Woodblock prints were first produced in Japan during the sixth to eighth century but it was not until the Edo period (1603–1868) that the full potential of woodblock printing as a means to create popular imagery for mass consumption developed. Known broadly as ukiyo-e, meaning “pictures of the floating world,” these prints depicted Kabuki actors, beautiful women, scenes from history or legend, views of Edo, landscapes, and erotica. Prints and printed books, with or without illustrations, became an integral part of daily life during this time of peace and stability. Prints produced from about the 1650s through the 1740s were printed in black line, sometimes with hand-applied color (see figure 1). These colors were predominantly mineral (inorganic) pigments supplemented by plant-based (organic) colorants. Since adding colors to a print by hand was costly and slowed production, the block carvers eventually hit upon a means to create a multicolor print using blocks that contained an “L” shaped groove carved into the corner and a straight groove carved further up its side in order to align the paper to be printed (see figure 2). These guides, called kento, are located in the same location on each block. They ensure consistent alignment as each color is printed onto a single sheet of paper.
From the 1740s to about 1765, the first block printed colors appeared on simple two- or three-color images (see figure 3). These benizuri-e (“red pictures”) utilized red, blue, or yellow; sometimes these colors were over-printed to create the secondary colors purple, orange, and green. From 1765 on, the skills required to use the kento registration system reached a level where several color blocks could be expertly printed and full-color nishiki-e or “brocade prints” such as those designed by Suzuki Harunobu (1725–1770) became the standard (see figure 4). Vibrant full-color prints designed by well-known artists such as Torii Kiyonaga (1752–1815) and Kitagawa Utamaro (1753–1806), produced and marketed by the great publishing houses of Tsutaya Ju‑zaburo‑ and Nishimura Yohachi, defined the period from 1781–1801, which is often referred to as the Golden Age of the Japanese woodblock print (see figures 5 and 6).
Japanese Woodblock Prints at the MFA
The Museum of Fine Arts (MFA), Boston, holds a collection of over 50,000 Japanese woodblock prints and illustrated books. This represents the largest number of such art works in a single location outside of Japan.
In 2010, a major five-year project to accession, image, and re-house this vast collection of Japanese woodblock prints was completed with cataloguing ongoing. This project placed information about each print along with its image and the translation of any Japanese text, signatures, and seals on the print into the Museum System collections database (TMS), enabling our current research as well as numerous exhibitions and publications. Of additional significance, the collection holds a number of multiple impressions of a single image, thus providing an ideal setting to identify, survey, and understand the organic and inorganic colorants used in traditional Japanese color woodblock printing.
In 2013, the Asian Conservation and Scientific Research divisions began a large-scale survey of the colorants used in the MFA collection of Japanese prints. This effort uses only non-destructive techniques, which means no samples are required of the prints that are formed by minimal levels of colorants absorbed into their paper fibers. The first two techniques used for this study are standard methods used in museum labs: X-ray Fluorescence (XRF) and Fiber Optic Reflectance Spectroscopy (FORS). A new and previously little-used technique of Excitation-Emission Matrix (EEM), or 3D, fluorescence spectroscopy was also used to successfully characterize several additional dyes.
The plant-based red, yellow, and blue dyes long considered to make up the palette of Japanese woodblock prints are summarized in Table 1. The list is based on various published sources, including early Japanese literature and analytical studies, and may not be comprehensive. The inorganic pigments used in the prints, such as red lead, hematite, and orpiment, can be easily estimated by XRF (see Table 2). The traditional organic blues, dayflower and indigo, can be confidently identified by FORS in the visible and near-infrared ranges. Both XRF and FORS were used in the examination of the prints discussed in this article. The unique component of this study was the use of EEM fluorescence spectroscopy to identify the yellow and red natural organic colorants on the prints. Examples of the color and line contour maps obtained as results for the fluorescence analysis are shown in Tables 3 and 4.
