhis M.S. andPh.D. in soil scienceft-om the University of Minnesota. His researchfocuses on the behavior ofrnelals and radiomi-clides in soils and aquatic envimnments. Mark D.Fairchild is the Xenix Pmfessor of Color Scienceand director of the Munsell Color Science Lcihora-tory at fhe Rochester Instituie ofTechnolog\- (RIT).He received his B.S. and M.S. degrees in imagingscience from RIT and his Ph.D. in vision sciencefwrn the University- of Rochester He was chairof the Commission Internationale de L 'Eclairage(CIE) Technical Committee 1-34 vn color appear-ance models. He is author tyColor AppearanceModels 2nd Ed., which serves as a reference to thefundamentals of color appearance and the formu-lation of specific models. Address for Landa: U.S.Geological Survey. 430 National Center, Reston.VA 20192. Internet: erlanda(a)usgs.gov

ing a spectrometer? When it comesto the colors that we use in life, theanswer is not that straightforward. Itis common to say that certain wave-lengths of the electromagnetic spec-trum are a given color, but in truth, itis more correct to say that those stim-uli arr perceived to be of a certain colorwhen viewed under specific condi-tions. So without the human observer,there is really no color, and practicalstandards for color must take this fac-tor of human perception into account.

The work of mapping human colorperception began in earnest in the 19thcentury. This year marks the centenaryof the first atlas of color, by an artistnamed Albert Heru-y Munsell.

Munsell's vision of a systematic wayto communicate color appearance hashad an impact on essentially all mod-em systems of color measurement andspecification. Today a school-bus man-ufacturer, for example, has access to anumerical standard based on instru-mental measurements of the spectralreflectance characteristics of the paint.Undergirding these numbers are colori-metric computations based on a systemcalled CIE, for Commission Interna-tionale de L'Eclairage, or InternationalCommission on Illumination. These in-strumental specifications and tolerancesallow paint and pigment manufactur-ers to formulate paint that matches thedesired school-bus color more easily.CIE charts of these numbers are alsobacked up with physical samples thatshow the optimal school-bus yellow aswell as limits along the perceptual colordimensions of lightness-darkness, hue

and chroma (or how much the apparenthue differs from neutral gray)—the di-mensions first described and measuredby Albert Munsell.

Such systems are used to create andcontrol almost every colored product.Examples include the textiles clothesare made of, the inks used in printers,the encoding of digital television sig-nals, the colors of signal lights, the prop-er cooking times for French fries, thepearlescent paints on automobiles thatchange color with angle, the color of thebeer at the local pub and the colors ofcrayons that children carry off of schoolbuses to create art in school.

A Colorful HistoryColor description was an interestingproblem to scientists and artists in the19th and early 20th centuries, engag-ing scholars such as Germany's JohannWolfgang von Goethe and Willielm Os-twald, and the American printer andeducational reformer Milton Bradley,best known today for his board games.

Munsell's interest in the descriptionof color began in 1879, white he was astudent at the Massachusetts NormalArt School (MNAS, now the Massachu-setts College of Art), when he read thethen-just-published text Modern Chro-matics with Application to Art and Indus-try by Columbia University physicistOgden Rood.

Although Munsell was certainly in-fluenced and motivated by the colorsystems reviewed in Rood's book, hisown system would have several fea-tures that had not been previously de-scribed, most importantly the relations

436 American Scientist, Volume 93

Figure 1. Albert H. Munsel! was dissatisfied with what he felt were arbitrary definitions of color, soin 1905 the artist published a color charting system based on how people see color. Munsell intendedhis system for art students, but today versions of his charts are used in fields from soil science to artrestoration. Here, Roy S. Bems of the Rochester Institute of Technology performs a visual match ofpigments in t.eorges Seurafs A Suiidu}/ on La Cniiiile liitte-~tS84, using color samples from a Mun-sell Book of Color. Bems used color matching to confirm instrumental measurements of the painting.By calculating the effects of aging on the paint pigments, Berns was able to digitally recreate thepainting's original colors. As seen in close-up (aK the yellow paints, rejuvenated in the digital recre-ation (b), have browned with age in the actual painting (c). (Images courtesy of Roy S. Bems.)

between lij^htness, luminous intensityand maximum colorful ness.

