September 16, 2008

Introduction & Background

The codicology of ancient Chinese manuscripts is a vast and fascinating subject. For the purposes of this study, I have chosen to concentrate on the large collection of manuscripts discovered at Dunhuang in 1900. More than 40,000 artifacts have been recovered from a series of Buddhist caves, many of them Chinese manuscripts dating from the first century B.C.E. up until the late 10th century A.D.

These valuable documents were scattered all over the world, but a recent initiative called the International Dunhuang Project is using digitization to bring the entire collection together for the purposes of research and conservation. Their efforts have made more than 20,000 of the manuscripts available in digital form.

Who made the manuscripts?

The caves of Dunhuang were part of an extended Buddhist compound, and this religious slant is apparent in the manuscript collection. Of the digitized documents, only about 425 are secular works; the rest are various Buddhist texts including prayer sheets and canonical sutras. Many of the texts have colophons identifying the scribes; a number of these identify the writer as a “bhiksu” or a fully ordained Buddhist monk. There are also references to monastery libraries and a system of organization; some scrolls were labeled with descriptors such as “9th scroll in the 15th bundle”. At least one scroll (Or.8210/S.296) had been a part of two different monastery libraries.

The secular texts were likely written by educated laypeople; China at the time had a well-developed educational system, a civil service whose members were appointed based on rigorous government exams, and a flourishing scientific community. One of the more extraordinary finds at Dunhuang was a remarkably accurate star chart (Or.8210/S.3326) from 940 A.D. Other secular documents include medical texts and literary works.


Paper was an early Chinese invention, but the concept of the book in codex form did not appear until many centuries later, sometime in the 10th century (Giles, 1957). The majority of the manuscripts found at Dunhuang are in the form of scrolls, the longest measuring 99 feet. Sheets of paper were pasted together to form the scrolls, since the papermaking frame was a standard size. A wooden roller was attached to one end; the other had a silk ribbon or braid which was used to fasten the scroll once it was closed.

Some of the later manuscripts have been folded concertina style, making it easier to navigate through the text, and some are small booklets, but these were still the exception rather than the norm even in the late 10th century (Giles, 1957). By this point in Chinese history block printing was becoming widespread, and these new book formats were conducive to being printed. There are extant examples of block-printed scrolls, however, so even as printing began to take over book production the old methods of binding were still in use.

One transitional method of binding has been discovered – whirlwind binding (Or.8210/S.6349). This was composed of separate pages, bound at one side, that were rolled up and stored as a scroll. Examples of this binding are extremely rare, and thus far only reference texts have been found using this format; obviously it would have been much easier to find specific sections in such a book than in a long scroll so the format was perfectly suited to reference materials (Chinnery, 2007).

Tools of a Scribe

The “Four Treasures of a Calligrapher’s Study” are paper, ink, ink stone, and brushes. The first three are examined elsewhere in greater detail. Brushes used in making the Dunhuang manuscripts were of two main types: a hard brush made of deer hair in a wooden shaft, and a softer brush made of rabbit hair in a bamboo shaft (Fujieda, 1997). Brushes could also be made from weasel, fox, goat or wolf hair, and the very finest brush was said to be made from the hair of a newborn child (Bo, 2002, p.102). The brush was held upright, perpendicular to the paper, as this gave the greatest degree of control.

Image source:


Early Chinese manuscripts were written on silk (Or.8210/S.5719), or on strips of wood (Or.8211/1 Side A) or bamboo tied together, but “silk was expensive and bamboo heavy” (Tsuen-Hsuin, 1985, p.35). Luckily the Chinese invented paper sometime in the first century A.D. and it rapidly became the material of choice for manuscripts of all kinds.

Paper was made from the fibres of many kinds of plants; the Chinese used mainly hemp and plants in the hemp family, rattan, mulberry bark, and grasses like bamboo, rice and wheat (Tsuen-Hsuin, 1985, p.52). Hemp and mulberry were the most common in early Chinese history, with bamboo becoming significant from the 7th century onward.

The basic process of papermaking has not changed much since the first century; the raw materials were prepared by “soaking, pounding, boiling, washing, and bleaching the fibres” (Tsuen-Hsuin, 1985, p. 69) to remove the colour and the binding agents and break the plants down into fibres. Chemical agents like lime and ash were mixed in to aid in the breakdown of the plant material.

Once the pulp was obtained, it was formed into sheets with the aid of a screen. Early screens were made of woven grass and the pulp was poured over them to drain the water; later screens were made of thin strips of bamboo tied together and either attached to a frame or laid over a frame. These bamboo frames were dipped into the pulp mixture, and had the distinct advantage of being smooth so that the sheets of paper could be removed while still wet. The sheets were then piled together, pressed to remove the excess moisture, and placed on heated racks or walls to dry. This meant that one frame could be used to rapidly make many sheets of paper (Tsuen-Hsuin, 1985, pp.64-69).