For this project, it was important to obtain reference materials from documented sources for each of the materials to determine the best method for its identification. Once the reference samples were obtained, they were authenticated using Liquid Chromatography with a Mass Spectrometer detector (LC/MS). Afterward, each material was prepared by historical methods, then printed onto Japanese paper. For the materials listed in Table 1, all were available from documented sources as raw materials, except for Toringo crabapple (Malus sieboldii) and Coptis japonica, Japanese goldthread, a member of the buttercup family. Fortunately, the Arnold Arboretum generously supplied samples of branches from five Malus sieboldii specimens in their collection. In lieu of C. japonica, a sample of threeleaf goldthread (Coptis trifolia), a related species native to North America, was obtained from a biologist in Vermont.
An initial set of 213 Japanese woodblock prints were examined at the MFA by the combined techniques of XRF, FORS, and EEM fluorescence. These prints covered the time period 1700 to 1800 (see Table 5). The goal for the analysis of the prints was to obtain an overview of the colorants used by artists active in each time period; this goal was later expanded to include information on publishers, since they were probably responsible for the final colorant decisions. Our research is ongoing and is conducted as time allows. It is hoped that the analysis of a more extensive set of prints will provide definitive information on the relationships between colorants, publisher, artist, and period. (See Table 6 for the results from the example prints mentioned in this article.)
Even with this limited data set, several patterns of colorant use were consistently found for the eighteenth century time period. From the beginning, it was clear that the prints often contained more than one yellow, red, or blue colorant. Though it seems logical, since each colorant has unique tonal properties, this finding was significant in terms of analysis, indicating it was imperative to analyze multiple colored regions for each print. Because of the expansion to the use of mixtures and colorant variations from 1781 to 1801, it was common to find three types of yellow, two reds, and two blues in a single print (example figure 5). Thus, while 213 prints were studied, there were over 1,500 individual analysis points.
|Plant Source||Japanese Name||Common English Names||Part of Plant Used||Color|
|Caesalpinia sappan||suo, suwo||sappanwood||heartwood||red|
|Carthamus tinctorius||benibana||safflower||florets separated from capitulum||red|
|Rubia akane||akane||Japanese madder||roots, stems||red|
|Rubia tinctorum||seiyo-akane||European madder||roots, stems||red|
|Berberis thunbergia||megi||Japanese barberry||roots, stems||yellow|
|Coptis japonica; C. trifolia||woren||goldthread||roots, stems||yellow|
|Curcuma longa (syn. C. domestica); C. aromatica||ukon||turmeric||rhizomes||yellow|
|Gardenia jasminoides (syn. G. augusta)||kuchinashi||gardenia||juice or extract from fruit||yellow|
|Garcinia hanburyi, G morella||shio, te-o; kusa shio||gamboge||resin/powder||yellow|
|Malus sieboldii (syn.Pyrus toringo)||zumi||Toringo crabapple||bark||yellow|
|Miscanthus tinctorius;M. sinensis||kariyasu||silver grass||grass cut when flowering spikes form, then dried over the winter||yellow|
|Myrica rubra||yama-momo||mountain peach, red bayberry||bark||yellow|
|Nandina domestica||nanten||nandina, heavenly bamboo||branch||yellow|
|Phellodendron amurense||kihada||Amur corktree||inner bark of trunk||yellow|
|Styphnolobium japonicum (syn.Sophora japonica)||enju||Japanese pagoda tree||unopened flower buds||yellow|
As expected from the literature, safflower (Carthamus tinctorius) was the primary red and pink colorant used consistently for all of the time periods and methods of application. Surprisingly, however, the second most prolifically used red was madder. While the analytical methods used in these tests, could not distinguish between Japanese madder (Rubia akane) and European madder (Rubia tinctorum), one or both of these colorants were consistently found on prints in all four of the described periods with their use increasing from 20% up to 50% over the hundred-year period.
The yellow colorants changed significantly over the hundred-year period from the sole use of flavonoids and gamboge during the beni-e hand-applied color period (1710–1740s) to the predominant use of turmeric and orpiment (an arsenic sulfide mineral) for the elaborate designs and techniques used for full color printing from 1781–1801.
Inorganic pigments were found on most prints examined for each time period. While the use of orpiment (As2S3) increased significantly, the use of hematite (an iron oxide, Fe2O3) and lead (Pb) were constant. Other inorganic pigments were occasionally found, such as added decorations using ground metallic brass (figure 1). Vermilion, a mercury-containing pigment commonly used for paintings, has been mentioned as a definitive colorant in Japanese printing. However, this study found only two prints containing vermilion, indicating it may not have been commonly used.