After his graduation in 1881, Munsellbecame an instructor at MNAS, spe-cializing in artistic anatomy and colorcompositit.)n. In 1892, while sketchingin Venice with Boston artist DenmanRoss, Munsell noted in his diary thatthey discussed tiie need for a "systemat-ic color scheme for painters" to preparementally and sequentially for tlie layiiigof the palette (Ross, who taught art atHarvard, would go on to develop hisown color system with a nijie-step valuescale in 1907),

Munsell was not satisfied with colordescriptions of the day, such as "topazyyellow" or "Indian red." His quantita-tive research in the field seems to havebegun on a summer vacation in 18y8,when, to create a teaching aid for hiscolor-composition students, he workedon distiLitt designations of colors andthen display on wheels and spheres.

His first model Lised a child's globeas a base for rotary color mixtures. Withthe globes, Munsell was able to createand demonstrate colors of various huesthat had equal lightness and chromaticbalance such tiiat they mixed to a neu-tral gray when the globes were spun.This allowed Munselt to demonstratehis concepts for an impressed Rood,who stated that Mui-isell "put an artisticidea into scientific form." Mmisell alsolater created a portable visual photom-etei to allow him to an urately scale theperceived lightness ot chromatic colorsthrough visual matchiiig to a gray scale.

Munsell's artwork focused mostly onportraits and seascapes. His paintingof Helen Keller hangs in the AmericanFoundation for the Blind headquar-ters in New York City, A view from theRoosevelt summer home at Campo-bello, painted by Munsel! for the fam-ily ui 1890, is displayed In the office ofFranklin Uflano Roosevelt's mother at

Hyde Park, New York, thus, he waswell known in the art world, but as hedeveloped his color system, Munsellalso moved easily among the scienceand engineering faculties in Boston. Heconsulted with individuals such as thenoted physiologist Henry PickeringBowditch at Harvard Medical School,

In 1905, intent on the goal that "col-or anarchy is replaced by systematiccolor description," Munsell publishedthe first edition of his 67-page text AColor Notation, which described theMunsell Color System and its specifi-cation of color by measured scales ofhue, value and chroma, H. E. Clifford,Gordon McKay Professor of ElectricalEngineering at Harvard, provided anintroduction to this volume.

Although Munseli honed his colorsystem within a circle of scientists, hesaw its broad application in general so-ciety, and his color spheres reached awide audience. Eor example, around

2005 September-October 437

1901, William Filene of Filene's Depart-ment Store asked Munsell to show thecolor sphere to the Shopkeepers' Asso-ciation in Boston. In the 1920s, the FileneCompany underwrote the cost of a colorchart prepared by the Munsell ColorCompany for use in the clothing indus-try. At about the same time, another ear-

ly Munsell color sphere was on displayand available for study at the PhysicalLaboratory of Columbia University.

Munsell's health declined rapidly af-ter a 1914 trip to Europe. He underwentsurgery for appendicitis in May 1917.Later that year, he gave up the studio hehad occupied since 1901; several montKs

Figure 2. Brightness and colorfulness are per-ceptions of absolute values; lightness andchroma are relative perceptions. Newspaperand white paper (top) each have a lowerbrightness in dim light than in intense light.However, newspaper always has a lowerlightness than white paper, so will appeardarker even when it achieves a high bright-ness in intense light. Similarly, a beach ball(middle) will appear more colorful underhigh light than low light, but its chroma—itsdifference from a neutral gray in the samelight—remains about the same. Because onlyrelated colors can exhibit lightness, an unlitlight bulb (bottom) may appear gray, but anilluminated bulb always appears white.

later it was taken over by the painterJohn Singer Sargent. Munsell died onJune 28, 1918, at age 60. The MunsellColor Company, which was establishedcirca 1917, carried on the commercializa-tion of his work, selling color charts andother products under the leadership ofhis son Alexander Ector Orr MunseU.

