HAN DYNASTY IN SERINDIA BY STEIN

Volume II, Chapter XX
History and Records of Tun Huang line

In this chapter Sir Aurel summarizes the results of his research on the ground of the Tun Huang wall during his second expedition and adds comments and his conclusions based on reference also to written Chinese records and translations of some of the documents he found at locations on the wall and towers.

Section I - Extension of the Limes beyond Tun-huang -

Stein found that Chinese records help one learn details of the wall construction and thereby elaborate on the archeological and topographical information he developed on the site from his survey itself. Among the documents he found in refuse at several towers are the oldest Chinese manuscripts known. He asked M. Chavannes, a Sinologist scholar to translate and comment on these, which he did in his report "Documents chinois".

In 214 B. C. the Emperor, Shih Huang-tu who founded the Ch'in Dynasty linked several of the defensive walls built by previous ancient kingdoms. His wall from Shan-hai Kuan on the Gulf of Lio-tung went west as far as Lin-t'ao, now in the Min prefecture in southern part of Kan-su - 110 miles south of Lan-chou. A century later the "Great Wall" was extended north-west as a defensive wall. The Han wall was new not only in construction but also in purpose. It was not for strategic defense, but it went 1000 miles further west to the edge of the Tarim basin to protect a strategic offensive expansion into west-central Asia.

Chang- Ch'ien conducted missions 138-126 B.C. under Emperor Wu-ti with the initial objective of securing allies among the Yulh-chih (Indo-Scythians) against the Hsiung-nu who continually attacked China. But the missions revealed to the Chinese the great possibilities for new commercial relations also if the Chinese could open trade routes via the Tarim to Farghana- Sogdiana and the Oxus region.

Chang Ch'ien's report to the emperor is in the Ch'ien Han shu. It showed that communication west would be at the mercy of the Hsiung-nu from the north and the Ch'iang Tibetans from the south over the first section - that is prior to reaching the vast desert. It was at the location of Tun-huang that the passage was narrow and the two nomadic tribal groups could meet. From there westward the absolute desert of Lop protected caravans from attack.

Beyond the Lop desert there were two great routes across the Tarim basin. The desert oases were unsuitable for occupation by nomads. If the Chinese could establish control over the small local groups inhabiting these oases they could protect convoys. The locals would welcome Chinese power to provide protection and also greatly increase profitable trade.

So Tun-huang was the gate and base for expeditions to the West - one along the south side of the desert along easy and safe route along the foothills of the Nan-shan mountains which provided water. This is still a main road. Between Lan-chou and Liang-chou at the east end of the Nan-shan there is an easy pass into this corridor. Beyond Lang-chou there are no natural difficulties for a road through fertile valleys with plenty of streams.

The main rivers - the Kan-chou and Su-chou become together large enough to penetrate northward toward southern Mongolia. However, along this narrow belt west from Liang-chou to Su-chou the only practical invasion route from the north for large armies is along this river named the Etsin-gol. It is by this route that Chingis Khan eventually invaded Kan-chou in 1226 A.D..

So the narrow corridor, protected on both flanks was the only route west. On the south-west side the Richthofen Mountains, high and always snowy, was a rampart with few alpine routes through which small Tibetan nomad groups might raid. And these were easily defended and fortified. On the north-east flank the Ala-shan mountains, low but barren and with extensive desert beyond were a good defence also.

Beyond Su-chou a narrow agricultural belt of oases was enough to provide food support for caravans and defensive garrisons clear to Tun-huang.
Thus the topography shows the reason for the location of the Han wall.

In 121 B.C. general Ho-Ch'u-ping defeated the Hsiung-nu sufficiently to establish Chinese control over Kan-chou and Liang-chou. In 115 B.C. the Chinese created a command center at Su-chou and began establishing military colonies along the route west. Then another command center was established at Liang-chou. In 111 B.C. two command centers divided the region with headquarters at Kan-chou and Tun-huang. In 119 B. C. the Chinese passed north of the Huang-ho to set up military colonies of 50,000 to 60,000 men. The Chinese annals describe the building of the wall west from 121 B.C. but not the details of specific local sites.

Stein remarks that prior to 1914 he had not looked for the wall east of Mao-mei on the middle Etsin-gol. He repeatedly compares this Han wall with Roman frontier walls in Africa, Middle East and Europe.

