SERINDIA

Sir Aurel Stein KCIE
Macmillan and Co
St. Martin's Street, London, 1912
5 volumes, index, illustrations, color plates, maps

The table of contents with outline of the chapter headings is here. Listing of the maps Stein made during this exploration are here and here. This is my summary of the monumental book

Volume I Pages i to 548

Introduction

Stein states his purpose - "to provide a full record of the explorations, archaeological in the first place but to a large extent also geographical, which my second Central-Asian expedition, undertaken under the orders of the Government of India, enabled me to carry out during 1906-8 in widely distant regions of innermost Asia. The plan of these explorations was directly based upon the results which had rewarded my first journey in Chinese Turkestan in the year 1900-1, and their start fitly coincided with the completion of Ancient Khotan, my Detailed Report on that pioneer expedition." Stein continues to describe his concept to combine archeology and geographical study in order to enhance knowledge of the history of the region in which influences from India, China and the Hellenized Near East met and interacted. He notes that he has left the story of the hardships endured to his Personal Narrative published as Ruins of Desert Cathay. in 1912. He writes that this expedition lasted for 2 and a half years during which he walked or rode over 10,000 miles.

Stein describes the administrative and financial support he received from various government agencies, naming the individuals involved and expressing his gratitude for such support. He notes that this report contains 94 topographic maps. And that his Notes on Maps Illustrating explorations in Chinese Turkestan and Kansu contains information on the methods employed. The Survey of India has rendered great assistance in this effort.

He acknowledges the assistance rendered by his Indian staff and local Chinese officials. Then gives here a brief but complete summary of the entire expedition and the results obtained. He notes that this report was not completed for publication before he departed on his third expedition. This turned out to be an advantage because during that expedition he obtained new information which has been incorporated at appropriate points in this text. Likewise, the time he was allocated after the third expedition for work in the British Museum on all the thousands of artifacts collected enabled him to further describe these materials. He gratefully acknowledges the individual scholars whose expertise has informed the appendices related to specific subjects.

Addendum and Corrigenda

List of Abbreviated Titles - a bibliography

Table of Contents Vol I - III

List of Illustrations in Vol I - 144 of them

Chapter I -

Across Swat and Dir

Section I - Alexander between Kunar and Indus

In the first two chapters Stein is not yet out of India. However, the reader will learn from these chapters both Stein's two main objectives and his methods. His professional background is that of a scholar of Sanskrit who is an expert on Buddhist art in north-west India (now Pakistan). This artistic style was established during the period of Greek influence and developed during the reigns of the Kushan kings - Indo-Scythian rulers. The only contemporary literary descriptions of the region are found in the accounts of Chinese pilgrims to Buddhist India who describe in great detail the geography and topography of their routes and the customs of the peoples they meet. Stein wants to confirm the accuracy of these accounts for their relevance to our understanding today. His second objective is to trace and establish what can be known about the spread of this Graeco-Buddhist art style through Central Asia toward China. His method is to combine the skills and techniques of historian, anthropologist, ethnologist, archeologist, art critic, literary scholar, nusmaticust, geographer, and topographic surveyor into a multi-dimensioned analysis of these two themes. And the skill of the topographic surveyor are doubly important since it is for the purpose of obtaining accurate maps of the critical regions north of India over which the 'great game' is being played that the Indian Government is willing to fund a share of the expedition expenses.

Stein always had a special fascination with the exploits of Alexander the Great, especially in Central Asia and India. He carried Arian with him on many travels. And he finally did locate Aornos - Pir Sar on the Indus river. In this opening chapter he describes his theories about Alexander's movements in NW India (now Pakistan) including the Swat River valley and mountain passes on each side. For his second expedition (typically) he wanted to travel to Kashgar via a new route and for this required special permissions from the local khans in the mountain valleys and the king of Afghanistan. He records his appreciation for receipt of these special favors including from H. M. Habib-ullah, the King. His route was to be via Swat, Dir, Chitral, Hindu Kush, Oxus River valley, Wakhan corridor and over Pamirs to Kashgar. This route required careful timing - too early would be in impossible snow but too late would find the gorges filled with rushing high water from the snow melt. He was to cross the Lowarai Pass in snow. The distance from Peshawar to the Chinese Pamirs would be 450 miles, much of it in snow covered mountain passes. The Sanskrit name for the area to be traversed was Udyana which name he naturally adopted.

On 28 April 1906 Stein departed Kashmir, marching past Uch, Katgala and the Talash valley to cross the Swat River at Chakdara.

Section II - Early Chinese Pilgrims to Udyana

In this section Stein focuses on the accounts of medieval Chinese Buddhist pilgrims to Udyana, including Fa-hsien in AD 403. This monk visited Kashgar and then went west into the Pamirs and then south through the narrow gorge to Darel. Stein compares the monk's description of topography with his own observation and comments of the accuracy. The monk passed through the Yasin Valley to the Darkot pass (which naturally made Stein want to use it also). Stein refers to his Ancient Khotan in which he described the campaign of Chinese general Kao Hsien-chih in 747 AD through the Yasin Valley to the Darkot Pass and Baroghil saddle to Sarhad. He points out that Fa-hsien went from Darel to Udyana via the narrow valley and three gorges of the Indus but that no European has passed this way, which naturally makes this a 'must do' for Stein (to be accomplished during his third expedition). Darel also, he notes, is inaccessible to Europeans, another opportunity he hopes to achieve.

In 518 AD Chinese pilgrims, Sung Yun and Hui-sheng who traveled from Khotan to Sarikol and then in 519 across the Pamirs to Wakhan. From there they crossed into Udyana (in the Swat Valley) and left excellent reports on the topography of the regions traversed. Stein compares the reports of every medieval traveler he can find with modern estimates given in the literature as well as with his own observations. His descriptions and commentary are far more extensive than can be repeated here. He states from these that the best route from Badakhshan to Swat is across the Hindukush south of Zebak and up the Sanglich valley. From there one can cross either the Dorah pass at 14,800 feet into Chitral or over the Mandal Pass at 15,300 feet into the Bashgol Valley.

Section III - Udyana in Chinese records of T'ang times

Stein references the T'ang Annals throughout his expedition. There is the excellent report from this period by Hsuan-tsang from 630 AD. (Note - the spelling of this name in Serindia varies between Hsuan and Hsuang). Stein follows the pilgrim's itinerary and topographical descriptions made during his return trip from Udyana to China. He proves from his own direct observations over familiar ground in India that Hsuan-tsang's descriptions of specific locations in north-west India are exact. This establishes the veracity of the memoir that Stein will refer to constantly - and go out of his way to follow on the ground in many cases.

Stein refers several times to a little known Chinese military campaign into Gilgit and Chitral. In AD 747 the Chinese general Kao Hsien-chih invaded Gilgit with 3000 men over the Pamirs and through these passes in the Chinese effort to keep the Tibetans and Arabs from uniting. The Chinese campaigns continued into AD 751-52. Stein marvels at how the Chinese general was able to move and resupply his army through this terrain. Of course he wants to tred the same ground.

Section IV - Through Talish and Dir

Stein appologizes for not accomplishing more research before moving north out of India. On 28 April Stein rode to Chakdara and explored a hamlet, Gumbat, for his first active archeological work on this tour. As usual he combines archeological with ethnological work and compares it all with his historical documents. He reports in detail about a ruined temple that he had visited in 1897 which was since then being quarried of stone for new buildings. Nevertheless Stein prepared a site plan with measured dimensions, another example of his methods that will be used throughout Turkestan. His discussion is not limited to what he sees but includes also reference to whatever other places he has visited or read about. For instance: "But the survival in the Gumbat porch of remains of the trefoil arch furnishes by itself a very characteristic indication. This architectural feature was long considered peculiar to the style of the Old Kashmir temples, where it first attracted attention. But its presence is obvious in the far older remians of Gandhara Viharas (shrines) and their sculptural representations, and M. Foucher, in his masterly analysis of architectural art in Gandhara, has proved that its true origin must be looked for there." The paragraph continues to elaborate on the whole issue. He concludes with his estimate that the structure dates from the 7th to 8th centuries.

Over the following two days Stein rode through Sado and along the Panjkora to Dir and again apologizes for lack of detailed results due to the rapidity of his transit. He noted in passing an absence of hill top fortifications while such were abundent in the valley. He was 'forced' to remain in Dir for two days. He didn't waste the time. While there he collected very old coins and made anthropometrical - linguistic measurements of representative ethnic types.

Chapter II - Through Chitral and Mastij

Section I - Chitral in Ethnology and History

Again Stein collects ethnological information (including photos of the faces of various ethnic groups and information about their languages). On 4 May he reached Chitral by crossing the Lowarai Pass in heavy snow. (More complete description of his actions and thoughts are in the Personal Narrative - Ruins of Desert Cathay). He notes that this was the main trade route between Central Asia and India with easy access to the Oxus valley and Badakhshan. He comments that the inhabitants of Chitral had little local produce to trade but made a living from duties on those passing through. The area was held by descendents of Iranian invaders. Again he digresses into discussion of the composite racial makeup of the local tribes and into their various obscure languages. He also discusses the medieval history of the region from Chinese and Tibetan sources, local history being too limited.

Section II - Ancient Remains in Chitral

In these two sections we see Stein's typical organizational principle. He separates discussion of history and language from that focused on the physical geography, topography and architecture. On 7 May he rapidly examined various archeological remains including forts, watch towers, and mosques and shrines and including Buddhist stupa from a list prepared ahead of time by the local British agent, Captain Knollys. The remains of an old fort were near the Agency. The long-time ruling dynasty, Mehtars, had a castle with massive square towers in Chitral town, which of course protocol prevented him from inspecting. Near Jughor village he saw both ancient and modern forts. He reports lack of time to visit two more forts at Noghorghi and Gankorini. At Pakhtoridini he visited a Buddhist rock carving that depicted a stupa. This leads him to describe the standard architectural features of a stupa from square base to dome and unbrella. He looks ahead and mentions similar finds in Turkestan. He translates an inscription carved below the graphic - "This is an offering to the divinities from Raja Jivarman" - and discusses palaeography and the language itself. He expands analysis noting that the carving is proof of ancien local worship surviving into legend.

On 10 May Stein rode on from Pakhtoridini through narrow gorges and past locations that figured in the conflict in Chitral to enter Mastij. On the 11th he was shown another scale carving of a stupa. Another inscription refers to Raja Jivarman. At this he again draws a wider conclusion that will be repeated throughout Turkestan. The Mohammaden conversion of an ancient Buddhist shrine or holy place of any sort into an occasion to establish their own object for worship is characteristic of the region. He remarks: "Had a tradition of sanctity lingered about the spot even during the long period of occulation, or has Muhammadanism so little affected the subconscious beliefs of the population that they are ever ready to reassert themselves at the old places of worship?"

Section III - Historical Accounts of Mastij

The section again is Stein's discussion of historical accounts. He describes Mastij as the valleys drained by the Yarkhun River above its cofluence with the Drasan. It is in effect an extension of Chitral north along the Yarkhun river, but topographically separated by the difficult gorge in which strong defensive positions are maintained. It preserves political independence under the long-time rule of the Khushwakt dynasty. But the ease of movement east over the Shandur Pass at 12,250 feet has enabled the Mastij rulers on occasion to control Yasin. The valley is mentioned in T'ang Annals where it is named Chu-wei or Shang-mi. The Chinese traveler, Wu-k'ung passed through Mastij in 751 or 752 AD. Hsuan-tsang also described Shang-mi without actually visiting it. Stein disputes others' opinions that Shang-mi refered to Chitral but such broader interpretation is incorrect.

Section IV - Old sites in Mastij

On 11 May Stein went to Mastij on the upper Yarkhun River valley. He found an ancient fort and small villages. He found an 'old site' in Sanoghar where the Chitral Scouts were in training. On a rugged precipice above the village was the ruined debris of the ancient fort, also named Sanoghar. He comments that the region suffered from much warfare between local tribes. He continued up the valley on 13 May past an old fort at Brep on top of an artificial mound to a total height of 34 feet. The remaining wall at the top was an oblong with long sides 180 and 183 feet; and short sides of 103 and 133 feet. The walls were based on large slabs with upper section constructed of sun-dried bricks. There was a bastion and tower as well. Inside the fort walls were the decaying remains of houses. Stein describes the construction in detail but could give no estimate as to its date. He spent that night in Miragram where he found the most interesting building was Obaidulla Khan, the ruler's old house which he photographed and described.
By 16 May he reached the foot of the Darkot Glacier and pass, which he was determined to see due to its historic role.

Section V - Kao Hsien-chih's Expedition and Darkot

Stein made camp at the foot of the pass. As usual Stein was carefully following his Chinese records. He wanted to see just where General Kao Hsien-chin crossed in AD 747 with 3000 soldiers. Stein describes that campaign in detail including its strategic and diplomatic background. The Chinese strategic objective was to block contact between the Arabs and Tibetans both of which groups were seeking to expand control into the Tarim Basin. They both succeeded - the Tibetans for a century or so and then not exactly the Arabs themselves but the Mohammadens via conversion of the various Turkish groups such as the Kirghiz and Uigurs. But the Chinese expedition into Gilgit and Hunza was successful in pushing the Tibetans at least temporarily east. What ended Chinese control was the defeat of Kao Hsien-chin not south of the Darkot but rather west of the Pamirs by the Arabs.

Stein narrates Kao Hsien-chin's campaign with constant reference to the Chinese names for locations compared to their modern designations. It was indeed quite a military and logistic exploit. The report indicates that it took the Chinese three days alone to reach the crest in the pass.

Stein wanted to cross the watershed by both the north-east and north-west routes. But the sever weather and remaining deep snow on the eastern route forced him to use the western the Baroghil Yailak, but not before personally ( with a handful of hardy guides) assending the eastern part of the Darkot to the crest and gaining a view into Wakhan from there. All these travails are described in his personal narrative, Ruins of Desert Cathay. After struggling up onto the Darkot and back down again across the glacier he climbed again through the Baroghil saddle.

Chapter III - From the Oxus to Khotan

Section I - Early Accounts of Wakhan

Stein mentions his life-long wish to visit Badakhshan and the Oxus River valley. What is interesting in this chapter is a comparison with the same discussion in his 'personal memoir'. Badakhshan and the Wakhan are of course in Afghanistan. In his personal memoir,Ruins of Desert Cathay, he devotes much space to describing the extraordinarily hospitable reception he received immediately on crossing the pass into the Wakhan from Afghan military and civil officials specifically ordered to meet him by their ruler. He descended from the Baroghil Pass on 19 May. Yet in this chapter there is no mention. This may be due to his subsequently being denied entry into Afghanistan (the government was on the outs with the British). However, in the Introduction he does acknowledge with gratitude the special permission he received from the Afghan government. Ironically he was finally invited into Afghanistan in 1943 and promptly died in Kabul, where he is buried.

In this section he focuses on mention of the Wakhan in Chinese sources. It was on one of the most direct routes from Iran to China over the Wakhjur Pass to Sarikol and then to the southern rim of the Tarim Basin. The region is mentioned as far back as the Former Han Dynasty. There are more mentions during the T'ang Dynasty. The first detail is found in the memoir of Sung Yun and Hus-sheng and their pilgrim party who went to India in 519 AD via the Wakhan from Sarikol to Udyana. Stein provides these accounts. The area also was included in Chinese accounts after their conquest of the area in 658 AD, and they continued to control it in the 8th century also. Most important for Stein, his pilgrim, Hsuan-tsang, passed through the valley during his trip to and back from India. Stein mentions the dates of T'ang entries. The geographical descriptions of Chinese visitors of those centuries correspond accurately with Stein's own observations. Marco Polo also crossed part of the Wakhan enroute to China.

Section II - Historical sites in Wakhan

Once in the Wakhan Stein wished he could take the opportunity to turn west down the valley to visit Badakhshan but could not spare the time, nor was he invited to do so. He mentions three forts in the upper valley. But he did take half a day on 21 May to climb a ridge to investigate the fort called Kansir at Lien-yun that the Tibetans had built to defend the access route to the Darkot and Baroghil Passes. This fort was captured by Chinese general Kao Hsien-chih after a battle with Tibetans in 747 AD. Stein greatly admired the tactics used and even more the logistics required to move the Chinese army further south over those passes. General Kao divided his force into three converging columns and essentially bypassed the fort. Leaving the ponies at the base of a steep ridge, Stein and Naik Ram Singh climbed for an hour to the fort situated on a sharp spur. His diagram shows a single wall 400 feet long with three bastions across the relatively easy side of an oval space having steep cliffs at least 1,600 feet high on three sides. Parts of the wall still stood to a height of 30 feet. The wall was constructed of rough stones with a solid brick facing. Stein comments that the construction methods were in common use by both Chinese and Tibetans. But the orientation and location of the fort would serve the Tibetans defending the Baroghil rather than the Chinese coming the other way. Having satisfied his curiosity, Stein rejoined his party and rode up the Wakhan from Sarhad to the Little Pamir and the Wakhjur Pass, the watershed between the Oxus and the Tarim Rivers on 27 May. He describes the difficulties of this route in Ruins of Desert Cathay. While traversing the mountain valleys Stein investigated several ruins which he described in detail. The first of these Stein came to conclude was originally a Buddhist Vihara (shrine) and later turned into a Moslem tomb.

Section III - On Hsien-chih's Route to Kashgar

After crossing the Wakhjur Pass, Stein was at last in China. He was met by local officials, guides with yaks and supplies as he describes in Ancient Khotan, but still endured many difficulties as even in May at such elevation there was deep snow and freezing rain. He used the same caravan route down the Taghdumbash Pamir to Tash-kurgan the capital of Sarikol. (See both Desert Cathay and Ancient Khotan.) This time Stein visited two ancient ruins. One mentioned by Hsuan-tsang in an elaborate legend, fully narrated by Stein, was connected with a fortified hill top above vertical cliffs, called Kiz-kurghan (tower of the daughter or princess). The ruin was near the confluence of the Taghdumbash and Khunjerab Rivers at the eastern end of a rugged spur at an excellent location to defend the river gorge from 700 feet above the river. It was protected on three sides by vertical rock walls and on the fourth by a wall. Stein of course climbed to the top along with the two Ram Singhs. They were confronted by a massive bastion 25 feet square. They were able to get through a gap and along a 60 foot long wall into the fortress. One wall was 100 feet long, another over 190 feet long and another bastion 15 feet square. Some walls were still 20 feet high. Inside the walls he found two large cisterns. The walls were well built to a thickness of 16 feet of bricks 15x12x5 inches. The bricks obviously had been carried up the ridge from somewhere else. Stein was delighted to have personally confirmed the accurate detail of a description rendered by his 'patron saint', Hsuan-tsang. Stein noted that the ruin had survived at all due to the dryness of the climate. Stein resumed his ride through the defile of the Taghdumbash River to a hamlet called Pisling to camp. On 31 May he continued for 40 more miles to Tash-kurgan. He passed the remains of an old fort, 58 feet square with a moat outside the walls just below Pisling and another at Ak-tam some 60 yards square 5 miles before Tash-kurgan.

Stein left Tash-kurgan on 3 June toward Kashgar via the most direct caravan route across the spurs of giant Muztagh-ata, which he had climbed during his first expedition. He managed to reach Kashgar over 180 miles in 6 days rather than the standard 10 days of caravans. Naturally his motivation was largely to see the same route that Hsuan-tsang had followed in 642 AD. Stein confirmed the pilgrims descriptions of topography in every detail. On June 4th he passed through the Shindi gorge and entered the desolate, high Chhichiklik Maidan plateau, deep in snow, at an elevation of 14,800 feet. Toward the center of the plain Stein found a mound in which was a square stone enclosure of 35 yards each side. Despite indications that the place had been used as a Muhammadan tomb, Stein chose to confirm that it was the very same shelter that his pilgrim had mentioned. He comments that the Muhammadan take over of a former Buddhist shrine is typical throughout the Tarim Basin. Stein tells the story of Jesuit priest Benedict Goez who also crossed this plateau in 1603. He uses the Jesuit's report to describe the terrain further on, narrow gorges, raging rivers and all through which he passed. In the Tangi-tar gorge Stein recognized the very place at which Hsuan-tsang had met a band of robbers. As a plus his keen eyes spotted a series of holes carefully cut into the sheer rock walls, cuts he believed were made to support a bridge over the slippery boulders.

Section IV - About Kashgar and Yarkand

Stein arrived at Kashgar to a welcome by his friend Sir George Macartney, then the political representative of the Indian government. He had stayed with Macartney in 1900 and had made all arrangements for this expedition through Macartney's good offices. Stein spent the time assembling his team and supplies for the coming work in desert and mountains. Macartney recommended to Stein the services of Mr. Yin Ma Chiang (Chiang Szu-yeh) who became an indispensable secretary, translator, teacher, and friend throughout the next two years. Stein defers to his Ancient Khotan for his typical digression into the history of the place, in this case Kashgar. On this occasion he visited several ruins outside the city that had escaped him in 1900. On 21 June he rode north toward the Artush Valley and T'ien-shan mountains to visit a ruined stupa known as Khakanning-shahri (town of the great khan). This monument of sun-dried bricks and plaster was 32 feet high on a square base 32 feet per side. Nearby were two small forts. The walls were much decayed but there were remnants of towers as well. Further along the road Stein found holes cut high up into the rock wall in which a painted head of a Buddha was visible from below.

Stein jumps in the narrative to Yarkand, reached after a 5-day ride (June 23-27). He inserts historical information from M. Chavannes recent translation of more Chinese records. These indicate that Yarkand was not significant during the T'ang era but was very important as a kingdom during the Han Dynasty period. At one time it actually ruled the entire Tarim region. Various individual kings ruling Yarkand are mentioned along with their military and political exploits against the Huns. Stein provides details of dates and events. All together the period was one of continual conflict between the many 'kingdoms' (that is towns) around the Tarim Basin and the Huns, Yueh-chih, and Chinese Empire. Stein remained at Yarkand for 4 days without obtaining any information about its ancient predecessor. He then rode over a new route along the Tiznaf River to Karghalik. On 4 July he visited an ancient ruin at Kizil-jai but found nothing of interest.

Section V - Along the Westernmost K'un-lun

Stein continued through Karghalik oasis on 6 and 7 July toward the western end of the K'un-lun at Kok-yar. This was the main caravan route on south past the head waters of the Tiznaf and Yarkand Rivers to the Karakorum Pass. He remained at Kok-yar for 16 days (July 9 - 24) (see Desert Cathay) to avoid the summer heat while finishing editorial work on Ancient Khotan.

This example is amazing to me in two ways. One is that he could organize his total program to include taking manuscripts from a previous expedition along in carefully lightened baggage with planned expectation of working on it amid all the other activities. The second is that even while being, what one assumes to be, isolated in desert and mountains he is continually serviced by 'dak runners' that is the 'pony express' that managed to find him without having any fixed address. He regularly receives large bags of correspondence and send out the same to addressees everywhere between India and England.

Naturally he didn't stop other work. Ethnography and historical geography were always in mind even as he performed anthropometric studies, linguistic analysis, topographical surveys, archeological digs and dealt with daily administration. He reminds that he identified Karghalik as the Che-chu-chia named by Hsuan-tsang. The same town called Chu-chu-po or Chu-chu-pan in T'ang Annals. Since Kok-yar also figures in these medieval accounts, Stein is at pains to study the geography and assess the agricultural potential. He analyzes the many names found in Han and T'ang accounts to connect Karghalik also with the name Hsi-yeh.

Stein took time to photograph as many local inhabitants of the various mountain valleys as would agree to assemble. He was keen to create data for ethnic studies. While at Kok-yar he induced hillmen from Pakhpu to be studied. He wanted to add to the limited results achieved on the first tour (see Ancient Khotan) in which he connected these valley dwellers with the Alpine (Galcha) types in Sarikol and Wakhan. This he established to his own satisfaction. He also interrogated these Pakhpu mountain men about current and ancient legends. The summary of these he includes in this section. The informants also told him about jade mines confirming the origin of the stone considered precious by Han rulers.

Stein moved on to Khotan between July 25 and August 5 via a route through the outer foothills of the K'un-lun in order to accomplish new topographical surveys. On July 31 he reached Puski where he was directed to a local 'mound' that turned out to be another ruined stupa which he describes in detail.