Through 1946, 15 editions of A ColorNotation were published by the MunsellColor Company. Other books followed,including the Atlas of the Munsell ColorSystem published in 1915, and A Cram-mar of Color, published posthumouslyin 1921. No other color system from thistime period has been as long-lived, com-mercially successful or influential.

Dorothy Nickerson, who began hercareer as A. E. O. Munsell's secretaryand laboratory assistant in 1921, didmuch to adapt the Munsell Color Sys-tem to commercial applications and todescribe its scientific basis. Her workincluded extensive involvement on theOptical Society of America ColorimetryCommittee and historical publicationson the Munsell System. Nickerson au-thored more than 150 papers and roseto prominence within the Inter-SocietyColor Council, which established itsNickerson Award in 1980.

The legacy of color research pioneeredby A. H. Munsel! is today honored bythe Munsell Color Science Laboratory,founded at the Rochester Institute ofTechnology in 1983.

Defining ColorIn any field of stvidy, it is important tohave a standard \'ocabulary. In the studyof color this vocabulary is often mud-dled, as terms such as "lightness" and"brightness" are casually interchanged.Even in education, treatment of color isinconsistent. To the grade-school child,color might be made up of three prima-ries: red, blue and yellow. The printeris taught that the three primaries arecyan, magenta and yellow, whereas thetele\'ision engineer is taught that color ismade up of red, green and blue. Finally,the physicist might be taught that colorsare divisions of the visible portion ofthe electromagnetic spectrum. Althougheach of these concepts of color is correctin its context, it is incorrect in others.

In the field of color science, the de fac-to standard for vocabulary comes fromthe International Lighting Vocabulary,published by the CIE, which clarifiesthe distinctions in various dimensionsof color experience. These terms scien-tifically and unambiguously define the

438 American Scientist, Volume 93

Figure 3. Variants of Munsell's color system are in wide use in many areas of science and technology. Botanists use Munsell Plant Tissue Chartsin plant taxonomic classification and for the identification of plant diseases that lead to tissue discoloration, such as potassium deficiency in agrape plant (left). Fabric dyers use color charts to select colors and to keep their products consistent from print to print (right).

perceptual attributes that Munsell n otated in his system. As such, the ftxrus ison the object-surface color descriptors oflightness, chroma and hue that are thosedefined in Munsell's system. But firstone must defijie the term "color" itself.

Few petiple, when asked, can give aprecise definition of what exactly coloris. It is almost impossible to do withoutusing an example. Even the CIE resort-ed to the inclusion of color names in itsdefinition of color. The most critical partof the formal definition of color to keepin mind is that it is an "attribute of vi-sual perception"—a description of colorappearance to human observers.

However, it is known that color per-ception varies somewhat from person toperson—and within individuals duringtheir lifetimes. For instance, as the eye'slens hardens from aging, it absorbs andscatters more short-wave length colors,so the visual system looks through a fil-ter that yellows increasingly with age.

With highly variable human observ-ers having such a central role in colormodels, one might wonder how accu-rate color measurements can be made.Careful experimentation has quanti-fied the human visual response to colorand defined mathematical equivalentsof average human color perception,creating "standard observers" that canbe used to simulate the human visualresponse to color computationally.

Such systems have been standardizedfor nearly 75 years and are very suc-cessful in both scientific research andpractical applications.