Chinese records from 108 B.C. relate the line of posts and forts from Chin-ch'uan or Su-chou as far as Yu-men (the Jade Gate) so named because precious jade was imported through this gate. Jade comes from yu=jade. Stein locates the Jade Gate at tower Txiv from 96 B.C. or somewhat earlier.

But he considers that prior to 103 B.C. the key gate may have been located further east and he proposes two possible locations. He did survey the wall between Su-chou and An-hsi in 1914.

One possible alternate location for the Jade Gate is near Shih-esh-tun, 15 miles north of Yu-men-hsien, where the wall now from north of Su-chou reaches the Su-lo- Ho river west at a bend of the river. If the wall indeed had a headquarters at that place the marsh south and along the right bank of the Su-lo Ho river would provide good flank defensive protection. In 1914 the route Shih-esh-tun to Su-chou north of the hills was practical.

The other possible location for the Jade Gate in 103 B.C. is the defile between Bulungir (Pu-lung-chi)_ and An-hsi where the Su-lo Ho passes the foot of low hills at Wan-shan-tzu on the left bank and a ridge on the right bank. This is the only defile along the Su-lo -Ho after it leaves the mountains and it is easy to watch and defend.

In 1914 Stein traced the wall on the right bank - north side - across the Su-lo- Ho - a decayed section - to hills that have a tower. On the steep south side the wall must have continued, but Stein found no remains. A series of watch towers from a later period was found on successive spurs where the high road from Bulungir to An-hsi passes. This shows that the defile was specially guarded even much later after the Han wall was abandoned. Twelve miles further west of the Hsiao-wan oasis the remains of the Han wall and towers are again found on a gravel glacis toward An-hsi.

The defile at Bulungir was suitable for a main guard station before the Han wall was built. Twelve miles east of Wan-Shan tzu defile is a large now abandoned circumvallation at Bulungir. During Manchu era this was a major garrison. Below the point where the road from Bulungir ascends the Wan-shan tzu spurs near the left bank of the river is a group of Chinese temples at Lau-chun-miao which might indicate that this area was near a major gate.

Whereever the Jade gate was in 103 B.C. a few years later it was moved west to tower T xiv north of Tun-huang. The Former Han annals record the success of Lu'Kuang'li on his second expedition against the Ta-yhuan in 102-101 B.C. and that the wall was moved westward -forward - beyond Tun-huang at that time. The western peoples sent delegations to the emperor then to increase diplomatic relations and trade. To protect these caravans military posts were built from Tun-huang westward to the Salt Marsh.

Stein points to documents he found dating from 98 B. C. to 96 - to 94 B.C. at various towers. So the 'limes' reached Txiv the Jade Gate by 96 B.C.

The wall was built rapidly during a few years, but Chinese military organization was well prepared for the task. In 104 B. C. Lu' Kuang-li took 10,000 men from China into the Lop desert route but few returned. In 102 B.C. on his second expedition he took from Tun-huang 60,000 men plus camp followers, 100,000 cattle, 30,000 horses and returned in 101 B. C. at the Jade gate with only 10,000 men.

The Chinese development of their wall defense was similar to the Roman limes. The topography shows the local reasons for site selection. Wu-ti's engineers made the wall to Tower Tiv on the edge of the marsh filled basin of the Su-lo Ho. This was the best place for flank protection. The wall ended where it used the west border of the lake and marsh. Likewise the line of watch towers beyond the end of the wall - Ti Tii - was intended to provide added security for the exposed western end of the wall. Advanced posts made it easy to watch the main route from the west and send a warning. The line of watch towers was similar to the Roman line in the African desert.

(Stein discussed this in an article about Third Journey of Exploration in the Geographical Journal number xlviii page 127). In 1914 Stein discovered near the NW edge of the salt sea the ruins of a fortified camp at the location of the exit of the route from the sea (Discussed in Innermost Asia). He also found more watch towers in 1915 along the Kan che-darya toward Korla on the road beyond Lou lan.

Section II - The Tun-huang Limes since its construction

Stein notes that there is no direct data on Chinese records apart from general history, but in the documents he found from 68 B. C. to mid 1st century they show Chinese control of the Tarim Basin, then Lou lan was reduced in 77 B. C. There was a Chinese protectorate general in 60 A. D. But no documents were found with dates between 39 B. C. and 1 A. D. But during the reign of Wang Mang A. D. 9 - 23 there was another change. The Late Han annals dated 6 B. C. to 5 A. D. show a break up of the western regions. There was much border defense activity during the Wang Mang reign. Wall towers Txiv - Tiv were abandoned.