Of interest to us were the compositions for purples and greens. Mixtures or overprinting transparent colors were noted in many prints from the 1740s on. In all analyzed purple regions, our results showed mixtures of safflower and dayflower (see figures 4 and 5). The presence of this mixture throughout the history of color printing seems to indicate that the tone obtained by mixing dayflower blue and safflower was preferred over other possible mixtures of reds and blues (for example, indigo and madder) to yield purple. The green regions varied more often, with earlier prints showing overprinting of turmeric with dayflower (figure 3) while later prints showed a more vibrant green made by mixing orpiment and indigo (figures 4 and 5).
|Mineral Source||Japanese Name||Common English Names||Chemical Formula||Color|
|red lead||tan, entan||red lead||Pb3O4||red|
|orpiment||kio, sekio, shio||orpiment||As2S3||yellow|
Two aspects of the results in this study seemed unusual. First, no examples of either gardenia (Gardenia jasminoides) or berberine colorants (e.g., Berberis thunbergii, Coptis japonica, Phellodendron amurense) were found in the analysis of 557 yellow spots in 213 prints. These plants, which generally grow in the highlands, have been described as common Chinese colorants that were used in the Japanese islands. The colorants were mentioned in the literature as being used for eighteenth century Japanese woodblock prints, but were not found on any prints analyzed in this preliminary study.
Second, madder was found on 142 red analysis locations in 90 out of the 213 prints (42%). While madder was available in Japan and was used prolifically as a textile colorant, it has not been previously mentioned as a possible colorant for printing. Madder may have been used as a substitution for more costly reds such as safflower (benibana) in order to keep the market price of an individual print affordable. With most of the print collection in the MFA, it is not always known whether the prints are first editions or later runs. Thus, there is always the possibility that madder was used for later editions, even though the date of the print is listed based on its initial production. Further work will be done to compare impressions and examine the paper fibers and formation methods to clarify the time periods of the madder use.
|Date Range||Woodblock Print Style||Number of Prints||Percentage of Prints Containing that Color Within that Style|
|Turmeric||Flavonoid||Gamboge||Safflower||Madder||Sappanwood||Red lead (Pb)||Red ocher (Fe)||Orpiment (As)|
|1650s–1740s||Hand-colored prints, Beni-e and Urushi-e||16||0||75||75||63||19||6||12||19||0|
|1740s–1765||Limited-color prints, Benizuri-e||21||62||24||19||38||33||14||14||10||14|
|1766–1780||Full-color prints, Nishiki-e: First Period||80||36||28||23||78||39||26||25||20||50|
|1781–1801||Full-color prints, Nishiki-e: Golden Age||96||72||7||1||91||51||11||20||14||71|
Additional Information on Specific Colorants
The florets of safflower (Carthamus tinctorius) produce a wide range of colors from cherry red to pink (figure 7). Native to northern India and the Near East, this popular dye plant was widely cultivated throughout Asia and Europe by the end of the thirteenth century. It is a tender annual with spiny leaves and composite flower heads containing many yellow to orange disk florets. The florets are picked, then dried and crushed into a paste. The paste is washed with water to remove the non-lightfast yellow chromophores including several quinochalcones. The red colorant, primarily carthamin, is then extracted in an alkaline bath. The deepest reds are obtained through several initial washings to remove all of the water-soluble yellows.
Red regions containing safflower were usually seen as bright fluorescence during the preliminary examination of the prints with a hand-held ultraviolet (UV) light. Thus, it was no surprise that the EEM fluorescence technique provided a unique and definitive pattern for safflower, even when it was visually observed as a faded brown tone. In addition to the fluorescence for the red chromophore, the pattern often contained an additional peak for the yellow chromophore that was supposedly removed in the preparation of the red colorant but often needed several washings for complete elimination.