Munsell resisted some suggestionsto tie his system to physical scales ofwavelength because he realized thiswas not feasible. Instead he focusedon developing a system that quanti-

tatively and systematically describedthe overall appearance of colors. Colorscientists to this day work to createmathematical simulations as accu-rate as Munsell's quantification of vi-sual observations. That is why, despitesignificant advances in the theory andtechnology of color measurement, hu-man obser\'ers are still used as the final

Figure 4. Dorothy Nickerson (left) was a tireless advocate and researcher for the color systemestablished by Albert Munsell (right}. On her death in 1985, noted color scientist David L.MacAdam of the University of Rochester wrote "Alas, the prophetess of color—inspired byMunsell—is silent." (Photographs courtesy of the Munsell Color Science Laboratory.)

www.arriLTicanscientist.org 2005 September-October 434

"I r0.0 0,2 0.4 0.6 0.8

relative luminance1.0

Figure 5. Color systems based on percep-tion have a complicated relation to physicalcolor measurement. Many combinations ofwavelengths can sum to the same perceivedcolor; for this reason, Munsell concentratedsolely on color perception for the basis of hissystem. Munsell value, or lightness, does notincrease linearly with luminance. A samplethat the human visual system perceives as amiddle gray (with a value of 5 in the Munsellsystem) has a relative luminance of about 20percent (100 percent would look white). Therelation between the other two dimensionsof the Munsell system, hue and chroma, iseven more complicated to relate to physicalmeasurements of light, such as wavelength.

judges of colored products to confirminstrumental results.

Of the properties defined by Munsell,hue is perhaps the easiest of the colorterms to understand. It is defined as thesimilarity of a perception to one of theperceived colors—red, yellow, greenand blue—or to a combination of twoof them. Hue is often described with ahue circle, and this is indeed the way it isdescribed in the Munsell system.

One important point of this descrip-tion, and the definition given by theCIE, is that there are unique hues—^again, red, yellow, green and blue—thatfollow the opponent color theory firstpostulated by Ewald Hering in the late1800s. Hering noted that certain hueswere never perceived together, such asreddish-green or yellowish-blue. Thisformulated the fundamental notion thathuman color vision is encoded into red-green and blue-yellow channels.

Miuisell's hue designation is consis-tent with this theory but adds a fifthprincipal hue, purple, to achieve anotherdesired property: equality of perceptualhue spacing for the entire hue circle. Thisequality of spacing means that the per-ceived change in hue is equal for eachequal numerical step in the Munsell huedesignation. Another reason Munselladded purple as a principal hue is thatthere are a lot of perceptible hue stepsbetween our perceived unique red andunique blue hues. In other words, wecan discriminate many purple hues butnot so many yellowish-green hues.

The attributes of brightness and light-ness are very often interchanged, despitethe fact that they have distinct defini-tions. Brightness refers to the absoluteperception of the amount of light of astimulus, while lightness can be thoughtof as the relative brightness. In otherwords, lightness is our perception ofthe brightness of an object relative tothe brightness of an object that appearswhite to the human eye under similarillumination. The human visual systemgenerally behaves as a lightness, not abrightness, detector, which can perhapsbe better described with an example.

A typical newspaper, when readindoors, has a certain brightness andlightness. When viewed side by sidewith standard office paper, the news-print often looks slightly gray, whilethe office paper appears white. Whenthe newsprint and office paper arebrought outdoors on a sunny day, theyboth have a much higher brightness.Yet the newsprint still appears darkerthan the office paper because it has alower lightness. The amount of lightreflected from the newsprint outdoorsmight be more than a hundred timesgreater than the office paper was in-doors, yet the relative amount of lightreflected when comparing the two hasnot changed with the change in illumi-nation. Thus, the difference in bright-ness between the two papers, or theirlightness, has not changed. The Mun-sell value scale is a scale of lightness.

As an interesting note, only relatedcolors can exhibit lightness. Related colorsare colors seen and judged in relationto other similarly illuminated colors.This is the reason that there cannot be agray light bulb. When illuminated andviewed in isolation, the light bulb is thebrightest stimulus in the field of viewand thus appears white.

The definitions of colorfulness and chro-ma are very similar to those of brightnessand lightness in that colorfulness is anabsolute perception, whereas chromais relative. Essentially, colorfulness de-scribes the amount, or intensity, of thehue of a color stimulus. A gray stimulushas no colorfulness, while a vivid redwould be high in colorfulness (and ofred hue). Similarly, chroma is to colorful-ness as lightness is to brightness.