Chapter IV - Remains of the Khotan Oasis

Section I - Old sites near the oasis

Stein returned to Khotan on 5 August after an absence of 5 years. His two visits during this second expedition were brief, in August and again in September. Interestingly he here completely omits discussion of his aborted effort during a month (last part of August and first days of September) to reach the headwaters of the Kara-kash River. He was pleased to see the rapid extension of the irrigation that was opening new areas for agriculture even though in some areas this would cover the wide debris fields known as 'Tati'. He rode south on 11 August in hopes of completing topographic surveys to the headwaters of the Kara-kash in the high K'un-lun south of Nissa and Karanghu-tagh. Sand Buried Ruins of Desert Cathay contains the account of this arduous effort in which the hill dwelling locals caused him so much difficulty. On the way he stopped briefly again at Yoktan, then continued to Aiding-kula and Naghara-khana. The latter, a mound, was now more exposed due to increased cultivation around it. He was determined to link the buildings formerly on it with a monastery mentioned by Hsuan-tsang. Yoktan continued to show some results of local efforts to search for gold under the ground but the increases in price of agricultural land and cost of labor was making such searching less profitable. He also visited Kohmari Mazar the sacred site from Buddhist accounts of Khotan as Mount Gosringa. Across the river opposite Faizabad Stein observed a grotto carved into the rock wall. Three miles beyond Faizabad Stein came upon a ruined, irregularly shaped, quadrilateral fort located on a narrow plateau. One wall was over 300 feet long and another 245 feet long. The 8 feet thick walls were constructed of sun-dried bricks measuring 18x12x6 inches. Near the apparent entrance the wall retained a height of 15 feet. Stein could find no datable evidence. The purpose for the fort, however, was clear - to guard the route from the mountains down the Kara-kash River valley.

Section II - Antiques acquired from Yoktan and At Khotan

In this section Stein describes all the antiques he collected at Khotan during both visits, 1906 and 1908. He devotes the full text section to description and analysis of some of the artifacts listed in the following section. These were mostly purchased from the merchants who bought them from 'treasure seekers' during the off seasons. He carefully distinguishes the provenance of each item by assigning a Yo to those he is sure came from Yoktan. The new collection, he remarks, greatly exceeds the previous supply. (Obvious from the next section.) Yoktan was occupied for centuries. The mud build structures decayed and new ones were built on top raising the occupied level. (This is well known throughout the world). Items in daily use were broken they were left and buried in this process. Wood and fabric and paper would decay in the damp soil, leaving mostly pottery shards to be found now. Stein remarks that some items, such as plates, jugs, amphora and other vessels, are surprisingly still more or less intact. He describes the decorations. Many of these are male or female heads among which he detects examples of Aryan (Alpine) representations. Many are artistically derived from Gandhara style. There are many terra-cotta animal figures of which monkeys predominate, plus camels, peacocks, yaks, boars and horses. There are models of the guitar, syrinx, drum, harp, flute and cymbal. Other objects are evidently imported. There are bronze, glass and gold items. Among obviously imported artifacts are busts or heads looking like Alexander, Persians, Romans, Indians and figures likely made in Sogdiana or the Sassanian Empire. Coins are described in more detail in Appendix B. In this section Stein limits discussion to chronology. The collection includes coins from the Sino-Kharosthi era to the Sung Dynasty.

Section III - List of antiques acquired from Yotkan and Khotan

This detailed 24-page list in small type testifies to Stein's acquisitive efforts even when visiting a place from which he already had obtained a huge collection.

Section IV - Desert sites in the North of Jiya

Returning to Khotan on 10 September, Stein was busy with long-range preparations. On 15 September he rode north through Jiya department. The next day he reached the previously excavated (1901) site at Rawak Stupa. (See Ancient Khotan). In the meantime the high sand dunes had shifted to the south-east burying some formerly open areas and opening others. However, the north-west wall was now covered more deeply than before. Unfortunately he also found deplorable evidence that Chinese 'treasure seekers' had destroyed much of the stucco figures he had so carefully preserved by burying it under sand. This is a clear example of the correctness of his efforts to protect precious ancient art by removing it completely to safety. The shift of sand had also uncovered much of the stupa itself. Even so, the volume of sand around the stupa made excavation impractical. Stein managed to uncover one corner to reveal the artistic decorations in the base story. With little more work possible at Rawak Stein moved on over 40-foot high dunes to the south-west to another ruined stupa 12 feet high and 24 feet square at its base. The sun-dried bricks measured 15x19x3 inches. There were other remains nearby - mud walls and typical pottery debris. Further south he came across dry irrigation canals, a water tank, embankment and yet another ruined stupa. Eventually they reached a house known to the guides. Stein identified it as a rectangular temple cella built of timber and plaster - 27 feet 3 inches by 24 feet 10 inches. This was enclosed by an outer wall that left an inside passage 7.5 feet wide. A few preserved sections of the walls showed colored decorations. Stein assigned a date between 4th and 7th century. He estimated that the site originally covered an area 8 miles north to south and well irrigated but now totally covered with sand dunes. On 17 September Stein moved on to Kine-tokmak, another 'old house' his guides had found for him. This was a brick structure some 34 by 40 feet with the remaining walls only 2 feet above ground level. This too Stein identified as a Buddhist temple. The stucco remains scattered about showed the same style as the Rawak Vihara. Next Stein moved on northward to other 'old houses'. These were simple dwellings constructed of wood and wattle. Only a few pottery shards were found. The party moved on to the south-east to a wide 'Tati' covered with pottery debris. The remains of fruit trees revealed it had been an orchard. Several 5th century coins turned up here.
Stein includes in this section a short 2-page list of the items found during this exploration

Section VI - The sites of Ak- terek

On September 18 Stein moved south to Ak-terek. He recognized the guides had brought him to the same place he had excavated in 1901 - The stupa of Arka-kuduk-Tim. All about were fragmentary remains from temple wall decorations, but no remaining structure appeared. Stein set some diggers to work who luckily hit the top of a brick wall at a depth of only 2 feet. Further clearing revealed many more small but well preserved fragments of wall decorations. Stein was especially pleased when experienced diggers showed that some decorations actually retained flecks of gold gilding such as none of the Yoktan relics had shown. The all camped some 2 miles away at Ak-kul farm. Returning the next morning with a larger team augmented from the farm, Stein could see how the sand dunes marched on propelled by the wind. With 60 men he was able rapidly to clear the temple wall about 53 feet long plus an adjoining passage 5.5 feet wide with a wall 3 feet thick that originally extended clear round the enclosure. Excavation continued on 22 September but was eventually stopped when the high sand dune continued to defeat the clearing. By then Stein had recovered a large quantity of relievos from within the central enclosure. The condition of many artifacts showed Stein that the structure had been burned prior to wind erosion and burial. The fire had 'fired' some of the clay into preserved terra-cotta fragments. Consideration of the several distinct levels being excavated showed Stein the chronological sequence of events that had produced layers of relievos, deposited loess, pottery fragments, more loess from the original floor up to the deep covering of sand. The tiny fragments found at Ak-terek could not have been understood as pieces of very large statues and relievos but for the find of such original contemporary figures still intact at Rawak. At Ak-terek Stein found very many body parts such as fingers, ears, noses, toes, and others indicating that the walls originally were lined with full sized Buddhas and Bodhisattvas. In between these there were halos, aureoles and elaborate decorative flowers and the like. On 19 September Stein found more evidence when his guides took him over several miles to another remains of a Buddhist temple 23 by 25 feet in size with an outer passage 6 feet nine inches wide, which they called Siyelik and the more decayed remains of two stupas nearby. Clearing parts of this structure again revealed fragments of relievos some identified as coming from the very same moulds as those at Ak-terek. At this location the discovery of human bones indicated that the former Buddhist site had been used later as a Muhammadan grave yard. All together Stein estimated the extent of the ancient site measured over 12 miles east to west under which undoubtedly there are still undiscovered ruins.

Section VII - List of antiques from Ak-terek and Siyelik

All together Stein's efforts at these two adjacent ruins resulted in detailed descriptions of artifacts filling 13 pages.

Chapter V - Ruined sites near Domoko

Section I - Shrines of Khadalik

In this initial section Stein, as usual, narrates what he did and when and how with general descriptions of what was found. On September 22 Stein departed from Khotan on his lengthy journey clear into China at Kan- chou. The first stop would be Khadalik. Since his first visit in 1900 Stein's local associates, in particular Badruddin Kahn, the (Ak-sakal) chief of the Afghan trader community, had been requesting all locals with interest in the desert to bring in any 'finds' they may uncover. Thus Stein through these intermediaries had already received a few Brahmi documents from a site known to be near Domoko. The finder, Mullah Khwaja had been brought to Khotan to act as guide. He was a village official of some standing and very reputable. He owed back taxes to the Keriya Ya-men so hoped to clear his accounts by selling such finds as his followers could uncover or to which they would lead him. Thus it came about that he found Khadalik and had already sold some documents to Indian merchants or at the Ya-men. Stein hired him as guide. On 23 September Stein continued on from Chira toward Khadalik, north of Domoko. On the way he noted the expanding cultivation around Gulakhma and Ponak visited in 1900. Next day with laborers hired at Malak-alagan he reached Khadalik. Again, pottery was strewn over the clay for half a mile. The site did not look promising as Stein presumed that locals would have already exploited it as Mullah Khwaja had. But digging quickly uncovered part of a frescoed wall plus paper documents in Brahmi script. Under 2.5 feet of sand the first important documents were recovered. Then more and more came to light. Fragments of painted relievos and panels appeared. Stein hired more workers from Domoko for the next day. Gradually they uncovered the remains of the building itself. It was a quadrangle some 73 feet long, north to south and 75 feet east to west. In the center was a cell 28 feet square and inside that was the platform 10.5 by 9.5 feet for a statue. Then the remains of intermediate walls forming concentric internal corridors appeared. The ruin was so large that excavation was not completed until late on September 26th. It became clear that much of the original wood of the walls had been removed for other uses. Stein records that out of 900 feet of wall only 4 feet remained. Mounds of wood and plaster showed that the takers had torn the wood pieces apart on the site before carting them off. Stein dated the destruction to the end of the 8th century due to coins found. Another ruin was a Buddhist shrine 24 by 25 feet inside a passage all around of 8 or 5 feet width uncovered nearby. The walls of this structure were brick but again the wood supports were gone. Again, the remains of the frescos covered the floor. Evidence of fragments indicated that some original figures were life-size. Stein recovered some of the moulds used in creating figures. Stein also found a very early packet of leaves from a Sanskrit Pothi written on birch bark from Northern India, plus wooden tablets on which was cursive writing in Brahmi script presumed to be of the old, unidentified language of Khotan. There were numerous votive offerings. Among these were Chinese coins some of which dated to T'ang Dynasty the most recent were dated 766-779 AD. Another building adjacent to the shrine was a hall 47 by 42 feet. He also found Tibetan documents. Stein continues with descriptions of more buildings found buried within the Khadalik area.

Section II - Antiques excavated at Khadalik

Stein organizes his material by reserving the more detailed analysis of what was found to a separate section, such as this one. Here he relates the types of material found at Khadalik to material found elsewhere around the Tarim Basin. But he cannot include the manuscripts pending much more scholarly study. He does mention that a few Buddhist texts are among them. He describes the artistic material. That found at Khadalik is similar to his finds at Dandan-oilik. (See Ancient Khotan)Stein comments that the use of Plaster of Paris moulds when creating decorations may indicate that the style is old, predating the construction of the shrine itself. At Khadalik he found plaster representations of Buddhas in various poses. Among the surviving wood scraps there are also painted Buddhas, Bodhisattvas and legendary figures. The fragmentary remains of carved wood work and painted frescos Stein describes with comments about how the destruction of the originals have deprived us of valuable examples of fine Khotan art.

Section III - List of objects excavated at Khadalik

The 29 pages of descriptions of each item indicates the value to Stein of his finding Mullah Khwaja as a guide to a site that despite being so near an inhabited location such as Domoko was nearly totally buried.

Section IV - Minor ruins near Khadalik

Mullah Khwaja knew much more about the Domoko region than only Khadalik. He took Stein next to a place called, Kighillik (the dung-heap) which precisely reflected its name and to another named Balawastge. The these also were not apparent from the surface. Mullah also led Stein to more sites scattered about Domoko. Stein describes all these, what he did to excavate them, what he recovered and what was its significance. The detailed list is only two pages long.

Section V - The Domoko-yar and Remains of Mazar-toghrak

On October 3rd Stein left Khadalik riding south to Domoko Oasis in search of another likely site reported by Mullah Khwaja. Stein also was eager to study the irrigation and cultivation in progress at Domoko since his visit in 1901. Stein draws favorable attention to the extensive scientific work of Professor Ellsworth Huntington on hydrology throughout Asia and Europe. He cites Huntington's book, Pulse of Asia, as the definitive account that includes the results of Huntington's lengthy study around Domoko in 1905. Stein devotes several pages to his observations of conditions there and in particular changes he notes since 1901. He was impressed by the engineering skill involved in the creation of a 200 yard long dam designed to raise the water level of the river so it could be diverted into irrigation canals. Throughout his expeditions Stein always made careful note of the way in which the desert rivers changed their channels and what the results were for local life. He also was convinced that for centuries a process of desiccation had taken place with drastic impact on the population. At the desert oasis tied to the rivers flowing from the K'un-lun there were two types of water - 'black water' was that coming out of springs which received their water from channels flowing underground from the mountains and 'white water' which was that created by the melting of snow and glaciers in the spring and early summer which then resulted in a summer flood.

South of the great dike lay the shrine Mazar-toghrak and 150 yards to the west was the location that Mullah Khwaja knew as a location for old relics. Stein began work there on 4 October. In the mix of animal bones and refuse fragments of wood tablets with Brahmi script appeared. Then came paper documents and fragments of various types of textiles. Eventually 50 wooden documents of various sizes and shapes were salvaged. The texts on some were in Chinese on others in the Iranian language written in Brahmi and some with both. Floors were uncovered at various levels one above the other. Stein estimated the documents dated from end of the T'ang era. Stein believed that the ruins on both sides of Domoko (one north and the other south) had been abandoned at the same approximate time, end of the 8th century. Stein discusses theories advanced as reasons the sites north and south of Domoko were abandoned about the same time as Dandan-oilik, some 50 or more miles further north into the desert on the same river system. He accepts the theory that lack of water could be a cause. But he points out that a cause could also be lack of manpower required to maintain the elaborate canal irrigation system necessary to make use of what water was available.

The two-page detailed list of the artifacts found around Mazar-Toghrak is included in the section

Chapter VI -The Niya Site

Section I - Return to the Ruins beyond the Niya river end

The discovery and excavation of the ruins north of Niya was perhaps the most spectacular and famous result of Stein's 1900-01 expedition. He wanted to return to complete work missed at that time. So he departed Domoko on 6 October and rode eastward. As whenever possible he chose a different route across the desert itself and past Achma oasis. Along the way he found more 'Tati's. He then spent until October 13th at Keriya, among other tasks buying seven large camels. Two days' ride then took him to Niya.
Stein left Domoko on 6 October headed eastward and passed the road to Keriya. He stopped for quick visit to Achima, a new oasis 6 miles further east on the edge of the Domoko cultivated area created by a new flow of water. He recorded visits to various debris sites in the desert. He had an administrative delay in Keriya until October 13. At Niya he met with his old helper crew from 1901 including Ibrahim Beg who had located the ruins in the desert. He took 50 men northward into the desert with supplies for 4 weeks on a 4-day march to the Niya ruin site. This ancient town was visited by HsuHsuanan-tsang on his journey to and from India. Stein digresses as usual into detailed discussion of the history and then a description of his findings. These included the remains of orchards and fruit trees and ;poplars. (plate 18 and 7 and fig 75). He found three tablets in the Prakrit language written in Kharosthi script. There were too many ruined dwellings for him to excavate in the available time, even with the large crew.

Section II - Northwestern Group of Ruins

On October 20 he started work with a divided party having sent Ram Singh to search for more locations in the desert to the north (which turned out a failure) . Stein excavated first ruins Nxiii (plate 7) the same building he had worked on in 1901. This time he found more Kharosthi documents complete with original seals. Illustrations 47-48 on page facing 217 show the wooden remains well - south room Nxiv and illustrations 49 to 52 show more remains of ruined buildings. Some records were of Chinese notes attached to missing presents between local royalty. From the names Stein connected this dynasty with Chinese annals. He guessed also the name Chu-mo was the same as Chinese Ching-chueh from the Han annals. On 22 October he was still clearing Nxx (plate 12 - fig 53). He provided more details of his work in Desert Cathay and Ancient Khotan.

Section III -Records of Hidden archives Nxxix

On 24 October he shifted his camp to another section where he soon made even greater finds. In a previously partially excavated room his assistant found a trove of written records hidden in a cache under the floor. These included contracts such as deeds and agreements still preserved with original seals. He had to wait for detailed translation back in London.

Section IV - Exploration of Nxxvi and of south eastern group of ruins

In this section Stein wrote more detailed descriptions of contents of various buildings. Among these was a piece of leather scale armor. He began work at Nxii which he had not examined in detail in 1901. There were a few small artifacts and one excellent piece of carved wood 5 feet long with floral designs of Gandhara style. He shifted next to Nxxv and Nxxvi. The last was a large home containing many rooms in one of which he found 8 Kharosthi tablets. He also found more Kharosthi tablets in other rooms. Stein returned to Nv to recover pieces of leather scale armor that he had failed to recognize in 1901. They were right where he had left them.

Section V - Exploration of southern most ruins and general observations on site

Stein excavated yet more buildings including Nxli (plate 18). On 30 October he moved back to the shrine at Iman Ja'far-sadiq, the small populated locale from which he recruited his workers. Stein concludes the section with his commentary and conjectures' about the significance of Niya and his finds. He notes, for instance the complete absence of any documents written on paper. He records the find of 10 Chinese copper coins dated AD 25 - 220. He noted that the administration of Niya was conducted in the ancient Indian language and script. He guesses that the ancient idea about the inhabitants having come from Taxila in NW India may be true. The documents may be in the ancient Indo-Scythian language. He guesses that Niya was abandoned in the middle 3rd century AD. He discusses various theories presented by other scholars and himself as to the cause of the abandonment. One idea he rejects is that this was due to gradual reduction in the water supply as the river dried. But he has shown that he place was abandoned hurriedly at one time in an emergency, not gradually. He connects this with the loss of Chinese control of the Tarim in 3rd century.

Section VI - List of antiques excavated at Niya

One has to wonder at the remarkable energy and meticulous effort Stein went to each day in recording exactly what, where and when each of the hundreds of articles were found. Many are shown in illustrations. This list fills 11 pages.

Chapter VII - Ancient Sites of Endere

There is excellent discussion of these topics in Chapters XXVI and XXVII of Desert Cathay.

Section I - Ruins of Bilal Konghan

After leaving the Niya ruin Stein stopped for a day at Iman Ja'far Sadiq Mazar. Then on November 1st he again started east directly across the sand to return to the Endere ruin excavated in 1901. The site contained a fort and stupa. He stopped again at Yar-tungaz Tarim to meet again with Abdul Kaim Akhun and collect more supplies. He continued on 4 November. Enroute he was told by a local at the Mazar, Sadak, who had brought his a Kharosthi document from a site in the desert. Stein was skeptical that such an old document would be from that location. Before even reaching the Endere Stein found a set of ruined buildings in the Bilal Konghan depression 5 miles to the west. They were within an oval fortification wall made of mud 16 feet wide at its base and 8 feet high on which was a platform and parapet. The gate was 11 feet wide and still retained its door beams. The outer side of the rampart was stamped mud all looking rather primitive. But Stein concluded the ruin was not all that ancient. His conclusion was based on absence of significant wind erosion which always was a characteristic of truely ancient locations subject to centuries of attack by wind-driven sand and gravel. Nevertheless he set his large crew of workers to excavating two buildings. The sand had protected much of the structures. The workmanship revealed at the walls were uncovered was not up to the standards shown elsewhere. Most annoying Stein didn't find any of the refuse dumps that contained much of his 'finds'. He decided the whole settlement was a relatively recent Muhammaden colony. In connection with this estimate he discusses the manner in which the Endere River (like others) has shifted its course over the centuries back and forth from east to west as silt piles up in one course causing the water to shift around it. But through the centuries during which Niya and Charchan were thriving an intermediate station on the Endere River was important. The Muhammaden effort was of a like causation, but in this case it failed rather quickly, the evidence of wide spread fire brings to mind another possible cause. Another is that the site may have been the location for a prison camp.

From local settlers at nearby Endere Tarim he learned that the difficulty in maintaining an agricultural settlement stems not only from the manner in which the river shifts its course, requireing a shift in the fields and homes, but also from the way in which the large spring flood can wreck the irrigation canals necessitating more repair labur than may be available.

Section II - Excavations around and within T'ang fort of Endere

On November 8th Stein reached the stupa land mark of the Endere site he found in 1901. His rushed visit then prevented him from doing extensive excavation and survey work. The same grafiti on the wall around the stupa remained - as Chiang Szu-yeh confirmed with date 719 AD. He quickly went to the mound identified by Sadak as the location at which he had found the Kharosthi script. This proved to be another lucrative refuse mound. It immediately disgorged another Kharosthi text prompting Stein to reward Sadak on the spot. The mound consisted of a small brick structure at the bottom full of debris and then a layer of stable refuse above it, indicating two separate periods of usage. Stein concluded that the fort was originally occupied by Chinese in the early 8th century and then abandoned when the Tibetans controled the area.

He here refers to his report from 1901 in which Hsuan-tsang's report that when he passed this area in 645 AD he found an already ruined ancient scene was cited as evidence of the antiquity of the ruin. He was pleased to find this Kharosthi evidence that Hsuan-tsang was correct, the site had been occupied centuries prior to 645 (and abandoned in the 3rd century) and occupied again shortly after Hsuan's visit. It became an important Chinese, T'ang, fort in the second half of the 7th and the 8th centuries and subsequently also a Tibetan outpost. Further exploration revealed more buildings including another stupa and a large tower. There was extensive pottery fragments and some coins on the clay ground. Digging into the rampart that dated from late 7th century Stein discovered that it also had been built right over the refuse of the much earlier Tu-huo-lo settlement. And right in this refuse layer Stein pulled out a leather document with Kharosthi script. Further clearing revealed that the bank of refuse was 6 feet thick (high). Stein was at pains to lean that in the fragments of textiles from this refuse there were none containing cotton. He now had ample labor available to clear the large buildings inside the fort walls. In this he found excellent woodwork. And the wood was from the cultivated Eleagnus tree rather than the wild Toghrak. Parts of the brick walls retained their thick plaster cover.

Section III - Survey of earlier remains at Endere

While clearing of the fort continued, Stein expanded his exploration around it. He found clay mounds buried in sand and concluded these also dated from the early Tu-huo-lo era. He also found to the north another enclosure measuring 540 feet south to north and 340 feet across. As so often was the case the insesant north-east wind had almost destroyed these walls but some bases remained 30 -35 feet thick Some parts of a rampart with wall of sun-dried bricks set on clay remained. Although more ruined by its longer exposure to the wind, this fort had originally been built of stronger walls. He notes that the results of continual erosion may leave coins, glass, pottery and other debris from very different era lying next each other on the remaining hard surface of 'tatis'.

Stein also uncovered another small but massive fort south of the main site with solid walls of stamped clay 8 feet thick and 18 feet high in places. The entrance gate was protected by a bastion. Inside was a court 48 feet square. Stein concluded that this fort also was constructed well before the T'ang era. The wall was considerably overtopped by a 38 feet high Tamarisk cone that must have grown up after the wall was built. While leaving Endere and heading south Stein found more isolated ruins of buildings in the desert. He made his last archeological effort around Endere on 13 November.