The wall at right angles from the Jade Gate to Yang barrier was for active defense. The Chinese needed to reduce the extent of the wall defenses being occupied in order to concentrate their forces. Chinese power was less after 25 A. D. They lost control of the Tarim Basin and held only the wall east of the Jade gate. During A. D. 58-75 the Hsung-nu twice attacked Tun-huang and the area clear east to Su-chou and Kan-chou. In AD 73 Chinese ;policy of offensive was renewed under General Pan Ch'ao. In A. D. 73-102 the Chinese had victories and reigned control of the Tarim basin. This reduced the requirement for large garrison wall defense. This advance was based on the northern route to Hami established in 73 AD. The line An-hsi - Hami since then has been the chief route west.

After 102 A. D. the Chinese control was again lost. 153 A. D. is the last date in the Han annals for Chinese operations west. In AD 153 also was the last dated for a document that Stein found. So the wall was abandoned but not the caravan route via Lou-lan. At Lou-lan he found documents from AD 263 - 330. Then the monk, Hsuan-tsang, made his pilgrimage to India passing through in 630 A.D. Then the Jade gate was near An-hsi. In the T'ang era the Jade gate likely was at Txiv again as it is mentioned in some T'ang memoirs.

Section III - Main Features of the Remains of the Limes

The wall was kept close to the route of Su-lo Ho river to its end in the extensive marsh. The river was used as a natural defense north of the wall. The wall was uniform in design and construction methods. The gap in the wall where the river and marsh made attack impossible shows the defense was adequate and parallels Roman wall design. Construction methods made use of natural resources. The use of fascines placed cross wise in regular layers alternating with stamped clay and gravel helped protect the wall from wind erosion. The fascines of reeds or tamarisk branches (wild poplars) which were nearby were easy to obtain. They were the same thickness as the wall - 7.5 to 8 feet. Water was used to reduce the stamped clay to cement, but it came from a distance. Behind the wall but close to it was the chain of watch towers for guard and to transmit signals.

The towers were NOT designed for strengthening the defensive capacity of the wall. Most towers had quarters for small detachments of watchmen and patrols. The distances between towers varied greatly according to the characteristics of the terrain - extent of view and vulnerable terrain. The section NE of Tun-huang - towers Txxxi - Txxxv - were at 3/4 mile intervals between each other. The distances between towers to the west were usually greater - in once case 4.5 miles, but on the opposite shore of Lake Khara Nor.

In the SW flank which was well protected by natural defenses - marsh - there was no wall line, but only watch towers 5 miles apart on high ground. Along the wall line towers were on high ground, clay ridges or mesas. But towers Txviia - Txix - Txxi were only guard stations on veery high ground. The watch towers were all solid (no interiors) and square. They tended to narrow toward the top and vary in size and height and material. The tower bases were from 16 to 24 feet square. Decay prevented absolute measurement of many heights. Some were 30 feet - such as Tvvid - others had floors of a guard room at 8-13 feet height, related to the height of the ground base; the elevation of the base and range of vision.. Material depended on local conditions. Bricks - sun dried - were used when water was not available. Some used clods of salt impregnated clay - tower Tx fig 174, for example. Soil for this clay contains brackish, undrinkable water.

The wall was always strengthened by insertion of layers of reeds, usually after 3 courses of bricks, or stamped clay or clods. The masonry was reinforced by insertion of solid Toghrak timbers. The excellent construction resulted in less erosion from the terrible wind-driven sand over 2000 years. There were small guard rooms on top of some towers - such as Tvic - Tvi - Txxx - Txxi - Txxiiia - All towers must have had shelter and a parapet and at least access to the top of course. The remains of stairs at some towers were visible. At others there were foot holes - the men used ropes to climb up. In the room on top of some towers defensive large stones were found.

Defensive methods were the same along the wall and in fortified villages and farms near Tun-huang - Su-chou and Kan-su. Even the walls of Ch'sen- huang -gate castle at Chia- yu kuan had stores of stones for missiles. (see Desert Cathay). But the watch towers were not strong defenses - just for small garrisons. Some towers were close to the wall at 19 feet but some as much as 24 yards from it. They were not for defense of the wall. Towers were plastered and painted white wash which is visible now were the wall was protected by adjacent quarters. Towers were made more visible at a distance this way. The quarters were very small. An example is Tower Tvib. They could accommodate less than a dozen men. Quarters had only one small entrance. The walls were of brick, same as the tower.