Figure 6. Munsell's first color models used globes (ii) as a base for rotary color mixtures, such that the five hues of equal lightness (called"value" by Munsell) and chroma appeared as a neutral gray when the globe was spun. Later versions used a "tree" in which value increasedfrom bottom to top, and chroma on every hue "branch" increased from center to edge (b). A Nickerson color fan, behind the Munsell tree,shows the maximum chroma for 40 hues at different values. Munsell published his original volumes in 1905 and 1915, for which he createdcolor samples from his own visual observations. Later editions (c, facing page) involved more detailed experimentation and contained a more

440 American Scientist, Volume 93

As with lightness, the human visualsystem generally behaves as a chromadetector. When we take a colored ob-ject out into bright sunlight from a dimroom, it appears more colorful But itschroma remains approximately con-stant, since the brightness of a whitestimulus under the same conditions in-creases along with the colorfulness ofthe stimulus.

Making Color a SystemMunsell chose to create his color sys-tem around the perceptual attributesof lightness (called Munsell value), hueand chroma because they are the threedimensions that relate most closely toour everyday experience of the colorsof objects. Our visual system adaptsto incredibly large changes in the col-or and level of illumination to helpus perceive object colors as relativelystable. Thus a perceptual color systemsuch as the Munsell system is mostuseful if it is enumerated in terms ofthe perceptual color dimensions mostclosely related to these relatively stableobject perceptions.

One representation of the Munsellsystem is called a Munsell tree. The"trunk" of the tree is the value scalewith dark at the bottom and light at thetop. Each "branch" represents a differenthue, and the "leaves" on the branchesincrease in chroma with distance fromthe center of the tree. In the Munsell sys-tem, each color is designated by its hue(with a set of letters and numbers), itsvalue (with a number ranging from 0 to10) and its chroma {with a number rang-ing from 0 for neutral grays to numberson the order of 15 or so for the highest

chroma samples; the actual maximumchroma available depends on the par-ticular hue and value chosen).

For example, a deep red sports carmight have the Munsell designationof 4R 3/12, where 4R hue designatesa hue just slightly more bluish than aunique red, value 3 indicates a slight-ly dark color (value 5 is perceptuallymidway between white and hiack inlightness), and chroma 12 indicatesthat it is a vivid red.

An important part of Munsell's ap-proach to building a color system wasthat he created each scale of a perceptualcolor dimension to have equal numeri-cal increments perceptually. He dividedthe hue circle into 1(K) equal perceivedhue steps and anchored it with fiveprincipal hues (red, yellow, green, blueand purple) and five intermediate hues.For each of those 10 hues, he created 10additional smaller increments. In addi-tion, Munsell chroma steps are abouthalf the size of the value steps, such thata change of one step in value appearsabout the same magnitude in color dif-ference as two steps in chroma.

Although the colors of the Munsellsystem are specified by their appear-ance in terms of value, chroma and hue,once samples for each designation arecreated, the system can be recorded andreproduced using physical metrics ofcolor. Specifically, the spectral reflec-tance of the samples and the spectralpower distribution of the illuminationare used together with standard humanresponse functions to designate physicalcolor coordinates known as tristimidusvalues, which are directly related to thestimulus wavelength and energy. These

coordinates ultimately define the sys-tem and allow reproduction of nomi-nal-color samples. Such numerical colorspecifications allow the system to con-tinue to be recreated even if the samplesof a current embodiment should fade orbe otherwise damaged.

The visual uniformity of the Munsellsystem has been one of the most sig-nificant attributes in making it one ofthe most important and influential colorspecifications of the last century. Thefirst color samples in his 1915 atlas werecreated by Munsell himself thr^^ugh hisown visual obser\^ations. The 1929 Mun-sell Book of Color represents a more sys-tematic definition of the color samplesthan the 1905 atlas through more de-tailed experimentation. It can really bethought of as the classic edition of theMunsell book that brought the systeminto the forefront of color science.

