INNERMOSTASIA - CHAPTERS II

Detailed Report on explorations in Central Asia, Kan-su and Eastern Iran

Sir Aurel Stein K. C. I. E.
Oxford, At the Clarendon Press,
Oxford, England 1928
Volume II Text

In this volume Sir Aurel Stein continues with his official report beginning with his movement from Barkul to Guchen and concluding with his return to India via Iran. For Volume I go toAnd for the photographs from all the volumes go to . The first 9 chapters of this volume pertain to Chinese Turkestan. The remaining 6 describe Stein's continuation through the Russian Central Asian area in the Pamirs and then south eastern Iran (Sistan). The chronology of the entire expedition is here. And here is a listing of the maps. For his first and second expeditions Stein published both an 'official report' and a 'personal memoir' but, unfortunately he did not write a similar account of this, third, expedition, rather a later summary account of the entire effort, leaving out the part related to Russia and Iran. Unfortunately, the maps do not include his routes in Russian Turkestan and only a general map of Iran. I have included here links to many photos and some of the map sheets.

Table of Contents

Volume II

Chapter XVI - To Guchen and across the T'ien-shan - 549-565
Section I - From Barkul to Guchen - 549
Section II - The Site of Pei-t'ing and the Posterior Court of Chu-shih - 554
Section III - Across the Mountains to Turfan - 560

Chapter XVII - The Turfan Territory: Some aspects of its Geography and History -566-58
Section I - The Geographical Position of Turfan and its Earliest Historical Notices - 566
Section II - Turfan from Later Han to T'ang Times - 572
Section III Turfan under the Uigurs - 581

Chapter XVIII - At Ruined Sites in Turfan - 587-641
Section I - Among the Ruins of Ancient Kao-ch'ang - 587
Section II - List of Antiques excavated, or acquired, at Kara-khoja - 596
Section III - Search among the Ruins of Toyuk - 609
Section IV - List of Antiques excavated, or acquired from, Ruins of Toyuk - 620
Section V - Work at the Sites of Murtuk - 633

Chapter XIX - The Ancient Cemeteries of Astara - 642-718
Section I - Seventh Century Tombs in Group I - 642
Section II - Figurines and other Sepulchral Deposits in Groups II-V - 650
Section III - Intact and other Burials in Tombs of Groups VI -X - 660
Section IV - General Observations on the Astana Burials and their Textiles - 667
Section V - Relics of Textile Art from the Tombs of Astana - 672
Section VI - List of Antiques from Cemeteries near Astana -680
Section VII - Conclusion of Work at Turfan - 710

Chapter XX - Explorations in the Kuruk-tagh - 719-748
Section I - From Turfan to Singer - 719
Section II - To Po-ch'eng'tzu and Sindi - 725
Section III - To Ancient Graveyards by Kuruk-darya - 732
Section IV - Mian Afraz-gul's Supplementary Surveys - 741

Chapter XXI - On the Ancient Route along the Konche-darya - 749-785
Section I - The Ruins of Ying-p'an - 749
Section II - The Ancient Course of the Konche-darya and the 'Town of Chu'pin' - 761
Section III - Watch stations along the Ancient Road to Korla - 768
Section IV - The Territory of Wei'li and the Modern Kara-kum - 777

Chapter XXII - From Korla to Kucha - 768-802
Section I -Along the Foot of the T'ien-shan - 786
Section II - The Seat of the Protector-General - 790
Section III -From Bugur to Kucha - 797

Chapter XXIII - Kucha and some of its ancient Sites - 803-829
Section I - The Oasis in its Geographical Aspects and the Position of its Ancient Capital - 803
Section II - Ruined Sites West of Muz-art River - 807
Section III - Remains South-east of Kucha and List of Antiques found or acquired - 818

Chapter XXIV - From Kucha to Kashgar - 830-841
Section I - Old Remains within the Bai District - 830
Section II - Past Ak-su and Marl-bashi to Kashgar - 834
Section III - A T'ang Itinerary from Ak-su to Kashgar - 838

Chapter XXV - Across the Pamirs - 842-862
Section I - Preparations at Kashgar - 842
Section II - Along the Alai Valley - 844
Section III - Along the Western Rim of the Pamirs - 856

Chapter XXVI - In the Region of the Upper Oxus - 863-895
Section I - Old Remains of Wakhan - 863
Section II - Through Ishkasham to Gharan - 871
Section III - In the Valleys of Shughnan - 877
Section IV -From Roshan to Darwaz - 884
Section V - From Kara-tegin to Bokhara - 891

Chapter XXVII - By the Eastern Marches of Khorasan -896-905
Section I - From Askhabad to Meshed - 896
Section II - Past the Perso-Afghan Border - 897
Section III - Into the Helmand Basin - 902

Chapter XXVIII - The Sacred Hill of Sistan - 906-925
Section I - The Historical Interest of Sistan - 906
Section II - The Remains of Koh-i-Khwaja - 909
Section III - Remains of Mural Paintings - 913
Section IV - Remains on the Hill-top - 921

Chapter XXIX - Ruined Sites within the Oasis of Persian Sistan - 926-942
Section I - Remains at and near Shahristan - 926
Section II - The Band-i-Sistan and the Ancient Name of the Helmand - 930
Section III - The Site of Zahidan and Later Ruins to the North-west - 932
Section IV -List of Pottery Specimens and other Small Objects from Later Sites in Northern Sistan - 938

Chapter XXX - In the Desert Delta of Sistan - 943-981
Section I - Ruins Ancient and Modern - 943
Section II - Remains of Prehistoric Settlements - 949
Section III - List of Objects found at Sites of the Southern Helmand Delta - 957
Section IV - Ruins of an Ancient Border Line - 972
Section IV -From Sistan to India and London - 979

Appendices - A to R - 983 - 1090
Index of Objects found - 1091-1117
Plates Described or Referenced to in the Text - 1118-1119
Plans Described or Referenced to in the Text - 1119
Figures Described or Referenced to in the Text - 1120-1121
General index - 1123-1159

Sir Aurel's work described in this volume was complex and involved visits to many locations (having unfamiliar names) with treks criss-crossing the northern side of Chinese Turkestan and then through then-controlled Russian Turkestan (now Tajikistan and across Uzbekistan, Turkmenistan ) and finally south-eastern Iran (Sistan). The reader will be greatly assisted by frequent reference to the maps, but they depict only the Chinese Turkestan area. This brief summary may also be of assistance. Stein with part of his team is found in chapter XVI at Barkul, on the north-eastern slope of the T'ien-shan mountains' far eastern extension at the town and lake Barkul. ( map43das ) He then moves west along that northern slope to Guchen. ( map31cas ) He briefly visits Pei-t'ing and then crosses the T'ien-shan -north to south via the Pa-no-p'a Pass to Turfan town on the northern edge of the Turfan Depression, ( map28aas ) which is far north-east of the main Taklamakan Desert and due north of Miran and Lou-lan. Meanwhile Muhammad is bringing a survey line directly west from Hami along the southern edge of the same mountain range and Ibrahim is bringing the caravan with heavy baggage along the main route from Tun-huang and An-shi to Turfan. Stein spends the next several months working in and around the Turfan Depression while Lal Singh conducts another survey back eastward across the Kurugh-tagh to Singer and Lou-lan, then north-east. In Chapter XX Stein next returns south-eastward to trace the dry river toward Lop and sends Afraz-gul to complete another survey along the Lop Sea. Lal Singh surveys another route west to Korla. In Chapter XXI Stein returns westward to Korla and in XXII on west to Kucha. Korla is an oasis on the northern edge of the Taklamakan in the foothills of the T'ien-shan which here extends to the south. It is approximately due north of Charchan. Kucha is another similar oasis slightly east of due north from Keriya After completing more archeological digs Stein then rushes west through Ak-su and Marl-bashi to Kashgar where he spends a month packing his 'antiques' for the long road to India.

Chapter XVI - To Guchen and across the T'ien-shan

Section I - from Barkul to Guchen ( map31cas )
Stein notes that he remained at Barkul for 4 days of rest needed by his team and himself after the difficult crossing. He thanks Mr. Li Shuy-jung, the district magistrate for his hospitality. And he appreciated Mr. Chen-t'ai, the garrison commander, for accommodation at the local temple . During the brief stay, the survey team kept busy collecting information about Lake Barkul and the nearby mountains. - Li Tu-lao-yeh told Stein about the ruin at Pei-t'ing located just north of the pass through the T'ien -shan that provided a direct route to Turfan. Stein comments that the Chinese garrison at Barkul was exceptionally large, apparently due to concern about Tungan - Mongol - incursions along the key routes to the south. Already the Chinese were populating the region north of the T'ien-shan with Han colonists. The town itself claimed only 2000 families and the old walled town to the east was in ruins.
The Chinese authorities were also very concerned by the influx of Kazak nomads pushed south after fighting with the Khalka Mongols who now occupied the grazing land north of the Dzungarian desert. The Kazaks were strictly located to the mountain slopes between Urumchi and Barkul and not allowed east of Barkul. This area controlled Chinese line of communications between Urumchi and the east. He recalls the ancient Chinese 'fear' of the 'barbarian nomads' . If these nomadic Kazaks could not be settled, it was feared that they might start pressing against other tribes (either for grazing land or loot) and set off a huge migration that could disrupt all Chinese Central Asia. Stein was happy to receive the assistance from one of the Chinese official measures to control the Kazaks which was offered him by the authorities. Namely, this was the government impressment of Kazaks to supply pony transport for official business throughout the region. This assistance, for which Stein as always paid well, enabled him to travel the 200 miles to Guchen in 9 days and along the way to study the Kazaks. He pronounced them to be "without exception fine upstanding men of brave bearing." And his keen appreciation for ethnic features noted that they were ,in contrast to the Kirghiz) of a strongly European type rather "free from marked "Mongolian" characteristics.. He also noted that they were well dressed, apparently more wealthy that Western popular opinion believes, and also dissatisfied with the grazing land allotments received from the Chinese and eager to regain their better territories in the Altai (which would require fighting). Stein only briefly (he claims) describes the route to Guchen, since it is already well traveled by Europeans. But 'brief' for him is nevertheless vivid and extensive for the reader now. His main point, repeated frequently, is that the northern slope of the T'ien-shan is well watered, rich in grassland vegetation, and forested. He stopped at Ch'uan-tzu-chien on the main Hami-Turfan road. From there the route south crosses the mountains in a low saddle at 5,600 feet elevation. As always he compares his assessment of the topography and vegetation with that found in the Ancient Chinese Annals.
On 13 October he crossed the barren plateau between Ta-shih-t'o ( north-center in map31cas ) in a blizzard. At Mu-li-ho (in north-west corner of map31cas ) he found a Yarkandi trader (British Indian subject) who had moved there from the Altai along with his client Kazaks. This gentleman, Ibrahim Akhun, provided Stein with much specific information on the terrain and economic activities in the region. He also noted the great increase in Chinese agriculturalist settlers. And this has generated greatly increased summer employment of labor from the other side of the T'ien-shan - namely the Turfan basin. Stein remarks that this must be similar to the conditions in the ancient "Anterior and Posterior Chu-shih' era.
On 16 October Stein passed through more rich agricultural land to reach Guchen. ( map28abs ) Guchen (Turki) or K-ch'eng-tzu (Chinese) was a major market town with a massive city wall (unfortunately not shown) and very Chinese in appearance. It is also the hub of the major international trade routes into Mongolia, Siberia, and China as well as across the T'ien-shan. Stein was told that an ancient ruined town was located to the north. He remained at Guchen for 2 days with a Kashgari trader. In the Bazaars, which he always sought for evidence of the populations, he met many Mongolians and Turfan visitors. A brisk trade in cotton, fruit, flour, sheep, felt and more caught his eye.

Section II - The Site of Pei-t'ing and the Posterior Court of Chu-shih
Stein describes his motives for selecting the separate travel routes for himself and Lal Singh across the T'ien-shan by different passes. Lal Singh used the Ku-ch'uan pass traversable by camels, while Stein used the Pa-no-p'a Pass (only accessible by mule and pony) in order to visit the ancient capital ruin at Pei-t'ing. He describes the historical connection of this capital city with Yar-khoto in the Turfan as recorded in Chinese Annals. Both movements were hurried because of approaching winter and Lal-Singh's pending mission to survey westward through the Turuk-tagh while the cold weather would provide ice.
On 19 October Stein started with Afraz-gul the 13 miles from Guchen to Urumchi. at Jimasa he visited the Pei-t'ing ruin. ( in north-west corner of Guchen oasis on map 28abs) He digresses as usual to inform the readers that the town was noted in a T'ang Dynasty inscription and a 10th century note that the place was in Later Han times the seat of the Posterior king of Chu-shih during a period in which the kingdom of Kao-ch'ang (Kara-khoja) in Turfan was controlled by China. In AD 702 the area was changed into the Protectorate of Pei-t'ing, which was one of the 'Four Garrisons' through which the Chinese administered Central Asia.
On 20 October Stein visited the place. En route he passed the modern ruins of villages destroyed during the Tungan Rebellion.
The outer walls had enclosed an area about 2,160 yards north-south and 1,260 yards east-west. The walls were greatly decayed and some completely eroded. Their plane table survey revealed some of the former connections. Stein estimated that originally the walls were 30 feet wide and 20 feet high with massive bastions in the corners. There was also an inner enclosure with similar walls. Stein noted that the interior had clearly been 'mined' by locals for useful material including manure. He found the remains of a Chinese temple in which he found some pieces from relievos that he assessed as being from Ming or later periods.
On 21 October Stein departed Jimsa southward to cross the moutains. ( map28aas ) Moving back wards they came to and stopped at Ch'uan-tzu'chieh (at southern edge of Guchen oasis. ) . This, he found, was a Chinese prosperous trading center with visitors from all around. Among these Stein particularly mentions again the European features of the stalwart Kazaks. The section concludes with a page containing a list of objects recovered at Pei-t'ing.