Section IV - General Observations of Endere and old Tu-huo country

In this section Stein focuses on the main facts he can determine from suveys at Endere and to the east. He believes the area was occupied during two distinctly separated eras. The earlier era is reveiled by the mound of refuse under the T'ang fort walls and by the smaller fort to the south. The large stupa and walls near it likely date the same. This period coincides with that at Niya - that is it was abandoned at end of 3rd or early in 4th centuries AD. To the later era belongs the T'ang fort and temple with the inscription dated 718 AD. He declines to state conclusively why either settlement was abandoned. He notes the general drying of the climate and reduction in river flow from the K'un-lun. But that is not the only possible cause for abandoment of agriculture. He once again pleased that archeology has proved his pilgrim, Hsuan-tsang so accurate in his descriptions. But he denies that Hsuan-tsang's mention of the name, Tu-huo-lo proves (as other commentators believe) that the Tochari ruled this region before appearing in Bactria.
He refers again to Chinese historical records and Sung Yun's report that the Hephtahalites controled a huge area from the middle of the 5th century until they were dislodged by the Western Turks a century later. Their main base was "Tokharistan' on the Oxus. Stein dates the first occupation of Endere to that of the Niya and Lou-lan sites which would connect Endere with the Tukhara. He wants to connect this with Hsuan-tsang's mention of the name Tu-huo-lo (Hephtalalite) control given to him by his guides for the ruin he passed.

Section V - List of antiques found or excavated at Endere

This list is only 4 pages long.

Chapter VIII - From Charchan to Charkhlik

Section I - Early accounts of Charchan

Stein rode for six days from the Endere River to reach Charchan on 20 November. He likes to point out that this was the same travel time as his pilgrim, Hsuan-tsang. He comments that the area likely had not changed much since the previous journey as it was the direct line between the gravel plain north of the K'un-lun and the vegetation belt next to the desert. He continues, that another account of this route between Khotan and Sha-chou is in the T'ang Annals, which also mentions Chu-mo (Charchan) and Yu-t'ien (Endere fort). Marco Polo's description is more detailed than that of Hsuan-tsang. Stein remained at Charchan, which he found was now a growing oasis, for two days. The town was favored by increasing water from the Charchan River. The volume of glacial runoff from the K'un-lun makes this river the only one east of the Khotan River that succeeds in flowing all the way east to the Tarim River in the joint delta marsh. However, Charchan is separated further from the next nearest settlement in any direction than any other oasis in the basin, which inhibits an influx of people despite the ample land suitable for irrigated agriculture.

The earliest record of Charchan that Stein has read is in the Former Han Annals in which it is called Chu-mo, a small kingdom boasting 320 soldiers. The same record mentions many other of the towns and villages all around Chu-mo. Chu-mo is again described in the Later Han Annals and in the Wei'lio. Li Tao-yuan, who died in 527 AD, commented that Chu-mo received water from the Charchan River. Stein continues to cite medieval accounts by referencing that of Sung Yun in 519 AD. He notes the absence of mention not only of Charchan but of travel on the southern route during the later middle ages and right up to the 19th century when Chinese expansion began again.

Section II - Ancient Remains around Charchan

Stein found only traces of the ancient Chu-mo such as a canal and ruined walls. And there were fragments of pottery and bronze all about. On 22 November Stein visited other potential sites showed to him by locals. Again, all he found was pottery and other small artifacts strewn over the hard clay.

Section III - The Charchan River route to Vash-shahri

Stein departed Charchan on 23 November in his hurry to reach Charkhlik. This leg of the journey began with 5 days travel along the right (south) bank of the Charchan River. Along the way he met a local hunter, Ismail, who guided him to some nearby ruins. On 25 November Ismail showed Stein a ruin on the north bank, which Stein believed to be a stupa made of large bricks. Ismail led Stein to two ruins on 26 November one consisting of three brick buildings and the other of over a dozen similar structures. Stein found enough evidence to show these were Muhammadan cemeteries. On 29 November Stein turned south-east to an oasis known as Vash-shahri near a medieval site which he surveyed and explored. Over a broad area he found pottery debris which later analysis dated to the Sung Dynasty. He also found bits of bronze, coins, glass, buckles, arrow-heads, hair-pins, and stone wear. He collected eight copper coins, three of which date starting from 681-627 AD, and four from the Sung 1023-1101 AD. There also were the ruins of a few buildings. Stein quotes from Marco Polo about the location of such a place. Stein comments on the recent efforts to expand cultivation at the new Vash-shahri which he also visited. Like Charchan there was plenty of good soil and water, but the isolated location was not enticing to settlers.

Section IV - The Oasis of Charkhlik and its old remains

After two more days and 51 miles, Stein reached Charkhlik on 2 December. He planned to base his further operations including the trip to Tun-huang from this relatively large oasis. These operations included lengthy explorations of the distant Lou-lan, Miran and the terrain all around the western and southern edge of the Lop-nor salt sea. He hoped to hire 50 laborers and many camels plus the food supply to last the whole party for 5 weeks and for his own team for another month. All the work and travel to and from would have to be completed by March in order to enable the following move to Tun-huang before the change of weather would prevent it. At this small oasis he could find few camels, enough to make a full complement of 21 counting those he already had. He planed to create an intermediate depot at the small fishing village at Abdal on the Terim River. That he was able to accomplish anything despite the strong reluctance of the locals to brave the desert was due to diligent assistance of the local Amban, Liao Ts-lao-yeh. The arrival of two experienced desert hunters from Abdal, Mullah and Tokhta Akhun, who ranged all over the Lop-nor area and had actually been with Dr. Hedin in 1900 was a great help as well in raising the morale of the Charkhlik men. But they had not been to Lou-lan over the direct route from Abdal.

Stein found some time while at Charkhlik to also investigate what ancient remains there might be. His study of historical records and of the geography had already convinced him that Charkhlik was the principal seat of local government and most important oasis in the south-east corner of the Tarim. This had to be the well-known Lop. He notes that Prejevalsky was the first European to visit Charkhlik in modern times (1876) which at that time had only a few free families from Khotan plus a Chinese convict labor colony. But he also reported on finding extensive remains of a medieval wall. Stein found 300 families in residence. The location was valuable from a business view point as well as governmental and military strategy. Was at the convenient cross roads of the east -west caravan route from Tun-huang to Khotan and the south - north migration route for Mongols and Buddhists in general between Tibet and the north. The relatively extensive agricultural area could support supplies for passing caravans. (A factor that Stein counted on.) Sure enough, Stein quickly located the existing remanent of the medieval town wall, locally known as Sipil. The remains of the mud wall extended for more than half a mile, north to south while the width was about a third of a mile. The remaining ramparts in places still rose to a height of 20 feet. It was now in the midst of cultivated fields but Stein found that ancient clay bricks were in current use in people's homes. In the center of the interior field Stein found the remanent of a Buddhist stupa. Outside this former enclosure Stein also found a larger mound some 50 feet high. This also was formerly a stupa situated on top of an accumulated pile of debris far more ancient.

Section V - List of Objects found or excavated at Charchan and Vasli-shahu

This two and a half page listing indicates the relative paucity of the artifacts Stein obtained in these two locations.

Chapter IX - Historical Notes of Lop, Shan-shan and Lou-lan

Section I - Marco Polo's Lop and Hsuan-tsang's Na-Pu-Po

It is important to recognize again in this chapter that Stein is attempting to marshal evidence to support his own arguments during a period in which explorers and European scholars were debating the location of various ancient places and their current equivalents. He wants to tie together analysis of the terms use in Chinese and Tibetan document spanning centuries with his own archeological analysis of excavated sites and comparisons of ancient geographical descriptions with his own observations of terrain and topography.

Once again Stein expounds on the written sources to discuss the history of an area. His favorite two authors, as we have seen repeatedly already, were Hsuan-tsang (his patron saint) and Marco Polo, whose fabulous account Stein wants to prove accurate from comparison with the geographical reality. The chapter title indicates the specific region to be discussed. He begins with Charkhlik the largest (although still very small) oasis in the area. He states that it 'must have been' the main place in the ancient and middle eras as it is now. This is Marco Polo's "Lop". Marco Polo also describes the caravan route east to "Sachiu" - Sha-chou or Tun-huang. Marco Polo's 'province of Lop' must include the entire area around the terminal basin of the Tarim River, as it means today. Stein allows that Lop could potentially refer to Vash-shahri, Charkhlik or Miran since these are the only oases with sufficient agricultural potential to support a significant population. And, indeed, all three do have ancient ruins. But of these only Charkhlik could have provided the kind of support that Marco Polo indicates. Vash-shahri is not the 'last' oasis on the route. But Stein claims that his excavation of Miran (to be described in Chapter XII) proves that it was abandoned centuries prior to Marco Polo's journey. Stein also points out that his examination of the remains of the Charkhlik's wall (decrepit fortification) also proves his point. It was still functioning when Marco Polo passed but was abandoned in the 14th century, after the decline of Mongol-Chinese control. By the time of Shah Rukh's embassy the southern caravan route was not used. Still later accounts mention that there were only ruins in this region, plus wild camels. (Note that today there is a large wild camel reserve nearby to the east). Stein notes that when he visited the area Lop was the designation of a Loplik fishing village 36 miles north of Charkhlik. Even though he found no specific reference to an oasis near Charkhlik in historical records between the end of the T'ang Dynasty and Marco Polo's visit, Stein notes that there are numerous accounts of embassies from Khotan to the Imperial Court during these centuries which must have made use of a oasis there. Stein refers to T'ang Dynasty accounts that describe the route between Sha-chou and Khotan. This account also mentions Shah-shan with Stein identifies with Lou-lan. Stein turns to Hsuan-tsang's account of his return trip from India in 644 AD on which he passed from Khotan eastward through Chu-mo ( Charchan) to Na-fu-po (Lou-lan). Stein again points to his own discoveries of Tibetan records at Miran, to be described in Chapter XII. These refer to the era in which the Tibetans exercised control over the Tarim, from the last of the 8th century to the last half of the 9th century. The Tibetans refer to a place called Nob. Stein believes that 'Great Nob' refers to Charkhlik and "little Nob' refers to Miran

Section II - Shan-shan between T'ang and Han Times

Stein backtracks chronologically to take up the issue of Charkhlik and the Lop region generally during Han and T'ang eras. He notes that the Lop-nor area was important for travelers even during periods in which Chinese power was absent. (That is between Han and T'ang Dynasties and subsequent to the T'angs. ) He mentions the report of P'ei Chu in 607 AD that a caravan route existed between Shan-shan south of Lop-nor and Khotan. Then there is the story of Sung Yun who was at Shan-shan in 519 AD having come from the south at Koko-nor. Next (going backwards) Stein turns to Fa-hsien, who with 4 monks in 400 AD journeyed from Tun-huang west across the desert to Shen-shen. For the history of 77 BC Li Tao-yuan mentions that the Han emperor replaced the king of Lou-lan and renamed the area Shan-shan and established a Chinese garrison at a town, I-hsin. Li - Tao-yuan provides much more description of the whole geographic region around Lou-lan - Shan-shan. Stein states that he has found no specific references to Shan-shan in the Chin Dynasty records (265-419 AD). But the Wei lio written between 239 and 265 AD (Three Kingdoms era) does describe the 'three caravan routes' from Tun-huang west. The southern route passes through Chu-mo (Charchan), Hsiao-yuan, Ching-chueh (Niya) , Lou-lan; all dependencies of Shan-shan. and then through kingdoms as far as Khotan. Stein insists that this Shan-shan must coincide with modern Lop area with capital at Charkhlik.

Section III - Shan-shan in Later Han Annals

Again moving backwards in time Stein seeks more information in the Han Annals (AD 25 - 220). Although his expedition visited both Charkhlik and Lou-lan prior to his discovery of the Han wall and its large cache of records near Tun-huang, he is writing this report well after returning from the trip with all the subsequent information at hand. He frequently throughout this report to the official account rendered in AD 125 by general Pan Yung after the Chinese conquest of the Tarim Basin around AD 73. These records mention Shan-shan and locate it on the eastern border of the Basin and west of the salt sea. Two caravan routes are mentioned: the southern one along the foothills of the K'un-lun to Yarkand, along which Yu'men, Shan-shan, Chu-mo (Charchan), Ching-chueh (Niya), Chu-mi (Chira-Keriya), and Yu-t'ien (Khotan) are listed. The northern route was along the North Mountains (T'ien-shan) to Kashgar. Stein notes that the Later Han records do not provide topographic details about Shan-shan. But they do report on political conditions circa 58-75 AD; namely that the kings of Khotan and Shan-shan controlled the entire southern region between them. Shan-shan is also mentioned for 45-46 AD at which time it was conquered by the king of Yarkand. In 73 AD the Chinese began to regain control of the Tarim by defeating the Hsiung-nu -Huns, establishing suzerainty over Shan-shan and going on to gain control over both Khotan and Hami clear to Kashgar. In 94 AD Pan Ch'ao with contingents from seven local kingdoms including Shan-shan gained temporary control over Kara-shahr, which controls the pass through the depression in the T'ien-shan. In 105 AD the 'Western Regions' revolted followed by Hun expansion southward. In 119 AD another Chinese general, So Pan, was sent to Hami, which caused the king of Shan-shan to submit. Shortly after this the Huns destroyed So Pan and his army forcing the king of Shan-shan to seek support from Tun-huang. Then Pan Yung, son of Pan Ch'ao prepared a strategic plan for the empire. The Huns continued to raid as far as Kan-su. At this point in 123 AD Pan Yung was appointed to lead the Chinese detachment of 500 men to establish a military colony at Lou-lan. The key strategic location and purpose for a garrison at Lou-lan was well described by Pan Yung in the Annals. It would provide support to the south for Shan-shan and the full route to Khotan, to the west it would control the roads to Kara-shahr and Kucha, to the north it would suppress the Hsiung-nu, to the east it would support Tun-huang. Pan Yung at Lou-lan did obtain the submission and cooperation of the kings of Shan-shan, Kucha, Ak-su, and Uch-Turfan. With these allies Pan Yung defeated the Hsiung-nu at Turfan and established a military colony at Likchun. In 125 Pan Yung, again with allies from Shan-shan conducted an offensive across the Kuruk-tagh against Hun allies and defeated them at Guchen. Kara-shahr was again conquered in AD 127. But after AD 132 - 134 Han power decayed throughout the Tarim and these local kings turned to warfare on each other.
Stein then writes that he plans to provide in the next chapter the archeological proof that Lou-lan was indeed established as a fortified garrison.

Section IV - Earliest Records of Lou-lan Under former Han

(It seems to me that Stein's reversal of chronological order makes the whole effort of historical proof more difficult.)
But that is what he proceeds to do in this section. For this era he relies on Wyle's translation of the "Notes on the Western Regions" from the Former Han Annals. In this document there are many references to the kingdom of Shan-shan for which the original name was Lou-lan. This is due to the great strategic and commercial importance for the Chinese this Lou-lan region had as the gateway to the West. Its capital was at Yu-ni reckoned at 1,600 li from the Yang barrier. The kingdom is described as fielding 2,912 soldiers governed by various named officials. Yu-ni was near the southern shore of the Lop-nor marsh (which Stein believes may be Miran). Stein notes that the relatively large population of Shan-shan given in the Annals indicates that they must have relied greatly on animal husbandry since the agricultural produce could not support this number.
Stein refers to the events subsequent to Chang-Ch'ien's embassy to Sogdiana and his return which determined Emperor Wu-ti to organize an offensive against the Hsiung-nu circa 121 BC. In support of this he extended the wall north west of Tun-huang. But local powers including Lou-lan sided with the Hsiung-nu in attacking and harassing Chinese envoys and expeditions across the Tarim to Ferghana. Thus general Chao P'o-nu was sent against them. In 108 BC he captured the king of Lou-lan conquered Ku-shih and dominated the adjacent kingdoms. Stein notes general agreement that Ku-shih means Turfan. Therefor Lou-lan must lie on the route to Turfan which in turn means in the Lop-nor region. Stein again points out that Khara-shahr lies in the key depression that affords easy passage from the extensive grazing areas north of the T'ien-shan across those mountains to Turfan in the south.

Section V - Lou-lan reestablished as Shan-shan

Stein continues with the Han Annals. Chao P'o-nu's campaign in 108 AD resulted in control over Lou-lan. But then the Hsiung-nu attacked the kingdom to which the king responded by sending one son each to as hostage to the Chinese and Hsiung-nu. At that point the Chinese had yet to extend their wall westward from Su-chou as Stein claims his exploration has shown. When the Chinese general again campaigned west to Ta-yuan the Hsiung-nu laid an ambush at Lou-lan against his return trip. Then the Chinese sent another expedition that captured the king of Lou-lan and forced final submission. Stein's point in all this discussion is to establish that Lou-lan corresponds to Shan-shan and Lop.

Stein continues with the Han Annals. He discusses Chao P'o-nu's expedition in AD 108 during which the general obtained the submision of Lou-lan. But the Hsiung-nu retaliated by attacking Lou-lan. Later Emperor Wu-ti sent another expedition to Farghana and again the Hsiung-nu attempted interference by sending cavalry to Lou-lan to block his return. The historical record shows that this Chinese expedition failed with great losses. At this the Emperor tasked the commander at Yu'men kuan (Jade Gate) to capture the king of Lou-lan. The result of this was that Lou-lan again submitted and became subsidiary to China. All this, Stein claims, shows that Shan-shan (also called Lop) coincides with the territory of Lou-lan. He continues with Chinese campaigns in 101, 99, 89 BC. He states that he has proven the location of the direct route between Tun-huang and Lou-lan during his exploration of 1914 but this information has not been published. (We can refer to the remarkable account in Innermostasia.) The identification hinges on the correct understanding of what the Annals refer to as "White Dragon Mounds". Stein continues with extensive discussion of the viscidities that beset the Chinese and the several claimants to the Lou-lan throne, which he feels necessary to support his contention about Lou-lan's location and that of the central supply route. He continues to marshal evidence for his contention. He repeats that the capital of Lop was in the Charkhlik - Miran region. He writes that the name of the kingdom was changed from Lou-lan to Shan-shan in 77 BC along with the creation of the Chinese garrison on the north edge of Shan-shan. Stein disagrees with the contention of other authorities that the capital of Shan-shan was moved from south to north of the terminal basin of the Tarim River. Stein insists that the capital of the subordinate kingdom of Shan-shan remained at Charkhlik but it was the new Chinese fortified garrison town called Lou-lan that was build north of the Tarim on the northwest edge of the salt sea. Actually the Chinese wanted to distinguish between the two. From the Chinese strategic view point the most critical area was north of the Lop-nor salt sea along the foothills of the Kuruk-tagh where their main route west to Turfan and along the southern foothills of the T'ien-shan was the most vulnerable to Hun attacks. But the descriptions of the capital town of the Shan-shan (formerly Lou-lan) kingdom indicate it must be south of the Tarim and Lop-nor salt sea as only the oasis at Charkhlik and to a lesser extend Miran could support such a population.

In summary Stein writes - the names Lou-lan - Shan-shan refer to the entire region - Lop - extending between the Kuruk-tagh on the north and the Altin-tagh on the south, and with the broad area into which the Tarim, Charchan and Konche Rivers flowed into their terminal marsh and ultimately into the salt sea and desert. The capital was Yu-ni at Miran and the Chinese colony, I-hsun, at Charkhlik. The region was connected to China proper through Tun-huang by two caravan routes; one across the foothills of the Altin-tagh and the other through the desert on the recently revived Tun-huang-Charkhlik route as far as Kum-kudak. From there one branch veered southwest to Yu-ni - Miran and I-hsin (Charkhlik) and the other veered west-north-west passing the eastern end of the salt march and to the terminal of the Konche-darya which is now dry. It was at that point that the fortified garrison at Lou-lan was established. From that point it continued further north-west to the north edge of the Tarim Basin (foothills of the T'ien-shan. This 'center' route was the more important for commercial travel and also the main route for diplomatic and military expeditions even though the center of local economy and political life was at Charkhlik.

Chapter X - Through the Lop Desert

Section I - First visit to Miran

Stein spent three days in Charkhlik organizing supplies, transport and men for his continued journey clear to Tun-huang. But he had two major diversions to accomplish before completing that phase - namely to explore Miran and Lou-lan. On December 6th he set out with 50 hired laborers and 21 camels, plus his Indian and local team. He packed supplies for 5 weeks and ice. The first objective was Lou-lan, some 100 miles to the north-east of Abdal and 70 miles from the nearest potential for water. The critical variable was to take a large enough team of workers to do the work in time but manpower that could be supplied with water (ice). But the water supply was dependent on the number of camels and the hoped for freezing of the spring at Altmish-bulak. The plan was further complicated by Stein's knowledge from Prejevalsky's expedition in 1876 that there was a ruin at Miran. And Tokhta Akhun brought Stein a bit of Tibetan writing from Miran. Stein reached Miran after a 2-days ride where he camped near a stream and jungle. On December 8th Stein made reconnaissance of Miran ruin which revealed the remains of a stupa and other ruins scattered over a flat clay surface (fig 111) plus a ruined fort ( plate 30) in the distance. Stein began with excavation near the east wall. Under the sand they found a series of small apartments made of brick with Toghrak posts supporting a roof. The workers immediately found deep piles of refuse in which over 130 Tibetan documents were recovered. Stein also recognized many pieces of leather armor - described in a later chapter. He also made a quick dig at the stupa which turned up stuccos Buddhist artifacts. Clearly the stupa and its adjacent buildings was much older than the fort - what to do? Stein was overcome by a surfeit of wealth in unearthed artifacts. He realized that a thorough, professional archeological excavation at Miran would require many days, days not available if he wanted to spend time at Lou-lan. He decided to rebury everything pending a return after completing work at Lou-lan. Nevertheless, when he returned he found that members of another caravan had attempted some 'treasure seeking' and damaged things. With Abdal very close he could establish a base there to serve all this effort and also support the following move to Tun-huang.

Section II - Past the Terminal Lagoons of the Tarim

On 10 December Stein left Miran for Lou-lan. With R. S. Ram Singh suffering from rheumatic fever he had to ride a camel rather than walk once the ponies were left behind. This required significant reduction on the other loads. Stein had to reduce the work force to 35 laborers plus his permanent team of 15. He dismissed the least efficient of the workers and sent them home to Charkhlik. As usual Stein describes the flora and topography of the route mile by mile to the Tarim river in great detail. At the river he found the tiny colony of Loplik fishermen. Even though the place was small, he established a depot there in order to lighten the supplies taken on to Lou-lan with Chiang Ssu-yeh in charge. His secretary was not physically up to a journey on foot through a desolate desert. Stein built a raft to form a ferry to shift the camels and supplies north, across the 40 yard-wide Terim River. From there the march to Lou-lan required 7 days. See discussion in Desert Cathay. But, Stein notes, his subsequent exploration in 1914 produced so much new information that full analysis should await that report - see Innermostasia. On 11 December they traveled along the Terim to Ak-kol where there were reed huts of a few Loplik fishermen. There they cut ice into slabs that the camels could carry in wool bags. Nine camels were then sent on along with 30 donkeys also carrying ice to establish an intermediate depot along the route. Then the main party continued on. They had to thread their way among the marshes and lagoons created by the summer floods. On 14 December Stein sent the ponies back to Abdal. The remaining trek to the north-east would be on foot. At another lagoon Stein left two laborers with rations to be forwarded when the donkeys returned from the main depot. Stein benefited from having Dr. Hedin's sketch map from 1900. Both Mullah and Tokhta Akhun had been to Lou-lan with Hedin, but from the north rather than south. That day the terrain became more and more difficult for the animals, reducing the achieved distance to 16 miles. They were crossing an area of flat surfaces cut by the wind into deep ravines all in the direction east-north-east to west-south-west..

Section III - Across an Eroded ancient Delta

On December 15th Stein created another depot where he deposited the ice carried by the 30 donkeys. Once unloaded the donkeys were sent back with two men to the previous depot where they would reload the food, ice, and other supplies and bring it up to the new depot. Meanwhile once the party reached Lou-lan some camels would return to this depot and then bring the ice and supplies to the camp. In his careful planning Stein shows excellent understanding of military logistics methods. Along these routes he carefully built cairns to mark the way for these resupply columns. They continued on by compass reckoning over difficult terrain full of ravines, decayed jungles of ancient trees and hard plateaus. Stein found that the beds of the ancient river showed that at some point there had been enough water reaching this region to support inhabitants. This was confirmed soon when they began to find Neolithic implements and artifacts lying right on the hard clay surfaces.

Section IV - List of objects found in Desert Marches north of Lop Nor

This five-page list records artifacts picked of the desert surface during the move to Lou-lan. Stein notes that in the rush to reach that site the men had to stick to a narrow path which meant they could not search widely even though they knew they were passing through an area full of such remains. Most of the items described are Neolithic or Paleolithic in origin.