Although the 1929 Munseli Book ofColor was certainly a success, it was rec-ognized that the color chips embodyingthe system did not accurately match thetheory for all of the samples in the book(generally about 1,500 color samples).The Optical Society of America's Color-imetry Committee—with critical con-tributions from Dorothy Nickerson—performed visual experiments, usingmore than 300,000 visual obser\'ations,to create a more accurate specificationof the system in terms of instrumentalcolor measurement. Their results werepublished in the 1940s and referred to asthe Munsell renotation, which is also thecurrent definition of the samples in theMunsell Book of Color.

The Munsell system evolved simul-taneously with instrumental color-

systematic definition of color samples. In the 1940s the Optical Society of America undertook another revision of Ihe Munsell color samplescreating the current standard. Modem representations of Munseli samples (d) include a computer rendering of the Munsell tree shown herewith a more recent color system, the OSA Uniform Color Scales, which describes the same colors in a different mathematical way. Rather thantreating each of the three dimensions of color separately, as Munsell did, this system is designed to provide equal perceptual spacing in allthree dimensions simultaneously. (Images courtesy of the authors.)

www.americanscientist.org 21X)5 September-October 441

MO. 1 BOX. i-^_L/<U-^ 6 -1

I N S E L L f^ CRAYONSFIVE MIDDLE COLORS

WITM QRAY AND BLAOK

OP RED, VELLOW AND BLUE.

WADSWORTH, HOWLAND « CO., l o c

I1

'

ii

1ii

i

i1

f 1

?" • .

\ ; 1

Figure 7. One of the first color products that Munsell had manufactured was a set of crayons.This set showed colore with a mid-level value and chroma. Munsel! sold his crayons to thecompany that now manufactures Crayola crayons, whose color names do not reflect MunseU'ssystematic approach to color. (Photograph of Munsell crayons reproduced by the permissionof The Huntington Lihrary, San Marino, California; photo-montage courtesy of the authors.)

measurement techniques (the CIEsystem of colorimetry). The Munsellsystem was improved through the useof instrumental techniques to specifyand produce modem Munsell samples,and instrumental colorimetry benefitedfrom the visual data on color-attributescaling (measurements of quantitativemagnitudes of our perceptions) provid-ed by the Munsell system.

Today, the most widely used com-putational color system for color mate-rial specification and tolerances is theCIELAB color space, originally pub-lished by fhe CIE in 1976. This colorspace has three dimensions fo describeohjecf color appearance. In rectangu-lar coordinates, the dimensions are L*(lightness), a* (redness-greenness) andh* (yellowness-blueness). An alternative

Figure 8. Munsell Soil Color Charts let soil scientists visually analyze samples in the field.In this view of a site in Maryland, the matrix of the water-saturated lower layer is black(designated 2.5Y 2.5/1 for hue value/chroma) due to the presence of the mineral pyrite, andthe areated zone above has pale yellow streaks (5Y 7/4) of jarosite, a pyrite-oxidation product.(Photographs courtesy of the authors, Martin C. Rabenhorst of the University of Maryland,College Park and Ted Callender, U.S. Geological Survey-retired.)

representation of the same color spaceis in fhe cylindrical cotirdinafe sysfemof L* (lightness), C%t> (chroma) and h ,(hue angle) that corresponds with thedimensions of the Munsell sysfem.

CIELAB has been successful in its 30years of existence, but color science hasnot stood still. Current research aimsto extend the CIELAB mathematicalmode! to improve our ability to pre-dict fhe appearance of colored stimuliand images under a wide variety ofviewing conditions. Mathematicalmodels attempting this complex taskare known as color appearance mod-els. The most recent, internationallyaccepted, mode! of color appearance isknown as CIECAM02, for the CIE 2002Color Appearance Model. Modem com-puter-graphic renderings of both theMunsell system and other systems ofcolor specification in proposed math-ematical color spaces are a great aid inthe formulation of these new modelsof color measurement.