Section III - Across the Mountains to Turfan
On 22 October Stein enjoyed his travel through the mountains in the Pa-no-p'a pass. Soon the track narrowed, excluding carts as it crossed the stream . While resting over night at the way station Pa-no-p'a he met a well armed group of bandits from Kara-khoja. He describes their activities in an enjoyable footnote. ( map28aas )
On 23 October he crossed the watershed which he measured at 12,280 feet. - The temperature at noon was already 30 degrees. The descent led south-south-east. It lay through a very narrow stream filled gorge only passable due to walls built of stone and then opened into a valley. Stein was impressed by the stark contrast between the barren hills here and the heavily forested hills on the north side of the range. The utterly dry climate of the Turfan held sway. He noted also that the wide extent of the valley they moved through indicated that when there was snow-melt the river must have expanded greatly. (also on maps 28 aas and abs). They arrived at Yoghan-terek roadside accommodations at 6,400 elevation after a day's trek of 30 miles. He notes that the mountain stream, very large at this point, disappears 12 miles further down into the rocky Sai, slope. From there it's underground passage enables the Karezes and springs that provide the essential irrigation for Turfan.
On 24 October he continued down the left bank of the stream. At 3000 feet elevation they came to Shaftulluk oasis whose orchards and arbors were watered by a lively spring.
On 25 October a further march of 19 miles brought them to the edge of Turfan town. Stein then compares his personal observations of the route from Jimasa to Turfan with the text in the T'ang Annals. He considers the correspondence exact. He makes the same conclusion or travel reports during the Uigur era.

Chapter XVII - The Turfan Territory: Some aspects of its Geography and History

Section I - The Geographical Position of Turfan and the Earliest Historical Notices
On 25 October Stein arrived at Turfan prepared for his planned winter program of exploration of the Turfan Depression. He explains this elaborate and ambitious program in detail. While he will stay mostly in the Turfan Depression Lal Singh will undertake another solo survey - back eastward across the desolate Kuruk-tagh to Singer. ( This index map shows the relationship of Turfan to the area from Yangi-hissar to Hami and Singer to Barkul mapindex9f ) Afraz-gul will survey toward Lou-lan, and Stein will survey another route in spring. Meanwhile Muhammad Yaqub will arrive shortly having surveyed another route from Hami to Turfan. Together these will fill a huge gap to complete the circuit around the Taklamakan. Stein gives as his reason for not writing more on the details of archeology and geography of the Turfan the extensive surveys by several German expeditions since 1900 that have taken a very large amount of artifacts to Berlin and are publishing expert reports. He mentions other expeditions as well, some not so favorably due to destruction plus he deplores the looting of the sites by vandals and local inhabitants. For one thing and understandable, they are mining the ruins for manure. But all this attention by European explorers has alerted the locals to the profits to be had by digging up relics and manuscripts that they can sell at Urumchi and along the Tarim caravan routes.
He claims that the massive quantity of archeological material remaining must be saved as rapidly as possible from this destruction. For these reasons he proposes to restrict the archeological report, but concentrate his personal observations. But first, in this section he will provide a general picture of the geographical and historical environment (making special use of Chinese records from Han and T'ang periods) of the Turfan necessary to appreciate the significance of individual archeological finds. He begins with the Ch'ien Han shu from the Han era. His main theme is the close political and economic tie between the Turfan Depression and the Guchen- Urumchi area to the north across the T'ien shan. From an economic view point this was the result of the remarkable climatic difference prevailing on the two sides which enabled people on both to profitably exchange surpluses not available on the other. The very warm weather in the deep Turfan Depression enabled cultivation of cotton and fruit in two seasons a year, but no space for grazing. While on the northern side of the mountains there was ample grazing for animals and water for grain crops. And exchange was enabled by the existence of the several reasonably low passes through the very high range. These passes also enabled political connections between "Anterior and Posterior Chu-shih" The Huns living to the north-east could also make use of easy routes to prey on settlements on both sides of the T'ien-shan. Thus Chinese records of the warfare between Hun and Han provide much valuable information. Campaigns in 89 BC, - 87-74 BC, 68-68 BC and 1-5AD are discussed. He describes the conflict between the Hun Shan-yu versus the Chinese usurper emperor, Wang Mang. With Wang Mang's death in 23 AD all Chinese power over the Western Regions was lost. Stein writes an extensive analysis of the names and descriptons of locations, distances mentioned in the records and his personal measurements to identify ancient towns with their current equivalents. .

Section II - Turfan from Later Han to T'ang Times
Stein continues with his historical narrative. During the Later Han dynasty the Turfan (that is Chu-shih) remained under Hun control. In AD 45 Hun extortion caused the king of Posterior Chu-shih together with the rulers of Shan-shan (Lop) and Yen-ch'i (Kara-shar) to ask for Chinese assistance which could not be provided. When Hun power declined the independent 'kingdoms' of the basin began war with each other. The Chinese attempted to regain control in 76 AD but failed. Chinese famous general Pan Ch'ao defeated the Huns in 89 AD. By 91 AD the Chinese had regained not only Hami but also Chu-shih. But by 107 AD the Chinese again were forced to withdraw and Posterior Chu-shi again. Multi-sided warfare continued. The Huns were threatening even Tun-huang and Kan-chou. To prevent them from linking with the Chiang nomads in the Nan-shan Emperor A'ti sent Pan Yung in 123 AD into the Tarim basin. In 131 he regained Hami. In 125 he won against the Posterior Chu-shih. Warfare continued, but Han power gradually declined to downfall in 220 AD. Stein continues with mentions of individual Chinese campaigns during the following 4 centuries, which I will skip over. He wrote that Chinese power began a more effective resurgence under the Sui around 608 AD. By then the Huns had been replaced by Turkish tribes. With the T'ang Chinese power again expanded, and Stein describes events. A Chinese army conquered the Turfan in 640 AD using siege engines to take Kara-khoja (Kao-ch'ang). The capital of this An-hsi Proectorate was transferred to Kucha in 658 AD. Stein notes that all this Chinese strategic campaigning indicates their recognition that Turfan was the critical location. The Chinese simultaneously occupied the area north of the T'ien shan to maintain complete control. Stein's favorite pilgrim Hsuan-tsang, passed through Turfan on his way west to India in 630 AD. Since by then Turfan (Kao-ch'ang) was already an administrative part of China, Hsuan-tsang did not describe it. But he did note that the sister of the local governor, Ch'u Wen-t'ao, was married to the eldest son of T'ung She-hu. supreme Kagan of the Western Turks. The An-hsi Protectorate supervised the 'Four Garrisons' - Kucha, Khotan, Kashgar and Tokmak) .After 670 AD the Tibetans invaded, defeated the Chinese, and reigned supreme throughout the region until 692 AD, in which year the Chinese regained control. But the Arabs destroyed the Chinese army at Tashkend west of the Pamirs in 751, ending Chinese expansion in that direction. The Tibetans again struck and captured Tun-huang and Kan-su in 766 ,severing Chinese communications routes to Turfan and An-hsi. Stein refers to his Ancient Khotan for more details. But the Chinese held out in Hami and Pei-t'ing. In 790 the Uigurs claimed to be relieving Pei-t'ing but their demands caused the remaining Chinese garrison and civilians to accept Tibetan control.
It is difficult to imagine today that the Tibetans could have been so militarily powerful as to defeat the Chinese and occupy the entire Tarim Basin, but Stein's excavations at Miran and elsewhere show the results.

Section III - Turfan under the Uigurs
But the Tibetan control of the Tarim didn't last long. By the middle ninth century the Uigurs were pressing south, another tribal group being forced to migrate south by more powerful nomads, in this case the Kirghiz. The Uigurs dominated as far southeast as Kan-su until about 1031, when they were replaced in turn by the Tangut (Hsi-hsia). But the Uigurs continued to rule over the western Tarim and T'ien-shan for centuries, despite the advance of Islam with the Karluk-turk dynasty from Kashgar (And they remain there today). Stein notes that the Uigurs selected the Turfan for their capital (but also resided in Pei-t'ing) in order to control a wide area. He mentions the Visit by Wang Yen-te to the Kagan Arslan in 982 AD as a source of information indicating the Uigurs controlled the area as far south as Khotan and west to Kucha. At that time Wang Yen-te saw 50 Buddhist convents and indicated the extent of Buddhist religious and cultural influence. He also reported extensive horse breeding operations. When Chingis Khan began the Mongol expansion the Uigurs became allies thus preserving themselves. During the Mongol supremacy Buddhism and Taoism (and even Nestorianism) were retained to some extent, but towards the western side Islam was gaining converts. It appears that by the end of the 14th century Islam had taken over most areas but Buddhism persisted longer in Turfan. The result of this is the existence of so many Buddhist structures in this area. It was Chinese and then Uigur political power that enabled these buildings to survive longer and without massive destruction. Also, the unusual topography, which provided water, largely flowing underground from the T'ien shan into the Turfan depression enabled continuous agriculture and therefore occupation. The Buddhist ruins all lie near villages and towns that have been continuously occupied. There are no ruined sites similar to Niya or Lou-lan.

Chapter XVIII - At Ruined Sites of Turfan

Section I - Among the Ruins of Ancient Kao-ch'ang
Stein returns in this chapter to his present undertakings. He spent the first 6 days in Turfan organizing his projected projects and doing reconnaissances of likely sites. As always he made a point of immediate visits to the local Chinese officials to secure their assistance. He comments that he found 3 month's worth of mail awaiting him there. Again, a remarkable example of the mail system. Naik Shamsuddin, Ibrahim Beg and Li Ssu-yeh arrived from Mao-mei with the heavy baggage and shortly after Muhammad Yaqub arrived, completing his survey line from Hami, and Lal Singh arrived from Guchen. (Muhammad's route shows on maps linked above south of the T'ien-Shan.
On 12 November he moved the team to Kara-khoja. ( map 28acs shows Kara-khoja Chong-hassar and all the other places in SE Tarim ). Where he remained until the 14th. Lal Singh was to perform another solo survey south to Singer and then back east to Altmish-bulak. Stein is effusive in pointing out the extreme difficulties Lal Singh would encounter during this difficult travel. It was hoped that from there he would be able to observe known peaks in the Kun-lun to establish survey control, if dust did not obscure the distant mountains. (It turned out that Lal Singh made an unusual mistake necessitating corrections, as we find out in Stein's memoir on the map-making process). Then he was to survey the Kuruk-tagh north-east between Altmish-bulak and Hami.
Stein focused his own attention on archeological sites around Kara-khoja. (map28eas ) He noted the serious destruction caused by previous explorers and even more by the locals digging for artifacts they could sell. Khara-khoja itself had suffered great damage since Stein's visit in 1907. Stein as always cites the work accomplished by other archeologists and explorers and explains that there was relatively less for him to do on unexplored sites around Turfan, but much survey work still required to establish the geography and topography. But he had little available time, anyway, and was so busy he could not accomplish all the archeological work he would have liked. For instance, he made only a few experimental digs at Idikut-shahri.
He makes an interesting remark about local labor. Here in Turfan and along the northern rim of the Taklamakan labor was in short supply because it was needed for the extensive agricultural work, while water was relatively abundant. While along the southern edge - Yarkand to Charkhlik - labor was in plentiful supply but lack of water limited his ability to engage large teams of workers in the desert.
Muhammad Yakub and Afraz-gul prepared a survey sketch of Idikut-shahri - Archeological work began at the building marked I in that plan. There he found Manichaeans texts in Sogdian script and a few Chinese fragments. He found fragments of frescos at the base of one wall which still held a few other decorations. Extensive digging revealed more wall to a height of nearly 12 feet. He was able to remove surviving frescos preserved in the lower 3 feet of this wall. He excavated another structure 50 yards to the north. Two groups of sepulchral monuments outside the town walls appear and another group is in . There a large stupa was excavated revealing a thick outer wall and tomb chamber plan 25. In the passage between the outer and inner walls he recovered fine frescos that had fallen. He describes this structure in detail along with the many artifacts recovered in it. The result was to show the extensive Buddhist presence north of the Taklamakan well into the 7th century and after the expansion of Islam into the Tarim basin. He found more paper sheets with Chinese and others with Brahmi, Sogdian, Runic Turki, and Uigur writing, some these were Manichaeans texts, some brought there from the West. A large 'hoard'' of metal objects (including iron, bronze and copper) was discovered close to the outer wall. Of the 61 copper coins found there some dated as late as the Sung Dynasty (1102-7). Illustrations of many relics are shown in plates in Vol III and detailed descriptions are at the end of this section.
While at Kara-Khoja Stein was 'informed' about another 'ghost city' high in the hills. He sent Afraz-gul to investigate, but it turned out to be another of the mythical towns such as he had not found in 1908 near the Inchike-darya (described in Serindia).

