No doubt this lack of awareness in the West reflects the scarcity of extant examples of this work and the unusual secrecy surrounding the practice. In 1887 an Imperial Ordinance banned the use of this type of watermarking by those outside the Oji mill. The technique, like analogous European techniques from which it may have derived, was developed principally as an anti-counterfeiting measure for paper currency. Its remarkable aesthetic qualities make it worthy of rediscovery for use by contemporary papermakers.    We first learned of this technique during graduate studies at the University of Iowa Center for the Book. Center Director Timothy Barrett, reviewing the exceptional detail in his collection of chiaroscuro watermarks from the Italian firm Fabriano, noted the even greater detail in his Japanese watermarking samples, some of which duplicated those he had seen in Thomas and Harriet Tindale's rare work The Handmade Papers of Japan.1 In that work, the Tindales discuss their exposure to tesuri and include samples of artistic applications of the technique.    We tried to achieve similar results using contemporary materials and tools, applying our knowledge of Japanese papermaking and letterpress typography. We started with the Tindales' description, the only information about the technique easily accessible to Westerners.    During an official visit to Japan at the time of the U.S. occupation after World War II (1948-1952), the Tindales visited the Oji Paper Mills and were shown examples of suki-ire-bijutsu-shi (artistic watermarked papers), which included both tesuri and the Western wire screen suki-keta-ho method. They learned that the former technique had been invented at Oji in 1889.    The Oji technique, as recorded by the Tindales, involved engraving an illustrative design on a plate made of wood, copper, bronze, antimony, or celluloid. Wet sheets of mitsumata paper, formed using either the tame-zuki or nagashi-zuki method (longer fibers than were used for the chiaroscuro papers of the West), were couched individually and left to weep for five to ten minutes. The wet paper was then placed directly on the face of the plate. Using a three-step burnishing process, the papermakers rubbed the damp sheets against the plate until the paper accepted the image. A dry cotton cloth was stretched on a wooden frame and placed over the wet sheet, then rubbed with a piece of corrugated celluloid. The cloth was removed, and the paper was rubbed again, this time directly, with more finely corrugated celluloid. The wet sheet on the plate was rubbed a third and final time, again directly, using an orchid leaf wrapped around a roll of dry cloth.    After burnishing, the wet sheet was placed between damp cloths. Following pressing, it was pasted on a wooden drying frame with nothing touching the design on the sheet. When dry, the sheet was cut out of the frame with a bamboo knife, the edges were trimmed, and the paper was rolled lightly through a calender press.    Varying shades in the paper reflected the varying depths of the engraved cuts, allowing a wide range of light and darkness to appear in the watermark. The engravings were so skillfully crafted that raised areas above the plate were administered to attain extremely light areas in the paper. This allowed superior detail in the tesuri watermarks compared to the best made using the Western chiaroscuro method. Moreover, the longer fibers used in tesuri resulted in much stronger sheets than Western watermarked sheets.    The genesis of tesuri occurred at the peak of the Meiji era, when modern Japan was being born. For three hundred years prior to Commodore Perry's arrival in 1853, Japan was an isolated island under the feudal Tokugawa Shogunate. Beginning in early 1868, the Meiji Restoration began the return of political power to the Emperor, marking Japan's transformation to a modern nation state. Despite political struggles and limited financial resources, the new government endeavored to build power through military and economic development on Western models. In many cases, this approach resulted in the integration of modern European processes with traditional Japanese culture and craft.2    This historical context is significant for understanding the development of tesuri. We know from numismatists that, beginning in the early 1870s, Japan contracted with German lithographers to help them develop new paper banknotes.3 We also know that chiaroscuro watermarking was a technique common among European currency makers by then. In the 1850s, William Henry Smith, of England, had developed the technique of engraving an image with varying gradations into a sheet of wax, from which a copper electrotype was made. A woven wire mesh screen was then embossed from the electrotype. This embossed screen was attached to the papermaking mould surface, causing fibers to accumulate in varying amounts during sheet forming. In 1874 another Englishman was contracted for several years by a leading Japanese paper mill to teach Western paper and watermarking methods.4   The Japanese also had indigenous techniques that may have influenced the development of tesuri. Wahon-hyoshi and kabe-gami both involve a type of embossing, the latter pressing wet sheets onto a wooden engraving. The decorative papers mon shoin employ a type of watermarking using stencils laid onto the su. This paper was used in making sliding doors.    