Ruins separate from the wall have been discussed already. From the east there was the ancient magazine at Txviii at a special location on the route to Lop. A document shows the date to the 1st century B. C. It was built at the same time as the wall. The traffic there went to Lou-lan. Next is the Yu-men - Jade gate - a small but massive fort, Txiv, on the route to Lou-lan. It was chosen to meet the need for a headquarters station. The extent of the refuse heap there shows extensive use of this site. Txva, north of Txiv, also had extensive use to control the passage through the wall. The next tower, Txii, was located where ground made it easy to have a police traffic control post on the Lou-lan route. Tower Tixa was well beyond (north of ) the wall for control of foreign ground and protected by the marsh and terminal tower Tiv. Probably there was an entrenched camp on the caravan halting place. The line of the watch towers on the SW flank and separate from the wall included Tvib, a small station as a section headquarters for the flank towers.

There probably never existed on our globe a system as systematically organized and guarded on a border line stretching over desert ground which was as barren and forbidding as that traversed by the Great wall beyond Tun-huang must have been all through historical times. Two conclusions are clear. First, we must realize that on such ground the constant maintenance of considerable detachments or bodies of troops along a line which for great stretches was devoid even of water, would have presented most serious difficulties about supplies and transport. This makes it obvious that regular pickets along the wall were small. for maintenance of signal and patrol service and protection from few raiders. Large forces were stationed only at the Tun-huang and Nan-hu oasis. Second, it is improbable that on such ground other structured remains not seen by Stein and team were likely as they crossed and recrossed the area for months.

Section IV - Military agricultural colonies

This section is Stein's analysis and commentary based on documents recovered at the various wall towers, especially document 60 found at Tvib. This is a specific 'imperial order given to the governor of the command, Chiu-ch'uan', to establish a military agricultural colony. The document prescribes that 2000 soldiers from the garrison command of Tun-huang should be established in a colony that will provide agricultural products to feed the troops assigned to the frontier posts. But the location for this agricultural enterprise is not specified. Stein considered the possibilities based on the terrain, water supply, and distances to the wall. Since it would be impossible to grow anything in the terrain adjacent to the wall itself, he considers that the only possibilities would be the oasis at Tun-huang or at Nan-hu. The latter being closer to the wall, he opts for that location. Considering that the document was found at tower Tvib he believes it relates to the the extension of the wall built west of Tun-huang. He notes that his examination of the actual wall shows how closely it conforms to the specifications in the Imperial edict. He dates this edict to later than 111 B. C. It refers to the mentioned command at Chiu-ch'uan which is Su-chou which was at that time the location of the main advanced base.

Other documents at T vib date as far back as 68 B. C. so at least this edict is dated after 102-101 B. C. which is the time the wall was extended westward as planned to reach the natural line of defense to the south-west where T vib was located. Since Tvib was the location of a very large refuse dump of documents, Stein believes it was also an office.

Stein then is mostly concerned with the location of the agricultural colony prescribed in the edict. He notes that such colonies were an essential part of Emperor Wu-ti's program and necessary for logistical support of the wall garrisons. They were also necessary for support of the caravans and expeditions traveling further west, beyond the wall. The original small inhabited places conquered from the Hsiung-nu and Yueh-chih could not support such extensive requirements for food and fodder. But such agricultural communities could not be established right near the wall, either, because it was located in totally barren terrain.

Stein here refers to the vivid descriptions of the towers and their locations already given in prior chapters. He comments, "Cultivation was never possible within historical times either in the marshy depressions which skirt the terminal course of the Su-lo Ho or on the gravel plateaus which over look and divide them". and "Nor could there have been, during historical times, a sufficient permanent supply of water at the foot of the hills west of Nan-hu to cross the huge gravel glacis, partly covered with high dunes, and assure irrigation south or south-west of the Limes."