Section II - List of Antiques excavated or acquired at Kara-khoja
This section is 13 pages of small type font descriptions of the many artifacts Stein recovered around Kara-khoja. Many of these are shown in the plates published in Vol III. The extensive list shows what Stein could accomplish in a limited location already the subject of other archeologists, even when he claimed lack of time and laborers

Section III -Search among the Ruins of Toyuk ( east of Kara-khoja map28acs )
On 11 November, with the two main surveyors gone on their solo treks westward toward Lou-lan, Stein left Khara-khoja to explore north-east of Turfan into the foothills. He stopped again (before in 1907) at Pichan to visit the Chinese magistrate. Among other things he wanted to examine more of the 'karez' irrigation. A Karez, found also in pre-Mongol Iran, was an underground tunnel dug out to form a aqueduct to carry irrigation water safely to avoid evaporation in the intense heat. In the T'ien shan foothills these were particularly effective in catching the runoff from the melting snow and transporting it below the rocky sarai to the fertile oases further south. He fully describes the terrain crossed during this short trip. Plan shows a Mazar encountered along the route. He found that this Muhammadan Mazar had been built into a pre-existing Buddhist shrine. Nearby was a much larger ruin of a Buddhist shrine and monastery. In the cliff above were caves also used as shrines. There were also two large, ruined watch towers built of bricks 13x8x4 inches. One tower was 19 feet square and 30 feet high with two flights of stairs leading to a room 8 feet square. The other tower on a bit of high ground was 16 feet square. Further toward Toyuk Stein came upon another tower "tower of Sirkup'. and Rather than a watch-tower, this was another Buddhist shrine. It had a base 48 feet square at the base and 10.5 feet high built of bricks 14x9x4 inches. Above this were six remaining receding stories each with arches and niches for Buddhist images, some of which were still visible. He could see red paint remaining in some niches and presumed that the structure was originally highly decorated. The structure was too fragile for safe climbing. Stein estimated its remaining height at 50 feet.
On 23 November Stein reached Toyuk and remained there for 15 busy days. Despite the many previous visits by archeologists, he found much remaining to exploit. The area was rich in vineyards and orchards able to support extensive shrines with many pilgrims, both Buddhist and Muhammadan. The Buddhist cave shrines that extended along both sides of a gorge for a mile, were replaced by the famous Muhammadan Mazar of Asahab-Kahaf (Seven Seekers). Stein was welcomed by the Mazar attendant, Kare Akhun Chiraghchi. The Buddhist caves had been ransacked by vandals and treasure seekers. Even so, the several German archeologists, whom Stein always gave credit, had found extensive art work. Stein then focused on locations that appeared to remain fruitful. He began on the northern side . He excavated rooms in which he found manuscripts buried under debris. He then shifted attention to a group of Shrines above the right bank of the stream 3/4 mile from Toyuk Mazar. and and . He began again on the other side. The caves were much damaged. . This was a small cella 12 feet square. He describes the remains in detail and provides illustrations in Vol III. Moving on he finally found a cave in which mud and debris had preserved some wall and ceiling frescos, which Naik Shamsuddin successfully removed. The room also contained silk and manuscript fragments. But further on he again found debris and fragments that indicated vandals had already taken out what they could. shows one of these. More important was the painted ceiling which shows "two rows of drawn and painted small figures all haloed and seated and grouped amidst exquisitely designed floral tracery." In the center was a Bodhisattva. Stein believed that the central figure represented the sun or moon and the 28 surrounding figures represented Naksatras or lunar mansions. With his special skill and great care Naik Shamsuddin managed to remove the entire ceiling to be installed in New Delhi. One important document found at Toyuk, Stein notes, has been identified as dating from 599 AD and mentions the name of the king of Kao-ch'ang.

Stein then moved on to the ruins at Murtuk ( map28acs northeast of Turfan and north of Khara-khoto)

Section IV - List of Antiques excavated at, or acquired from Ruins of Toyuk
Another 13 pages of densely printed notes about each of the antiques found in a location Stein assessed as already being thoroughly worked over by German archeologists plus hordes of local vandals and treasure-seekers. Remarkable what one man's perserverance could accomplish.

Section V - Work at the Sites of Murtuk
On 9 December Stein departed Toyuk for Murtuk. He had already visited the cave temples and shrines at Bezeklik, south of Murtuk. He notes that this was the largest of the Buddhist shrines and that Professor Grunwedel had devoted two months to its study, during which he removed many panels to Berlin. Stein graciously credits the extensive expert description of the frescos that enabled him to select the most important for removal. Stein spent a total 5 days here during two visits. But thanks to the training he had given Naik Shamsuddin and Afraz-gul, these two stalwart assistants were able to work there for 2 months removing wall paintings and carefully packing them for transport, while Stein explored elsewhere. Stein provides a plan and a few photos and and and . Stein meanwhile from Dec. 14, 1914 to Jan. 7, 1915 traveled the 115 miles north back across the T'ien-shan to Urumchi to visit his old friend, P'an Ta-jen, who was now Financial Commissioner of the Province. He does not fail to describe this journey in the dead of winter.
Stein explains that the visit was not entirely social. He had learned of expanded Chinese official concern about his activities, both archeological and surveying. He was anxious to forestall official edicts that would have curtailed or totally prevented his work. All this was generated by the new regime in Peking since the 1911 Revolution. The Chinese were becoming even more antagonistic to foreign 'looting' of their historical treasures.
He remained in Urumchi from 18 Dec. to 3 January and took advantage of the time to visit also the Russian Consulate, the Belgian Mission and the China Inland Mission, whose leaders he gratefully acknowledges. He was reassured by the Russian resident surgeon that his leg would heal fully. But he recognized that the two senior Chinese officials - the Governor General and his Foreign Affairs adviser, were not that friendly.
Returning to Murtuk, Stein was pleased to find that 50 cases of frescos and other artifacts were ready for transport by camel to Kucha and then Kashgar with more in preparation. The sooner his relics were safely on their way the better - and without him around them to bring undesirable attention.
Until 17 January he remained at Murtuk to study Bezeklik shrines and select more frescos. He found a few more interesting sites. These marked as pliv pl vi in Plan 29 again. The section concludes with 4 more pages listing details of antiques found around Murtuk.

Chapter XIX - The Ancient Cemeteries of Astana

Section I - Seventh-century Tombs in Group i
With the precious camel loads on their way west for the long trek to Kashgar, Stein turned his attention to one last project near Murtuk. This was the extensive ancient cemetery north of Astana. ( in same oasis as Khara-khoto map28eas ) The sketch plan of the cemetery is . Naturally, these graves had been opened and looted already in previous centuries, plus they attracted the attention of recent explorers and archeologists and local treasure seekers. Nevertheless, Stein was determined to search them for himself. He spent two weeks at this activity. The local daroga supplied both an expert guide and work party to do the digging. Plan 31 again. As usual, despite all the previous looting, Stein managed to find much to report. One interesting example is that some of the buried had Byzantine or Sasanian gold or silver coins. As usual, Stein provides a massive amount of detail about both his process and methods and the content of the graves. And he draws many conclusions from his finds. One set of graves he can date to around 640 AD from a few coins.

Section II - Figurines and other Sepulchral Deposits in Groups ii- v
In this section he focuses on some of the items found in the tombs and supplies many illustrations and and plan again. Some of the 'monsters' found there are . There were also many models of mounted warriors. Among the documents found there are various texts related to mounted service, horses, transport and the like. In this group of graves he found fragments of documents with dates from 690 to 709 AD. These also related to the employment of horses.

Section III - Intact and other Burials in Tombs of Groups vi-x
Stein here shifts to another group of graves which also had been ransacked by grave robbers. He provides results similar to those in the preceding sections. and and

Section IV - General Observations on the Astana Burials and their Textiles
In his summary he notes that the inscriptions found on various graves dated from 608 to 698 AD and the dates on such documents that had them corresponded. The manner of burial indicated Chinese customs, which in turn indicate that Chinese culture was strong in the area of Turfan even before it was annexed to the Chinese Empire. He describes the manner in which the burials correspond to historical texts relating to Chinese burial customs, such as placing the bodies of man and wife in the same tomb. Another was the ancient practice of placing gold or silver coins in the deceased's mouth. The type of garments and provisions included are also Chinese.

Section V - Relics of Textile Art from the Tombs of Astana
In this section Stein focuses more attention to analysis of the textiles mentioned in the previous sections. He notes that the dates of the graves situate the silk and the very few other textiles between those of the later T'ang era found at Tun-huang and those of much earlier era from Lou-lan. Burial custom was to use worthless textile for shrouds. Silk was not produced at or near Turfan. Yet silk was the textile used for these wrappings. Ergo it must have been imported in significant quantity to Kao-ch'eng. Yet, Hsuan-tsang, Stein's patron saint Buddhist pilgrim, noted during his passage there in 630 that there was very little traffic on this route from central China. Stein concludes that the silk then was likely imported from Khotan and from Sogdiana, that is Ferghana, Samarkand and Bukhara. If so this is indeed significant. We must note also that Hsuan-tsang traveled west as far as Samarkand before turning south to cross Afghanistan to India. This comment only touches the surface of Stein's analysis and conclusions drawn from study of these textile fragments. Some of the silks showed definite Chinese designs, but many were clearly of Sasanian design. These silks date fully three centuries earlier than those recovered at Tun-huang. Stein describes 'Sasanian design', both as to the weave and the color patterns in detail. Study of this lengthy section would be well worth the time of experts interested in the history of 'Chinese' silk manufacture.

Section VI - List of Antiques from Cemeteries near Astana
This is another detailed list, filling 31 pages. It is another example of Stein's industrious and skilled archeological methods. To produce so much of value from sites already thoroughly ransacked over centuries is remarkable. Notice also that this work on an open plain was conducted in the dead of a Central Asian winter.

Section VII - Conclusion of work at Turfan
Stein was always thinking of and preparing for more than one project. He notes here that during the work at Astana Lal Singh arrived back on schedule from his ardorus survey route through the Kuruk-tagh to Singer and Astin-bulak (approximately due north of Charkhlik and northwest of Lou-lan) and back during which he filled enough plane table sheets for three large maps and established triangulation by sighting across 130 miles to peaks in the Kun-lun range (south of Charklik). ( map shows the relationship to the Taklamakan ) Stein photographed his team. After a brief rest Lal Singh set out again to survey a route west from Singer along the T'ien Shan to Korla.
On February 6 Afraz-gul was then dispatched with the seven best camels and Abdulmalik as guide to complete more surveys toward Lou-lan and meet Stein on 11 March at Ying-p'an. Muhammad Yaqub continued to make a detailed survey within the Turfan Depression for a map at scale one mile to the inch.
( mapindexx4bs ) shows the overall relationship of all these places.)
Stein prepared to follow to Singer. But first he wanted to return to Yar-khoto, which he had explored briefly during his second expedition. map 28 And before that he had to deal with the local magistrate about edicts issued from Urumchi.
On 6 February Stein was pleased to see his caravan of 45 camels with 145 cases of artifacts depart for the six weeks journey to Kashgar with Ibrahim Beg in charge. The local Tungan official was happy to assist Stein and warned about coming government controls.
On 9 February Stein went the 4 miles from Turfan to Yar-khoto.
But on 11 February the official message from Urumchi arrived with an edict that he should stop digging.
On 13 February Stein had to curtail his work and depart Yar-khoto to disappear into the Kuruk-tagh hills. He had completed one sketch and some preliminary excavations. He describes the ruined town as one of the most interesting and largest ruins near Turfan. It was built on a plateau above ravines "Yars" providing it with natural strength. It was called Chiao-ho in medieval Chinese of the Han and T'ang eras and was the capital of Anterior Chu-shih. The plateau "extends for over a mile from north-west to southeast.' The yars, which completely surround the town, and are 100 feet or more deep, generally have water. The sides are cliffs providing natural defense. There are only two practical approaches, both difficult and defended. A Buddhist shrine was photographed for Serindia(279). It had 80 small towers around a central stupa. Stein also found two Viharas (shrines). and in which he found some Chinese and Uigur text fragments. Stein determined that this was a Buddhist monastery still occupied in Uigur era. The floors of the rooms were dug below ground level to provide cooler accommodations during the hot summers. There were cemeteries on the plateau north of the building complex. This was divided into two sections ( shows the larger one that contained dwellings. Plan II Is a large temple. . Within the temple was a tall structure similar to the tura at Sirkup with niches in rows on all four sides containing statues of Buddha. and This large Vihara, Stein believes, was the principal Buddhist temple of the capital of Turfan. He found a second shrine 13 feet tall cut out of the clay ground with walls 44 by 34 feet and 17 feet high composed of three distinct layers. shows additional buildings. shows the buildings had deep basements for summer and were several stories high for winter. Stein found wells as much as 100 feet deep within some houses. Stein notes that market trade was likely performed in suburbs east of the town, since the city itself was inaccessible to camels or carts. The capital was a Buddhist center plus residence for wealthy officials and rulers. The panorama shows that the plateau was not completely occupied by houses. There was also a deep cave dug into the side of one road. shows the southern part of the town. Stein decided that by Moslem times Kao-ch-ang (Khara-khoja) only 4 miles northwest was the new capital and most important city.