Printed examples of the safflower colorant can be seen in figures 1, 3, 4, and 5. In figure 1, the Kiyotada I hand-colored print from the beni-e period, safflower was found on the base of the umbrella and the flowers on the woman’s kimono. Looking at the later Harunobu print from 1767–68 (figure 3), safflower was found on the purple robe of the kneeling child. In the 1788 print by Kiyonaga (figure 4), safflower was used for a pink collar, purple sleeve, and orange frame.
Madder: akane and sieyo-akane
The roots of madder plants (from Rubia tinctorum, Rubia akane, and many others) produce a deep true red color that was widely prized throughout the world (see figure 8). All madder family (Rubiaceae) plants give strong red dyes with the colorants concentrated in the parenchyma of the roots and stems, even though the plant and flowers do not exhibit any red colors. For processing, the roots were typically harvested in the autumn after a minimum of two years growth. The plants were marketed as whole dried roots rather than as a powder.
Madders, a general name applied to anthraquinone-containing dyes extracted from plants from various genera and species, often exhibit strong (orange) fluorescence in a work of art when examined under ultraviolet radiation. The strong fluorescence is usually stated to be associated with purpurin, a common anthraquinone found in many types of madder, including both Rubia tinctorum and R. akane, a plant native to Japan. It is not certain when R. tinctorum was first utilized in Japan, or from where it would have originated, but it does not appear to be possible to distinguish it from Rubia akane based on its EEM fluorescence pattern.
Printed examples of the intense red madder colorant can be seen in figures 3, 5, and 6. Figure 3, the Kiyomitsu I limited color print from 1762, has madder as the sole red colorant and it was used for the man’s robe, face, and feet. Looking at the later print by Kiyonaga (figure 5), madder was used for the red cloth under the musicians. When it was found, the madder EEM pattern was very distinct and its color was a bright deep red. Madder was not found in secondary colors such as purple or orange in the prints examined for this study.
|Torii Kiyotada I (Figure 1, MFA # 11.13273)|
|(Analysis Points)||(Analysis Interpretation)|
|Blank – womans face||__|
|red on umbrella||Safflower|
|pale robe man’s chest||Flavonoid|
|woman’s robe near collar||Flavonoid|
|orange corner man’s robe||Gamboge|
|skirt yellow and black||Brass flakes*|
|Torii Kiyomitsu I (Figure 3, MFA # 11.1903|
|Blank – background||__|
|green bell||Turmeric overprinted with dayflower**|
|Suzuki Harunobu (Figure 4, MFA # 21.4463)|
|Blank – white snow||__|
|green brush||Orpiment* mixed with indigo**|
|brown robe||Safflower mixed with dayflower**|
|orange leaf on robe||Flavonoid mixed with sappanwood and safflower|
|Torii Kiyonaga (Figure 5, MFA # 11.13921)|
|Blank – face||__|
|red ground cloth||Madder|
|yellow pants on musician||Flavonoid|
|green grass||Orpiment* mixed with indigo**|
|green knee||Orpiment* mixed with indigo**|
|pink collar on woman||Safflower|
|pink shoulder on man||Safflower|
|orange scroll border||Safflower mixed with turmeric|
|purple cuff on man||Safflower mixed with dayflower**|
|yellow at bottom||Turmeric|
|Kitagawa Utamaro I (Figure 6, MFA # 11.14364)|
|Blank – face||––|
|center yellow flower||Flavonoid|
|green in hat||Flavonoid with dayflower**|
|brown in hat||Safflower|
|blue on kimono||Dayflower**|
Red dyewood: suo
The insoluble red dye from sappanwood (Caesalpinia sappan) and other types of red dyewoods (sandalwood, barwood, narrawood, padauk, camwood, Brazilwood, etc.) were prepared as colorants by pounding chips of the heartwood into a paste mixed with a little oil (see figure 9). These were formed into cakes or bars for storage and sale. The red colorant was so popular in the seventeenth and eighteenth centuries that many of these species are now extinct or endangered. Its color was said to be orange-red, brownish-red, or cinnamon-like.
Using our references, the EEM spectra could easily distinguish the sappanwood fluorescence pattern from safflower and madder. However, the reference spectra for a few other types of red wood dyes, such as sandalwood and Brazilwood, produced similar but not identical fluorescence spectra. Thus, it was difficult, if not impossible, to differentiate between the various red dyewood sources.