Chapter XI - The Lou-lan Site

Section I - Excavation of Ruined Dwelling LAi

For Stein, Lou-lan - and the region of Lop-nor - was a central objective of his entire expedition. He had read the reports by Dr. Sven Hedin of his 'accidental' discovery of the ruin in March 1901. One can see the sense of rivalry that Stein shows in his rush to get to Lou-lan and to Tung-huang before any other European explorer can mess up the pristine archeological site. He pointedly notes that Dr. Hedin only spent 3 days in work on the eastern part and one day on the western part of Lou-lan. And he only had 5 additional, untrained workers. Moreover, much of his results has not yet been made available.

On December 18th he sent the camels northwest with Tokhta Akhun to a salt spring known as Hangi-bulak, near Altmish-bulak. Five other camels returned to the half-way depot to bring the supply left there to Lou-lan plus fresh ice by then delivered by donkey caravan. Rai Ram Singh took some other camels to conduct a short survey in search of ruins Sven Hedin had reported from his journey in 1889-1901. Stein first made a survey of the various structures to determine which should be excavated first. There were various timbers sticking up out of sand banks all around. The terrain was quite different from that at Niya. There the buildings were in deep sand dunes. But at Lou-lan the ground was bare, hard clay with sharply eroded ravines leaving mesas (Yardangs) in between. Excavation and recovery of artifacts were by now standard procedures not only for Stein but also for his lead assistants, Naik Ram Singh and Ibrahim Beg. Work began with site LAi near the stupa. This building was 60 yards south-south-east from the stupa on the top of a terrace. It appeared to be a well-built house originally much larger. Entire rooms had fallen away down eroded slopes. The construction methods resembled those at Niya and Khadalik. The foundation consisted of large beams laid in the clay. In these there were sockets cut into which the square vertical posts for walls and to support the roof were set. There were bundles of reeds laid horizontally tied to the vertical posts. These were then covered on the outside by mud plaster. The structures were oriented so that one wall faced the wind from the east-north-east. The largest room in this building was 31 by 13 feet. On the floor under a few feet of sand Stein recovered the first documents - in Chinese. One of these was dated to AD 330. Next Stein found more wooden documents with Kharosthi script. Some of these were in the familiar wedge shape with a covering and under pieces similar to those found at Niya. Paper documents with Prakrit language in Kharosthi script were found on the surface. Stein noted that this indicated that the Chinese officials continued to use local agents. The building also contained two fragments of woolen carpet with the colors still brilliant. There were other items including imported fine Chinese silk. Stein established from the dimensions of a roll of yellow silk that the standard width for exported silk rolls was 18.75 inches (50 centimeters). At this point Stein looks ahead and states that he found at the Han wall near Tun-huang two wooden measuring rods with marks showing an 'inch' of 22.9 millimeters and the length of 22 Chinese 'inches' at 19.83 inches. There were also various household items including glazed pottery in these rooms. Coins of the Emperor Kuang Wu-ti (AD 26-57) were recovered as well.

Section II - Exploration of Ruined Dwellings LAii - vi

Stein continues with excavation of each building in sequence moving to the south-west. First of these was LAii and then LAiii. This building was among those searched by Dr. Hedin, who reported its complete and thorough recovery. Stein makes a point that nevertheless he was able to find more valuable documents overlooked by Hedin. Here Stein found some documents dated to 265-74 AD. He established from the numerous official documents uncovered at LAii and nearby LAiii that it was the actual administrative headquarters for the station. Even his local laborers recognized that this was the Ya-men of the Kone-shahr. LAii, LA iii, LAv and LAvi were saved from worse erosion by a thick wall. Stein describes each of the rooms in detail, their physical appearance and construction, he archeological methods in searching, and the results obtained over many pages of text. He moved on to LAiv on December 19th.

Section III - Discoveries in an Ancient Refuse Heap LA vi-ii

This one measured about 100 by 50 feet - a real mine, as he noted. The most numerous type of relic recovered from this heap was Chinese records on paper and wood. Stein comments that wood predominated and the paper had been brought from China. At Niya, a contemporary site, but in the desert fall to the west, he found no documents on paper. And he notes that in this particular dump there were few documents in Kharosthi script. He was particularly excited to find one unusual document, an example of Early Sogdian writing and language. The refuse dump also contained masses of broken pottery, wooden implements, furniture, textiles, bronze items, lacquered vessels, and more. He found masses of the same kind of material laying about in the open. But the special value of this material was that it was found in layers adjacent to dated documents, thus it could also be dated. The whole pile dated from the second half of the third to the early part of the fourth century. Numerous Chinese copper coins were found as well. As many of these were 'clipped' that is debased, Stein noted that the process of monetary debasement set in early.

Section IV - Remains of a Walled Enclosure

December 22 was occupied with clearance of more buildings with relatively less results. But while investigating more buildings, Stein came across the remains of long walls already eroded down to 4-5 feet. The first one he noticed was at a bearing of N 65 degrees E to S 245 degrees W. This was in conformity to the direction of the prevailing wind. He then searched for and found walls to the north and south of the LA site about 1,020 feet apart. He had found parts of the outer defense wall of the Lou-lan station. But he didn't have enough remaining time to search further. The temperature was now down to 46 degrees below freezing. But in this report he looks ahead and writes that during his return expedition in 1914 he was able to complete the survey, which confirmed that the station was enclosed in a square of 1,020 feet on each side. (See Innermostasia) Most Chinese towns are built square with the walls oriented along the compass lines - north-south and east-west. But at Lou-lan the power of the prevailing wind and its terrible erosive impact forced the Chinese engineers to orient their station to meet this reality.

Section V - The Ruined Stupas of LA

Stein leaves this exploration to last although it was the stupa that he visited first. He numbered it LAx. It was toward the north-east quarter of the total site on an isolated clay terrace 18 feet above the eroded ground. The stupa still attained a height of 33 feet. Again, he comments "That it was a Stupa and not a watch-tower as supposed by its first discoverer, could clearly be recognized at a glance by anyone familiar with such ruins." (Guess who that 'first discoverer was). He goes further in comments about Hedin's book Central Asia in which Stein points out that the very illustrations show that the accompanying text discussion is erroneous. Worse, he caustically comments that Hedin's workers severely damaged to structure. This stupa was like all the others that survived erosion and their features are described in his previous books such as Ancient Khotan. The standard stupa had a square base rising in three diminishing levels, surmounted by a cylindrical or octagonal drum on which rested a dome. This particular stupa had the lowest level 51 feet square only a foot high, the second level 44 feet square set back 4 feet with a height of 3 feet, and the third story 33 feet square 12.5 feet high. The drum here was octagonal and 7 feet high with a circular plinth 1.5 feet high on top. On this was the cylindrical dome 22 feet in diameter and remaining only 7 feet high. The digging into the structure revealed some of the wooden beams used to support the structure.

Following this discussion Stein again disputes Hedin. He comments that Hedin's single line of levels could not accurately indicate the true level of any terrain at an ancient period because the effect of wind erosion was neither uniform in its effect on localities nor uniform in its impact over time periods.
He then describes some structures found outside the confines of the enclosure. One was another stupa LAxi. It too had been dug into by 'treasure seekers'. Stein describes several more as well miscellaneous artifacts found scattered about the whole Lou-lan site.

Section VI - Remains of a Buddhist Temple

Stein concluded by the evening of December 22nd that he had found everything he practically could in the available time. (He found more remains to the north-east in 1914). However the Surveyor had found more ruins to the west. Stein was becoming concerned about the water supply. Tokhta Akhun returned with a report that the spring at "Yangi-bulak" was so salty that the water did not freeze and the camels could not drink. Fortunately the camels that Stein had sent south to the supply dump did return with water and food. He then used all the available camels to move his camp west to site LB which was a ruined stupa prominent in the desert. On December 24 he started work uncovering the lower sections of the stupa. Again, this was a building that Dr. Hedin had perfunctorily examined. There also were two dwellings adjacent to the stupa. He numbered the three LBi, ii, and iii. Stein describes all three in detail. LBi was of interest only because it was well enough preserved for him to study the construction methods. LBiii contained some Chinese coins from Han era and various fabrics and artifacts in bronze plus wooden furniture. At the shrine itself, LBii, Stein found the woodwork left by Dr. Hedin's team lying on the surface. Despite the damage Stein recognized carved decoration in the Gandhara style similar to woodwork at Niya. He recovered other woodwork in excellent condition from beneath the sand. He measured the foundation beams which were dovetailed at the corners at 19.5 by 18.5 feet. The roof beam was 17.5 feet long. The various structural beams had been attached by dowels. Some posts remained attached to the beam. The carpentry work and carved decorations were excellent. The decorations proved the building was a Buddhist shrine with representations of Buddhas and Bodhisattvas. Despite the ruined condition of the remains Stein was able to mentally reconstruct how the various wooden members would have fitted to form the original building. Not far from the shrine Stein found the ruin of another stupa with base 26 x 18 feet.

Section VII - Finds in Ruins L.B iv-vi

On December 25th Stein began work at sites a mile to the east-south-east that he numbered LBiv to vi. These produced a wealth of new 'finds' despite the fact that Dr. Hedin had spent a day there in 1900. There was a dry river bed nearby. The deep pile of sheep dung all over the floor indicated that the place had been used by shepherds years after it was abandoned by its original owners, but this 'cover' had helped protect the remains. These included excellent carved wooden furniture and decorations, textiles including woven rugs and a slipper, and tablets containing Kharosthi script. Implements included a piece of an iron knife and a wooden bird arrow. Stein as always was careful to re-bury everything he didn't remove. By nightfall of 28 December he was prepared to move on. By then some of the workers were sick from the hard work, meager rations and water, lack of adequate shelter at night and exposure to freezing wind. He had to abandon his plan to search for the ancient caravan route eastward across the Lop salt flats (which he did accomplish in 1914 as described in Innermostasia). At that point Stein departed as described in the following chapter.

Section VIII - Chinese documents from the Lou-lan Site

But first, as usual, he divides his description and analysis of important artifacts from the discussion of the excavation and architectural remains provided in the previous sections. In this section he combines analysis of the documents at Lou-lan with information from Chinese historical records. This subject is continued in section X. Stein again credits M. Chavannes with the ability to translate the numerous Chinese documents found at the 'Ya-men' building and throughout Lou-lan. His first conclusion is that Lou-lan was a small fortified garrison built to protect the Chinese main route from Tun-huang to An-si and the north rim of the Taklamakan. He repeats that the Chinese opened this route around 110 BC and continued to use it during the Han Dynasty era. He writes that the strong archeological evidence that he has uncovered would not be sufficient to prove this point without the independent documentary evidence. And the records that describe the expansion of Chinese power into the Tarim are too generalized to pinpoint the exact location of such a garrison base as Lou-lan. The Annals state clearly that the main route led across the northern rim of the desert just south of the foothills of the T'ien-shan rather than along the southern side of the basin. They also state that protection was necessary in defense against the Huns raiders. Geographical study also shows that the shortest route leads in this direction but also that there is a lengthy stretch over which water is either difficult or impossible to find. But back in Han era and before, Stein believes, the Konche-darya still carried sufficient water to provide for the western section and that it passed near to Lou-lan. Even so, the eastern section between Lou-lan and Tun-huang was 120 miles over which there was no water. (His desire to find the exact route over the salt sea was a major factor in his Third Expedition.) He stresses that the evidence of drying of the Konche-darya is clear. Again, he refers to the documents, noting that 15 of them bear dates corresponding to years between 263 and 270 AD (the close of the Wei dynasty and first years of the Western Chin). The Chinese record indicates that Emperor Wu-ti of that period (265-289) restored Chinese control in the Tarim. The annals mention many embassies from as far west as K'ang-chu (Samarkand) to the Chinese court. And there is a record describing the son of a Shan-shan or Lop ruler going to the court in 283 AD. Two other documents are dated 312 and 330 AD toward the end of the time of Chinese power. By that time, Mr. Chavannes writes that Chinese communications and central control had been lost. The documents include orders from the imperial court to the Chinese commander of the whole Tarim region to be delivered via the local commander at Lou-lan. Another document dated 324, Stein believes, was sent from Lou-lan to prepare for a meeting of the Chinese commander with the king of Kara-shahr. Stein discusses a number of other documents bearing on the issue including reports about military movements and actions. The content of the majority, however, deal with local issues of supplies, transportation, land use, and armament. Their importance is the information contained about details of administration and daily life on the frontier. The storage and issue of grain was a very significant issue in the documents. And these also indicate the requirement for numbers of clerks and accountants. Records provide information on the size of daily rations and methods for issue and accounting. Supplies had to be issued not only to the local garrison but also to members of passing embassies and other units moving west. It was Chinese policy to make such outposts self-sustaining. The Lou-lan records indicate the extensive nature of the effort to achieve this in the inhospitable area next to the desert. This was to be a military colony in which men would perform both agricultural and military service. Documents refer irrigation and growing of vegetables for winter storage. There are inventories of farm implements such as hoes, spades, saws and the like. The critical nature of canals and berms is reflected in other documents. An official was designated at the chief of the water supply. Military concerns are the subject of reports about cross bows, arrows, sword-blades and leather for armor. There are inventories of medicines. Personnel records show that some at least of the garrison were Indo-Scythians (Yuch-chik) mercenaries, the traditional enemy of the Hsiung-nu. According to Chinese Annals the Hsiung-nu drove the Yuch-chik from north of the T'ien-shan westward and southward into Sogdiana and then Bactria, from where the Kushan rulers extended their power eventually into north-west India. Ordinary commercial and mercantile topics are discussed more in the private letters found at Lou-lan. Letters from the west frequently relate to Kara-shahr. But the Lou-lan documents do not include much information on the higher levels of Chinese imperial diplomatic and military strategy. The term "Lou-lan" seems to have been applied to the region in the 3rd and 4th centuries.

Section IX - Kharosthi Documents from Lou-lan site

Stein's organizational methods appear again as he separates his discussion of the Kharosthi from the Chinese language documents. His first conclusion is that this is the same language that he found at Niya and that it showed continued Chinese use of Indian speaking individuals to conduct local administration and business. But their appearance so far east at Lou-lan raises questions. For answers Stein relies on the expert translations provided by Professor Rapson. In form and content the Kharosthi documents at Lou-lan are identical to the much more numerous ones Stein recovered at Niya. They are in early Prakrit dialect with some Sanskrit terms included. Personal names mentioned are clearly Indian or Buddhist, including that of the rulers whose reigns are referenced. The content is personal; deeds, wills, contracts, birth announcements, plans for visits, questions about sales. Stein believes the local region referred to, Kroraina, is the same as the Chinese name, Lou-lan.

Section X - The Lou-lan Site in Chinese Historical Records

In this section Stein wants to discuss what information directly from Chinese records may add to understanding about Lou-lan and what the archeological evidence he found at Lou-lan may illuminate topics in the Chinese archives. He refers to his previous discussion in Chapter X. He writes that the Chinese use the name, "Lou-lan", for the entire Lop region prior to 77 BC and then use the term, "Shan-shan", for the same area. The political center of this area was toward the south, but the principal Chinese interest was along the north side of the Tarim between the Kuruk-tagh and the Tarim River marshes. The Former Han Annals do not describe exactly the location of the caravan route north-west. He again notes that his exploration in 1914 solved this issue. But, as we have seen from the descriptions in the last several sections, he found no datable proof that Lou-lan was an active site as early as Han era. The Later Han Annals mention Shan-shan (Chapter IX) but do not provide sufficient detail as to location. A proposal by Pan Yung in 119 AD may refer to establishment of a post at Lou-lan. A commander in chief to be sent to Lou-lan is expected to exercise power over both Kara-shahr and Kucha on the north side and Shan-shan and Khotan on the south side. The purpose of the proposed military colony was to guard the caravan route between Tun-huang and Lara-shahr. Stein presents a lengthy and elaborate analysis of the names Lou-Lan and Shan-shan. He then refers again to the Chinese Annals to state that in 124 AD Pan Yung was himself appointed Chang-shih - that is commander, and it was at Lou-lan that he received the submission of the kings of Shan-shan, Kucha and others. Then he used Lou-lan as his military base to launch his conquest of Turfan from the Huns. During the era of the Three Kingdoms (AD 221-65) that followed the end of the Han era, the documents Stein found were prepared. He finds references to Lou-lan in the Wei lio, written by Yu Huan between 239 and 265 AD. It is in the Wei lio that we find some specific information on the three main caravan routes west from Tun-huang. These routes start at the Yu-men kuan, gate near Tun-huang and go west. Formerly there were two but now three such routes. The southern route Stein notes is along the foothills of the K'un-lun (called also Altin-tagh) to Charklik and then through Shan-shan, Chu-mo, Hsiao-yuan, Niya. He defers discussion of the 'new' northern route to Chapter XIX. In this section he focuses on the more central route. Stein quotes the ancient Chinese text fully including printing the Chinese ideographs. The Wei lio specifically mentions ancient Lou-lan. He focuses on the text mention of the "Dragon Mounds" or as the Han Annals called it "White Dragon Mounds". Looking ahead again to his Third Expedition in 1914 he states that on that trip he proved where and what these 'mounds' were. Reading Innermostasia we see what a trying experience that was - crossing the difficult salt sea from Lou-lan going east and then south east to the caravan route north of Tun-huang in winter. The route west from Lou-lan, Stein declares, followed the then water filled bed of the Kuruk-darya toward Kucha.

Stein turns to the text written by Li-Tao-yuan prior to 527 AD upon which he places considerable reliance. This document describes the several rivers that flow eastward from north and south of the desert and converge around Lop. Stein's argument is extensive and complex, filling several pages. His whole point is to prove that the site that Dr. Hedin found 'by accident' and that Stein thoroughly excavated during two expeditions - namely Lou-lan - is the military colony established for the purpose of supporting Chinese continued power across the Tarim to the Pamirs. And this military agricultural colony depended on irrigation methods for survival. He writes that the description given by Li-Tao-yuan accurately describes the engineering methods used to achieve this irrigation. This construction was annual due to the spring floods and demanded a very large labor force. Stein compares this with contemporary irrigation projects he observed throughout the Tarim. Stein is convinced and here reports that the ancient "Town of the Dragon" refers to the great expanse of high clay ridges, mesas really, with vertical walls north-east of Lou-lan that he explored in 1914 with great difficulty. I noted above that Stein had wanted to conduct this exploration in 1908 but could not due to lack of water. His preparation in 1914 was more extensive. He claims this identity is 'proven'. In addition to the correspondence between text and geography, Stein of course refers to his own observation of these fantastic natural formations during 27 February 1914. He writes that "Town of the Dragon" is an apt description that would occur to ancient imaginations.

Section XI - The Abandonment of Lou-lan

Stein reports that he has not found any Chinese mention of Lou-lan later than that of Li-Tao-yuan. Chinese history records that after the end of the Eastern Chin Dynasty (AD 371 - 420) the empire was divided into southern and northern parts at which point all Chinese efforts to retain power to the west ended for several centuries. It was revived during the Sui Dynasty (AD 589-618). He considers it proven that Lou-lan was abandoned shortly after the final dates on the documents he unearthed, that is in the 4th century, as the last of these is dated 330. Likewise the Chinese coins, of which Stein found over 500 during the two expeditions around Lou-lan, not one is dated later than Han or Western Chin Dynasty. But the cause of the abandonment is unknown. One possible cause would be directly related to the above mentioned political change. Stein doubts this. He prefers to ascribe the problem to change in the climate causing increased dryness and decreased water in the Kuruk-darya. At all times the water supply was precarious and its available volume required extensive engineering effort to enable irrigation. Or the cause may relate more to the simple inability of the Chinese administration to continue the exertion required to maintain the water supply. He notes, finally, that it appears that the buildings remained in some use by, for instance shepherds, for some time after they were abandoned by the official military establishment.

Section XII - List of Antiques from Lou-lan

This section is another of Stein's detailed descriptions of each and every artifact found, no matter how small. It fills 23 pages in small type.

Chapter XII - Return to Miran Site

Section I - To the Tarim's Delta of Charchan-darya

Stein completed work and left Lou-lan on 29 December 1906. He divided his team into two sections. The larger group of exhausted laborers returned directly to Abdal with the sick Rai Ram Singh in charge. Stein, always seeing new 'finds' along different routes, went south-west, back into the desert in search of the delta of the Tarim River and a site mentioned to him earlier, Merdek-tim. This excursion required 7 days to reach the Ilek River and lagoons. He crossed the several dry beds of the Kuruk-darya delta (which he surveyed in detain during his third expedition in 1914). Along the way he chanced upon a Chinese coin, a piece of bronze mirror and a bronze spear point. He also found some neolithic objects. On 4 January 1907 they reached the Koteklik-kol lagoon of the Ilek River at a point a day's march south of where he intended to be. On the 5th he went back north along the frozen Ilek and on the 6th found a local Loplik fisherman who guided them to the hidden ruin, Merdek-tim. It was a small, circular fort with stamped clay and brick rampart all overgrown with reeds. The rampart was 29 feet thick at the base and 14 feet thick at top. The fort was 132 feet in diameter. On the south side there was a gate opening 6 feet with the rows of Toghrak posts that would have held a gateway remaining. Inside a few beams were found. There were two types of bricks in the walls, one 18 x 11 inches with 4 inch width and the other 14 x 10 inches with 3.5 inch thickness. Chinese coins dating from Later Han era were found on top of the rampart. Two were dated to Wang Mang (AD 9 -22). Sub-soil moisture decayed any remains of wood, documents or clothing. The fort occupied a strategic location near a former branch of the Tarim River. Stein disputed the theories advanced by Sven Hedin on the ancient location of the Tarim and its delta lakes.

Stein left Merdek-tim on 7 January to move south toward Charkhlik and Miran. For 6 days he rode south to Kotek-shahri near the Charchan-darya. That river ends in the desert around the fishing station, Lop. On January 11th he rode on from Lop across the delta of the Charchan -darya. He describes the topography of the area over which he rode to Charkhlik while finding Stone Age artifacts along the way. In Charkhlik from the 17th to 21st of January he engaged more laborers for the work at Miran and organized transport and supplies for the long journey to Tun-huang. As always he received gracious assistance from the Amban, Liao Ta-lao-yeh. On the 22nd Stein started back to Miran. The following day he met Chiang Ssu-yeh with more workers from Abdal.

Section II - Ruined Fort of Miran

Stein was eager to get back to Miran with enough workers and time to do a thorough excavation, the results of which as it turned out he had only hoped to achieve. On the 23rd he established his camp with almost 50 men next the walls to mitigate the effect of the freezing wind. As usual in his report he begins with description of the physical aspects - the structure of the ruin. The fort was on a bare pebble-covered plain east of the Miran river. The walls formed an irregular quadrangle with the two longer sides facing east-north-east-east and north-west. The shorter sides faced west-south-west-west and south-south -east. The longer sides were 240 feet each. The western was 168 and the southern was 200 feet long. There were oblong bastions at the corners, two were 22- 24 feet high and a third was 28-29 feet high. There also was a bastion projecting from the face of each curtain wall. That on the south was was largest (fig 113) built of solid clay to a height of 41 feet. On top was a parapet 7 feet high and 2 feet thick made of stamped clay with tamarisk brush wood between layers clay. The curtain walls were constructed differently. The greatly eroded south and west walls were of stamped clay below with a thick parapet of brickwork and tamarisk above. Its rampart probably originally was 12 feet thick. The south wall was originally 18-20 feet high, the north wall had already been eroded down to 10-15 feet height. The east wall was 24 feet high in places with a solid, brick and clay parapet six feet thick in sections. The wind had pushed the sand outside into a pile nearly to the height of the parapet. At the same time the wind in other locations, such as along the west wall, had scoured out the sand and gravel into pits or ravines 10 feet below the surface level. For this reason the quarters were constructed along the inside of the east wall and not along the exposed west wall. An interesting feature which Stein noticed was that the clay used in the walls already had pottery fragments and potsherds in it, indicating that the the location had been occupied long before the fort. He found piles of large stones clearly not used for construction. Later he concluded that these had been ammunition.

Section III - Excavations in Miran Fort

From description of the fort Stein turns to describe his methods of excavation. On 24 January he continued the preliminary effort begun on 8 December along the inside of the east wall where the Tibetans had built their quarters. The quantity of refuse indicated a long occupation during the 8th and 9th centuries. The quarters consisted of a double row of small irregularly shaped and sized rooms built against the fortress wall. (Plan plate 30) The row closest to the wall was protected so that its roof remained. And the refuse nearly reached these roofs. Apparently some rooms served as dumps for an extended period and were then simply closed. He comments that the odor of ammonia was overpowering. He remarked also that the Tibetan soldiers must have been the dirtiest of the folk he found in Turkestan and that was saying a lot.