Chapter XX - Explorations of the Kuruk-tagh

Section I - From Turfan to Singer

On 16 February, after negotiating with the local Tungan magistrate for suitable diplomatic replies to Urumchi that would satisfy the magistrate's official needs but prevent interference with Stein's work, he departed his accommodation with the Russian Ak-sakal in Turfan for the Kuruk-tagh and on to the Lop sea. He traveled directly to Singer to meet his guide, Muhammad Baqir, Abdurrahim's youngest brother, and visit several more ancient sites. Then he intended to continue to survey areas past Yardan-bulak to ancient cemeteries noticed on previous crossings. The main objective of Stein's Lal Singh's and Afraz-gul's repeated surveys was to determine the actual courses of the Kuruk-darya and Konche-darya river beds to Lou-lan and the Lop salt sea. At that time this was a controversial geographical question significant for determining the source of water for Lou-lan. Stein believed that the now-dry Kuruk-darya had in early medieval times flowed past Lou-lan and into the Lop sea toward its northern edge, but had dried plus the flow had made a sharp turn south so that the Tarim river then flowed toward Miran and the southern corner of Lop. (From modern maps it appears that the river again flows east and into the Lop sea toward the northern edge.)
Stein took the shortest route, due south through the basin to Singer. On the way he examined the ice formations in the Toksun river from which he supplied his party. Crossing the low hills of the outer Kuruk-tagh, which also forms the southern rim of the Tarim Basin, he reached the spring at Achchik-bulak. From an elevation below sea level they soon reached a height of 4,300 feet. ( map28gas and map29aas) Stein as usual describes the topography including the various salt-springs and the caravan routes between Lop and Toksun. This he relates to the probable routes used in Han times.
By 21 February he was deep into the successive ridges of the Kuruk-tagh. The valleys generally had salt marshes since there was no external drainage.
On 22 February they crossed the third range, again at elevation of 4.600 feet. ( map29aas ) shows the western Kuruk-tagh.
On 23 February they halted at Singer oasis, the only inhabited site in the Kuruk-tagh. - The oasis is formed by a single spring and was occupied by one family. They survive on the grain and orchard fruit they can sell to the few travelers plus the valuable meat of wild camels and other game. Muhammad explained to Stein that his grand father had settled at the oasis to build the small area of cultivation which his father, Yusuf Salich had extended. But his four sons found the meager agricultural production insufficient for four families, so Abdulmalik had moved to Deghar and Abddurrahim had moved to Tikenlik, leaving Muhammad to care for the family realm. But at the moment all four were now guiding different of Stein's explorers. Stein notes that Singer lies on the direct route between Turfan basin and Lou-lan. He then discusses the current and medieval climate conditions and their relationship to the possibilities for Chinese travel through the region.

Section II - To Po-ch'eng-tzu and Shindi
On 24 February Stein departed Singer. ( map29abs ) Po-ch'eng-tzu was an abandoned ancient Chinese way station currently occupied by one small family who were digging in a lead mine. He noted the evidence of ancient mining shown by old slag heaps. The station held also the remains of an old fort about 80 yards wide by 40 deep. The remaining walls were of cut slabs of clay about 3 feet thick with square corner towers. Water was supplied by a short stream eminating from springs. Stein noted sufficient water for a much enlarged area not being cultivated or lack of labor.
On 26 February he continued toward Shindi. (map25aas ) All the terrain and vegetation along these routes are amply described, even down to the size of individual elm trees. Clearly, he always took detailed notes each day. He passed hills whose elevation he estimated at 10,000 feet. He now was on map 25 crossing a pass at 5,800 elevation. Further on via valleys he found the well-marked route between Singer and Shindi. While passing he measured the water flow here at 14 cubic feet per second, (I mention just as an example of his practice whenever finding flowing water). There was evidence of repeated use of the area by Mongols for grazing. And here he met again Abdurramim, eldest of the 4 brothers, who had moved his family here and undertaken an effort to cultivate the area, but hampered by the lack of labor available from Tikenlik ( which still exists considerably further south on the Tarim river and new highway between Ruggiang and Korla). Abdurrahman guided Stein through the ruin at Shindi. This consisted of two small buildings, one a tower, which Stein describes. From Shindi Stein took a circuitous route through the mountains south via Ala-tagh and Azghan-bulak to reach a section of the Lop-Turfan route ( map29acs ) to Toghrak-bulak at 3,600 feet elevation.
On 3 March Stein turned south-east toward Yardang-bulak. They reached this spring early on 4 March. It was an ice covered water about 500 by 150 yards and provides both ice to make reasonably fresh water and grazing, so Stein spent a day refreshing his team and the camels. Stein, as usual, spent the day writing.

Section III - To Ancient Graveyards by the Kuruk-darya
On 6 March Stein departed Yardang-bulak toward the Kuruk-darya. He left most of the baggage at Yardang-bulak in order to carry more ice. After a few miles walking they found the location of Lal Singh's earlier camp but no water. So Stein had to halt there a day and send the camels back to Yardang-bulak to bring more ice. The following day the camels returned and Stein set out again to find the graves near the Kuruk-darya sited by Lal Singh the previous year. Eleven miles of further walking brought them to the wide, dry bed of the Kuruk-darya. Following Lal Singh's survey they soon found cemetery 1. Next day they begain excavation of the cemetery. - This is another example of Stein's persistance - to travel far out into the desert just to explore a few graves seen the previous year, but not studied for lack of time. He found that the central grave had been burnt. He described each grave and managed to recover a few artifacts listed with the text. In all 6 graves were opened. All were very ancient. Stein concluded these were not Chinese but the indigenous population of Lou-lan area. The burial methods were the same as those used in the graves at L.F. at Lou-lan. Next day the team moved about 5 miles east to Lal Singh's marked cemetery 2, which contained about 22 graves. Stein excavated 6 of these and found only skeletons without coffins. Stein determined that these also were the graves not of Chinese but of the indigenous nomadic people.
On 11 March Stein turned back south-westward to survey a section of the Kuruk-darya. They explored into the desert to the south and then returned north-west to the dry river. They found water, but too salty, at a depth of only 5 feet. The section concludes with another page listing finds at the two burial grounds.
The

Section IV -Mian Afraz-gul's Supplementary Surveys
On 13 March Stein returned to Yardang-bulak and remained there waiting for Afraz-gul and let the camels rest again. Sure enough Hassan Akhun arrived with the best camels followed shortly by Afraz-gul and Abdulmalik. Thus two of the brothers were reunited after long separations.
This section recounts his efforts and their results. He had trekked from the edge of the Tarim river at Chainut-kul ( map30cas ) past L. M at to the Kruk-darya (That is well to the south east of Stein's survey area.
Stein summarizes Afraz-gul's report.
On 6 February he departed from Kara-khoja to Lukchum via Toyuk. From there he went to Deghar. With Abdulmalik (younger brother in the clan) as guide he crossed the Kok-dawan at 2,260 feet on the third day. They then turned south-south-west across barren plateaus at 3,400 feet.
On 14 February they crossed the main range of the Kuruk-tagh. They halted at Kuruk-toghrak-bulak and then.
On 15 February they reached Altmish-bulak. ( map29ccs and map29cbs ) (The important spring at which Stein had kept is camels in the previous year.) There Hassan Akhun was left with the camels again.
On 16 February Afraz-gul took two men on foot to the ruins north north east of L.E. fort, which they had skipped excavating the previous year. For this they had to carry eqipment including the plane table, ketmans, and food and ice for four days supply.
On 17 February then returned to the mesa Afraz -gul had found in 1914. L.Q. map 32 - The graves proved to contain bodies similar to those at L.F. - that is indigenous people. Bronze items were recovered in the vicinity. On another mesa he returned to a tower previously sighted. He then returned to Altmish-bulak. After a day of preparation, he set out for another mesa seen the previous year. From there he surveyed the western 'shore' of the Lop sea toward the south.
On 23 February with Abdulmalik he found the track that Stein had made in 1914 toward Tun-huang. From there they turned again south-west along the 'shore'. The plane table results appear in maps 29 and 32.
By 1 March, having crossed the survey routes from 1906 and 1913, they reached Chainut-kol ( map30cas ) with fresh spring flood of the Tarim river all about.
On 6 March he started back west for Yardang-bulak via site L.M. He found many artifacts on the open ground as he passed. The section concludes with a page listing of items found by Afraz-gul

Chapter XXI - On the Ancient Route along the Konche-darya

Section I - The Ruins of Ying-p'an
On 17 March Stein set out for Ying-p'an. The following day they reached Ying-p'an ruin where they met a team from Tikenlik sent to bring them supplies. ( map25bas ) Stein was amazed to meet in this group a Panjabi he had known 14 years previously. He connects this occasion with a conclusion that this was typical of the way Indians and other Westerners in medieval times would have penetrated throughout the basin as traders. Next day he turned his team to work on the Ying-p'an ruin first discovered by Sven Hedin in 1896. ( map25bbs ) Right away Stein came upon a group of stupas . The ruins are shown in on a plateau 28 feet high above the Shindi river. The plateau is shown on . The main stupa is and . Its diameter was 18 feet, resting on a square base 27 each side and 7 feet high, but Stein could not determine its original height. This was constructed out of sun-dried bricks 15x13x3 inches. Between layers of brick there were layers of stamped clay and gravel 5-6 inches tick. Around the stupa was a brick wall about 3 feet thick and 61 by 50 feet in length and width. Stein quickly found his rubbish heaps, overlooked by previous searchers. Among the debris were found Kharosthi wooden 'documents'. There were 9 smaller stupas around the central one. The whole site was typical of a worship shrine found at the head of a life-giving river. Stein visited two more small stupas a few miles north along the stream. The better preserved of these measured 15 feet square at the base of 7 feet height topped by a 8.5 foot high dome. They were built of sun-dried bricks 15x13x3 inches. That the site had been occupied past Buddhist times was shown by Stein's discovery of two Muhammadan cemeteries one containing about 33 graves and the other 23 graves. There was a small fort. shows the walls around the site. This was circular with diameter of 194 yards, built of stamped clay and 24 feet thick at base. The better preserved sections were 18 feet high. Thirty foot wide gates were in the east and west sides. Stein dated the fort to Han times since it corresponded to the fort at nearby Merdek. But coins found there indicated it was occupied still on T'ang era. Outside the west gate Stein found another shrine containing a stupa and . It had been partly ruined by some previous careless archeologist. This shrine had a platform some 46 feet by 40 and 13 feet high built of bricks 15x12x4 inches. The stupa was about 17 feet in diameter on a base 23 feet square. He found the remains of seated Buddhas formerly at the corners. Stein found yet another grave site west of this stupa. Stein opened four of these graves and describes their content in detail. He believed these graves were much earlier than the Muhammadan graves, but not so old as those at Lou-lan, probably T'ang era graves. The section ends with a 3 page list of artifacts found here.

Section II - The Ancient Course of the Konche-darya and the 'Town of Chu-pin'
Here Stein once again discusses the controversy over the ancient course of the Tarim River - its tributaries the Konche-darya and Kuruk-darya. This was a argument between explorers- geographers of the time, now overtaken by changes. The question was where did the Kuruk-darya /Konche-darya flow into the Lop seas, more toward the north or to the south. Stein expressed his view also in Serindia. The main opposing view was that of Sven Hedin. The historical connection was how did Lou-lan receive sufficient water to irrigate agriculture and provide for its inhabitants. During his second expedition in which he surveyed around Lou-lan in 1906 he found evidence in the dry river beds that the Kuruk-darya must have flowed nearby at the time the station existed. He returned to Lou-lan in this third expedtion in 1914 and conducted further surveys of the area including a short distance west. But at that time he then traveled east across the Lop sea to Tun-huang. Therefor his performing multiple survey lines, by himself, by Afraz-gul and by Lal Singh in February- March 1916 was to connect the western origins of the two river beds conclusively with the Lou-lan area. He affirmed that these surveys proved that the delta of the Konche-darya was indeed east and south but near Lou-lan. Moreover, he found that the Kuruk-darya, now dry, had received its water from the adjacent Konche-darya which still flowed. Stein credits the Russian explorer, Colonel Kozlov, with finding in his surveys of 1896 that the head of the Kuruk-darya had been in marshes near Ying-p'an and that the bed of the Konche-darya still flows near the same place. But no explorer had yet proven the connections. He states that this was the purpose of both his expanded surveys of the river and examination of Ying-p'an. He identifies Ying-p'an with the ancient Chinese name Chu-pin on the then high road between Lou-lan and Korla. He points out that the size of the ruins at Ying-p'an shows that this area was much more populated and had much more agricultural production in Han and T'ang times than today.
Stein is at pains to disprove the theory advanced by Sven Hedin that the Kuruk-darya had taken the entire volume of water from that side of the Tarim including that of the formerly flowing Konche-darya by a bend to the south and far south of Lou-lan. Then the modern Tarim river took all this water toward the south-west edge of the Lop Sea. Stein also quotes at length ancient Chinese records which Dr. Hedin had not seen to corroborate his theory. ( On map25eas and map25das ) he depicts the geography as far as the current headquarters at Kara-kum. ( map25fas ) The Konche-darya now flows more south near Tikenlik and provides water to the Tarim River. The flow of the Konche-darya is made more uniform throughout the year by the existence of Lake Baghrask which acts as a reservoir to hold excess spring flow of melted snow for later discharge into the river. This in turn meant that when the Konche-darya flowed into the now-dry Kuruk-darya it provided the steady flow of water needed at Lou-lan. Stein continues by noting that the lack of gradient throughout the whole region resulted in frequent shifts of the bed of the rivers. Moreover, this phenomena could have caused the Inchike-darya and even the Yarkand rivers to the west also to provided water to the Konche-darya.
All Stein's extensive efforts at physical study on the spot and mining of ancient Chinese records is academic now. A look at a recent map of Chinese Turkestan shows that there are many channels now. The major river, the Tarim, flows across the northern edge of the desert and then turns south- east before passing Tikenlik (map25bcs ) and the turns full southward to end in its delta west of Abdal and Miran. It does receive water from the many rivers flowing out of the T'ien-shan between Korla and Kucha and further east (for instance the Shindi river) as well. But these rivers continue also to flow in multiple channels east past the location of Lou-lan ruin and disappear into the desert at the north-western edge of the Lop Sea.
Stein, having shown the significance of the connection between the Konche and Kuruk daryas around Yin-p'an, returns to narration of his expedition.
On 20 March he conducted another reconnaissance to link the Shindi river with the others by showing how the river beds indicated shifts.
On 21 March he then set out westward to find the Konche-darya. From there Afraz-gul would turn south to Tikenlik to link the survey line with that developed by Lal Singh from the south in January 1913 then return west along the Yarkand-darya and adjacent high road through Kara-kum to Korla. Meanwhile, Stein planned to go across desert northward to find the Ying-p'an route to Korla at Kurghan. The results of Stein's descriptions of the topography show on the map.
On 22 March he camped at Kurghan.