Munsell's Extensive ReachAlbert Munsell originally set out tobuild a color sysfem as a way to im-prove his teaching about color to hisown art students, but if is clear that hissysfem has had a significant impact onthe art and science of color. Some areaswhere Munsell products are currentlyused include imaging-system calibra-tion and characterization, evaluationof color vision for color deficiencies,botany for the classification of certainplant diseases and nutrient deficien-cies, color coding of electrical wire andcable, food sfandard colors to assess thequality of French fries, olives, tomatoes,cherries and pumpkins, American Na-tional Standards Institute (ANSI) safetycolors and the description of skin, hairand eye color in forensic pathology.

Even crayons have an interesting re-lation to the work of Munsell. The colorsof the crayons themselves are standard-ized and controlled using colorimetrictechniques similar to those describedabove for school buses. In addition, oneof the first color products that Munsellhad manufactured was a set of crayonscarrying the systematic color designa-tions of his sysfem. Early on, Munsellsold his crayon business to Binney &Smith, the company that now manufac-tures Crayola crayons.

Had fhe sale gone anofher way,American children might now leam avery systematic and technically accurateway to describe color from their cray-

442 American Scientist, Volume 93

ons. In some countries, such as Sweden,where the Swedish Natural Color Sys-tem (NCS) is a national standard, chil-dren are faught a systematic method ofcolor specification at an early age, andcommerce (such as buying paint in ahardware store) is carried out using pre-cise numerical color specifications thatare also intuitive to those so trained. Inthe United States, instead of asking theirplaymates to share the 4PB 3/10 crayon,American children are left to ask for the"cerulean" crayon and then to ask theirparents "what's a cerulean?"

Despite this one near-miss for popu-larizing the the Munsell system, thereare examples everywhere of productsthat have benefited from some form ofcolor-measurement system. The brew-ing of beer has provided an applicationof instrumental color measurement fornearly a century (and probably visualcolor measurement for many centuriesbefore that). The brewer Joseph Lovi-bond is credited with creating one of thefirst color-measurement instruments,the Lovibond Tintometer (a version ofwhich is still available). The tintometeris an instrument in which a coloredsample, such as some beer, is matchedwith a mixture of cyan, magenta andyellow filters. The amounts of thesestandard tintometer filters required tomake a match served as a color mea-surement. Nowadays, tintometer val-ues and CIE color specifications can becomputationally interconverted with-out the need for a visual assessment.

A recent example tif use of the Mun-sell system for visual coiorimetry can befound in the work on the spectropho-tometric and colorimetric analysis andreproduction of rare works of art andother cultural heritage by Roy S. Bemsof the Rochester Institute of Technol-ogy. Bems performed visual matchingof painted areas in Ceorges Seurat's ASunday on La Grande latte~1884 at theArt Institute of Chicago, and the Star-Spangled Banner in the Smithsonian'sNational Museum of American Histo-ry—the tlag that inspired Francis ScottKey to write the poem that eventuallybecame the American national anthem.Given the value of these two pieces, itis not feasible to make repeated instru-mental measurements in contact vviththeir surface or to illuminate themwith the light intensities required fornon-contact measurements. Samplesfrom a modem Munseli Book of Colorwere used to quantify visually the cur-rent colors in both pieces and' to check

the few instrumental measurementsthat were allowed. The resulting colorspecifications were used to analyze thepigments in the Seurat painting andcreate a digital image representing theappearance of those colors when thepainting was originally made. For theStar-Spangled Banner, which is nowbeing stabilized for exhibition with apolyester backing, the measurementswere used to help the Smithsoniandesign a new exhibit for the flag bysimulating what it would look like af-ter cleaning and with illumination byvarious types of light sources.