Braving all these conditions, Stein found hundreds of Tibetan documents on wood, paper or leather. In one small room he found the only example of Runic Turkish he collected in Central Asia. The majority of the documents were on wood. Some examples of a higher quality paper turned out to be Buddhist religious texts most likely imported from distant monasteries.

Section IV - Miscellaneous finds in the Miran Fort

In addition to documents the Miran fort contained many other artifacts including bits of clothing, implements, armor, and personal items. Many were the pieces of leather, scale armor (described in more detail in section VII). The scales came from different suits of armor of different sizes. All the scales were oblong and no examples of round-headed scales were found. Other implements included spinning and weaving tools and fishing gear. There were many fragments of cloth, woolen, cotton, felt, and a few of silk.

Section V - Tibetan Documents from Miran Fort

Stein reserves this section for discussion of documents. Stein notes his relative ignorance of Tibetan which forced analysis of the documents to await their examination by experts in London. But from their quantity and location he presumed they were mostly official records and reports. At the time of this report only a summary of the general content has been made. He confines himself to discussion of those documents so far translated that refer to the Miran fort itself or events related to it. In these the names - Nob - Little Nob - and Great Nob appear. He proposes that Little Nob is Miran and Great Nob is Charkhlik.

Section VI - Records in Turkish Runic script

The unique rarity of this document prompts Stein to devote a whole section to it. Stein quotes the analysis provided by Professor Vilhelm Thomsen, a famous scholar of Turkish writing. Professor Thomsen proposes that the document is a register of persons who may be entering or leaving the fort's jurisdiction -a kind of passport. This is similar to the continued use of such documents throughout India, Persia, Central Asia, Mongolia. Stein writes that the official term used is "Yariq" and that this is an unknown word. I should point out the Mongol term - Yarlik - designating the official grant of office by the Kipchak- Mongol khans to Russian princes. In any event Professor Thomsen believes the term designated an official sent on a mission or given some duty. In an extended analysis Stein connects all this with the history of Chinese, Tibetan and Uigur relations and periods of control of eastern Turkestan.

Stein concludes with a discussion of the strategic location of Miran and the reasons the Tibetans established their fort there. He notes that from ancient times to the present this location is at the crossroads of the east west caravan route he has been following and continues to follow in his expeditions with the historic route between Lhassa and Mongolia. But for the Tibetan military the fort had strategic value only during a period in which they succeeded in power over the Tarim. For the Chinese and Mongols Charklik was a more convenient strategic center.

Section VII - Descriptive list of Antiques from Miran Fort

This is another 8-page detailed listing of the items fount at the Miran Fort. I mention several of particular interest to those wanting information on armor:

M I 0068 - Leather armor scale: Oblong, for type see Ancient Khotan i pp xvi 374, 411. Lacquered inside red, and outside red with thin top layer of black. For lateral lacing, by one long side, hole at each end in extreme corner; by the other, two pairs of holes nearer middle. For vertical lacing, one pair of holes parallel to, and .5 inch from one end. No thongs remain. Condition good 3 1/8 by 3.2 inches

M I 0069 - Five leather armor scales, detached; for general type see M. I 0068 and Ancien Khotan. Thong-holes generally as in M i ix 002, but those for vertical lacing are irregular. Lacquered behind black over red (but one scale dark red with a pattern resembling M I xxiv 0042-0) in front, very thin black or dark red, over red, over black, leaving surface dark black and crimson. Orn. with red line along one overlapping side and top. and two curved fig. one above the other towards one long side. Beyond these, three scales show red ellipse with black center. Orn. is scraped out after application of all the coats of lacquer, so that those beneath the surface are shown in concentric rings down to bottom of orn. Largest scale 3.5 x 3.25.

M I viii 001 Fr. of leather helmet, lower edge, curving in to neck, then out to rim; inside lacquered red; outside black (over layers of red and black), with thin coating of red in band along rim, and remains of two red bands above. parallel to rim is bare strip .25 wide from which lacquer has been removed; two holes for attaching thongs, in which are now two silk strings, buff and gamoge, knotted together. Lower rim and upper edge are orig, edges of fr, side have been cut

M I ix 002 Leather scale armor, twenty pieces, all now separate except two pairs which have long sides attached; cf M I 0068, 0069, xxiv 0040, etc. Lacquered inside with thin coat of red over black; outside black, over red, over black; later coats visible where orn scraped away, as in M I 0069 etc. here orn. consists of two (or three) comma-shaped figs. one above another, about .5 inches from one long side. Pair of holes near each corner, parallel to long edge, for lateral lacing; another pair (or one large hole) parallel to short edge, and about one-third down from it, for vertical lacing; instead of latter, four longer scales have two holes, one-third of way respectively from each end. Method of attachment of overlapping long sides is as follows;- from the top of the previous joint the thong is brought from behind through the lower of the two holes near top corner, passed down in front and to back again through lower hole of bottom corner; then brought up behind and to front again through the upper of the top corner holes, lastly to the back through the lower top corner hole, through which it was first brought. It is then carried along behind for attachment to next scale. This method of lacing makes no use of the upper of the bottom corner holes, and it therefore prob. not the method orig. employed . condition good - Scale average 2.75 x 2 inches

M I ix 003 Leather scale armor, twenty-four pieces; cf M I 0069, Six scales are of ordinary size (average 2.75 x 2.25) three being still joined by long edges; seven of greater length (average 3 1/8 x 1 15/16) two being joined; seven of still greater length (average 4.25 x 2 1/16) three still joined; and the remainder are frs. Smaller scales much resemble M I ix 002, but have only one coat of lacquer ( black) with orn. of two shallow elliptical rings sunk in the black, and higher center lacquered red. Three holes at equal intervals down long edge for lateral lacing, and bronze rivet (apparently ornamental) half-way between the ornamental rings; lacing which remains in long sides too disjointed to show method employed. The longer scales resemble these; but for vertical lacing there are one (or two) pairs of holes placed along top end, and another hole or ;pair of holes parallel to this, towards middle of scale. The longest scales have additional rivets, orns. and holes in proportion to their length. Condition very fair.

M I ix 004r Fr. of leather armor scale, as M I ix 003; orn. (sunk in black lacquered surface with red centers) consists of an ellipse and a crescent with tips turned towards ellipse; crescent prob. a part of larger orn.
Three more entries;
M I xxiv 0040 Leather scale armor; three pieces, one consisting of six scales (four in one row, two in another) one of three scales, and one of two; same type as M I 0068, 0069 etc. Lacquered inside dull red, and outside dull red and black seven times alternately, the two coat being red; each scale orn. with three shallow cup-like depressions in line parallel to, and 5. inch from, the overlapping long side; these scraped out as in M I 0069. For lateral lacing a pair of holes are pierced towards each corner, by the long edge; and, for vertical lacing, two pairs of holes parallel to end (one pair being placed about .25 inch from top, and the other about half-way down scale). The scales prob. over lapped upwards. (See Ancient Khotan i p xvi). Method of attachment is as follows: - Lateral lacing; scales are first laid with long sides overlapping so that holes in lower exactly correspond to those in upper. Thong is brought from back through lower of bottom corner holes, passed up, and to back again through upper of bottom corner holes; then up behind and to front and back again through top corner pair of holes then across behind to holes in opposite top corner, through which it is passed to front and back (working downwards), and so on to second bottom corner holes; thence across to next bottom corner and again up; as described. Vertical lacing: - for this two thongs are used ,running side by side down the pairs of holes described. The method of starting or finishing off these at the top; is no clear, as the ends are now found loose behind. The thongs are taken down front of the scale, and passed back through the pair of holes half-way down; round a thong (which runs horizontally behind these holes along whole set of scales, never appearing in front), and to the front again through the same holes. Continuing down the front, they pass behind the overlapping top of the next scale, are brought to the front through its top pair of holes, and so on as described, being finally passed sideways through the two upper holes of the last plate to which they reach. Scale (average) 3 inch x 2 5/16.

Chapter XIII - The Ancient Buddhist Shrines of Miran

Section I - Sculptured Remains of Ruin M. ii

As the work at the fort was being completed Stein moved a team to the ruin he had found back in December. This was about a mile and half to north-east of the fort. Fragments of pottery were abundant on the open ground. The ruin appeared as a solid, two story shape in the sand, 46 feet by 36 feet and with first story 9 feet high. (Plate 31 ) The second story was about 17.5 x 15 feet with remaining height still over 11 feet. The lower sections of wall were protected from wind erosion by sand. These contained niches coated with plaster in which there were relievo decorations. The niches had contained stucco statues in relievo representing standing Buddhas. Stein describes all this scene in detail. The projecting sections represented pilasters similar to those at Persepolis and at Gandhara. Stein writes that the entire design follows that of Graeco-Buddhist art. They are also similar to examples at Lou-lan. On the north-east side Stein found among the debris a huge stucco head. The scene was similar to that he had uncovered at Rawak Stupa in 1901. The poses of the seated Buddhas conformed to those at Rawak and in Gandhara. Further excavation turned up a fragment of a Sanskrit document in Brahmi script clearly written in India. Dr. Hoernle dated this to about 400 AD. The entire stupa was enclosed by a protecting wall and this in turn led to another small, circular structure. Between outer wall and stupa Stein found evidence that the ruin, long after its abandonment, had then been used by shepherds. Stein devotes considerable attention to analysis of materials at this temple in comparison with other locations in his effort to date the place. He concludes that the Miran Vihara cannot be later than the 5th century. The absence of any Tibetan writing at the stupa indicates that it was abandoned before the Tibetan occupation that created the nearby fort in the middle 8th century

Section II - The Stupa Cella Miii and its Wall Painting

On January 31st (while the fort was still being reburied) Stein shifted his excavation to a location about a mile west-north-west of the fort. Removal of the sand soon revealed a building with a square base surmounted by a round dome. The dome superstructure had fallen into a pile 8 feet deep which protected much of the lower walls. As usual there was also evidence that 'treasure-seekers' had centuries before already dug a channel through the outer wall and into the stupa base. Once cleared the form and content of this shrine were revealed. It had windows (each at 2 feet 8 inches from the floor and 2 feet 3 inches wide) on three sides and a doorway on the remaining, west, side. The stupa itself measured 9 feet in diameter at the base. As the excavation proceeded Stein was startled to see the emerging paining of the classical Cherubs or 'angels'. Then a 4-foot long silk banner on which was Kharosthi script was uncovered. This indicated a date of 3rd to early 4th centuries. The art work was in tempera on stucco with a thin backing. The height of the fresco remaining on its wall was 3 feet 10 inches from the floor. Stein recognized that removal would be extremely difficult, but essential.

He spent February 1st collecting wood from the nearby forest of dead trees and then cutting pieces to make flat boards. He describes the physical effort in Desert Cathay. Prior to removal Stein made a careful written record of the location of each piece of stucco, whether already fallen or remaining attacked to its original wall. This frieze was continuous below the windows. This consisted on the remaining north-east and south-east sections of lunettes, each about 2 feet 2 inches wide on a curved surface, in each of which there was the head and shoulders of a winged male figure. There was nothing remaining on the south-west wall and only a section with two lunettes on the north-west wall. Above the series of lunettes was a black band to separate them from the paintings above. And below them was another band 9 inches wide.

Above the lunettes the painting was much more damaged or destroyed with its wall. The remains showed the feet or some of the lower parts of standing figures. Nothing could be salvaged from the walls, but amid the debris Stein recovered some material formerly on the walls. Even so he was able to determine that the walls contained a continuous series of scenes depicting events on the life (or former lives) of the Buddha.

Stein places particular importance in these art works so laboriously recovered for their study of Graeco-Buddhist art. He notes that in Gandhara itself only sculpture has survived, but no examples of painted surfaces from this early era. These works are in tempera on a thin coating of plaster of Paris placed on a backing of loess. A pale pigment of ferric oxide forms the ground over the white plaster. In a footnote Stein describes the technical methods used at the British Museum to preserve these fragile bits of plaster. He continues with much discussion comparing the art work from Miii with that he found in Mv, such as the story there of Prince Vessantara.

Section III - Remains of Painted Friezes of Miii

Stein continues in this section with detailed study of the subject matter. He begins with description of the representation of Buddha dressed in a dark purple-brown robe. To his left are six disciples in two rows dressed in bright green or red robes. Stein describes in detail the various ways in which the representation does or does not correspond with Indian artistic style or with that of Gandhara or Hellenistic models. He remarks on the large eyes which he connects with European features. And he finds quite significant the hand of a missing figure which is shown appearing from inside his robe in a style similar to Greek or Roman statutes in which a hand appears from inside a toga. The 'angles' also have heads reminiscent of early Christian art in Egypt.

Section IV - The Dado of "Angles" in Cella Miii

Here Stein focuses on the 'angles' mentioned in the previous section.
These 'angles' were painted in groups of six figures in each of the four circular sections between the door and windows. The north-eastern and south-eastern lunettes were largely intact, but in others the heads were missing and only wings and shoulders remained. He managed to rescue seven figures. The composition, Stein indicates, is of a heavenly group but the artist has managed to produce distinct individuals. The heads are all Western with a Semitic cast. All the faces depict young men with large eyes and rounded cheeks and chins. But details of the poses create great variety for the individuals. Stein's discussion continues for many pages.

Section V - Excavation of Temple Mv

Stein found yet another buried structure 60 yards north-west of Miii. He started digging there on 4 February. With the surface cleared Stein recognized that this was another square temple with a circular cell in the center that had contained another stupa. The outside square was 40 feet a side composed of sun-dried bricks. The walls were seven feet thick. Outside these walls was another passage 5 feet wide. The south wall remained to a height of 10 feet. In the center the stupa remained only to a height of 10 feet with a circular passage around it 7 feet wide. The stupa was also of sun-dried bricks covered with a thick layer of plaster. The circular base was 12.5 feet in diameter. Mouldings remained up to 7 feet. The top of the stupa was gone and also the high dome that had covered the structure. Stein estimated that this dome must have had a span of 26.5 feet. Unfortunately there remained no way to determine the architectural methods for supporting such a dome. Stein found in the debris various wood members that formerly had supported the dome and an an umbrella over the stupas. Most important, the wall surrounding the stupa still contained a frescoed frieze and dado. Other decorations including series of dadoes and lunettes similar to those in Miii collapsed when excavated. The preserved section depicted a youth on horseback and had two inscriptions in Kharosthi script nearby. Stein continued excavation despite freezing wind and dust storms. Two arcs of fresco remained, separated by the original door and by the passage hacked through the wall by the ubiquitous 'treasure-seekers'.

Section VI - Mural Painting of Buddhist Legend in Cella Mv

In this section he describes the fresco that filled the upper portion of the circular wall. A few illustrations accompany the text in which he specifies the excellent color which does not reproduce in the B/W photos. The lengthy fresco depicts scenes from the popular legend of Prince Vessantara. He is shown leaving his city (having been banished) with his wife, Maddi, and two children in a chariot.

Section VII - Painted Dado in Cella Mv

In this section Stein describes the dado that was below the frieze in building Mv. It was 2 feet six inches high, below a triple black border separating it from the continuous frieze above. He identified this as Hellenistic style similar to that in Gandhara. The entire dado below the upper frieze, he calculated, could have pictured 28 figures of which he could make out 20. These were standing figures show at 3/4 length. Some were wingless amorini and others were youths wearing Phrygian caps. The figures were not Buddhist but secular art. As he notes in the following section, the black and white photographs that resulted from his laborious effort don't do full justice to the art. His text describing each image fills several pages.

Section VIII - Kharosthi Inscriptions of Mv and the Remaining ruins of Miran

The Kharosthi inscriptions in the heading consist of an interesting detail in one of the frescos the signature of the artist was found, just like in western paintings - this one stated "This fresco is the work of Tita, who has received 3,000 Bhammaks." Stein immediately connects this 'Tita' with the Graeco- Roman name Titus. But this was a popular name also in Syria and Persia. He notes many other details which confirm the influence of Hellenistic art into Central Asia. Another inscription on the fresco reads, "This is Isidata, the son of Bujhami". This one identifies the figure below it, Prince Vessantara (Visvantara). Again, as at other sites, Stein complains of the physical difficulty he had in squeezing into the small space while trying to make photos. One can appreciate this when recalling that he was using a 1900 camera with glass plate photographic methods and in freezing weather and that could not be developed until many months later. Nor, of course, could he make color photos. Nor could he remove many of the fragile frescos without destroying them. With the carpentry skill of his Naik, they did manage to free some of the smaller examples. This effort he describes in more detail in Desert Cathay. As always he made sure to re bury the frescos in hopes of protecting them for future study. In a footnote he mentions that much of his protective effort was undone later by a Japanese explorer who rashly attempted to remove frescos but left a botched job. During his third expedition (see Innermostasia) he managed to remove a dado after 12 days careful work.

He also describes several other nearby ruins, such as Miv, Mvi, and Mv iii, three stupas. All these had been ransacked ages ago by 'treasure seekers'. He gives the measurements of these in the text. In particular he makes note of the architectural use of the 'squinch' a technique for placing a round dome on a square building. This, he points out, connects Miran with the architecture of Sassanid Persia. In details he discusses the history of square and round buildings. He laments again that he has been prohibited from accomplishing the essential archeological work in Bactria which would establish the physical links between the Near East and Central Asia. Although this work is the report on his second expedition we are reminded by his comments that it was prepared after his third and thus benefits from facts he learned later.

Section IX - List of Antiques from shrine of Miran

This section is 9 pages of very detailed descriptions of the fresco panels, stucco reliefs, and fabrics recovered at this shrine

Volume II- Chapter XIV Through the Lop Desert to Tun Huang

Section I - From Abdal to Tun Huang

Stein moved east on Feb 21. There are two direct routes between Lop, south of the Tarim with Tun-huang to China. One route was along the high barren slope of the eastern Kun-lun Range or Altin-tagh Mountains. The other route was through the desert depression between Kuruk-tagh on the north and the foothills of the Altin-tagh on the south. That one goes along the south side of the Lop Nor lake. The route passes through Abdal, then Donglik, then Chindailik and Lowaza - Koshe - Besh-toghrak to the Su-lo Ho. He always refers to written expeditions - and here comments that Marco Polo gave an excellent description. He quotes Henry Yule.

Section II - Chinese Records of the Lop Desert Route

Chinese records of the Lop Desert route show the Chinese move west from Tun-huang to Lou-lan. In the Han archives this is called Shan-shan=Lou lan. Stein mentions here his 1914 expedition described in Innermost Asia. This by virtue that he was editing the report on the earlier expedition while finishing the third one. The 120 mile trip covered a section with no water or fuel or grazing for animals but it can be used by carts. Map 70 The southern branch is easier but longer. Marco Polo used that one. Han records of the route started at the Jade Gate - Yu-men - Wei lio records this in 239-65 as better than the central route from Tun-huang to Lou lan across the Lop salt sea bed. The Jade Gate is at Txic fort. Map 74. The route passes through the terminal basin of the Su-lo Ho River which was opened by Chinese emperor Wu-ti and used for 450 years, then abandoned by the increasing dryness and lack of water. Chinese power was lost in Central Asia under the Eastern Chin Dynasty 317-419 until the T'ang Dynasty in the 7th century.

Buddhist monk Fa-hsien was at Tun-huang in 400 AD (Map 78). The route to Miran and Charklik was 380 miles and took 17 days to Abdal. Other travelers were Sung Yun a Buddhist in 519 and Hsuan Tsang in 630. The latter returned via Tun-huang from India in 645.

The Tibetan fort at Miran was noted by an embassy from Khotan to China. In 936-1126 the Chinese used the route from Lop to Tun-huang.

Section III - Marco Polo and Later Travelers on Lop Desert Route

Marco Polo in 1273 went by Lop=Charkhlik across the desert to Tun-huang. Stein gives extensive quotations. He cites his book, Desert Cathay. By then the northern route via Khotan -Ak-su - Turfan - Hami - Su-chou was used.

Section IV - First Remains of the Old Chinese Limes

- Stein found these on March 7 while going east (Map 74) when he noticed the first watch towers. Plate 33. He was going to the spring at Toghrak-bulak to camp - the first watch tower was Tii. The second mound, Ti, was on a defensible hill location. A dry river bed suddenly became impassible with spring flood. It was a tributary of the Su-lo Ho.
On March 8 he found tower Tiii (Plate 33) The plan is fig 49 - plate 36 - the tower was 20 feet square of solid masonry of sun-dried bricks - He then found a straight wall - 8 feet wide - to the east. The wall was built of alternate layers of 2-4 inches thick bundles of fascines of reeds and 6-7 inches thick clay and gravel. The reeds were 7 feet long and also used as revetments along the wall. See photos 157 - 158. He then noticed a chain of watch towers each on commanding ground and visible to the others.

He visited again on March 17th in 1914. The towers had foot holes and holes for beams for ropes used for the soldiers to climb to the top. The beams were 3.5 feet apart in pairs and 4 feet up. A tower was 25 feet tall on commanding knoll and 22.5 feet square of solid masonry with each brick 14x7x5 inches and reeds after each 5 layers of bricks. (fig 173).

Section V - Ruins by Su-lo Ho marsh

- The long line of the wall was parallel to the Su-lo Ho. Stein continued on 9 March (fig 178) to tower Txi. Then the wall went further to the north east. He found another tower fort - Txiv - 30 feet tall (fig 183) plate 40 and then another tower (fig 184) with a building with 3 halls - 440 feet long (plate 41) inside with a massive wall enclosure. It had towers on the corners. He decided this was the famous "Jade Gate".

On March 10 he received fodder for the ponies out from Abdal. He found Khara-nor - the 'dark lake'
On March 11 he reached Tun-huang oasis and its northern limes. This is the walled city of Tun-huang noted by Hsien. He stayed there 3 months

Chapter XV - Tun Huang Oasis and its Northern Limes.

Section I - Geographical Features of the Lower Su-lo Ho Basin

On this expedition Stein was there from 12 to 22 March. He had been there also in 1907. In part I he describes the geographical facts about the region and especially the lower Su-lo Ho basin. And the caves at Chien-fo-tung (map 78). There had been a major Tungan revolution in 1862-73 during which much of the local population was killed and homes destroyed. Stein had to distinguish between ancient ruins and recent ruins. He describes the Su-lo Ho flowing 200 miles east to west in the main corridor from interior China to the west. The corridor was the highway from Lanchou , capital of Kan-su, along the northern slope of the Nan-shan - Liang-chou - Kan-chou - Su-chou - Yu-men-hsien. Thus the only road west was through this narrow corridor along the watershed. The Han emperor, Wu-tu, (121 BC) defeated the Huns in 121 BC and pushed out of the valley. A similar campaign again in 1877 pushed Chinese control westward. One route was via Tun-huang and along the south side of the desert, the other turns north west from An-hsi (Kua-chou) along the edge of the Tine-shan through Hami - Kumul. In the Han dynasty Tun-huang was famous center of the largest cultivated oasis between Su-chou and Khotan a distance of 1200 miles. The Tang-ho River from western Nan-shan peaks over 20,000 feet elevation sent snow melt water to the Su-lo-Ho (map 83 - 84) and helped the defense wall along its south side. The section of wall north of the river to a crossing at Wan-shan-tzu ridge (map 83) shows on the south bank to the marsh at the end of the river. The 'caves' were 12 miles SE of town (map 78 ).

Section II - In search for the old wall north of Tun-huang

On March 16th Stein went to the 'cave of the 1000 Buddhas' - Discussed elsewhere - then on 23 March he started due north of Tun-hung to go back for more detailed study of the towers he had seen on the way to Tun-huang. There were many local small, local forts for refuge against the recent Moslem rebels - Tungan raids - so Stein had to separate those out from the ancient ruins. Some show nevertheless on his map. On the left bank of the Tang Ho there was a walled village, Shih-pan-tung. On 24 March he found the tower Txxiv (map 78 - D3). It was made of quarried clay blocks and was 20 feet square and 18 feet high. Tower Txiv (fig 152) plate 34 was 26 feet square, 20 feet tall with a room on top and appeared to be more recently built. Finally he found the wall towers with a Tungun guide. They marked the road from An-hsi to Lop nor.