Section III - Watch-stations along the Ancient Road to Korla
Kurgan turned out to be a ruin of a small watch tower, which Sven Hedin had found in 1896. It was a massive tower 34 feet square at base with an enclosing wall, about 10 feet high and 3.5 feet thick at its top, at a distance of 76 feet on all sides. Both structures were constructed in similar fashion to the towers along the Han wall near Tun-huang. It was built of sun-dried bricks 15x7x3 inches with reed layers 2-3 inches thick between each course, and covered with plaster. The tower still had a height of 29 feet with a chamber 12 feet square at the top. At a height of 20 feet the tower wall was 7 feet thick. The southern face showed a breach 5-6 feet wide, where there probably had been the entrance. The tower, then was hollow, in contrast to many solid towers at the Han wall. Stein found the remains or wooden rafters inside, indicating that there had been several stories of rooms. There were loopholes (6 by 4 inches) in the southern face, which faced the outer gate and in the surrounding wall. These were set in two rows 2 feet apart vertically and 5-6 feet apart horizontally, thus providing for maximum firepower. Both external wall and tower were reinforced by Toghrak timbers. Several walls and the tower showed exposure to fire. Stein found evidence that the site had been reoccupied much later than the destructive fire, during the T'ang era. But this tower and the others Stein found dated from Former Han period, as noted in Annals from 101 BC. at that time they were needed for defense against the frequent Hun raids.
On 23 March Stein set out once again toward the north-west. After moving 16 miles they reached the next tower, YII, described by Dr. Hedin. This one was also similar to the Han wall towers and also had adjacent quarters plan 38 . It was on an artificial platform 12 feet above its surroundings. The tower remained about 20 high and originally was probably 20 feet square. It was built of the similar bricks 15x7x8 inches with reed layers each 16 inches. The reinforced masonry facing was about 2 feet thick but fallen on the northern side. The nearby quarters measured 27 by 19 feet. Stein found a few worth-while artifacts in the thick layer of refuse. The platform was extended by a well-built revetment and there were two approach ramps which Stein considered excellent examples of Chinese engineering skill. Running out of water, Stein sent the animals over night south to find the river, They did find a fresh water lagoon and refilled the two tanks. Next day Stein abandoned search for springs and towers apparently missed in passing and set out for the next visible tower ,Y III. It was a pyramid with base 55 feet square and 20 feet at a remaining height of 30 feet built of typical bricks 15x8x3 inches. But there were posts jutting up from the top indicating that the original height was at least 10 more feet Stein noted strong construction from strong timbers. It was designed as a watch and signaling tower. As usual, Stein searched the deep pile of refuse and found a Chinese document and a few wooden items.
About 5 miles further, he found another tower YIV . This one was similar to the last but had decayed much more but still remained to a height of 30 feet. Up to 10 feet the construction was of single layers of bricks with reeds between them. Further up the reeds were only between every 5 or 6 brick courses. And there were vertical and horizontal timbers.
After moving on for 4 miles Stein found another tower Y V. It originally was likely 24 feet square, built of alternating layers of reed fascines 3 inches deep and earth layers 2 inches deep. It was very badly ruined, standing only 12 feet high. The whole tower had slide down the tamarisk cone on which it was built.
On 25 March they followed a track to the Konche-darya during which move they met Ibrahim, a local hunter, who then served well as a local guide. The river was full of clear water flowing at 2 feet per second and 40-50 yards wide. Ibrahim guided them to a very badly damaged tower Y VI, set among tamarisk cones, that they otherwise would have missed. Still, it was 22 feet high.
On 26 March Stein departed camp at Gherilghan-kol to investigate two more towers (called Sanje and Yar-karaul) before reaching district headquarters at Kara-kum. ( map25fas ) Sanje, Y VII, was constructed of solid brick masonry. Only the northern face remained in tact to a height of 25 feet. Stein found that there was an inner core some 35 feet square at its base with an outer addition to 57 feet square. Both parts were of bricks 15x8x3 inches with reed layers between each 4 courses. Stein assessed this tower as from Former Han dynasty era. In the refuse he found several pieces of lacquered leather scale armor. A few empty graves were found nearby.
They moved on to the watch tower called Yar-karaul YVIII located on one of a series of mesas 50 feet high and 112 yards long (yars). and plan 38 again. Only the southern wall, 4 feet thick, remained standing to a height of 10 feet. The tower was about 19 feet square at the base and built of the same size bricks. They continued on to reach Kara-kum after dark, having traveled 31 miles that day even with the two stops to excavate the two towers.

Section IV - The Territory of Wei'li and the Modern Kara-kum
Stein remained at Kara-kum for two days. He associated this area with the ancient Chinese Wei-li kingdom, mentioned in the former Han Annals, just north of the Shan-shan and Chu-mo kingdoms. The Kara-kum district extends from near Korla along the Konche-darya to below Tikenlik. At that time it held 2000 families. The town is not mentioned in the Annals of the Later Han but its king is listed among others defeated by Pan Ch'ao in 94 AD. It is again mentioned in the Wei lio written between AD 239-265. and again in the T'ang annals. Its importance stems from its location as a cultivated link between Charkhlik, Tsaidam and Tun-huang to the south and Korla and regions north of the Taklamakan. (And today it is still on the main highway between those places.) Stein was particularly interested in the several shifts for the location of the district headquarters that resulted in its placement at Kara-kum. He was able to interview the retired district magistrate, Huang Ta-lao-yeh, a Tungan from Urumchi. This official explained the details of local cultivation and the problems with expanding colonization toward Lop. Chief among them was not the water supply, but rather that irrigation and cultivation soon brought salt up from below the surface. Another problem stemmed from the character of the 'colonists' brought from distant oases around the Taklamakan. For the most part these individuals were either lazy or independent souls used to roaming and not sedentary heavy agricultural work. These folk would drift away after a couple years of expending the initial government payment given to attract them. At the same time these colonization efforts were a source of profits for the Yamen officials who extracted their part from the same government payments. Stein examined the agricultural plots and interviewed many farmers to confirm the assessment from the Amban. He found another ruined watch tower nearby. There was a tower at Suget-bulak Y IX.
Stein summarizes his conclusions about the line of towers and its construction as part of the Chinese expansion across the northern route between the Taklamakan desert and the T'ien shan mountains. He mentions also the two similar towers he found near Ming-oi in 1908 as described in Serindia. He notes the close geographic connection between the towers near Korla and those he had just visited west from Ying-p'an. The Kara-shahr valley was a strategic passage between the Chinese colonies being developed south of the T'ien-shan and the Hun grazing regions to its north. ( map21xx8as )
From Suget-bulak Stein passed through excellent and prosperous agricultural areas fed by a canal from near Korla. ( map21xx9as . ) On 30 March he reached Korla. The section concludes with another detailed list of objects found at the six watch towers.

Chapter XXII - From Korla to Kucha

Section I - Along the Foot of the T'ien-shan
Stein's three surveyors all joined him at Korla. Lal Singh had surveyed the rugged western Kuruk-tagh, Muhammad Yaqub had completed his survey in the Turfan depression and then begun survey around Lake Baghrash, and Afraz-gul had surveyed the route from Tikenlik to Korla. Stein was busy there preparing for the next phase of the expedition. He comments that he has already inserted the information about Korla itself in Serindia.
On 6 April the teams set out westward over the 938 miles to Kashgar. ( map21xx10as) Lal Singh was sent north to survey a route through the foothills and as high up in the T'ien-shan as he could. Muhammad Yaqub was sent south to survey the Konche and Inchike Daryas and then the Yarkand darya clear to Yarkand. With him went the best camels in order that they might be rejuvenated before being sold. Afraz-gul will travel with Stein to do what plane-table survey time will permit. Stein rushed along the main routes in order to reach Kashgar by May. Stein notes, that in view of his rush and that the places along the high road had been explored and reported on by many other writers he will confine himself to general remarks about what he sees.
Just past the edge of the cultivated area near Durbil Stein found the first small fort, with walls 16 feet thick of stamped clay still reaching 26 feet high. At a further 16 miles he found a ruined watch-tower reduced to a mass of clay, but its base platform was still 53 feet square with a height of 18 feet. On this was the remains of the tower 26 feet square and 8 feet high built of the same size bricks as found at the other towers. Surrounding this platform was the remanent of a enclosing wall. Stein estimated it was from the Han Dynasty. The next day they reached Charchi after traveling 22 miles.
On 8 April they continued to Eshme.

Section II - The Seat of the Protector-General
Ten miles from Eshme Stein reached Chadir, a larger town. ( map21xx1cs) A further 10 miles from the edge of Chadir cultivation brought them to Yangi-hissar oasis with a population estimated at 800 families. This oasis owed its importance to a route from it leading north to the Yulduz plateau in the mountains; a favored Mongol trade route. Lal Singh surveyed this route up to 11,800 feet in the pass. At Yangi-hissar Stein surveyed north to Ak-tam and then to Bugar while Afraz-gul surveyed south to Aghrak and then to Bugar.( both appear here map21das ) Afraz-gul found the ruin of a small fort at Aghrak. Stein found two small ruins at Ak-tam, the crossed the Kizil River to the south-west and reached Bugar. ( map21abs ) They halted for a day a this district headquarters, a sizable town governing some 4000 families. The Bugar 'old town' ruin was an enclosure 300 yards square of decayed earthen ramparts mostly only 10-12 feet high. But part of the western side still rose a full 18 feet with its top 22 feet wide holding a parapet. Stein considered it was from Muhammadan era. Afraz-gul took a southern route to Kucha along which he found another, similar fort some 260 by 240 yards in circumference, called Koyuk-shahr .
At this point Stein briefly stops his narration to address again the ancient Chinese annals at length from which he concludes that the Bugur area was the location stated for the residence of the 'Protector General' who was to command Chinese garrisons circa 100 -60 BC.

Section III - From Bugur to Kucha
On 12 April Stein left Bugur moving directly west to Kucha while Afraz-gul again took a more southerly route. A few miles on he came to a massive watch -tower inside a walled enclosure known as Lai-su-tura. This one was 48 feet square and still 47 feet high constructed of the same size bricks. A hundred yards to the north he found the remains of a fort whose walls had mostly eroded but originally was about 192 yards square with the remaining wall segments 10 feet thick. One corner tower remained in the south-west corner 39 feet square at base and to a height of 26 feet. Stein's main purpose in finding these watch towers and forts was to confirm that he was on the ancient direct caravan route to Kucha. Another 2 miles west he found the remains of a stupa and another enclosure also on plan 39) 102 by 84 feet in circumference. This one he identified as Muhammadan. He halted to camp at Yangi-abad, a hamlet with 18 families.
On 13 April he set out again for Kucha. He soon found another watch-tower (KV) 32 feet square at its base and 29 feet high with a 13 feet square guard room having 4 feet thick walls. (on plan 39) The bricks again were the same as other towers as well as the timber reinforcing. Further west he found yet another walled enclosure KVI. (also plan 39) This one had one enclosure 57 by 48 yards within another one. Both four-foot thick walls were of the same size bricks. At a further 2.5 miles distance west Stein found the next ruin (KVII) (on plan 39) a small enclosure 22 feet square with walls 10 feet high of the 18x8x3 inch bricks. Another 800 yards west was a larger enclosure 94 feet square (KVIII) with walls 5 feet thick and 13 feet high in places. The bricks were 12x6x3 inches. There was a gate on the southern side, protected by an outer wall. There were 12 foot square bastions at the southwest and east corners. By nightfall they reached the eastern edge of the Kucha oasis.
On 14 April Stein reached Kucha town, a further 17 miles through the cultivated area. He was greeted by the Ak-sakal, Sahib Ali Khan, and his Pathan merchants. ( map17fas )
Stein again digresses to discuss the written records from the T'ang Dynasty relating to the route he had just traveled. He guesses that the T'ang route was a small distance south of the modern road.