He notes that the existence of wooden documents in excellent condition after 2000 years shows how dry the area has always been. Nor could he discover any traces in the ground of agricultural occupation. Thus he concludes, as described above, that the only possibilities were the two oases, Tun-huang and Nan-hu, both quite a bit south of the wall. He notes further that his conclusion indicates that the defense of the wall was entrusted to colonies of soldiers who acted also a cultivators, and were settled some distance from the wall. This conforms to the clear evidence that the wall and towers were occupied only by very small detachments (since the accommodations there were very small). Thus the permanent organization, which lasted for over two centuries, was well organized in every respect in conformity with the realities of the topography and climate.

As he comments, "No system could have been better adapted to overcome them than one combining the provision of a permanent local force for the Limes with the production of the food supplies which it needed. It is obvious that such a system was workable only on the basis which would allow the bulk of the soldier colonists to remain near their lands except in times of emergency while the actual guard and signal service along the desert Limes were carried on by small detachments in turn."

Stein then provides lengthy examples of similar methods used by the Romans along their Limes (which Stein consciously copies in his terminology) - and by the Austrians along their "Militar-Grenze" in the Balkans. But the small number of documents prevents detailed study of the actual Chinese organization and allows for only 'general observations'. But document number 60 is not the only one that provides some valuable information. Document number 63, found at the same tower, is another imperial edict concerning three companies that garrisoned watch-posts on the extreme south-west end of the Limes.

Section V - Officers and soldiers of the Limes

In this section Stein continues with more detailed study. We can only touch on some of this here. He has examined and analyzed the large pile of documents retrieved from all the towers as a body and formed his conclusions based on this unified study. Plus, he repeatedly refers to the studies of noted French Sinologist, M. Chavannes, who was able to evaluate this collection with reference to his extensive study of Han Dynasty archives. (Stein throughout his books always gives extensive credit to the many specialists with whom he collaborates). In this section Stein begins at the top, so to speak, - with what can be known of the high command along the wall.

At the time of Emperor Wang Mang apparently there was a high commander who in turn controlled four commands between Tun-huang and Liang-chou. Another document refers to an administrator in Chiu-ch'uan (Su-chou). The local command was entrusted to an officer based in Tun-huang itself. In turn, this commander (governor) controlled several different sections of the Limes. One of these was the Yu-men barrier.

There were also local commanders at Tun-huang and Yi-ho. Stein describes quite a few details of the command structure, including the Chinese titles for various ranks or offices) gleaned from the documents.

He notes,. "The fact that the great mass of our Limes documents has been recovered at small watch-posts explains why references to officers subordinate to the 'tu-wei' (military commandant) are far more numerous." The company commanders were called tui-chang. Documents describing many of their routine actions, such as receipts for food supply, are numerous. Each company had its distinctive name and each held a specific portion of the Limes walls and towers. From the documents Stein presumes that each 'company' may have been associated with a specific 'station' but that these 'stations' could not have been along the wall itself.

The t'ing was an administrative subdivision comprising 10 li each of 25 families. The t'ing was a small area within the oasis the population of which was expected to provide a company tui for guards on the Limes. The t'ing provided not only food but also troops. "Thus the t'ing was both a permanent recruiting area as well as the supply base in ordinary times for the company contingent which it furnished." Thus there were two officials - one the t'ing-chan was commandant of the area -and the other, the tui-chang, was commander of the troop unit that it supplied. This t'ing would have about 250 families and raise a body of men for a full company. One specific document (number 198 from the same tower) describes the a as having at least 145 men (likely a full strength unit then would be 150 men).

Stein continues with discussion of ch'eng - walled towns - garrisons. He notes that the remains of these are still seen through out the Kan-su marches housing villages or serving as refuges.

He discusses the duties of the tui. The daily duty was to provide detachments in rotation to occupy specific towers and provide patrols along the wall. This service is mentioned frequently in the documents. The commandant of a watch tower was designated a hou-chang. He was responsible for all activities at the tower and of the team there. The main duty of the men at a watch-tower was observation and then signaling of the approach of unauthorized or enemy forces.

The wall and towers were not designed for defense against a large enemy invasion, which in general could not be mounted across the Su-lo Ho anyway. But rather it was to interdict raiding parties. In addition to unit commanders apparently there were officers (wei) who were charged with control of 100 li and they had subordinates who inspected various lengths of the frontier. So the units had external inspectors. The overall impression one has from reading Stein's analysis is that the Chinese bureaucracy related to the Han wall was very elaborate.