Section III - First Discovery of Dated Han records at Txxvii
dates from the Han records found at Txxvii

- On March 27 he moved south east to Towers on the wall line - Watch tower Txxvi (fig 150) was 25 feet tall, 20 feet square, and made of hard clay interlayed with thin layers of brush wood. At tower Txxvii he found the dated documents he so much wanted to find. The document was of Chien Wu, regnal title of Emperor Kuang Wu-ti. from AD 25 who founded the Eastern Han Dynasty. The document was dated 150 AD. On March 28 he moved to tower Txxviii. He was obtaining excellent results from search of refuse dumps. He found more documents from AD 35-61 including military orders. The ultra-dry weather preserved materials under very thin cover of sand.

Section IV - Search of Ruined Limes Stations Txxviii and Txxx.

He found more records and official documents - many used wood stationary. And there were many domestic articles plus military items such as cross bow arrow heads. Plate 34 is a plan of a tower (fig 154) He found ceramic ware from Sung era 963 AD at a Buddhist shrine. Tower Txxx was the western most tower.

Section V - Survey of limes toward An-hsi

- Towers Txxxi and Txxxv. he was moving back north east. He found a long east to west stretch of low salt bogs parallel to the Su-lo-Ho. He repeatedly complemented the Han engineers for maximum use of terrain and topography both to enhance the strength of sections of wall and also to enable them to skip sections where the river and marshes would prevent incursions. On March 31 the temperature was still 39 degrees below freezing. He moved east to a line of towers. The wall itself curved north in a demi-lune shape in front of each tower. The towers were not ON the wall - that is physically integrated with the wall itself but behind it a few yards. Clearly they were for signalling purposes, not for major defense, nor was the wall capable of resisting a major attack. It was a barrier to small-scale incursions. The line of the wall went east - north east to west-south west. The distance between towers varied from 3/4 mile to 1 1/4 mile. Fig 157 shows layers of fascines 6 inches thick. The wall was 6.5 feet wide at the top but originally over 10 feet wide. He noted again that the Chinese engineers were experts. He returned to his original camp on 3 April.

Chapter XVI - Oasis of Nan-hu and the Yang barrier

Section I Remains between Tun-huang and Nan-hu

This also includes the route south west to Nan-hu. Stein returned to Tun-huang and then moved south-west to the small village oasis of Nan-hu. There he found also the ruin of a much older previous version of the village. From there he moved back north along a different route to the limes. He started on 5 April at the old town Sha-chou of T'ang era and passed many walled enclosures and old villages. On 6 April he reached Nan-Hu oasis (maps 79 and 78).

Section II Nan-hu oasis and its present resources

Stein refers readers to his book Desert Cathay. - Plate 35 figs. 159 and 160 - He found 4-5th century coins plus arrow heads and pottery. One Sung coin of 1038-40 and T'ang era coins also. Nan-hu was sacked by the Tungun rebels in 1866 and had not fully recovered. He considered that the wall he found going north might be identified with the Yang barrier of Han times - the Yang-kuan defense to protect the limes from attack in the rear. At that time the Chinese garrisons and caravans needed protection not only from the Huns to the north but also from the Tibetan tribes from the high plateau to the south - the K'un-lun mountains.

Section IV - Position of the Yang barrier

The Yang kuan of Han records - he used the old Chinese records - the location specified the old Chinese name - S'hou-ch-iang from Han founded in AD 521. The "Jade gate' at tower Txiv (map 74_ was north west of Nan-hu 36 miles. The Chinese commanders realized the strategic need to hold Nan-hu and the road along the slope of the Altun-tagh. Nan-hu has the only water from the Altun-tagh.

Section V - Abandoned village north of Nan-hu

On April 11 he moved north from Nan-hu and found recently ruined villages. On 14 April he reached watch tower Txii (fig. 181) on the limes. He then went 5 miles west to reach the same camp he had used on his first day visit to the wall.

Chapter XVII - End of the Chinese Limes

- Section I - Terminal station of the wall

- Now he was finally able to move west to find the very end of the Han wall. April 13 it was now getting hot. He describes this period in more detail in Chapter LVIII and LXII of Desert Cathay. The west end of the wall is at the marsh that forms the terminal basin of the Su-lo-Ho. The wall and detached towers here made excellent use of advantages of the terrain. (map 74 - plate 33) The wall here runs due west from Tvii on top of a narrow plateau - from Tiv it turns south to marsh ground and the line reached the extreme north east corner of the terminal basin of the Su-lo-Ho which was filled with lakes that continue west to 92 degrees 55 min on Map 70 for 300-400 square miles. The wall rests its flank for 30 miles on a line south west on ground that was impractical to cross. Further south are dunes and the Altun-tagh. The signal towers cover ground to the south edge of the basin. The map of the Han wall - limes - shows the western stretch Tvii, Tiii, Tiv on a ridge - North of Tiv is the end of the Su-lo Ho. Stein's camp 154 was at Toghrak-bulak ( a spring). This is where the caravan route to Lop crossed. There were additional watch stations Tiva, b, and c west of the end of the wall. On 30 April Stein records finding more military documents, orders, rations lists. He gives details of the wall and tower with military appraisal of the size and skill of the builders. Also found were traces of a fortified camp behind (inside) the wall, where the route to Lou lan left. So a large garrison held the west end of the wall. On 2 May he made a reconnaissance north west toward Lou lan. There were signal towers along the caravan route north and outside of the wall. Towers include plate 33.

Section II South west defense flank of the limes

- Stein found 5 watch stations detached from the wall in positions south west of its end. On May 3rd he visited tower Tv (plate 36) and found document dated from 39 BC. He explored this region in March, April and May of 1907 and 1914. In the spring a powerful wind came from the north-east due to the rising very warm air over the Tarim basin and remaining cold air over Mongolia. This generated a blinding sand storm which over the centuries was the main force eroding the towers, walls and structures. He noted that in areas where there was no sand, but only a flat rocky surface the erosion was less. The following chapters are devoted to detailed descriptions of the methods and times Stein devoted to study of each tower and the walls in between. Tower Tvi (fig 169) was located on a conspicuous site on a narrow ridge. It was 18 feet square and 15 feet high made of sun-dried bricks each 15x7.5x5 inches with layers of reeds between every three courses. Toghrak trunks 13 feet long were embedded in the masonry as vertical supports. And long trunks were fixed horizontally as a frame work inside. There was a staircase on the east side.

Section III - Ancient station Tvib and its records south of Tvia. (fig 171)

There were two towers on a mesa behind the line of the wall. The location commanded a view of the marsh and covered the entrance route from the west between the Su-lo Ho ending marsh and the sand dunes to the south. Map 74 plate 37 - The main tower controlled the wall towers - its base was 21 feet square made of sun-dried bricks 14.5 x 7 x 5 inches with reed layers every few courses. Next to it was a small building for troops quarters and a massive wall. It apparently was an office as documents were found inside - one dates from May 10, 68 BC. There was a fireplace in one room. Stein found an iron sword blade, wooden bowls, other records from 63-57 BC. The documents were made of wood strips tied together. Apparently this was an official archives. 256 out of 708 documents total were found in this one location. They were Chinese and published in Chavannes "Documents" and included tablets of calendars for years 63-59 BC for correct dating of official correspondence. Four documents were imperial edicts on border line and troops. They included the emperor's orders for establishment of an agricultural military colony at Tun-huang region and methods to be followed in construction of rampart for guarding the border. Stein deduced that the border construction likely began around 68 BC. The towers were a flanking defense section .The usupor Wang Mang (AD 9 - 23) likely let the wall section deteriorate and abandoned towers Tiv - Txiii. One document refers to imperial edict on the organization of three companies - Ling-hu, Yen-hu, Kuang-ch'ang. Ling-hu is mentioned in numerous documents so this company likely was the garrison. Other documents were orders to officers, one was an inventory of cross bow ammunition. Some were orders from Ta-chien-tu, the headquarters of the west end of the line. (see chapter XX) Stein also found bundles of brush wood piled ready for use as signal fires. There were records of fatigue duty including collecting fire wood.

Section IV - Last wall towers of the limes Tvic and (place 37)

- These were outlying signal towers three miles further west on flat top isolated mesas 150 feet above the low ground. The last tower west from Tivb had an unbroken view west and south. It was 20 ft square plus quarters and made of dried bricks each 14.5 x 7 x 5 inches to a height of 16.5 feet. There was a room 7.5 feet square on the top also unusually well preserved. The adjacent quarters contained an anti-room along the south wall of the tower. The tower originally had a coating of white wash plaster Here Stein found a Sogdian text and guessed that some of the garrison could have been Sogdian mercenary troops. He commented that it was very likely that garrisons along the wall included foreign officers and men, mercenaries hired for these garrisons. On May 7 he visited camp 172 with tower Tvid - the last of the line to the south west. It was 20 ft. square standing on a small hill and 30 ft high. The top was 13x14 foot brick parapet. This was the furthermost watch tower at the end of the limes. Sand dunes block the route between Nan-hu and these stations.

Chapter XVIII - The Westernmost limes wall

- Section I - Western end of wall to tower Tviii

- He returned to the western end of the wall itself, east of Tivb. (fig 149 - plate 36) Tower Tiii was the next one on the wall on the north edge of a plateau. The wall went three miles straight to Tvii. Stein at this point noticed a straight depression 8-9 yards inside the wall on a line made by centuries of patrols walking along the wall. He caught sight of this when the sun light angle showed a shadow of the depression, which was only 1.5 to 2 feet wide and 5 inches deep. But it was all along the wall and he then looked for this along other sections of the wall. It is remarkable that such a track would be so well preserved after 2000 years. South east of Tower Txiia 3/4 mile on an adjoining segment of wall he came to tower Txii (fig 181). The southern end of a gravel ridge next to a wide marsh depression bog on each side of Txiia. Here the caravan route to Lop and Lou lan passes on the south side of Txii. The watch tower was far from the line of the wall. The map 74 (D3) shows tower Txii well placed to guard this route and watch traffic and customs - immigration post and border control like the modern fort at Chia-yu kuan at Su-chou. But the main control was at Yu-men tower Txiv. The remains of Txii are fig 181 plate 38. The broken tower is 21 feet square and 18 feet tall built of bricks 15x8x5 inches. Layers or reeds placed between every 3 courses of bricks. There was an out building next to the tower. There was marsh east of Txii and deep northward. The line of the wall again near Txiia was on opposite side of the marsh for 2 miles to Txiii (fig 180) - It showed some damage - there were quarters next to it - 23 feet square and 24 feet high. The wall had plaster where it was preserved by debris. Plate 38 shows rooms 13 x 8. Fig 180 shows excavation and stairs to the roof of the quarters. Toghrak wood was in stairs. There were dated records from 56 BC and 5 AD for Tang-ku company and letters on silk. 70 yards south were more stacks of fascines. East of Txiii the ground dips into a depression 15 feet lower. The wall is preserved well in fig 176. For 200 yards the wall was still 10-11 feet high..

Section II - The towers Tixx and marsh sections of the Limes

Sections of the wall and towers from Tx to Txvii show excellent engineering to make the wall and towers conform to the terrain and make maximum advantage of natural obstacles. The line was far north of the caravan route between Lop and Tun-huang in this section. The terrain is very complex with a mix of marsh and depressions and gravel ridges. The gravel plateaus (plate 33) are like a coast line with tongues between bays and inlets. The depressions are due to erosion from drainage coming from the south underground but now there is only marsh from springs open water and beds of reeds. The marsh is impossible for horses or camels.

Section III - The Ruined watch stations Txi and Txiia

- Txi is fig 178. It is half a mile from the lake - plate 33. It is in detached location on a knoll over looking the marsh. North of Txi the plateau is cut by two ravines. To the east part of the wall remains toward the marsh. The tower is decayed. It is 16 ft high, and probably was 24 feet square. The layers are made of clay lumps 2 feet thick. There is a guard room on top. The west base side has a small room 21 x 12 feet with clay walls. There is an outer enclosure wall 75 feet in diameter. This tower likely sheltered people into the Sung Dynasty era. But the refuse dates to the Han era - a record of arms and armor - a calendar for AD 153. A text from a book of 48-33 BC.

These were the western most towers on the limes.
Tower Txiia was decayed but remained 8 feet high and 23 feet square. It was made of sun dried bricks each 17 x 8 x 5 inches with 3 layers and then reeds. There were quarters next to it. There were many fragments including Chinese records and Sogdian documents on paper and Chinese silk documents from AD 1. There was a personal record from AD 20-21. Records of the Kuang hsin company from the Wang Mang period 9-23 AD. All towers west of the Jade Gate Txiv were from the Wang Mang era or before. The route to Lop and Lou-lan passed near Txi, where there was the last drinkable water to Togruak-bulak. The marsh east of Txi cannot be crossed. The wall, again, appeared on the edge of the marsh toward Txiia - fig 177

Section IV - The Early Sogdian Documents from Txiic and their paper

Ts'an lun invented paper in AD 105. The documents Stein found dated before AD 153 indicating a rapid expansion of paper use. Also the date shows the garrison was able to use paper. The writing was in Aramaic script and an Iranian language There was a document - a letter - in an early form of Sogdian cursive - Aramaic intermediate to Sogdian-Uigur.

Section V - The Watch stations Txii and Txiii

South of Tower Txiia is a set of mounds of reed fascines - laid out probably for use in signal fires. Or possibly they were for use in repair of walls. South east 3/4 miles from Txiia on a adjoining segment of the wall he came to Tower Txii (fig 181). It was on the southern end of a gravel ridge next to a wide marsh depression with a bog on each side of Txiia. Here the caravan route to Lop passed on the southern side of Txii far from the line of the wall. The ancient route to Lou lan crossed here also. (Plate 33 - Map 74 D3) show Txii well placed to guard this route and watch traffic from well behind the line of the wall. Txii was a road side traffic and customs control and immigration post for border control like the Modern Chi-yu-kuan at Su-chou. But the main control post was at Yumen at Txiv. The remains of Txii (fig 181) Plan 38 show the broken tower 21 feet square and 18 feet tall with bricks 15 x 8 x 5 inches in layers with reeds after every three courses. There was an outbuilding next to the tower. The marsh east of Txii was deeper to the north.

The line of the wall appeared again near Txiia on the opposite side of the marsh 2 miles from Txiii (fig 180) which had some damage. But there were quarters next to it made of the same bricks - Tower was 23 feet square and 24 feet high. Its wall had plaster where preserved by debris. Plate 38 shows rooms 13x8 feet. Fig 180 shows excavation and stairs to the roof of the quarters made of Toghrak wood. Records with dates 56 BC and 5 AD were recovered from the Tang-Ku company with letters on silk.
Seventy yards to the south were more stacks of fascines. East of Txiii the ground dips into a depression 15 feet below. The wall is well preserved (fig 176). The wall for 200 yards is 10-11 feet high. Next to this wall is another depression track made by foot patrols.

Chapter XIX - The Jade Gate Barrier.

Section I - Ruined site Txiv

He describes tower Tvii to which he returned along the wall after first sighting it on his first way to Tun-huang. (Plate 38 - photo 168) The tower was 23 feet square made of bricks 14x 7-8 x 4-5 inches. And it had been white washed and plastered. The quarters next to the tower had 13 coats of white slip of AD 8. The Jade Gate at tower Txiv had none west of it later than Wang Mang in AD 9-23, which Stein believed meant the western end of the wall may have been abandoned then. This section of the wall is a division of the line and the caravan route it protected. The section is a mix of gravel plateaus and salt bogs and marsh. (maps 74 and 78) one can see the depressions. From top of Txiv the main terminal of the course of the Nan-hu drainage could be seen to south east to its junction with the Su-lo Ho. The reason Txiv was placed here (fig 179 - 183) was it was on top of a neck of raised ground like an isthmus across the depression. The passage of traffic between the reed marsh to the NW and salt bog to SE narrowed the route to-from Lop as it followed this passage and was guarded by this fort. It was a strong position with tactical advantage - two knolls were occupied by ruined forts on mesa terraces. Called Hsiao-fang. From the top of Txiii one can see all the watch towers from Txi to Txix and could at night see signal fires much further - probably from Tix to Txxii or 30 miles. This location also has grazing area and water.

The fort is massive in construction - at Txiv thick walls of stamped clay. (fig 183 - 184 - plate 40). An opening in the middle of the north wall is 13 feet wide but it was not the real entrance which is on the west wall and only 8 feet wide. The interior is 54 ft square. There was an inscription on a medicine case of Company Hsien-ming. This shows the use of medicine cases as well as the method of wood covers. There was a wooden label - Hsien-ming company of Yu-men - 100 bronze arrow heads of Ming type - attached to a container bag or box. There were many more items at Tviii due to the burial under a roof that caved in. Stein spent considerable effort excavating in a large mound north of Txiv. He found records from the Wang Mang era 9-19 AD and 48-45 BC. The hillock was 100 yards east to west. He found a tunnel which he thought might be a dungeon for prisoners. There were some wood records from 96-94 BC and 14 AD. He found some T'ang era coins 681-27 AD and Buddhist text from the T'ang. The last period before the Tibetan invasion closed the western route. But Hsuan-tiang (645-64) passed here in 645.

Section II - Position of the Yu-men headquarters.

He decided the hillock was a Buddhist shrine built over the remains of an abandoned Han place. Many remains of all sections lay in thick layers. Many written records dated mainly to two periods 94 BC when the limes were first built - AD 14. There were 5 documents AD 9-19 - 48 BC- 45 BC and AD 4-15. So the station was busy during this period up to Wang Mang. The Yu-men (Jade Gate ) - was mentioned in 10 documents. Chavannes published 672 of the documents from the total on the limes and the station is mentioned in 24 so half of these were documents at this site. This was administrative headquarters of the west end of the wall and during this century was the chief frontier post. The Emperor Wu-ti to the close of the Former Han Dynasty. Stein found a second line of wall from north of Txiv passing close by to the west and continuing S-SE toward Nan -hu. This wall was only 5 feet wide and starts SW on the edge of the marsh south of Txv and Txva. Due south across gravel - to the west of the fort it went 3.5 miles on to a watch tower (plate 40). The secondary wall connected Txiv to Nan -hu and its Yang barrier. Towers Tix, x. and marsh sections of the limes. Stein discusses this also in his first visit. Stein discusses the wall and towers Tviii to Tix. This was the best preserved of all the watch towers. It was located on a knoll above the steep eastern edge of a gravel plateau. Here the wind driven sand from the east could not attack the tower, so there was less erosion. Tower ix on a knoll 60 feet above the ground plateau had a view west and east but not such a good view to the north (dead ground close in). The area was not easily viewed from Tviii or T ix due to the deep clay terraces. So raids could approach the wall here. To protect this weak section they built a tower, Tixa, north of a nullah 2.5 miles from Tviii. Stein returned from the SW to visit Tixa - an outlying picket watch station north of the wall. Tixa was 18 ft sq of bricks 18 x 9x 4.5 inches with reed layers between every 5th brick course. The tower was enclosed by small defense wall 34 yards square. The wall from Tix north east is discussed in Chapter XIV. Tower Tx (fig 174) The wall continues to the edge of salt lake. Here the line extends east to Khora-nor Lake across marshes and small lakes south of the Su-lo-Ho. At the east side of the lake is Tower Txxiiib. The line of the wall was designed with careful purpose to supplement the natural defenses of the river and to save labor in construction and in the effort of guards. (plate 33). Stein shows the topography including the coastline of the depression to the north from erosion. The marsh was quite impassable for horses or camels. Towers were on commanding ground - Tx to Txvii over an 18 mile stretch. The wall was on each section of firm ground and able to be passed at the edge of the marsh. The engineers used all the natural obstacles. From Txvii to Khora-nor at Txxiic was 18 miles - all south of the Su-lo Ho. Two sections - Txix - xx and Txxiib - c .

Section III - Wall between Yu-men and Yang Kuan

- The transverse wall was not strong and not well built or at the original period but later. It was a possible inner line for control rather than a defense wall or was possibly to reduce the length of the border line needing coverage -it was 24 miles long. This wall may mean the western section of the line was abandoned in mid-century. When stations on the SW flank were abandoned the need came for another wall. The transverse wall from Txiv to Nan-hu - Yang Kuan was made then. After 16 AD the Hsung-mu (Huns) made a big attack on the northern border and western regions were broken up. In AD 23 power of the Chinese in the western regions collapsed. After the beginning of the Later Han (25 AD) the situation continued. The Huns controlled from AD 58-75 and reached Tun-huang and Kan-su area. Chinese power was again expanded by Pa Ch'ao (73-102 AD) and caravans continued west to Lop. Watch towers Txi and Txiia - (fig 178 - plate 33) - westernmost side of the line. Txi (fig 178) is half a mile from the lake (plate 33) in a detached location on a knoll over looking the marsh. North of Txi the plateau is cut by two ravines. Part of the wall to the east goes to the marsh. The tower is decayed due to water. It is 16 feet high and probably 24 feet square. It is composed of clay lumps 2 feet thick and has a guard room on top. The west side has a guard room 21x12 feet with clay walls. There is an outer enclosure wall with circumference of 75 feet. This was a shelter for people well into the Sung dynasty. But the refuse from Han era shows records of arms and armor - a calendar of AD 153 - text from a Chinese book of 48-33 BC. All towers west of the Jade gate - Txiv date from Wang Mang AD 9-23 or before. The Route to Lop and Lou lan passed near Txi which had the last drinkable water to Toghrak-bulak or to Tivb. The marsh east of Txi cannot be crossed. The wall again is on edge of a marsh toward Txiia to tower Txiia (fig 177). The wall takes a sharp turn to SE to edge of swamp. Tower Txiia is decayed - 8 feet high and 23 feet square of sun-dried bricks each 17x8x5 inches with reeds between every 3 brick courses. There are quarters next to it and many fragments - Chinese records, Sogdian documents on paper and Chinese records from AD1 on silk. There is a personal record from AD 20-21. He shows the wall at the lake shore on a knoll with wall on both sides of the lakes - but towers very close to the lake were reduced to lumps of clay and reeds due to water action from underneath. Stein found more records, arms and equipment - a calendar of AD 153. Txi was the last place caravans could get water until reaching Lou lan. Tower Txiia (plate 39). Stein found Sogdian document on paper and Chinese document on wood from AD 20-21. Tower Txiv - the Jade Gate before 137 AD - a paper mentioned it in 105.

Section IV - Txiva and refuse at Txva

- The wall reappeared on the NE edge of a marsh SW to NE across the plateau. Tower Txiva was on a ridge where the wall crosses the ridge and was 50 yards from the western slope of the ridge and 15 feet high and 24 feet square made of the same kind of bricks. (fig 188 0 Plate 39). There was a room on top 8 feet square. Sogdian documents - discusses the language and type of paper. Ts'an kun invented paper in China at AD 105. These documents before AD 153. This shows the last date for the garrison and the rapid use of paper this far west. There is Aramaic writing and Iranian language. One document is a letter in an early form of Sogdian cursive and Aramaic intermediate to Sogdian Uigur.

Section V - Relic of the Ancient Silk Route - stations Txii, Txiii (plate 39)

Stein found stacks of fascines in neat rows ready for use as fire signals. (Fig 181 - plate 33) Tower Txii was a road side post not part of the wall defense itself. The section of the wall and towers Tx - Txvii show excellent engineering to make the wall and towers conform to the terrain and make maximum advantage of natural obstacles. The line was far north of the convoy - caravan - route between Lop and Tun-huang. The terrain was a very complex mix of marsh and depressions and gravel ridges and gravel plateaus like a coast line with tongues between bays and inlets. (plate 33). There were depressions due to erosion from drainage coming from the south. Now only marsh from springs and open water and reed beds.

Section VI - New Route to the North -

Stein asks why there was such a huge refuse mound at tower Txva with 117 documents while at Txvi there were 80 and at Tvib 228. During reign of Wei ho - AD 239-255 there was discussion of routes west of Tun-huang. Previously there were the two routes and now there were three - 1 the southern route - 2 the central route and 3 the northern route. Route 1 left from Altun Toghrk on the southern slope toward Miran - Route 2 went to Lou lan - and Route 3 went from Yu men kuan via Heng keng to Kara Khoja in the Turfan basin and from there westward to Kucha. This is a route via Hami still a major road today. Stein describes this new route. It started at Txva and crossed the Su-lo Ho easily. (plate 33) He guessed there might be watch stations in the northern desert along this new route.