Chapter XXIII - Kucha and some of its Ancient Sites

Section I - The Oasis in its Geographical Aspects and the Position of its Ancient Capital
Stein begins his discussion of Kucha with the T'ang records and the significance of Kucha's geographic location. (Today, one can see via Google Maps, what Stein described as the largest cultivated area between the mountains and desert. The whole area is laid out in rectangular fields.) Kucha is watered by two major rivers that flow from glaciers high in the T'ien-shan, the Kucha River and the larger Muz-art-darya. And its position enables the water to reach the area and even further south toward the Tarim river without appreciable loss from evaporation or going underground. The rivers are diverted into numerous canals that enable irrigation over a wide area. He noted (again) that the chief difficulty of expanding agriculture was lack of manpower, the opposite problem from that along the southern edge of the desert. Moreover, the river brings down a considerable amount of fertile soil. To the south the large Kucha oasis cultivated area is protected from any expansion of the desert sand by the wide beds of the Tarim River system which cross the entire length of the line between desert and mountains, something lacking south of the Taklamakan. North of the main oasis the climate and soil conditions are favorable right up into the foothills and in the valleys. From the point of view of the Chinese expansion all this was enhanced by Kucha's location at a cross roads of trade routes from many directions. The east-west route was the main one providing access from China to the Pamirs and beyond into Sogdiana. It is about half way between Kashgar and Turfan and Lou-lan. For these reasons it was the location in which the Chinese established the military and political headquarters for the "Four Garrisons" during T'ang era and was the location for the "Protector General" during Han era. Stein notes also that the extensive remains of Buddhist temples and shrines and outposts attests to its importance. He mentions that the town still has an old wall that existed in sections around it with a rampart of stamped clay 60 feet wide at base and remaining height of 18 feet, with some sections to 23 feet. There were small, square bastions at intervals along the wall. A massive tower, named Pilang-tura was located 3/4 mile from the south-eastern corner. This was built on a stamped clay base 37 feet high out of bricks 16x8x3.5 inches . The tower was 82 long by 70 feet wide at its top plan 39. But the modern buildings around it prevented detailed study. Stein notes that German and French archeologists have already studied and reported on the ancient sites within Kucha, so he will devote his limited time to locations westward, across the Muz-art darya.

Section II - Ruined Sites West of the Muz-art River
Stein spent almost 3 weeks in and around Kucha. Considering his urgency to get to Kashgar that is good evidence of the extent of profitable work he found in this extensive location.( map17aas and map17bas )
On 20 April he began touring sites south-west and west of the cultivated area, guided by Mir Sharif. A few miles out of town they found Kosh-tura, a tower with remaining height of 54 feet. The tower was 95 feet wide on its northern side, 82 feet on the eastern side, but the other two sides were decayed. At 25 feet elevation the masonry recedes to form terraces .Stein pronounced this a Buddhist shrine. The bricks were 15x12x4 inches in size. A few yards to the south-west he found another mound 32 feet square and 36 feet high. Along the way, Stein measured the water flow in seven canals.
On 21 April Stein moved further west to the Toksun area to another ruin named Kalmak-shahr. It was a small fort some 100 feet wide with a wall 14 feet high and 13 to 3 0 feet wide at its base. Stein skipped several other ruins to follow the Khotan route southwest to Dash-tughemen. He found another ruined outpost named Ak-tiken-shahr about 90 yards square. Traversing 27 miles on the 21st he reached a new camp at Shahidlar.
On 22 April he turned again south to find Tonguz-bash. The walls here were about 168 yards on each side with various bastions all built of sun-dried bricks 15x8x3.5 inches. The walls were 18 feet thick and 18-20 feet high. There were gates on north and south sides protected by curtains and an outer court. Stein found no structural remains in this fort. But the refuse dumps contained the usual small artifacts, Stein dated to the T'ang Dynasty as an outpost to guard the route from Khotan. Moving another mile, Stein found a ruin (also) about 130 yards long that he determined were Buddhist shrines and monasteries. After this visit Stein returned to camp at Uzun-pichin.
On 23 April he moved to see several small ruins to the north. These included another enclosure 168 by 153 feet with a wall 15 feet thick and in places 20 -22 feet high, plus bastions. ( ) Further north he passed Topa-shahr to Wang-yari, where he found a cemetery. Another 4 miles away he found an unusual enclosure, named Och-kat, about a mile in diameter that had a triple ring of ramparts. The outer rampart was in places 78 feet thick and 15 feet high. The second ring was 52 feet thick. Stein could not account for this unusual fortification.
On 24 April he continued on to Tajik and Toghrak-akin. At Kosh-tura he found another tower 45 feet square at base and 34 feet high constructed of clay slabs. (This place is clear on Stein's map from Serindia map34 ). To the north 86 yards away was a ruined platform about 46 by 42 feet and 18 feet high. He found evidence of a shrine on the top. Mir Sharif informed him that in past years there had been walls with remains of painting. Some 60 yards to the east was another enclosure. From Kosh-tura Stein went west to Tajik .
He spent 24 to 27 April surveying Tajik. and . With assistance of more workers Stein cleared these sites. One was a Buddhist shrine in which he found many remains including the head of a Bodhisattva. At two miles further on at Toghrak-turn Stein found more Buddhist buildings and at the head of a gorge and in caves along the sides. Stein describes the caves and their remaining contents in detail.
On 28 April he started to return to Kucha.

Section III - Remains South-east of Kucha and List of Antiques found or acquired
On 29 April Stein continued back to Kucha gathering information on canals and Buddhist shrines along the way.
On 30 April he visited the ruin at Kotgluk-ordu.
On 1 May he moved on to Khanak-atam where he met Afraz-gul. The latter had found several more enclosures during his survey from Yulduz-bagh.
On 2 May Mir Sharif took Stein to another ruin named Chong-shahr, an oval earthen rampart 10 feet high and 340 yards on its major axis. Next to it was a mound 70 yards across and 30 feet high. Another small enclosure 398 feet square with walls 7 feet thick lay to the north-east.
On 3 May they continued east-north east to find yet another small, oblong enclosure 200 yards north to south with 15 feet high walls. Stein mentions more enclosures seen by himself or Afraz-gul. The section ends with seven pages listing antiques frond around Kucha. Stein also describes some of these, mainly those of bronze including arrow-heads and buckles.

Chapter XXIV - From Kucha to Kashgar

Section I - Old Remains within the Bai District
On 6 May Stein departed from Kucha along the high road to Kashgar.
On 8 May he visited the large group of Buddhist cave-shrines at Kizil Ming-oi located in ravines along the left bank of the Muz-art river at western edge of Kucha oasis ( map17eas ). The place had already been greatly exploited by German French and Russian archeologists, so Stein defers detailed descriptions of the art works there to other publications.
On 9 May he left the road to Bai to take a northern route to the ruin at Tezak-kaghe Ming-oi (far north-west in map map17eas ) He found the paintings on the cave walls had deteriorated but sufficient remained to identify the caves as Buddhist shrines. A ruined building of stamped clay 40 by 26 yards remained on the top of a small ridge. The end of this ridge contained remains of a walled village 140 by 100 yards, protected by a ditch across the ridge and a stone wall. Stein then remained at Bai visiting the local magistrate. He then visited another Ming-oi on the right bank of the Muz-art. He camped at Jigdalik ( map12gas )
On 13 May Stein visited the cave shrines. There were four cave-shrines close together and others scattered in side ravines. They contained only tracings of wall paintings. Plan shows Jig I where Sahib Ali, Stein's friend and Kucha Ak-sakal had dug up the mass of documents years ago and sent them to Sir George Macartney. The cave was 12 by 14 feet complete with door and window. There were still fragments of Brahmi documents that Stein retrieved. Another cave, Jig II in the plan, contained remnants of painted plaster images including a seated Buddha. Stein found more caves on the other, eastern, side of the ravine one of which, Jig III, he shows in the plan. Stein considered that it was the presence of springs in this location which had enabled to establishment of these Buddhist shrines. This was the final archeological exploration Stein undertook during this expedition.
On 10 May Afraz-gul set out to survey a more southern area toward Ak-su, a route Stein believed Hsiung-tsang had followed.

Section II - Past Ak'su and Maral-bashi to Kashgar
From Bai Stein stayed on the high road past Ak-su and Miral-bashi to Kashgar for the remaining 370 miles, which he covered in 17 days. Of course he had already spent considerable time at those locations during his second expedition. However, he had not then surveyed the high road itself, so he took this opportunity to do so. Ak-su also is a sizable oasis with both 'old town' and 'new town' locations. He described the historical record in Serindia.. ( map12aas )
On 18 and 19 May he visited with another old friend, Mr. Chu Jui-ch'ih, the Tao-t'ai. There also Lal Singh arrived from surveying high in the T'ien-shan ( map11aas ) and then departed to continue traversing as high in the range as possible past Kelpin to Kashgar. Stein then covered the remaining 150 miles in six days. He describes as always the topography and vegetation along the road. He found two ruined forts at Chilan, He found that irrigated agriculture had been greatly expanded throughout the region since his visit in 1908. He also noted that for this section the modern road differed from that of the T'ang era, which had passed watch stations at Chong-tim and Lal-tagh, due to changes in the location of available water.

Section III - A T'ang Itinerary from Ak'su to Kashgar
Stein notes that it required 5 long days for him to travel from Larl-bashi to 'new town' Kashgar. From translations of the T'ang era travel itineraries he determined that ancient Po-huan corresponds to Ak-su and Su-lo to modern Kashgar. Beyond that he is mostly at a loss to establish ancient Chinese place names with their current names. Thus I leave it to the reader of Innermostasia to deal with his elaborate but inconclusive discussion of ancient and modern names.

Chapter XXV - Across The Pamirs

Section I - Preparations at Kashgar
On 31 May Stein reached Kashgar. He was occupied throughout June with careful repacking of his tons of 'antiques' into 182 tin-lined cases to be carried for safe transport by 80 camels over the Karakorum pass 800 miles to India. For this work he was greatly aided by Khan Sahib Badruddin Khan, Ak-sakal of Khotan who came to Kashgar and also arranged for the hiring of skilled carpenters and others. There also he had the assistance of Chiang-Ssu-yeh in translating and transcribing many Chinese documents. Stein much lamented that this was the final meeting he had with his faithful secretary and associate from the second expedition. As always, Stein was also thinking about and arranging for his future work. This was the very unusual opportunity he received from Russian and British governments to conduct his cherished dream of travel through the Russian controlled Pamirs to the Oxus River and then to Samarkand and then across eastern Persia (Iran) to Baluchistan then controlled by India. But he was unable to secure approval to enter Afghanistan and had to content himself with looking at it from the north bank of the Oxus. At Kashgar Stein again met Lal Singh and Muhammad Yaqub' arriving from their surveying in the T'ien Shan and to Yarkand respectively.
On 6 July Stein departed Kashgar. Lal Singh completed another survey on the area near Muz-tagh and the headwaters of the Kashgar river before supervising the camel caravan back to India. ( map2fgas and 2fjas show Lal Singh's surveys west of Kashgar.) Near Opal Stein bid farewell to Chiang-Ssu-yeh with hope to see him again. But his friend died in Kashgar in 1922. Lal Singh and Stein met a final time at Bostan-arche, a Kirghiz at 10000 feet elevation and . Lal Singh was assisted by Naik Shamsuddin and Muhammad Yaqub to survey along the route up the Yarkand river and across the Karakorum to Srinigar. Afraz-gul remained to accompany Stein on his long journey. Since this travel was outside Chinese Turkestan, I greatly reduce a summary of Stein's observations in the following sections and chapters. Throughout the Russian areas he performed no archeological or topographic surveys. And the lack of detailed maps in his report with the trace of his very zigzag route through the Pamirs makes following his trail purely by his text descriptions very difficult. map

Section II - Along the Alai Valley
On July 19 Stein departed from Bostan-arche toward the Pamirs. He was once again in his favorite element, very high mountains with clear and refreshing air. His route was up the valley from the Ulugh-art pass which he crossed at 16,600 feet on 20 July. He was immediately in glacier territory with ravines to precipitous for pack animals. Continuing down past a Kirghiz camp at Sarat Stein was thinking of the descriptions of the region written by Ptolemy. At Kun-tigmaz he met Sir Percy Sykes who was returning to Kashgar.
On 22 July Stein turned north west toward the Altai. ( map2gxxas shows these on the southern route. )
On 23 July he crossed the Kosh-bel pass at 13,800 feet, then on 24 July he crossed the Kum-bel pass at 13,600 feet. Further on he crossed the Altai in the Shughnan valley at 14,000 feet to reach Por-dobe. There he met the Russian Customs Officer, M. Zampoin, and the Russian governor of the Pamir district, Colonel Jagello. Stein gives great credit to Colonel Jagello for extraordinary attention in providing transport and official introductions to the local head-men in the districts to be passed.
Stein's purpose in choosing this round about journey was to study the terrain and ethnology of this section of the ancient 'Silk Road' as it passed from Sogdiana - and the Oxus valley across the mountains into Chinese Turkestan.
On 28 July he departed Por-dobe to travel down the Altai valley, the ancient trade route used by Chinese silk merchants to reach Baktra and Samarkand. He halted at Yaman-karchin from which he could views the Trans-Altai range. Stein discusses the climate and resulting agriculture of the Altai valley in detail to support his appraisal of its role in history as recorded by Ptolemy and Marinus. This also is the route recorded by the Chinese pilgrim, Hsuan-tsang.