Stein then analyzes the documents relating to rank and file troops. Many documents found in the individual towers related to low level personnel issues. One interesting item is that the documents frequently mention the exact canton from which the individual came and many of these are far from the frontier while others came right from Tun-huang. Moreover the men from far distances were concentrated in some towers while locals were in others. And it appeared to Stein that soldiers stationed in the early era of wall construction came from a distance while those whose service indicated later times came from the locality. Plus, the men from a distance were mainly convicts conscripted for this difficult service in the far western desert.

Some of the tower guards were mounted. The records also include units passing through on their way westward who were issued food supplies from the local stocks. The records also indicate that non-Chinese individuals were involved. Stein notes that the extensive international trade brought many foreigners to Tun-huang. .

Section VI - Service Conditions of Life on the Limes

In this section Stein describes duties and supplies. He notes that a main duty of a tower detachment was to send signals by smoke during daylight or fire at night. He already noted the piles of fascines carefully laid out in readiness near some of the towers. He also noted that on the top of some towers there was evidence of burning. And the documents (such as number 552 and 84 - 87) discuss this as well. This method pre-dated Han era.

Several documents are detailed records of the time and date such a signal was received and the name of the man who received it. There were signal towers not on the wall itself, prepared to forward such signals. One document lists some 42,330 bundles of fascines ready for use.

The daily duties of individual soldiers are spelled out in some documents - for instance, one would be on lookout, two busy collecting more wood or clearing away brush outside the wall and one acting as cook. Careful records were kept of all such duties. Another duty was making bricks (with number of men so employed and number of bricks created). Plastering work recorded the exact areas accomplished. Records of parties sent to collect wood or hemp or anything included exact times and distances involved. There were records of the exact number of days an individual served.

But Stein did not find sufficient records to determine if the low ranking troops were also paid in cash, but officers were paid in coin in addition to receiving rations. Other records relate to postal service and include dates for sending and receiving of mail, including addresses. Records also relate to notice about individuals or animals passing through the wall going in or out. Such records were found most frequently at those stations next to official customs border posts, such as Txva and Txi that were on the caravan route.

Other records refer to supply; food, and arms and armor, and clothing. Food included wheat, millet and rice and some records detail receipt of specific food by specific individual soldiers and from whom it came. The central magazine at Txviii was a granary and Stein found records there for whole cart loads of food. This supply depot also issued the large quantities of food required for caravans heading west.

There were copious records of the arms and equipment on hand and turned in for repair. Stein notes, "Lists, orders, labels and other documents relating to the arms kept at various watch-posts are abundant. The principal armament of the Limes guards consisted of cross-bows; they are frequently mentioned and in several varieties." They were categorized according to the force needed for stretching the string by weight - 3, 4, 5 and 6 shih a shih being equivalent to 120 Chinese pounds. Cross bow strings (silk or hemp) were listed separately. Arrows, too, were inventoried in two types of bronze heads. Stein found many of these in various types and sizes. Arrows were stored in bags or boxes containing 100. But the typical issue to a soldier was 150 arrows. Interestingly bows (not cross bows) were mentioned only as arms of barbarians.

Swords were mentioned only twice in the documents Stein recovered. Shields and other pieces of leather armor or leather scale armor are listed. Stein found some pieces of actual scale armor at Niya and lacquered scales at Miran fort. Chinese records included inspection reports, lists of damaged weapons, which were routinely turned in to depots for repair.

The soldiers were provided also with clothing, such as tunics and vests. Most were of linen. Stein also found scraps of silk and leather or hemp shoes. Other items included hammers for driving tent pegs and much medical supply including a medicine case. The documents, too describe great attention to medical - health issues. Medical reports include prescriptions for individual cases and general recipes for disease. Some, Stein notes, may have been from a note-book of a local doctor.
Other documents refer to discipline issues including punishments.

Stein remarks on the extensive nature of Chinese bureaucratic record keeping. The clerks were quite busy. Documents reveal standard bureaucratic phrases and nomenclature. Orders included requirements that a order be posted in a prominent place for all to read. There were orders against graft or ignoring instructions. Some official communications include individual self-denunciations for shortcomings. There were reports of unauthorized absences. There were portions of calendars used for keeping accounts.
In addition to official documents, Stein found many private letters and poems and parts of well known books. Some letters describe the author's hardship from duty in the remote desert. Many documents are written on silk or wood - predating the invention of paper. Among the books was the famous Chinese text Chi'chiu chang, a lexicographical text composed in 48-33 B. C. and used for education in writing Chinese script.