In this study, the red dyewood fluorescence pattern was not often found, its use being limited to just a few artists and publishers. The print by Harunobu (figure 4) shows an example of the red dyewood. It was used for both the red and the orange regions, while the purple colorant was found to contain safflower and dayflower.
Described as the most popular yellow colorant in the world, the rhizomes of turmeric (Curcuma longa) produce a bright yellow orange dye that is commonly used for food and textiles (see figure 10). Native to India, turmeric is now cultivated worldwide. Though a perennial herb, the plant is often completely harvested, then the roots are cooked, dried, and ground into a powder. Turmeric is a direct dye with high tinctorial strength that began its use as a fabric dye prior to the tenth century and is still used today as a curry seasoning.
Yellow regions containing turmeric usually were brightly fluorescent during the preliminary examination of the prints with a hand-held UV light. The dye produced a very clear, consistent fluorescence pattern, likely because of its single primary chromophore.
In this study, the printed examples that contain turmeric are figures 3 and 5. In the print by Kiyomitsu I, turmeric is used as a clear, intense yellow for the trim as a contrast to the bright red. In the print by Kiyonaga, turmeric is used as a strong yellow background color. This colorant is fairly lightfast and retains its color better than the flavonoids.
Flavonoids—silver grass: kariyasu; Japanese pagoda tree: enju; Toringo crabapple: zumi
Flavonoids occur in most dye plants and their yellow colorants were discovered from the earliest times. While many colorants in this group are not lightfast, their abundance has resulted in their wide use. In Asia, the primary flavonoid-containing plants were the luteolin containing grasses, such as Miscanthus tinctorius (silver grass: kariyasu), that were cut each fall, then dried, and kept until the next spring for extraction. Other common Japanese dyeing plants include Styphnolobium japonicum (syn. Sophora japonica; Japanese pagoda tree: enju, see figure 11) and Malus sieboldii (Toringo crabapple: zumi).
Flavonoid-containing dyes tend to have numerous compounds. For fluorescence measurements, the emission positions were similar and tended to blend into a single elongated peak. This pattern tends to be weaker than the turmeric pattern and was often noted mixed in with the absorption pattern for the paper, thus making positive identification difficult. Additionally, it was difficult to make any consistent determination for the various types of flavonoid yellows. Since each contains similar compounds, but in different proportions, the excitation and emission maxima are similar, blending into an oblong mesa-type absorption area rather than a single peak.
In this set of analyzed woodblock prints, flavonoids tended to be a popular early colorant that later gave way to the use of turmeric and orpiment. One possible reason for the shift in yellow colorants is the poor light stability of most flavonoid yellows. The prints shown in figures 1, 4, 5 and 6 show examples of yellow flavonoid printed colors. Figure 6 was included as an example to show that even though it is difficult to distinguish between various flavonoid colors, this print does show two visually different yellow-colored regions that both produced slightly different fluorescence spectra even though both corresponded to flavonoids. It is possible that this print contains two types of flavonoid yellows, such as Japanese pagoda tree and Toringo crabapple. It is also possible that the print was exposed to uneven levels of light and that the yellows at the top of the print have deteriorated differently than those at the bottom.
Gamboge is a golden yellow colorant that is extracted by tapping resin from various species of the evergreen Garcinia trees, most commonly G. hanburyi and G. morella native to southeast Asia and India (see figure 12). The trees must be at least ten years old before they are tapped. The resin is extracted by making spiral incisions in the bark, and by breaking off leaves and shoots and letting the milky yellow resinous gum drip out. The resulting latex is collected in hollow bamboo canes. After the resin is congealed, the bamboo is broken away and large rods of solidified resin remain. Gamboge is marketed as solid pieces or as a powder.
Visually, under a hand-held UV light, the yellow regions containing this colorant appeared dark, as if absorbing the fluorescence. The lack of fluorescence was confirmed by the EEM fluorescence pattern produced for reference samples of gamboge. The pattern exhibited a complete absence of emission peaks; this also included an absence of the paper peaks indicating that the paper was covered with a blocking agent. Gamboge is the only known organic reference in the potential set of Japanese colorants that corresponds to this negative pattern. Thus, this is a unique material that visually appeared yellow, but without sampling, it could only be characterized by the absence of any measurable inorganic elements (e.g., Fe, As) by XRF along with the absence of any unique fluorescence pattern by EEM fluorescence.