Section III - Along the Western Rim of the Pamirs
On 2 August, having established in his mind that the Altai valley was the principal trade route between Baktra and China, Stein turned south to spend time and great effort traveling through the mountainous Pamir regions of Roshan and Shughnan to the upper Oxus River and the Wakhan valley.
Lets face it, Stein was simply enamored of his views of massive mountain ranges over 20,000 feet high. His adventures during the second expedition in the Kun-lun during which he was dangerously frost bitten had not for a moment cured him of this avocation.
That first day he traveled up the Tars-agar saddle at 11,500 feet elevation he halted at a small Kirghiz camp. shows the Sel-tagh or Muz-tagh. And shows a part of the glacier wall in front. From there he descended into the Muk-su valley. He had to avoid the old route because of spring floods from the Sel-dara glacier. Therefore he had to use winding trails to cross various ridges and camp at 14,000 feet before crossing the Kayindi pass at 16,200 feet on 5 August. From his next camp he visited the Chakur-jilga glacier at 14,600 feet.
On 7 August Stein turned south into Takha-korum-jilga and then crossed the Takhta-korum pass at 15,100 feet to camp at 13,000 feet elevation. At this point Stein again had to secure transportation and guides from the next Kirghiz group. This was provided by Kokan Beg at his summer camp - Kara-chim - elevation 13,700 feet. Among other things Kokan Beg informed Stein for the first time of a large lake in the Murghab valley created by an earthquake 4 years previously, and which blocked the usual route to the Alichur Pamir. This forced Stein to reroute via Saunab and the Roshan valley. Stein again did a quick anthropological measurement and .
On 10 August Stein back tracked across the Kizil-dawan and Kok-yar valley to the Tanimaz river to enter another area, cultivated by Tajiks. Further on he found some results of the giant earthquake and a huge barrage .
On 12 August Stein moved on to the Bartang valley.At Darband there was a large rock with a guard tower built to defend Roshan from the Kirghiz raiders. . From there he climbed over steep rocky ledges to reach Saunab and at 9000 feet. There he changed from Kirghiz to Tajik labor and transport. The baggage would be carried by porters rather than animals over the extremely difficult terrain.He noted that these were representatives of the 'pure' Homo Alpinus, the ancient population of the entire region. Before leaving Saunab he visited a small fort. Prior to the Russian control each tribe was in danger of attack by others, here from both Kirghiz and Afghans.
On 14 August Stein departed Saunab down to a village at Nusur He crossed the Bartang river on goatskin rafts guided by men swimming along. From there he reached the last hamlet at Barchidi and then followed a path above the Bartang river which was now hardly flowing due to being blocked further north.
On 16 August Stein ascended another ridge 1000 feet The entire region had been changed by the earthquake. He reached Shedau lake, another one formed this way. and The former Sarez Pamir was snow a lake. and . From there, despite the damage caused by the earthquake, the Roshani mountaineers carried Stein's baggage over a 13,200 foot high spur.
On 17 August they descended 2000 feet to the Yerkh fiord The further passage was even more dangerous, over shifting debris. The expert Roshani mountaineers built rafiks of brushwood and stones to create temporary ledges against sheer rock walls. At one place this process required 5 hours of labor to move one mile. They stopped at a homestead of 6 families 500 feet above the lake. It required a day's halt for the Roshani men to open enough of a path for ponies above Yerkh lake. This extemporaneous route was necessary due to the block of the old route by the earthquake. At Ushinch Stein was met by fresh Kirghiz transport sent by the Russian commandant. Moving on he camped next at 14,400 feet. From there he crossed the Langar pass above 15,400 feet to Langar village where camped at 12,300 feet elevation.

Section IV - By the Alichur and Great Pamir
Stein's detour had taken him further west than he planned. So from Langar he went back eastward to Yeshil-kol (about 12,700 feet) and the Alichur Pamir, then south to the Great Pamir. He notes that there had seen frequent visitors whose reports well known. He mentions that from the head of the Ghund valley he saw an excellent view of Shughnan. He concludes that this route across the Pamirs to Shughnan was that followed by medieval Chinese troops. (See his discussion of the Chinese campaign to Hunza described in his accounts of Hunza and Gilgit). Here he briefly mentions again the campaign of 747AD by Chinese general Kao Hsien-chin and several other reports by travelers. He passed the Buruman ridge that formed the new Sarez lake when a landslide blocked the former narrow river gorge Fig provides another view. The Buruman ridge is seen again from the Little Marjanai valley. A further mile on he passed Kamparchuk where the loads on the ponies had to be lightened. At the Great Marjanai valley he found several walled enclosures that he ascribed to Chinese military use. He passed a cliff, Sume-tash , over looking a stream in the Alichur Pamir at which he found a small shrine. This had contained a marker (since removed) to the Chinese victory over the Khojas in 1759.
On 22 August he left Sume-tash. Two days later across the Alichur Pamir he came to a Kirghiz summer camp at Bash-gumbaz-aghzi. He spent a day doing anthropological measurements again. From there he turned again south through the Bashgumbaz pass, elevation 16,300 feet, between the Alichur and Great Pamirs on 26 August.
On 27 August he halted at next valley over looked the western end of Lake Victoria and its outlet to the Oxus River. He made a panoramic view. Here he was nearing the borders of Russia and Afghanistan by Lake Victoria. He finally was seeing a life long view of the Great Pamir, so much tied to the travels of both Hsuan-tsang and Marco Polo. (see his account in Ancient Khotan.) Naturally, he proceeds to quote from both travelers. He provides another view of the lake. and Here Afraz-gul bagged an Ovis Poli - the great mountain sheep named after Marco Polo. Stein was anxious to find the pass by which Kao Hsien-chin crossed from the Pamirs to Wakhan en route to Gilgit. He located it on the south-western side of Chor-jilga Once across the pass there were two routes, to Langar and to Sarhad.
On 28 August Stein departed from Lake Victoria down the Pamir river to Kangar-kisht where it met the Ab-i-Panja. There he found a bridle path . connecting Langar-kisht with a road along the Alichur Pamir. Three days later he passed the Mats valley . To the south he could see the tops of the Hindukush (in Afghanistan). .
On 30 August Stein reached Langar-kisht 3 miles from the confluence of the Pamir and Ab-i-Panja where he was welcomed by a Cossack garrison. And here he records an interesting personal incident. He there met one Sarbuland Khan the Ming-bashi of the Russian Wakhan, who was the younger brother of Ali Mardan Shah, the old ruler of Wakhan and was acquainted with Raja Pakhtun Wali of Darel and Tangir (see Serindia). And it was one of his sons from Ashkuman who, with a Wakhis team, had 2 years previously helped Stein across the Chillinji pass.

Chapter XXVI - In the Region of the Upper Oxus

Section I - Old Remains in Wakhan
Stein expesses his delight to again be on the Ab-i-Panja, the main branch of the Oxus River. He notes that in 1906 he had only visited the river further east toward its headwaters above Sarhad and the Wakhjir pass. (see Serindia) He remained at Langar-kisht to accomplish more anthropological measurements.
On 1 September he reached the confluence of the two branches of the Oxus flowing from the Great Pamir and Sarhad. Nearby to the east on a rocky ridge he found the massive walls of a fort above the hamlet, Hissar.His sketch plan shows that the approach to the fort is on the south-west side, the others being too precipitous to climb. It was about 140 yards long and 75 yards wide at the widest place. The well built walls were 6 feet thick at the top. There were oblong bastions and small rooms within. The locals claimed the fort was built by 'Kafirs', that is long prior to the arrival of Islam. Stein estimated that the method of construction could well support that idea. A mile west of Hissar there was another hamlet, Zang, above which on a steep spur 1000 feet above the village was another ruined fort, called Zangibar. This one was oblong about 60 by 25 yards built of stone slabs . The lower 6-7 feet were courses set in hard plaster, above these the stonework was much rougher. There was a small square bastion on the northern face.
Stein commented that most of the Wakhis were of the Ismailia sect of Islam, whose leader was (and is) High Highness the Aga Khan. (These were the Assassins of the middle ages). The close connections maintained with India, despite the Afghan territory in between, enabled Stein to compose mail for transmission to India via Chitral.
On 2 September Stein continued down the valley closely matching the topography to Hsiung-tsang's description. He passed the Kala-i-Panja , the location of the main Afghan post. Next he passed Ishmarg, and from which he viewed the Hindukush far to the south. He stopped to camp at Warang - elevation 9,700 feet. Near there Stein visited another hill-top fort on a spur north-west. This one consisted of a stone wall across the narrow end of a spur defended on both sides by sheer cliffs. The enclosure was 108 by 20 yards with a tower.
On 3 September Stein visited several small cave dwellings. He could see that they had been occupied recently. The local head man confirmed that they had been used to protect the families from Afghan or Kirgiz raiders. Stein then moved, passing the view of Khandut (on the opposite Afghan side of the river), on to Wenukut where he visited Ihsan Karim Ali Shah, the chief of the Ismailias in Wakhan. Stein immediately connected Khandut with the Hun-t'o-to of Hsuan-tsang who wrote that it was an important Buddhist shrine. During his day at Yamchin Stein surveyed another hill fortress at Zamr-i-atish-parast. and It also was 1000 feet above the valley floor on a steep spur. The first line of defense is about 400 feet up the slope in which there is a gate flanked by round towers. There is a wall 4 feet thick and in places 11 feet high across the spur to the cliff. The wall ajoins the inner wall that is 6 feet thick in places. Round towers, one of which is 13 feet in diameter, guard this wall as well. All the walls are built of unhewn stone set in mud plaster. But the towers are of sun-dried bricks 12x9x4 inches with walls 6 feet thick containing loop holes 12 inches wide on the inside and 8 inches wide on the outside. The second wall line is from the edge of the Yamchin ravine. On the other side is another fort, Zulkhomar . From the cliff edge the main wall goes 450 yards across the width of the spur. This wall is 4.5 to 5 feet thick with loop holes and is 1-15 feet high in places. It has 17 towers many of which are round with 13 foot diameters. Fig shows a section of double wall and several towers. Interior walls and a quadrangular bastion that flanks the re-entering angle of the wall suggest provision for separate defense. The line of the wall curves to a massive tower at the top of the knoll and then turns NNW across a small dip in the western flank of the spur. The edge of the gorge there is protected by another tower 15 feet in diameter of bricks 16x11x5 inches . Below is a massive oblong outwork This wall has decorative bricks set on edge. Further north the wall curves round and is defended by three round towers. One of these still has sockets for the beams of a second floor. And it has a double row of loop holes. The wall continues up for almost 400 more feet to the southern corner of the triangular citadel . Plan 48 is a detailed sketch of this fort, extends a further 130 yards north. At the point where its two longer sides meet there is a kind of ravelin with a massive square tower that guards the approach from the plateau. That ground is separated from the citadel by a 120 feet deep ravine. The citadel walls are built of stone slabs set in plaster. Its 3.5 feet thick and 13 feet high outer walls are loopholed and have a 1.2 feet wide parapet. The walls also have circular towers and there are rooms inside. Opposie the eastern end of the main walls is a small rock island and plan 47 on which is another small fort with massive walls called Zulkhomar. Stein surmises that these fortresses date from the Zoroastrian era. Whatever their date, Stein notes that the extensive and excellent construction points to the ability of a much larger local population at that time then is now present.
On 5 September Stein continued down the Wakhan valley to Shitkhar where he met Qazi Qadam Shah, who helped him obtain samples at Ishkashm of the Galach language. Further on he came to a steep cliff At this dangerous point a demon was wont to kill passerby until driven away by a saint. From there they crossed a narrow canon by a bridge to reach Darshai. North of the bridge on another rocky ridge Stein found more ruins. This also was locally known as a "kafir" fort.