Section VII - Great Magazine of the Limes -

The wall between Txv and Txvii is a continuous link with notable ruins in this part. At Txv the ruined tower on a small hill 20 feet high at the east end of a basin was brick as usual 19 feet high with a small room adjacent. From Txv the wall was traced but was low across gravel plateau to Txiv for 1.5 miles. This one was on southern edge of a dry basin with a good view north and east. It was built of sun-dried bricks 14 x 7x 4.5 inches with reeds laid between each 3 courses. The height was 13 feet (plate 39) and it was 24 feet square. There was a 8 foot square guard room at the top. 50 yards south east there were stacks of reeds for signaling. There were 10 inscribed slips and records in the room dated from AD 68 to 77 plus one of 86-74 BC. From Txvi the wall went on gravel terrace and shallows to Txvii at the western edge of a basin crossed by the Su-lo-Ho with lakes and marshes. This tower was 22 feet square of the same brick but broken at 10 feet height. There were quarters next to it. Refuse records were dated 58 BC. The wall then continued east into the marsh a mile away to remains of a broken tower Txviia on a small hill 50 feet high. Then a lagoon of 3 miles to Txix with bogs to the north. 4 miles long by 2 miles wide. (Fig 186). The river was overlooked by Tower Txviii.

On April 2nd visit Stein found a high structure just behind the line of the wall between Txvii and Txix. This wall was 560 feet east to west with three large halls each 139 feet by 48.5 feet inside and an adjoining length wise face ran 5.5 feet thick. (plate 41) The building was on the northern edge of a clay ridge separated by 65 feet wide cut and all 15 feet above the adjacent ground. The building walls were still 25 feet high with no large windows, instead there were triangular openings 3 feet high in one row level with the floor and another at 14-15 feet up on the wall. These were used for ventilation. (fig 182) The right side shows the least damaged section of the enclosure. There were SW towers - left of fig 186 20 feet high. A NW tower is in fig 182. Plate 41 is a plan of the building and it shows 3 towers with traced walls built within the inner enclosure. These were posts to guard the contents - not to protect from outside. Besides the inner enclosure the remains of two mounds parallel to the south and north walls - the northern one 80 feet outside and the southern one 100 yards away. There was no trace of east- and west walls. Only one wood record dated 52 BC was found. Stein believed this was likely a granary as there were several records about grain release and several written works of the T'ang era mention this place.

Section VIII - Lake Section of Limes - Txix to Txxiii

- The location of the magazine was secure by a wide marsh north so no wall was needed. To the east the Limes reappeared on firm ground between the marsh and a lake 3 miles long and 1.5 miles wide. The gap between the marsh and lake was 1 mile wide but had 2 watch towers, Txix and Txx, and a connecting wall (fig 190) on top of an isolated steep clay hill 100 feet high. (plate 33) Tower Txix (plate 36) was at a naturally strong point at the northern end of a steep ridge with full view of the depression of the Su-lo Ho. It was made of the usual bricks, 22 feet square with a guard room at the top 8 feet square and on the eastern side a room full of much refuse including a label for a bag of 100 bronze arrow heads. Watch tower Txx was a ruin. (fig 190) along the wall 3/4 mile toward the lake on a hill north east at the end of a mesa 70 feet high. It was of brick 13 feet high with adjacent rooms in which many wooden records were found. The wall seen at NE end of the clay ridge on which Txx was located and the wall went to the edge of the lake - a good barrier -.There was no wall on the south shore for 7 miles. But the watch towers from Txxi to Txxia were on high mesas. Txxi was decayed tower at western end of the top of a steep clay ridge 80 feet above the depression. It was north of the caravan track and 3 miles SE of Txxd. Txxi had 10 feet height remaining made of bricks 17x8x5 inches. There were two apartments on the east side. Three miles ENE from Txxi on top of a small clay rise 80 feet high was ruin of Txxiia built of clay clods and Toghrak twigs - 13 feet high with no remaining quarters.

Going around the edge of the marsh NE Stein found firm ground (map 78 A3) and the north end of a well marked plateau spur from the Kuruk-tagh natural barrage that holds up the Su-lo Ho in the Khara-nor basin. The barrage was 2 miles wide. The river here was fordable and the line of the wall crossed the barrage from the western shore of Khara-nor to a wide marsh on the other side. But the line of the wall here was much ruined, only barely seen between towers Txxiib and Txxiic. The wall was 1.5 miles south of the Su-lo Ho. Tower Txxiib was decayed brick, 13 feet high, on a 20 foot high hill. One of the wooden records recovered here dated from AD 12. Another was a list of arms of the Ch'ing-Tu company of P'eng'-wang, controller of the wall.

Tower Txxiic was at the highest point on an isolated clay ridge away from the shore of the Khara-nor. The ruin was built of stamped clay and layers of reeds but decayed down to a height of only 10 feet. Stein found two dozen records and silk there. The earliest record was a daily unit duty roster mentioning guard, cooking, making bricks and the like. Looking from Txxiic across the Khara-nor to the east Stein could see only Txxiid about 5 miles away on the southern shore. This ended his visit as he had to return to Tung-huang.

However, he found two more towers - Txxiii and Txiiia - to the south on the way to Tung-huang, at the northern end of a plateau toward the Khara-nor while skirting around the south shore. Txxiii was on a narrow, steep crest 110 feet above the plain with a good view north and east. From there he could see Txxiid. Txxiiia was lower down. Txxiii was 16 feet square and only 10 feet high made of lumps of clay and fascines of wood between each 5 layers. About 150 yards north and down the ridge was Txxxiiia, overlooking the caravan route as it wound around the ridge in a defile and then went due west toward Yu-men and southeast toward Tung-huang. The tower was made of bricks 14x7x4 inches and reed straw and was 16 feet square, 8 feet high, with a guard room on top.

Chapter XX - History and Records of the Tun-huang Limes (This chapter is so important that I have created a separate file for it - see.

Section I - The extension of the "great wall' beyond Tun-huang

Stein found that the ancient Chinese records were a great help for him in learning details of the wall construction. Plus he was well trained in both archeology and topographical surveying itself. The oldest Chinese records were well known manuscripts and records described by M. Chavannes in his book Documents Chinois. (Stein did not read ancient Chinese). In 214 BC the Emperor Shih Huyang-tu-ch'in linked defensive walls from the old kingdoms. His wall from Shan-ha Kuan to the gulf of Lio-tung went as far as Lin-T'ao now in the Min prefecture southern Kan-su - 110 miles south of Lan-chou. A century later the Great Wall extended NW. A new wall going 1000 miles further west to the edge of the Tarim basin was to protect expansion into west-central Asia. The Chang Ch'ien mission of 138 - 126 BC under Emperor Wu-ti disclosed the western area. The purpose was the secure allies against the Hsiung-nu in Yueh-chik and from the Indo Scythians. But the mission showed a new idea of commercial relations. If the Chinese could open trade routes via the Tarim Basin to Farghana - Sogdiana and Oxus region they would profit. Chan Ch'ien's report to the emperor showed that communication through the area would be at the mercy of the Hsiung-nu from the north and the Ch'iaang Tibetans from the south. But passage was narrow between these two forces. However, at Tun-huang the Tibetans and Huns could be in contact. Then to the west was absolute desert of Lop that protected the route from attack beyond Lop. There could be two great routes.
Section II - The Tun-huang Limes since its construction
Section III - Main Features of the Remains of the Limes
Section IV - Military agricultural colonies
Section V - Officers and soldiers of the Limes
Section VI - Service Conditions of Life on the Limes

Section VII - List of antiques from Ruins of the Tun-huang Limes

This is a catalog showing the impressive amount of valuable artifacts that Stein was able to take to London and India. One has to wonder at Stein's attention to detail as he meticulously made a written entry in his notebooks each day describing exactly where and how each item was found.

Chapter XXI - The Caves of the Thousand Buddhas

Stein recounts his visit to this remarkable location and its historical connection with the Buddhist monks he follows.
Section I - General Description of the site
Section II - Inscriptions at the Caves of 1000 Buddhas
Section III - Wang-Tao - shih and his Restored Temple

Chapter XXII - Exploration of a Walled up Hoard

Section I - First Opening of the Hidden Chapel

Besides the open cells with their marvelous frescos and statues Stein was let in on a secret that he managed to exploit to world adulation except for Chinese condemnation - namely, he secretly took back to England and India a fabulous trove of early medieval manuscripts.

Section II - Finds of a Polyglot Library

Ch'ien-fo-tung is same as Caves at Tun-huang. The area was conquered by Tibetans in 759 AD and by 766 AD Tibetan power extended over all Kan-su. Tun-huang was the gateway for further conquest. In the 8th century the local governor, Chang I-ch'ao, in 850 broke the Tibetan power and returned allegiance to China. In the chapel are two Imperial edicts from 851. That of June 23 851 grants honorific and ecclesiastical title to Hung-j'en as chief Buddhist religious of Sha-chou and to Wu-chan a Buddhist control of Sha-chou. The second edict addressed to Hung-j'en conveys the emperor's eulogies for loyalty and help in submission of Chang I ch'ao. The point was how the Buddhist monks exercised influence over the wild Tibetans. Chang I-ch'ao's submission enabled the Chinese to re-establish partial suzerainty over the area. But with T'ang downfall in the 10th century China again lost control. In 938-42 AD a Chinese mission from China to Khotan passed through. By then Tang-hsian tribe gained power. A century later they raised the Kingdom of Hsi-hsia or Tanguts. By then the Uigurs were at Kan-chou. Tun-huang was the strategic crossroad of N-S and E-W routes and key for the Tibetans. Also, since Sogdian language Buddhist writing was found the Sogdians from Samarkand must have had influence. Stein also found Manichean texts. The importance of the find was increased by its exact dating. The walled library was closed up around 1000 AD before the Tangut conquest and was not discovered by the Buddhist monk caretaker, Wang - tao-shih, until 1900, which greatly enhanced the value of the find. Stein took 7 cases of manuscripts from the cave plus 5 cases of paintings and textiles.
Section II - Acquisition of Manuscripts and Art Relics

Section III - Subsequent Investigations of the Deposit

Chapter XXIII - Pictorial Remains from the Thousand Buddhas

He managed to photograph with 1911 equipment many of the frescos and illustrations on manuscripts - Many are printed in his book and I have copied a few of them

Section I - Recovery and Study of the Paintings

In this section Stein describes the methods used at the British Museum to preserve and analyze the paintings. Many depict Mahayana Buddhist beliefs in a Chinese style. There are paintings dating from 864 - 983 AD. Tun-huang was controlled by Tibet prior to 850 while T'ang power was weak.
Section II - Time and Milieu of the Production
Section III - Arrangement, Materials and Technique
Section IV - Scenes from the Legendary Life of Gautama Buddha
Section V - Buddhas and Bodhisattvas
Section VI - Lopapalas and Vajrapanis
Stein gives detailed description of Buddhist iconography. scenes include - Buddha mother Queen Maya. Prince Siddhartha - many Bodhisattva's. Stein was especially interested in the relation of these paintings to the art style of ancient Gandhara that that he had studied in India. There were many paintings of the Buddha. And paintings of Bodhisattvas - popular worship, One named Avalokitesvara the protector of mercy, -a 4 or 6 armed figure. Another was Vajrapani - another Manjusri and another Kssitigarbha with the shaven head of a monk. Avalokitesvari was guider of souls and held a willow branch. Four Lokapalas - guardians of regions=warrior kings with gorgeous dress and arms - The guardian of the north is Vaisavana with pike or halbard - The guardian of the south is Virudhaka with club - The guardian of the west is Virupaksa with a bared sword - The guardian of the east is Dhitarartra - with bow and arrow. Section VII - Divine groups and Assemblages
Section VII - Pictures of Buddhist Heavens
Section VIII - Miscellaneous Paintings, Woodcuts and Decorative remains.

Chapter XXIV - Textile Remains and Manuscripts from Ch'ien-fo-tung
Detailed analysis of the individual items and their overall importance
Section I - Decorated Textile Relics - their Material
Section II - Chinese Designs Decoration of Textiles
Section III - Designs of Sassanian Type and their Imitations
Section IV - Manuscripts from the Hidden Library in Brahmi and Chinese
Section V - Manuscripts in Tibetan, Sogdian, Turkish
Section VI - List of Sogdian and Turkish Manuscripts

Chapter XXV - Cave Temples and antiques of the Thousand Buddhas
More discussion of this remarkable place. It is now repaired and established as a major tourist attraction. Tun-huang even has an airport now. One can see much of this by doing a Google maps search.

Section I - Decorative Art in some Ch'ien-fo-tung Shrines
List of Paintings woodcuts textiles et cetera.

Volume III

Chapter XXVI - The Marches of old Kua-chou

Section I - Oases of Kua-chou and its Historical Importance

Stein left Tun-huang on 13 June heading east toward An-hsi. He meanwhile packed up 12 cases of art works to send to India. This part of the travel required 3 day's march - 55 miles along the foothills of the Nan-shan. He reached Kua-chou (maps 78 and 81) following the main caravan route across the desert. Then it was 15 more miles to An-hsi between the Nan-shan and the Su-lo Ho. He passed ruined walled villages and towns, most of them relatively new and ruined during the mid-19th century revolt. Kua-chou-ch'eng was a ruined city with walls of 1/2 x 1/3 miles circuit with gates and temples. Old town An-hsi was the ancient main headquarters for the Han era - its walls were crumbling. It was the last staging place on the road to Hami which crosses the Su-lo Ho there and then goes along the Pei-han stone desert hills for an 11 day march. Stein remained there for 6 days organizing his crew - until 19 June. The Chinese occupied Hami in AD 73. The route west through Hami was longer than the southern route but avoided the worst of the desert.

Section II - Old remains near An-hsi and Yumen-Kuan

Stein had already found traces of the Han wall to 35 miles west of An-hsi. On June 20th he found two more old towers - one 4 miles SW of An-hsi was 18 feet high and 14 feet square, like the others. Here he found Han era pottery. The second tower was a mile east and 22 feet high with 17 feet square base. The line of the wall was visible toward the east. He revisited An-hsi in October 1907. There the wall intersects the road from Tun-huang to An-hsi with 5 small towers, likely stations for reception of visiting officials. To the east the wall disappeared, but there was one more watch tower, 8 feet high and 5 feet in diameter. On 24 June at Ch'iao'tzu the wall again appeared 7 miles away to the ESE. The wall between Tun-huang and An-hsi was well away from the river. Stein found a walled town ruin 15 miles SSE of An-hsi, This was the earlier location of the town. The enclosure was 600 yards square with a wall 15 feet thick. Wind-blown sand had cut the walls mostly on the East and West sides. The modern wall of An-hsi also was eroded mostly on the eastern side where the wind was strongest. Map 81 shows the river and An-hsi and the caravan route north-west along the Pei-an-zu.

Section III - Ruined site at Ch-iao-tzu

On June 24th 1907 Stein departed to the south into the high mountains, the Nan-shan. His exploration method was to go into the desert during the winter and the high mountains during the summer. He describes this in Desert Cathay and in Innermost Asia. In the mountains he found a walled town and temples sited to use springs. (Map 83). The remains of a town had walls 5 miles long E-W and 3 miles long N-S. Another walled town is Plan 46 - So-yang-Ch'eng. There were ruined towers and other walled places. There was a massive wall of stamped clay 20-30 feet thick with round bastions at the corners and rectangular bastions along the walls. It had 2 remaining gates on north and west sides with outworks for protections. The SW corner walls formed a receding angle with a redoubt and inner enclosures. The northern wall of 670 yards long and southern wall of 493 yards. The east and west walls were 650 yards long. There were 2 walls facing the east side, The inner one was 200 yards from the outer, both broken by the wind. (fig 185). He found coins dating from the T'ang and Sung dynasties, and one from the Ch'in 1156-61. In the SW corner was a massive building with pottery from T'ang and Sung. There was a Buddhist stupa at Su-yang-Ch'eng 40 feet high and 27 feet in diameter from the Sung era 11-13 century.

Section IV - Grottoes of Myriad Buddhas

On 29 June Stein was moving into the mountains near Ta-shih village on the road from An-hsi. This is important route into Tibet. He came to the series of shrines in caves similar to at Tun-huang, but less extensive. Here were mural paintings. About 15 miles further on he found a stone wall built as a defense in a gorge protecting from the south. There were more caves and paintings at Wan-fo-hsi.

Chapter XXVII - Northwest frontier of Kan-su

Section I - To the Gate of the Great Wall

From July 3rd 1907 Stein left the cave of the Myriad Buddhas to explore the western Nan-shan for 2 months, covering 24,000 square miles - maps 82 to 94 - see also Desert Cathay. He found another ruined fort, Shih-pao-ch'eng (map 84) - a walled enclosure about 180 feet square with a keep in one corner. It was 50 miles west of Ch'ang-ma oasis and 7000 feet above defile on the Su-lo Ho in the Richthofen Range where it breaks through toward the Yu-men hsien. He found two more posts guarding the valley route (map 86) from Tibet. He explored the area, passed south of the Richthofen Range, and crossed the divide that separates drainage of central Asia from that flowing into the Pacific. He found the gate Ch'ia-yu-kuan, 25 miles from Tu-tafan River - This is he border gate of the Ming Great Wall and is restored today for tourists. It is located in the narrow valley between the Nan-shan to the south and the Pei-shan desert hills bordering on Mongolia to the north. To the north the terrain is steppe, stone gravel with many miles separating the valley from water. The valley is only 12-15 miles wide. Stein could see the fortress gate and line of the Great Wall still some 20 miles away to the east when he was at an altitude of 8000 feet in the Nan-shan. He descended to explore the Su-chou district.

The Ta-han-chuang fort was still occupied to watch the route along the mountains. It had massive watch tower and walled enclosure. On that day Stein reached the high road 4 miles west of Ch'ia-yu-kuan with mountains to the north and south. This was the route west to Tun-huang for 2000 years. He examined the Ch'ia-yu-kuan fortress walls and high towers and inner and outer gates. The valley here is crossed by the wall - south into the Nan-shan. The wall to the north was hidden by a ridge. He noted this is the ideal place for a barrier wall , with the valley only 8 miles wide. He noted a second line of wall and towers far away to the north east. (Map 88) - shows the wall SW and NE. It is not an extension of the wall at Ch'ia-yu-kuan. The gate fortress is medieval (not ancient), Emperor Ch'ien-Lung 1736-96 ordered it. It is typical of medieval Chinese fortresses. There are three successive gates through the massive bastions and inner defenses and a single road street east to west. Since Ming times it was a border control post -immigration and customs - not for strong military defense.

Section II - The walls of Ch'ia-yu kuan

Stein visited the wall to the south and north west. He surveyed the NW section of the wall where it starts at the NE corner of Ch'ia-yu-kuan and goes unbroken along the eastern foothills of the precipitous hills. The wall has a clay rampart 11 feet thick and 12 feet high with a parapet 4 feet high. Inside the wall are watch towers an average of 1.5 miles apart with a tower near the gate. (plate 47). The tower is 36 x 33 feet square built of layers of stamped clay each 4-5 inches thick. The top of the tower has loop holed walls 6 feet high and watch room. There is a double line of foot holes on the inside wall to help the watchmen climb using a rope. The brick built wall is at the same height as the main wall but 1/2 the thickness and forms an enclosure around the tower. There was a ruined quarters for garrison. Each tower was a defense point defended independently. Outside the line of the wall (that is north west of it) at a distance of 1-2 furlongs there were three detached towers - very modern looking and massive - on spurs of hills westward. The towers were 40 feet wide at base and 30 feet high with square entrenchments around them as outer works, guarding ravines. The wall at a mile from the 3rd of these reaches a hamlet, Huong-tsiao-ying at the mouth of Hao-shan-k'ou gorge. Then the wall crosses this half mile of rock spur on the left bank to 200 feet above the valley and beyond it the precipice is unscaleable. The wall ends at the mountain range which rises to the NW to 9200 feet 10 miles away. The ridge has deep valleys and steep cliffs forming an impassible barrier. Also the defense of the mouth of the Hao-shan-k'ou valley lies outside the main wall. Where this approaches a farm, a branch wall goes at right angles SW up a steep spur by the river and ends were a cliff is very steep. Then at the end of the valley a mile away, another wall of 200 yards goes across the valley with loop holes in it. The parapet is on the SW up the valley guarding a small valley another 100 yards up the gorge. Here is another older wall made of stamped clay 11 feet thick and 11 feet high with a parapet 2 feet high. There is still another wall at right angles to the NE. A watch tower is located north of the gate 1.5 miles away. The earlier one is of stamped clay, 8.5 feet thick at base and 8 feet high plus a parapet.

Section III - Su-chou and Central Nan shan

Stein reached Su-chou on 22 July and stayed 6 days. This was the Emperor Wu-ti's headquarters in Han times. Stein spent most of August 1907 in the Nan-shan south of Su-chou and Kan-chou rivers. He found many wide upland valleys and marveled that they were not being used, but noted that the Chinese didn't want nomads there. He reached the head waters of the 4 rivers at 11,000 feet elevation. Maps 88-89-92 show the Pei-ta Ho and Hung-shi pa Ho valleys between the Richthofen and Alexander III Ranges and the To-lai-shan Range. Su-chou is well irrigated and fertile ground. The descriptions of his travel are captions with the maps.

Section IV - From Kan-chou to Chin-t'a

Stein spent 6 days from August 27 into September at Kan-chou, 5100 feet elevation. It is a fertile area where the river bends sharply north. Marco Polo lived in Kan-chou for a year. Stein found a wall going on the left bank of the Pei-ta Ho or Su-chou river NE of the oasis. Starting again on the right bank at Ai-men it went NE to the desert range at the foot of the hills. The high road to Kanchou shows on map 91. Then the wall goes SE parallel to the line of the road on the right bank of the Kanchou River, to near Kanchou city. The Great Wall of Kan-su goes to the east and north of the road. The Han era wall did not pass the bend of the Su-lo Ho. The other walls are much later - Ming walls. Stein returned to Su-chou on the road along which he found more Ming era towers or forts along the right bank of the river, one was 60 yards square (fig 252). The Ming wall shows on his map 91. At Kao-T'ai the road passes through a narrow space about 5 miles wide between the Nan-shan and hills on the other side of the river.

Section V - The Han limes from Yu-men-hsien to An-hsi

On his way back from Su-chou to An-hsi Stein wanted to search for remains of the Han era wall north of the Su-lo Ho along its eastern section. Stein found the walls in 1914. His return in 1907 to An-hsi was a busy 12 days. During this period R. B. Lal Singh was doing more detached surveying.

He also examined the tower defenses on both sides of the narrow section of the valley. The wall from SW to NE has gaps and watch towers at 1.5 miles. Stein visited two that he found to be 25-26 feet square and 20-22 feet high of stamped clay layers. They had repairs made of sun dried bricks. The wall continued east. The line of the wall is on map 88 with a side wall to Yeh-ma-wan and sections toward Su-chou. A fort at Yeh-ma-wan is at the bend of the wall with a guard gate for a road to Hami. From Yeh-ma-wan the line of the wall goes SE and around a marsh for 7.5 miles. It is north of Su-chou. Stein visited on 26 July 1907. In 1914 going from Su-chou to the Su-lo Ho on this route he again found the wall. In September 1907 he saw that the wall goes on the left bank of the Pei-ta Ho or Su-chou River near the NE end of the Su-chou cultivation oasis. It starts again on the right bank near Ai-men and goes NE to the desert hills The same wall at Su-chou then goes east along the edge of the Su-chou cultivation area.

Stein made very clear that the well known line of walls following the route Kan-chou to Su-chou is not a border defense or from the Han conquest of western Kansu of Emperor Wu or an extension of the Great Wall. The original Han wall did not pass from the bend of the Su-lo Ho SE toward Chia-yu-kuan and Su-chou, but a really unbroken chain of remains show that it first continued east past the oasis Hu-hai-tzu or Ying-pan north of the An-hsi to Su-chou road. It then led far away toward the NE beyond Mao-mei on the united course of the Su-chou and Kan-chou Rivers .The new wall is much later, from Ming Times and its purpose then was exclusion and seclusion, to keep people in as well as out. As a defensive line it was close to the cultivated area at the foot of the Nan-shan and the high road had no purpose in T'ang era when China controlled Central Asia. Later with Tibetan and Turkish conquest of Central Asia it also had no use. When the Uigurs and Tanguts controlled Central Asia also the wall had no purpose nor did it during the Mongol Era.