The paper was sized with a glue made mainly of botanical ingredients, sometimes with animal materials or starch flour added.These mixtures were added into the pulp before it was formed into sheets. This helped the fibres stick together, adding strength to the paper, and also prepared the surface for accepting ink without “undue absorption and running” (Tsuen-Hsuin, 1985, p.73).

Inks & Dyes

The vast majority of the Dunhuang manuscripts are written in black ink.Red ink was used for signature seals at the end of a document, and there is one example (Or.8210/S.76) of a medical text where each section starts with a character in red ink. Historical sources tell us that the black ink was made from soot and animal glue, sometimes with additives of various botanical compounds (Zhan, 2007).

There have been a number of recent scientific studies attempting to analyze ink remnants using Raman spectroscopy; the first (Clark, Gibbs, Seddon, Brovenko, & Petrosyan, 1997) identified the red ink as cinnabar, with the additional presence of an organic red pigment that was likely madder; the second (Burgio, Clark, & Gibbs, 1999) found that a blue dye used to colour the paper was indigo. Paper was also dyed yellow, and this dye has been identified as coming from the Phellodendrum Amurense or Amur cork tree, along with other unidentified sources (International Dunhuang Project, 1995).

One distinct feature of Chinese ink was its form; the ink was formed into sticks, often with decorated sides. To write, the ink stick was mixed with water by rubbing it against an ink stone that had a reservoir for water at one end (as in this

Decoration of the text

The paper was often dyed, partly as protection against insects and partly to achieve an attractive glossy surface. Yellow was a favourite choice but many other colours were also used, including red, green, pink and blue (Tsuen-Hsuin, 1997).

From the earliest development of writing in China, calligraphy was considered an art form in its own right, and this may explain the lack of decoration on the vast majority of the Dunhuang manuscripts. A keyword search for “illustrated” in the IDP database returned only about 60 results, from over 20,000 manuscripts. When the manuscripts were decorated (apart from the calligraphy), it was mainly with line drawings (Peald 3c Recto) done in the same ink as the text. Drawings were often diagrams or representations of the text, as in this star chart (Or.8210/S.3326), rather than purely decorative.

Most of the manuscripts have visible gridlines; these were likely to assist in making the text uniform, but they may also have been considered decorative – in most cases the lines are as visible as the text. If they were purely functional you would expect them to be fainter so as not to draw attention away from the text.


The earliest Chinese characters were pictograms etched into bone fragments sometime before 1700 B.C.E. These beautiful characters, known as Jiaguwen or Oracle Bone Script, were inspired by nature and usually resembled the thing they were meant to represent. During the Shang Dynasty (c. 1750-1040 B.C.E.), a second form of the Chinese language was developed for inscribing historic deeds on bronze vessels; not surprisingly this was called Bronze Vessel Script (Jinwen). These were still reminiscent of natural forms, but they had evolved into ideograms that could represent abstract ideas. From these roots, Chinese script went through several more main stages of development, some designed to be easy to read, others to be purely an art form; the common thread being the beauty and grace of the characters as they grew more and more refined (Bo, 2002). Possibly because of their origin as pictograms, Chinese calligraphy has always been considered an art form in its own right, and thus the handwriting is a key measure of a manuscript’s attractiveness and value.

Evolution of Chinese Characters



Bo, S. (2002). Between Heaven and Earth: A history of Chinese writing. Boston: Shambhala Publications, Inc.

Burgio, L., Clark, R.J.H., & Gibbs, P.J. (1999). Pigment identification studies in situ of Javanese, Thai, Korean, Chinese, and Uighur manuscripts by Raman microscopy [Electronic version]. Journal of Raman Spectroscopy, 30, 181-184. Retrieved September 12, 2008 from Wilson InterScience database.

Chinnery, C. (1997). Bookbinding. Retrieved September 15, 2008 from

Clark, R.J.H., Gibbs, P.J., Seddon, K.R., Brovenko, N. M. & Petrosyan, Y.A. (1997). Non-destructive in situ identification of cinnabar on ancient Chinese manuscripts [Electronic version]. Journal of Raman Spectroscopy, 28, 91-94. Retrieved September 12, 2008 from Wilson InterScience database.

Fujieda, A. (1997). Chronological Classification of Dunhuang Buddhist Manuscripts. IDP News, 8-9. Retrieved September 13, 2008 from

Giles, L. (1957). Introduction to descriptive catalogue of the Chinese manuscripts from Tunhuang in the British Museum. Retrieved September 15, 2008 from

International Dunhuang Project. (1995). Conservation and science. IDP News, 3. Retrieved September 14, 2008 from

Tsuen-Hsuin, T.(1985). Paper and Printing. In J. Needham (Ed.), Science & Civilisation in China, v. 5 pt. 1. Great Britain: Cambridge University Press.

Zhan, X. (2007). A study in the manufacture of old Asian inks. IDP News, 30. Retrieved September 13, 2008 from