In light of this evidence, it seems reasonable to deduce that the Japanese integrated nineteenth century European methods with traditional Japanese craft to create something novel. Unfortunately, our real understanding of tesuri is limited by the apparent absence of the original plates used to create the papers. In the foreword to Tindales' work, Haruji Yoshida, Director General of the Government Printing Office in Japan in 1950, writes that all of the Meiji era examples were destroyed in a wartime fire, in 1945. The plates are the keys, being the source of the subtleties that make tesuri so unique and beautiful.    Our priority, in trying to replicate this mysterious method of watermarking, was to find the proper relief plate on which to burnish our Japanese paper. Materials prepared for letterpress printing�photo engravings, woodcuts, linoleum-block cuts, and photopolymer plates�all seemed promising. Most importantly, all of these plates consist of high and low levels, with the lower, recessed area cut, engraved, or washed away.    In letterpress printing, the raised surface produces the actual image because it receives the ink and is what gets printed. Unlike planographic printing (lithography), this dimensional plate can make an impression (a sort of embossing) in the paper when it is printed. We needed a plate surface that could reproduce various images and would consist of these three-dimensional levels so that, when we laid a recently formed sheet of Japanese paper on top of the plate and hand burnished the paper, the fibers would be displaced around the raised image. They would thin out over the high areas of the plate and accumulate in the low ones.    We began by burnishing damp paper onto photoengravings made from copper, zinc, and magnesium, since the Tindales' examples used engravings. The watermark images these plates yielded tended to be muddy and lacking in detail. We then tried woodcuts and linoleum cuts, with limited success.    Finally, we turned to a relatively new material used in letterpress printing: photopolymer plates. These consist of a thin layer (about a sixteenth of an inch) of light-sensitive photo emulsion laminated onto a thinner, metal backing. Any part of the plate that is exposed to light hardens to a polymer, leaving the unexposed area water-soluble so that it can be washed away. An image that contains no half-tones (e.g., handwriting, typography, a line drawing, a pixilated drawing, or a photograph reduced to black and white) can be translated to a photopolymer plate. However, we discovered that large areas of uninterrupted, raised emulsion can be difficult for the paper to accept without appearing somewhat washed out. For example, we found that white spaces in a laser-printed computer graphic needed to be converted into gray tones to ensure they would be pixilated throughout.   A negative of any of these types of image is laid on top of the unexposed surface of the photopolymer plate and then vacuum-suctioned into place. This is then exposed to ultraviolet light or sunlight, so that the clear areas in the negative let light through and harden the emulsion. The plate is then softly brushed while submerged in water, which washes away the unexposed, unhardened emulsion, leaving a raised image (where the original negative was clear) on the finished plate.   We had finally discovered a suitable plate material. We found the photopolymer plate's high and low surfaces to be equally smooth, which kept the wet paper from sticking. With polymer, we were able to reproduce even the finest details of drawing and type. As a material, it is quite resilient and stands up to repeated burnishing without any wear to its surface.    The success of our tesuri experimentation also relied heavily on the kind of paper we made. The Tindales' watermarks used mitsumata and a combination of chemical pulp. A paper made in Fukui Prefecture, fusuma-ban torinoko, has the same mix and appears upon inspection to be nearly identical to the Tindales' papers. Not knowing the make-up of the chemical pulp used, we have tried pure mitsumata. Mitsumata has the shortest fibers of the three most common Japanese plants used for papermaking and some believe it has the greatest transparency and best sheet formation.   