The historical record shows the policy change in Ming Dynasty around 1368. They wanted to close China to western influence. This is the reason for the new wall - unlike the Han wall, the Ming wall ignores strategic advantages in locations. It is a police line. The wall ends at Chia-yu-kuan, a place where in Ming times caravans from west were carefully checked. The Ming era supported by archeology also the state of cultivation then made a cheap wall possible. The Ming built a wall from Pei'ta Ho north to Hao-shan-l'ou and it is different and more recent. Some time after the Manchu takeover - maybe Emperor K'ang-hsi (1662-1723) - when policy changed again - the Ming wall decayed and was abandoned here. The Ming wall was replaced by a new fort wall at Ch'a-yu-kuan. The new wall had a military purpose, defense from tribes. The new wall had an advantage. Su-chou was the main military depot and headquarters for expansion again. At Marco Polo's time the wall and fort at Ch'a-yu-kuan was not yet built.

Chapter XXVIII - To Hami and Turfan

Section I - From An-hsi to Hami, Hsuan-tsang's desert crossing

Section II - The Historical Role of Hami

Section III - The ruins of Ara-tam and Lopchuk

Section IV - Visits to Ruined sites of Turfan

On November 13th Stein started down the narrow valley (map 59) to the ruin at Chong-hassar - Hassar-chaki at the east end of a marsh - salt lake - in the Turfan Depression. The oasis at Lukchan was known from Han times at 50 feet above sea level. The walls of Chong-hassar were 6 miles from Besh-tam. It was an oblong fort with an outer enclosure made of an irregular wall on a low terrace and built of sun-dried brick (plate 50). The NE corner of the oblong fort that measured 200 feet by 150 feet outside had a massive keep. There were small, vaulted chambers and casements. The rooms were from 10-16 feet long and 6.5 to 68.5 feet wide. The buildings were in Turfan style using the vaulting found in common usage. It is 360 feet below sea level and very hot in summer. There was a Buddhist shrine on the SW wall. (fig 266). The outer wall on the SE side was 24 feet high and 6 feet 4 inches thick, made of bricks 18x8x4 inches. The site has an excellent view to the distance. The fort was placed there also to protect local farmers and enlarged several times. It was added to a smaller fort. It guarded the routes to oases at Lukchun and Kara-Khoja. To the east the route went from Lukchun to Singer. The outer wall protected the enclosure of an adjoining fort to the NE that had inferior walls. In Uigur times Chong-hassar had an agricultural settlement. The area was a Buddhist colony during Uigur era. Stein left Besh-tam on 18 November going north to Toyuk (map 59) to see more Buddhist shrines and caves. He visited Kara-Khoja - the ancient Kara-ch'ang. Turfan was the capital during T'ang and Uigur eras - ruins dated from 272-277 AD. The grottos at Bezek-lik have mural paintings. He visited Yar-Khoto (map 54) which also was a capital in T'ang era. It has a strong position between 2 deep cut ravines. It was called Chiao-ho in Chinese (plan 49 and photos 273 - 275 - 276)). It covered a large area of clay buildings. He found copper coins there from T'ang (K'ai - yuan ( 618-627) and two other coins (ch'ien-yuan) from 758-760.

Chapter XXIX - Kara-shahr and its Ruined sites

Section I - Historical Topography of Kara-shahr

Stein left Turfan on December 1st for Kara-shahr. The road was via Toksun (map 54) then south through the gorge at Su-bashi in the hills that link Kuruk-tagh and T'ien-shan mountains south of Urumchi. It was 140 miles from Turfan to Ushak-tal. He found another small, ruined fort at Oc-tam NW of Toksun. It had massive walls of stamped clay. The route goes from Yar-Khoto to T'ien-shan to Toksun and Kumush for 56 miles. Hsuan-tang went this way and described it well and stayed briefly at Kara-shahr. Stein found a great lake (Baghrash) in the basin and the Khardu-gol River. The lake discharges into the Konche-darya. He noted that Kalmuk Mongols live in the mountains north of there. He visited another major fort 6 miles SE of Ushak-tal. The SW face was 270 yards and SE face was 308 yards - both of stamped clay with the wall still standing to 25 feet and up to 15 feet wide at the top. There were also towers of clay. He found one copper coin dated 258-260 AD. There was another ruin at Chak-kur (map 48) with a small enclosure and towers of stamped clay 25 feet by 19 feet. On December 8th Stein reached Kara-shahr village. The nearby ruins and Ming-oi near Sharchuk.

Section II - The Ming'oi site north of Sharchuk

On December 11th Stein left Kara-shahr via Korla road to Sharchuk 16 miles to the SW and visited more Buddhist shrines north of Ming-oi (means 1000 houses). He noted that Shar was Turkish term for salt efflorescence - also called chor or chuk. He took many photos of this area and made detailed plans. The buildings were on top of ridges and plateaus that were offshoots of the mountain range (plate 51). There were a large number of ruined buildings but he had very limited time for excavations. Kara-shar was under the control of Pan-tajen, the Tao-t'ai of An-su and Stein's old friend. So Stein had considerable official support in obtaining local assistance including large bands of Turki labor from Korla. This help was supervised by Naik Ram Singh and the interpreter, Chiang Ssu-yeh and R. B. Lal Singh. In 12 days they managed to clear many sites and 100 shrines from 4x6 feet to 80x80 feet. The walls were of sun-dried bricks 12x6x3-4 inches. Stein made extensive detailed descriptions of the buildings and damage from rain and snow there. Near Lake Baghrush the water caused destruction and decay. He was working in temperatures 42 degrees below freezing. There was also much damage from fire. He found many Chinese coins, but none later than the 8th century. But he found Uigur documents from the 9th and 10th centuries. Possibly the fire came from Muhammadan invasions by Karluk Turks in the 2nd half of the 10th century. Stein was especially interested in the Buddhist art in Gandharan style.

Section III - Relievos and Frescos from the northwest portion of Ming'oi site

Stein found large statues of Bodhisattvas in Gandharan style - too large to move. He found Chinese coins of the Wang Mang period AD 14-19 with coins from 766-783. Plate 51 shows a watch tower on a ridge overlooking the caves. It was 24 feet square and 25 feet high (fig 181). Another similar tower he found near Korla.

Section IV - List of antiques excavated at the Ming'oi site, Sharchuk

This section includes extensive pages of inventory of articles found and taken from the location.

Section V - The Khora site and defile of the Iron gate

Stein finished work at Ming-oi on 23 December and sent a convoy of antiques to Korla. He then went to Khora to the north and then west. He found more small Buddhist shrines and small cells with frescos and a stupa. He left there on 26 December to reach Shik chin and on 27 December reached the Korla oasis. He found another watch tower 32 feet square and 19 feet high with bricks of 12x8x3-4 inches with reeds between the brick courses. These were similar to the line of watch towers from Tun-huang to Lou-lan and on into the Tarim basin. This route had danger from Hsung-nu (Huns) The towers had inner and outer casings to strengthen them. Stein was at Korla in both Dec 1907 and again in April 1915. He crossed the river that drains lake Bagh rash and is guarded by a fort at Bash-eghin. The road passes through a 7 mile defile by the river. This is called the Iron Gate in T'ang annals.

Chapter XXX - To Kucha and the Keriya River end

Section I - Korla and its old sites

On December 27-28 Stein was at the oasis NE corner of the Tarim Basin where the large river that drains Lake Baghrash flows out and becomes the Kinche-darya. He noted the uniform quantity of the water. The Ancient Han Chinese annals called it the Wei-hsu. In 1915 he went along the Konche-darya from Konche to Karuk-tagh hills (map 45). Hew found another fortress at Ubuggen-bulak made of earthen ramparts, 12-15 feet high, and 380 yards N-S and 250 yards E-W. At Yan tak-shakr there was another smaller fort, 140 yards square, with same type of wall, 1.5 miles SSE of Korla. Shah Kalandor is the 3rd town 3.5 miles SW of Korla with a circular wall, 30-40 feet wide at base and 12-15 feet high. In January 1908 Stein went SW from Koral with guides into the desert for two days to reach the Konche-darya (map 45). Between the Charchak and Inchike Rivers he found another small ruin (map 45) with a rampart 180 yards round and walls 30 feet thick at base and 12 feet high. He heard many folk tales (ghost stories) of towns and treasure caches hidden in the desert and guarded by demons who would prevent anyone who chances to see them once from ever seeing them again.

Section II - From the Inchike River to Kucha

On January 9th Stein went to see another reported site and did find a ruin across the Charchak River 2 miles north of Inchike River. There was an old Muhammadan cemetery. On January 12th he went up the Inchike darya NW for 60 miles to the caravan route at Bugur. He saw more watch towers on the road to Kucha. (See the report of the 3rd expedition). He discusses the Han annals about the military operations here in 101-BC to 90 BC and 68 BC. In 68 BC general Cheng-chi established a garrison of 1500 men at Ch'u-li.

Section III - Through the Takla Makan to Kara-dag

This part of the exploration was one of Stein's most dangerous. On January 17th at Kucha he prepared to cross the Takla Makan directly going straight south to save time for his subsequent spring projects. This was pure desert with no water for miles and days. On the 25th he sent 24 loaded camels south up the Khotan River bed to Khotan. On the 25th he left Kucha town to reach Shahyar on the 26th. This was the town furthermost south before the desert. He crossed the frozen Tarim River on 30 January and into the desert (map 35). It was 180 miles due south (read Desert Cathay for his personal account.) (The route shows on maps 35, 30, 36). (Serindia maps) He crossed the Achchuk Darya at his camp marked 317 (Map 35) and obtained water from shallow wells dug in the dry river basin. After that there was no water until camp 320. He describes some of the rigors in travel in winter up a dry river bed. By February 6th going S - 190 degrees west he found a bed of dead forest in the delta of the Keriya River, camp 321 map 36. (which, remember, flows north from near Khotan fed by water from the mountains south of the desert.) There he could dig wells for fresh water, but it was for 6 more days of marching south before he sighted in the distance the shine of the ice sheet of the drying river. It turned out that the river had changed its course to the west of that found by Dr. Hedin in 1896. At camp 323 - 25 miles further south - he found flint and 4 miles further further on passed Kara dong without seeing it. He had been there in 1901. He went back to Kara-dong and found additional ruins since his visit in 1901 (fig 302 plate 95). This site is a mile long N-S and 1/3 miles wide East-west. He found timber buildings like at Niya. It was an ancient agricultural community from the 3-4 centuries.

Chapter XXXI - Ruined sites east and north of Khotan

Section I - The site of Farha-beg-yarlaki

On Feb 22 Stein left Kara dong going south up the river to meet a team coming north from Khotan. On Feb. 25th he met Turdu's nephew with men at Kuchkar-aghil (map 31). then went straight to the desert at Domoko oasis. There were ruined dwellings and Buddhist shrines. (see Ancient Khotan) On March 2 Stein went to Farhad-beg-Yaikaki - a large area 4.5 miles SE to NW with timber houses and old documents and pottery.

Section II - The shrine Kxii and Chronology of the site

Here Stein found a colossal stucco image of the Buddha buried in sand but too fragile to move. All he could do was write a description. He determined that the site had been occupied until a date earlier than the Dandon site and then abandoned in the 8th century.

Section III - From Domoko to Khotan

On March 13 Stein went south past Domoko to visit a Buddhist shrine at Kara-yantak. He then went to Keriya to arrange explorations in the summer in the mountains to the south in the Kun lun to Polur. He returned to Domoko on 19 March and then went NW into the desert to Ulug-Mazar again via Domoko. This was the last location where his entire party was united at the end of the winter explorations. On March 22 he went south to Chira, where he found 3500 skulls buried in a large oasis. On March 24 he went west to the edge of the Khotan district. From Lop-bazaar he went north to the ruin at Iman Asu Mazar and found a stupa- then through Sampala to Bizil. From Bizil he crossed to the west bank of the Yurung-Kash and by March 30 reached Khotan.

Section IV - Temple remains north of Khotan

On April 5th Stein again headed north toward Ak-su on the T'ien-shan. He visited locations found by local guides who showed him Buddhist remains. One wall was 34 feet long under sand had had huge painted figures of Buddhas and Bodhisattvas, but the sand then crushed the wall. He wrote a detailed description of Mayalik. On 11 April he moved down river to Islamabad (map 27), the northern most part of the Khotan district. He left Islamabad on 13 April going north.

Chapter XXXII - From Mazar-tagh to Maral Bashi

Section I - Ruined Fort on Mazar-Tagh

Stein traveled for 2 days from Miran to reach the hill of Mazar-tagh on the west, left bank of the Khotan River. This was north of the confluence of the Yurung-kash and Kara-kash rivers . The river dried up 5 miles further, but sub-soil water and pools were available. This is the shortest route from Khotan north across the desert to the Tarim. On April 16 he sighted the ridge. It is 24 miles long and 300 feet above the desert. The ruin of the fortress is well preserved including its high walls. (see figure 335, site plan, plate 59 and map 26). Figure 329 shows detached watch tower on the plateau 30 yards wide. The south face of the ridge is a precipice. The north side is an easier slope. This is a strong position for a fort. It is inaccessible from South and East. The tower protected from the west. The fort defense on the north west is a wall across the crest and two 16 foot square bastions. The curtain wall is made of flat pieces of clay set in mortar and layers of Tamarisk branches. The bastions project at either end of the wall of 10 inch layers. The posts and beams are of Toghrak wood inserted to strengthen the wall. The wall is 10 feet thick as are the fort walls on SE and NE. Later walls and inner buildings and court are built of coarse bricks 15x8x3.5 inches. The space inside the walls is a level area made out of the steep north slope. The inner court is 50 feet square covered with debris of other buildings. The charred timber shows the buildings likely were made of wood. Outside the wall on SW side they disappeared in fallen timbers down the slope. The wall fell due to its foundations that slid down the hill. The gate on this side was to the keep. (fig 330). The keep originally was of coarse bricks 4 feet thick and later strengthened to 8 feet thick on three sides. The main west wall of the fort was built of different material. It was originally more than one story but that is now gone. On the NE side an apartment between the keep and fort outer wall was 20 feet by 6 feet. The outer court was 20 feet lower down with stairs. The walls of the outer court were of sun-dried bricks same size but without layers of tamarisk and timber. It suffered from more damage. This court space was 90 feet by 29 feet. It was probably roofed as there are charred timbers all over. It was likely a stable. The gate through the 10 foot thick wall is on SE. The tower on the crest at 225 feet above the clay bank of the river bed was 60 yards from the west bastion of the fort. It was solid, like watch towers on the Han wall. It was built of flat hard clay brought from the river and layers of tamarisk at 10 inch intervals plus toghrak posts and beams in the masonry. The base was 25 feet on SW and NE sides and 22 feet on the other sides. The SE face was broken. The tower remained 20 feet tall with a view over a wide area along the river. Stein excavated this first. That required 3 days. He found Tibetan documents and tablets like at Miran and Tibetan paper records and other objects.

He found a storage pit 5.5 feet deep and 6 feet square, revetted with timber in a room. There were Tibetan records and moulds for statues. On his return visit in November 1913 he found the shrine. All this showed results of a great fire. He found coins dated from AD 758-760 and one from 766-80. Figure 334). The refuse outside the fort contained more artifacts. More Tibetan records and 1000 items including arrow shafts, some bronze and broken bows and wooden sheaths for swords and daggers, shoes, string sandals, wool clothing, wooden seal cases and keys and locks. The Tibetans held the area in AD 791 and the Uigurs in 860. There were Tibetan military reports and requisitions and statements about weapons and inventories. But the Khotanese language continued also and likewise Chinese documents on paper. An official certificate was dated 786. There were also Buddhist monastery documents from elsewhere.

Mazar was a key watch tower location controlling the river route, north - south across the desert. The Tibetans even captured Kucha and Pei-t'ing on the north side of the desert. Mazar was like the Miran fort.

Section II - Through Ak-su and Uch-Turfan

Stein continued on from Mazar via Khotan Darya. The march north of 150 miles along the dry bed of the Khotan-darya required 8 more days to reach the Yarkand and Tarim Rivers. and on April 28 crossed the Tarim a mile below (east) its confluence with the Yarkand and Aksu Rivers. (map 25) Stein inserted a digression about the history of the Ak-su area from the Han annals. It took him three more days to reach Ak-su where he stayed for 5 days to visit his old friend, Pan Tajin. who was instrumental in gaining Stein access to so much official assistance. He send R. B. Lal Singh to survey the outer Tien Shan west to Kashgar.

On May 6th he left Ak-su to visit Uch-Turfan and Kelpin in the mountains. He wanted to complete the archeology there before returning to Yarkand and Khotan. He found more forts in the hills above Ak-su. In two more days he reached Uch-Turfan. On May 8th he moved from the west bank of Tushkan -darya. (fig 337) The Chinese citadel was on a 250 foot high hill.

Section III - Through the mountains to Kelpin

On May 11th Stein left Uch and found another watch tower at Kong tai. He spent a day to see this at 8000 foot elevation and at Sarbel at 10,000 foot elevation. He found Kirghiz nomads who insisted on showing him a rustic Ziarat on a rocky hill. He went via pass to I dak-jilga to a real stone image in the center of an enclosure 2 feet 10 inches high by 12.5 inches wide at Chalkoide. It was carved in flat relievo - a male holding a sword next to a miniature stupa. The enclosure around the image was filled with votive offerings of Muhammadan shrines. And animal parts, bones, horns and such. Stein wrote that the Muslims had no idea of the real origin of the shrine.
Then he was two days to Kelpin - Sahit-kak down through gorges. The Kelpin oasis (map 14) is isolated but cultivated. Stein describes the people there and the irrigation system. He found a small fort of clay walls 166 yards by 90 yards with 6-10 foot high walls 6 feet thick. Local men brought him coins from Han to T'ang Dynasties.

Section IV - Desert sites north of Tumshuk and Maral-bashi

Stein left Kelpin going south (map 47) and it was already hot. He took 3 days to explore the area and found another tower built of clay with ramparts around it 110 yards square of stamped clay. Chang-tun fort - Was still 35 feet tall and 25 feet square. It was made of sun-dried bricks 11x10x3 or 14x8x4 inches. He found old coins from early T'ang 766-80 Ad and earlier ones from AD 14-19. Some were located in the desert on a direct line between Chilan and Ak-su. He continued on to Marl bashi and Kashgar. Another fort was found at Lal tagh (map 14).

On May 19th he moved SW and found more coins from Sung Dynasty - 1034-1068. He reached Tumshuk on the high road and saw Buddhist shrines. He then started to return to Khotan 350 miles via Yarkand. On May 21 he reached Marl bashi (map 15). Another 8 days was required to reach Khotan.

Chapter XXXIII - From Khotan to London

Section I - Preparation at Khotan

Stein reached Khotan on 9 June. He spent 6 weeks packing his treasures for movement over the Karakorum Pass into India ( a very dangerous route). He had tin cases fabricated which he packed himself. He strengthened them with glued cotton to protect the large frescos and the packed them tight between compressed layers of reeds. It would be an 8000 mile journey over mountain passes using camels and then yaks and ponies and then a cart and train and steamer to get them on to London. Stein discusses his preparations in detail. He planned to explore the Kun lun and Karanghu tagh mountains on his way to India. In 1900 - 1901 on previous exploration journey he found it impossible to pass through the gorges of the upper Yurung-Kash to find its headwaters. Now he was determined to succeed. So he decided to go east along the foothills and then turn south around the main mountain chain and up through valleys and back west to reach above the gorges. After that he would be able to meet his caravan on the mountain road to the Karakorum Pass. Along the way he would find the head waters of the Keriya River as well. He left Khotan for Polur, the last inhabited place in the foot hills of the Kun lun. From there on he would have no supplies but what he carried. He estimated it would take at least 40 days.

L. B. Lal Singh rejoined Stein on 20 July after 3 months of surveying the T'ien shan from Ak-su to Kashgar (maps 1,5, 6,9, 14,19,23). Stein surveyed the Kun-lun from Kilian and Middle Kara tash River (maps 16,17,21) and visited Yotkan to acquire more terra cotta. On Aug 1st the heavy convoy departed with 93 cases loaded on camels, via Sanja and then over the glacier to Suget on the Kara Kash. Stein departed on 3 August.

Section II - Across the K'un lun ranges

Stein spent 2 days going over the Tikelik tagh at 18780 feet elevation (map 28) then the valley of the Kara tash River. He found yet another wall, across the route - 1230 yards long with two bastions and 30 feet deep . Then he reached Chakar (map 28) along the river. On August 6th he reached the Nura (map 32) at 7000 feet (fig 343) in the midst of mountains 21,000 feet elevation. On the next day he moved to see another ancient site at Tun and found more pottery and a mound 36 feet across and 16 feet high, probably a ruined stupa. This was the last old site. He found Moslem tombs on the next day. He again consulted Chinese former Han annals for a description of the area. Then he turned west at Tost-Iman to Polur (map 32) at 8500 feet, on a side stream of the Keriya River. He spent 3 days there for final preparations and set out on Aug 12th (description is also in Desert Cathay). The next 500 miles were across barren rock and ice. He passed through gorges above Polur for 4 days to the northern plateau adjoining the K'un lun Range at 15000 feet. This area, he noted, had already been visited by European explorers since 1673. (maps 22, 28,29,32,323). He met a lone hunter who acted as a guide and gave directions. Then he saw the gold pits (map 29, 33) at Zailik valley and received important assistance from the miners there for 8 days. He then moved to the glacier basin where the eastern and longest branch of the Yurung-Kash rises - He followed Yak trails over heights of 18-19000 feet that probably no one had seen before. He traveled along the river to its head (map 33). He could see glaciers that had receded in recent times, which caused the reduced water flow down at Dandon and Niya on the edge of the desert.

Stein reached an easier area near Ulugh-kil Lake on 3 Sept., where a depot of supplies and transport was already assembled. Then he went over the Polur-Lanak-la route to the basin above 17200 feet where the Keriya River rises at glacier - the same mountain range that encircles the Yurung-Kash sources. Leaving the water shed of the Keriya at 18000 feet he turned west to explore the ground of the Aksai-chin .There was a high snow covered spur of peaks to 23000 feet. Between them were broad valleys at the entrance of which were a series of isolated basins between 15000 and 16000 feet holding mostly dry lakes. By this time the lack of grazing had killed his ponies and donkeys.

Another week of long marches from where he left the track to Lanak-la brought him to the large salt lake (map 22). Three more days of movement NW across the sterile basins and salt lagoons brought him to the old route. He found stone heap markers-cairns - from 1864-66 expeditions and was able to follow the old track to the pass. On 18 Sept. he reached the valley of the eastern tributary of the Kara-Kash (map 22) at Haji langer shelter. Two days later he reached the lower down Kara-Kash valley and found the men of Satip-aldi Beg's Kirghiz party with supplies and transport from Khotan. But then Stein wanted to trace the route that Haji-Habib ullah had used in 1863-66 up to where it crossed the K'un-lun Range above Karanghu-tagh over high glacier passes. So he went back with the Kirghiz and Yaks to Haji-langer out of curiosity.
On 22 Sept. Stein went up a glacier northward with R B L S for 10 hours to reach the crest at 20,000 feet by a peak of 23071 feet. He had great views but the temperature was 16 degrees below freezing. The result of his rash determination was frost bite.

Section III - Over Karakorum to England

Now Stein was in trouble and the dangerous and lengthy Karakorum Pass route still lay ahead. He describes how he was carried for 4 days to Portash to meet the baggage caravan that had crossed the Sanju Pass. He had to leave R. B. Lal Singh to lead the slow caravan, while he rushed ahead with a small party. He reached India from the Kara-Kash valley over the Karakorum Pass -crossing it on 3 October at 18,687 feet. He was met at Murghe defile by a rescue party of Tibetan porters sent out to meet him. Nine more days were needed to reach the first village where he was met by Reverend Schmitt who had rushed to cross the Khardang Pass to help. Stein reached Leh on 12 October. On 1 December he started for India and Dehra Dun and Calcuta. On 2 Dec. he took ship from Bombay and reached London on 20 January 1909.

There follow nine appendices and then lists of artifacts, maps, illustration, a huge index.