In the eariy part of the 21st century,the film industry is undergoing a transi-tion to digital cinema. The century-oldMunsell system Ls playing a role in thistechnological transition as well. Digitaldjiema projectors are capable of display-ing an enonnoLis range of colors, higherin chroma tlian those traditionally avail-able in film and television systems. Inutilizing these displays with existing con-tent, it becomes necessary to map the col-ors available from historical displays tothe extensive gamut of newer displays.One of the objecti\ es in performing thesemappings is to preserve the apparent hueof the colors in the images. The Munsellhue scales, as well as newer data on hueperception, have been used to derivemathematical color spaces that are nowimplemented in digital cinema systems.These color spaces have applicability togamut mapping, the prcKess of changingdesired aim colors to colors witliin thecapabilities of a display device while pre-serving the appearance of the inuges.

Color description using Munsellterminology is an important part ofsoil-profile characterization and soilmapping, providing field-discemableevidence of soil-forming processes. Forexample, color changes can indicateoxidation and reduction of iron com-pounds. When a high water table lim-its oxygen penetration info a soil thathas developed on parent material thatcontains the ferrous iron mineral pyrite(FeS2; typically black), that mineral isstable and tends to color the soil in thelower, water-saturated horizons. Whenpyrite reacts with atmospheric oxygen,sulfuric acid is produced, along with thepale yellow, ferric iron mineral jarositelKFe3(SO4).(OH)J. As a result of this"sulfuricization" process, one sees bothlow soil H and streaks of jarosite in theoxidizing zone above the water table.The presence of the distinctly coloredjarosite is readily described using the

Munsell Soil Color Charts. Other formsof oxidized iron such as the red, fer-ric oxide mineral hematite (FejO-,) andthe yellow-brown, ferric oxyhydroxidemineral goethite (FeOOH) can contrib-ute to color features. Jarosite, hematiteand goethite, either synthesized ormined from natural deposits, are usedas paint pigments,

Color is fascinating to almost every-one, and the scientific study of colormeasurement and perception has a richand interesting history. As with all fields,many have contributed, and modemscientists can see further by standingon the shoulders of the giants that pre-ceded them. Color science is perhapsunique in the degree to which expertsin a wide range of disciplines can makeimportant contributions. Albert Munsellis one such example in which an artistand teacher satisfied his curiosity andneed to teach by developing a systemthat had a profound impact on color sci-ence and commerce and will continue todo so long into the future.

BibliographyBems, R. S. 2000. Billmexier and Saltznum's Prin-

ciples of Color Tcchnolo^i/, 3rd Ed. New York:John Wiley & Sims.

Derefeldt, G. 1991. Colour appearance sys-tems, In The Pcnre}7tioii of Colour, ed. P. GOLI-ras. Boai Raton, Florida: CRC Press.

Fairchild, M. D. 2003. Color Appearance Models,2nd Ed. Chichcster, U.K.: Wiley-IS&T.

Kuehni, R. G. 2002. The early development ofthe Munseil system. Color Research mid Ap-pliaition 27:21)^27.

Kuehni, R. G. 2003. Coior Space and Ih Divi-sion:^: Color Order from Antiqiiiti/ to the Pres-ent. Hoboken, New Jersey: Wiley.

Long, J., and J. T Luke. 2001. Vie Neit' Mmist'llStudent Color Set- New York: Fairchild Books.

Munsell, A. H. 1946. A Color Notation: An iihis-trcited system defining all colors and tlieir relationstn/measured scales of hue. niluenndchroimii^thed.). Baltimore, Maryland: Munsell Color Co-

Newhall, S. M., D. Nickerson and D. B. Judd.1943. Final report of the O.S.A. subcommitteeon the spacing of tbe Munsell colors, journalof tlie Opticiil Societ}/ of America 33:?iS5-A\8.

Nickerson, D. 1976. History of the Munsellcolor system, company, and foundation, 1.Color Research ami Application 1:7-10.

Nickerson, D. 1976. History of the Munsellcolor system. Color Research and Application1:121-130.

For relevant Web links, consult thisissue oi American Scientist Online:

http://www.americanscientist.org/lssueTOC/issue/761

www.americanscientist.org2005 September-October 443