Gamboge was most often found in the hand-colored, beni-e, prints and is illustrated in the corner of the warrior’s robe (figure1), along with other non-analyzed points such as the kimono collars and the center frame of the umbrella.
Dayflower: aigama; awobama
Though rarely used elsewhere, dayflower (Commelina communis) blue was commonly used in Japan. Dayflower is an annual plant native throughout much of eastern Asia that bears one-day-blooming flowers featuring two large blue upper petals. The anthocyanin-containing juice extracted from the flowers was used by illustrators and printers for blue and green colors. Cloth or paper was dipped into the juice and dried; once needed the cloth or paper was dipped into water to extract the blue colorant (see figure 13).
The best analytical method for the identification of dayflower is fiber optic reflectance. An example of the difference in the FORS spectra for dayflower and indigo is shown in figure 14. As dayflower and indigo were the only two plant-based blues dyes used for woodblock prints, the FORS method could quickly and simply distinguish between the two materials.
Though dayflower was sometimes used by itself for blue areas (see figures 3 and 6), its poor lightfastness and its sensitivity to water were possible reasons that it was most often found used for greens and purples. Figures 3, 4, 5, and 6 show examples of the green and purple tones.
Indigoid dyes were used in Neolithic Europe, Pharaonic Egypt, and now in twenty-first century jeans (see figure 15). While the source plants provide slightly different hues, indigo has been regarded as the color of kings. Indigo producing plants, such as Indigofera tinctoria (a tropical shrub or subshrub), contain colorless glycocides that can be converted to the blue colored indigo on exposure to oxygen. To produce the dye, the fresh leaves are macerated in hot water, after which an alkali is added (such as lime) to ensure the colorant remains in a colorless soluble form. Once the colorant is extracted, it is either printed out or cast into cakes where the insoluble blue indigo precipitates as it reacts with oxygen from the air.
Of the prints selected for this article, indigo was found in figures 4 and 5 in the bright grass green colors. For both prints, orpiment, an inorganic yellow, was mixed with indigo to obtain the vivid, somewhat lightfast color.
The purpose of this paper is to provide specific information on the analysis and identification of natural colorants used in the production of Japanese woodblock prints. Three non-destructive analysis techniques were used so that no samples were removed from the prints. X-ray fluorescence (XRF) was used to determine the presence of any inorganic compounds, and fiber optic reflectance spectroscopy (FORS) was used to distinguish between indigo and dayflower in the blue, green, and purple regions. Additionally, methods were developed to successfully use a third technique, excitation-emission matrix (EEM), or 3-D, fluorescence spectroscopy, for the characterization of the red and yellow plant-based colorants.
The MFA collection of Japanese woodblock prints is an ideal venue for the use of three combined techniques for the identification of the colorants because:
- The palette used for woodblock prints is limited.
- The colorants and substrates for the print were prepared with consistent, often documented, methods that had minimal variation.
- The prints are flat and the size of the prints, even within their mats, is less than 1 square meter.
- The speed for all three types of analysis is fast and allows for easy analysis of multiple locations.
- The MFA has an extensive set of over 50,000 Japanese woodblock prints and illustrated books that allows for extensive surveys of the materials within each time period, style, publisher, and artist.
- The knowledge obtained from the colorant identification will promote the understanding of the light stability for each print, and thus help preserve its vibrancy.
The authors are grateful to Kathryn Richardson (The Arnold Arboretum) and Allaire Diamond (Williston, Vermont) for supplying known reference samples for analysis. Additionally, we commend Arianna McQuillen and Kaeley Ferguson for conscientiously analyzing the set of Japanese prints by EEM fluorescence and FORS.
Michele Derrick is the Schorr Family Associate Research Scientist, Joan Wright is the Bettina Burr Conservator in Asian Conservation, and Richard Newman is the Head of the Scientific Research Division, all at the Museum of Fine Arts, Boston.