We employ the nagashi-zuki method in forming our sheets and make them unusually thick, sometimes laminating two sheets together to ensure there will be enough fiber to form the image. Couching sheets into a post and pressing them in the traditional way would remove too much water from each sheet and would prevent the fibers from being manipulated, so we form sheets and couch them one at a time onto layers of Pellon. We then remove some of the excess water from the sheets by carefully placing cotton blotters on top of them and gently patting. When the paper is dry enough for easy handling (almost as dry as a post of pressed paper ready to be brushed for drying), we place each sheet on the face of an imaged photopolymer plate.    Handling becomes a central concern, because we move the sheets several times throughout the process. Stretching and creasing from improper handling at any point in our process would ruin the end product, regardless of how well the sheet was formed or how well it was burnished. A University of Texas conservation student, Heidi Nakashima, suggested we use the technique developed by Hyogushi (a Japanese scroll mounter) for handling objects with no wet strength. In this technique, a wooden bamboo stick or baton is pasted to the wet paper's edge for transport, and positioned at an angle as the sheet is set back down, ensuring no alterations to the form. In our recent experiments this technique has worked very well.    The burnishing tools that we have had the most success with are the broad sides of an etching burin and a bone folder. Our damp mitsumata paper is too delicate to burnish directly, so we have tried various materials as a layer between the paper and the burnishing tool. Thin Mylar is not flexible enough and cotton wrinkles easily if not kept stretched taut. Monofilament silk-screening material seems to work well and we can easily purchase it already stretched on a frame; its transparency helps us see where we were burnishing. We were surprised by the wet sheet's ability to withstand, without tearing, the force we use in burnishing. After burnishing, the plate and paper stick to the silk-screen frame. As in couching Western papers, we angle the frame while holding down the plate and paper, then quickly pull up to keep from stretching the image.    After we have burnished a sheet and it has taken on the image from the plate, we are faced with drying the sheet while preserving the image. We successfully have dried our watermarked sheets in a stack dryer, comprised of blotters layered between corrugated cardboard. The ideal way to dry them would be just as they were in Oji, on a wooden frame. So far, we have had only limited success with this method.    In overview, this experience with tesuri brought us closer to both the past and the future. Many of us working with traditional craft always seem intent on recreating past techniques. We wonder how the artisans were able to craft by hand such extraordinary, labor-intensive work as the plates used at Oji. But we also need to look forward. As we approach the twenty-first century, new materials and equipment enable us to take traditional craft into new areas. For example, the computer program Adobe PhotoShop enabled us to generate photographic images in minutes. The photopolymer plates we then made and used could pick up even the smallest dots, which would surely have taken a skilled wood engraver much time.    Through our experiments with tesuri-kako-ho we discovered we can never perfectly recreate past fine craft techniques. We can, however, integrate traditional craft with new developments and technologies to create something of unique quality and lasting integrity.   Notes   1.     Thomas K. and Harriet R. Tindale, The Handmade Papers of Japan. Tokyo and Rutland, Vermont: Charles E. Tuttle & Co., 1952. 2.     Nihonga, a style of Japanese painting developed in the Meiji era that used contemporary Western images in a traditional Japanese format, is a good example. Examples from outside the arts include the adoption of the Western calendar, European-style parliaments, public education, and Prussian military systems. 3.     Col. Joseph E. Boling. "Building a National Currency - Japan, 1868-1899," an exhibit of the American Numismatic Association, 1996. 4.     Narita, Kiyofusa. A Life of Ts'ai Lung and Japanese Papermaking. Tokyo: Dainihon Press, 1966. For an interesting discussion of another watermarking technique used in the production of banknotes, see also Sukey Hughes's Washi: The World of Japanese Paper (Tokyo: Kodansha, 1978), p. 196.   The authors thank Timothy Barrett for his inspiration and encouragement in this research; Knut Graf of Frog Design for assistance with computer graphics; and Chris Brown for review and comments on various drafts of this article.