Section II - Through Ishkashm and Gharan
On 6 September Stein continued down the Ab-i-Panja valley past Ramanit and Udit and at Sang he passed the boundary between Wakhan and Ishkashm, both on the right bank. They reached Namadgut, inhabited by Wakhis but part of Ishkashm. Much of historic Ishkashm lies south on the left bank of the Oxus, in Afghanistan (Badakhshan) now that an international border separated the territories on the two sides of the river. (Then Russian) now Tajik Ishkashm is at the great bend of the Oxus where it turns 90 degrees to the north. Stein was required to remain, much to his disappointment, on the north side.
On 7 September Stein remained at Namadgut and on the 8th he visited another fort, called Qala-i-Qa'qa. Stein describes this fortress in great detail and provides real, professional military engineering analysis of its purposes of which I only include a partial summary. This was also on a pair of east-west rocky ridges above the river and separated from mountains further north by a plateau. The larger (northern) ridge reaches 400 feet above the river and 225 feet above the plateau. The cliffs are extremely steep. The southern ridge shown here is shorter and separated from the parallel northern one by a depression Beginning at the eastern end, the outer wall is of sun-dried 14-15 x10 - 11x3.5-4 inches brick close up to the cliff but only 3-3.5 feet thick. The wall has both round and square towers of bricks on a stone foundation. The line of loop hole was low. They were 3 feet 3 inches up the wall on the inside and 2 feet 3 inches up on the outside, indicating they were designed for shooting down. They were 7-8 inches wide. Along the eastern and northern sides there was a parallel second wall 6 feet away. Toward the west the ridge was not so high and there the wall was much thicker (16 to 33 feet), a solid rampart of sun-dried bricks. In the corner was a massive circular tower 25 feet high. At the north west corner the rampart turned south south west to cross a gap between the two ridges. Two square bastions appear on the sketch plan. These, built on stone foundations extend outside the wall 20 feet. The wall here and the bastions were faced with solid bricks 16x9x4 inches with interiors of layers of stamped clay and thin layers of brushwood. The wall that closed this gap then continued to the south to the foot of the precipice at the western end of the smaller ridge. There, was a ravelin of massive brick work. From the eastern end of this out work the wall descends toward the river strengthened by two more towers.Further down the wall has three more towers that may have guarded a gate. There is an outlying tower 25 feet high. Much of the rest of the former wall is gone. Stein believes it originally extended along the river up to a traverse wall and up from a tower shown in fig. 414. At the tower on the left in 414 the outer wall leaves the river to ascend to a terrace on the main ridge and then to the east to a huge tower. The main wall then turns to the north with a gap for a gate. The tower has loop holes seen in fig 407. Fig 412 shows the southern ridge. Plan 49 shows this part of the walls also. From corner xii the secondary wall turns at right angles to the northwest and goes up the narrow crest to the citadel. This one is build of bricks 18x14x3 inches and is 8 to 10 feet high. There are three more round towers guarding the connecting wall. The walls of the citadel conform to the terrain. They enclose an area about 150 by 40 yards. On the highest point, 350 feet above the river there is another structure with two rooms, one 28x19 feet and the other 19x11 feet. Its well-built walls are 3 feet thick.
On 8 September Stein finished his survey of the fort to visit the Ziarat of Hazrat of Shah-i-mardan, nearby. Then he headed further down the river to Nut, the Russian outpost across the river from Ishkashm. There he was hosted by the Russian commander, Captain Tumanovich. Stein again collected anthropological measurements. Nut is located at elevation of 8,400 feet above sea level and 400 feet above the river giving it a fine view of the other side .
On 10 September Stein left Nut to follow the Oxus around its sharp bend to Shughnan. He came upon another walled enclosure on a crest 500 feet above the river about 7 miles from Nut. The walls were of stone slabs with a few loop holes At Malwach he entered defiles opened for passage by Russian engineers. Until then the hamlets both north and south had more communication across the river with Badakhshan than with each other. He continued for 3 days across Gharan to Shughnan.
On 11 September he camped at Barshor. Near Andarab he passed the famous mines that produced rubies or spinels mentioned by Marco Polo.

Section III - In the Valleys of Shughnan
Shughnan is on the right bank of the Oxus part way north along its channel and north of Sharan and south of Roshan. Below Andarab he encountered the worst of the defiles. He camped at Kharuk at 6,650 feet elevation where he remained on 13 and 14 September where he again met Colonel Jagello. Stein takes this opportunity to return to the T'ang Annals for references to Shughnan (then called Shih-ch'ih-ni, or Shih-ni, and Se-ni). There was an embassy from this area at the Chinese Imperial Court in 646AD and a grant to its king who accompanied General Kao Hsein-chih in 747 AD during the campaign into Yasin. The region was also mentioned by various pilgrims who crossed it between India and Central Asia. Living in such a restricted set of valleys, the Shughni are noted and feared by their neighbors as raiders.
On 15 September Stein departed from Kharuk, moving up the Shakh-dara valley to its head. Again, from Kharuk he could have continued along the Oxus, but he always wanted to go the hard way, over high mountain passes. He stopped at the village fort of Rachkala, 8,400 feet and the headquarters of the Mirs of Shakh-dara. His route led through more difficult defiles past ruined villages . Further on he passed more "kafir' forts and outposts. He once again was forced to unload the ponies and resort to rafiks. He camped at 10,100 feet.
On 18 September Stein continued through mountain valleys and past tiny hamlets of Roshani cultivators and also Kirghiz nomads. Here he found a Russian cart track that led back east to the Alichur Pamir. There was another ruined fort above Jaushangaz . There he turned north to reach the Dozakh-dara pass into the Ghund valley. On 19 September Stein crossed the pass at 14,000 feet to find several glacial lakes and terminal moraines. He camped in the Ghund valley.
On 20 September Stein continued down the Ghund valley passing forts of the Shughnan mirs at Sardim, Wang and Charsim. In Charsim he found an interesting house of the local Ak-sakal just like those in Mastuj. He camped at Shitam to prepare for crossing the next pass. Interrogation of the 'greybeards' revealed memories of past Chinese control into Shughnan.

Section IV - From Roshan to Darwaz
On 21 September Stein left Shitam to cross the Shitam into Roshan where he camped at 12,600 feet. Beyond that the ponies had to be unladen.
On 22 September Stein passed a series of glaciers where the party had to cut steps and cross dangerous ice for 3 miles. They reached the head of the glacier at 16,100 feet. As always he loved the spectacular view from the pass. and . The descent was along neve beds after which they reached vegetation at 13,900 feet. At Raumedh Stein met a new team of men from Roshan to relieve the hard working porters from Shughnan. Stein noted the two groups spoke different languages.
On 23 September Stein continued down the valley past another glacier. At Khaizhez elevation 6,800 feet they reached the beginning of the Bartang valley and camped. It took two more days to reach Kala-i-Wamar on the Oxus. and Which Stein could have reached quickly simply by continuing down the river from Shughnan. ) Views at and and show the difficulties of passing through the Bartang valley, including use of another raft. The Roshani men carried the baggage across rafiks against the sheer canyon walls. At Paghu Stein found an interesting village. His comment, "Alpine seclusion seemed to have kept this small corner of the world almost untouched by the change of ages." The same impression was created by the local men. and at Kala-i-Wamar. Stein's interest in ethnic differences, a very topical 'scientific' subject at the time, again shows in his detailed comments on the physical features of Roshani men and women, whom he categorizes as original and pure examples of the 'Homo Alpinus' European type. he devotes a lengthy footnote to description of the women he happened to meet and, naturally, ties this back to Arrian's commentary in the Anabasis. .
On 25 September Stein reached Kala-i-Wamar. During his travel down the Bartang he had used goat skin rafts on several occasions. and He remained there for only one day due to concern about timing his future travel over high passes before they were closed by snow. He was given interesting examples of carved wood and photographed others in place. He camped in a orchard next to the Shughnan Mir's castle. (From when deputy Mirs from Shughnan were appointed to govern Roshan.) And examined the home of the Ming-bashi, Mir Shikran. He described this as an example of late Hellenistic and Saracenic design. He was invited to inspect the inside which he described in a lengthy footnote, as being similar to those in Yasin and Chitral. The neighboring castle had thick outer walls of stone reinforced by heavy timbers.
On 27 September Stein left to reach Kara-tegin. (Again choosing to take trails across high passes rather than the road along the river and on to Bokhara.) He route led over the Adude pass across the Roshan range.
On 28 September he continued on from Shahji-shau-jai (11,500 feet) over moraines and a glacier at 14,500 feet. Continuing down he passed more moraines at 13,300 feet to find the first cultivation at 8,700 feet where nightfall forced him to camp.
On 29 September he reached Matraun (5,500 feet), a Yazgulam hamlet where officials from Bokhara were waiting. (Another of the many examples of how his prior planning always resulted in his being met with assistance.) He comments on the very sharp ethnic differences between these flat-land officials and the mountaineers. He immediately continued on toward Wanj. Yazgulam valley was so isolated from easy approaches that it was for long a kind of 'no man's' area between Darwaz and Shughnan-Roshan. But the Yazgulamis were Sunnis rather than Ismailis, reflecting their dependence on the population to the north. Stein, after taking the round-about mountain route arrived back at the Oxus and continued through the gorges of the Oxus below Yazgulam. On 30 September Stein reached Rokhar, the main village of Wanj at 5,600 feet where he found another ruined castle. On 1 October Stein again returned to the mountains, moving 30 miles up the Wanji valley to Sitargh at 6,900 feet. In this valley he encountered another different ethnic type, Persian speaking Tajiks.
On 2 October Stein was forced to remain at Sitargh due to heavy rain and snow.
On 3 October Stein was able to start for the Sitargh pass in which, at 12,400 feet, they again found moraines and crossed the pass at 14,600 feet after a 7 -hour climb. and . Further on he made . A further 13 miles from the pass at 10,500 feet he camped on Ziginzau plateau.
On 4 October Stein continued down to reach Pashmghar (8,500 feet) in the Khingab valley. Two more days of hiking past interesting villages brought Stein to Lajirfkh. He camped at Sangwar (7,400 feet).

Section V - From Kara-tegin to Bokhara
On 6 October Stein was again forced to halt, at Lajirkh (6,800 feet), due to heavy rain.
On 7 October Stein was again on the move, across the Gardan-i-kaftar pass at 12,200 feet to reach the Karashura river. and
On 8 October Stein moved on from his camp at Kulike to the Khush-kulak saddle. This shows part of the view. . The view toward the east included mountains he had seen over a month earlier from the other side. He was now back north about where the direct east-west route crossed the entire mountainous region. From the northern edge of the Khush-kulak plateau he descended over broad spurs down to the Surkh-ab valley. He passed fields at 8000 feet being harvested. The first village was the Turki speaking Kirghiz Oital at 6,100 feet. He camped at Kanish-beg,
On 8 October he continued for 2 days down to Gharm. Through the valley Stein noted that the semi-nomadic Kirghiz were being bought out by Tajik farmers. He remarked that this was an example of the Iranian people pushing back against the Turkic invaders.
On 10 October Stein passed Langar-i-shah to reach Gharm where he again received a warm welcome.
On 12-13 October Stein moved on along the Surkh-ab river to Ab-i-garm, the western edge of the Kara-tegin region. He takes this occasion for another discussion of historical references in Chinese sources, especially the memoir of Hsuan-tsang.
On 14 October Stein left the mountains at Ab-i-garm to travel 270 miles in 9 days to Samarkand. His path first led through the gorges below Sangardak Then he crossed the watershed between Ab-i-garm and Faizabad. He remained in Samarkand to refit and repair equipment. But meanwhile he visited the ancient ruins at Afrisiab, the ancient capital of Sogdiana and Alexander's Maracanda.
On 25 October Stein departed Samarkand by rail to Bokhara where he was for the first time allowed to visit the Ark and other medieval sections.
On 28 October Stein departed, again by rail, for Askhabad.

Chapter XXVII - By the Eastern Marches of Khorasan

Section I - From Askhabad to Mershed
Since this part of Stein's remarkable expedition was outside Chinese Turkestan, I reduce the level of detail in the summary provided here. But he did find fortifications worth noting. And he made photographs worth viewing.
On 29 October Stein reached Askhabad where he had to comply with frontier regulations.
On 31 October he crossed the frontier to Bajgiran in Persia where he received a mounted escort by Kurds from Kuchan. On 3 November Stein reached Meshed where he was met by the British Consul-General. Remember that this was during World War I and the British and Germans were in conflict to control Persia. Stein's mention of the impact of the Germans on his travel is a good reminder to us of the mostly-forgotten activity in Persia. The entire Perso-Afghan frontier was in chaos with bandits roaming at will. Stein spent a week there catching up on paperwork and preparing for the dangerous trek. He was provided with a Hazara militia escort for safety. And the Russians had deployed troops along the northern part while the British Indian army had the same along the southern part of the Afghan border.
On 11 November he set out on the 500 mile journey to Sistan.

Section II - Past the Perso-Afghan Border
At Fariman he met Mir Muhammad, a Tekke Turkoman who told Stein about the methods of the Turkoman raiders before the Russians put a stop to slave running. Along the way he passed ancient forts and shrines such as this madrasah built by Shah Rukh in 1444. and .
On 20 November Stein halted in Bamrud, a village said to be immune from Afghan raiders due to payment of protection money, since a current Afghan raid of the area was in progress. He passed a large fortified ruin at Tabbas-i-Mazena.

Section III - Into the Helmand Basin
Readers now who have served in Afghanistan may be interested to learn that the Helmand River, the scene of much fighting, flows south-west into Iran where it disappears in marshes.
On 27 November Stein passed another ruined fort at Duruh village and a larger one at Ghala-koh. He provides the usual detailed analysis including size of bricks and all. A few miles further he came to a 6,200 foot peak on which was Ghalakoh fortress.
On 28 November Stein continued toward Sistan. He covered 65 miles in two days to reach Bandan.
On 30 November he crossed a detrius fan and gravel to the shore of the Hamun, the terminal basin of the Helmand and then Nasratabad, where he met the British Consul.

Chapter XXVIII - The Sacred Hill of Sistan

Section I - The Historical Interest of Sistan
This is another interesting excursion into history.

Section II - The Remains of Koh-i-Khwaja
This section may prove of considerable interest. However I have omitted Stein's very detailed descriptions of each section and building in this large complex.
On 6 December Stein departed from the British Consulate at Nasatabad to visit Koh-i-Khwaja an isolated hill containing much-frequented Muhammadan shrines on its top. The ruin at its base was called Ghagha-shahr. The plateau is a mile long and nearly as wide. The cliffs are steep. and and . There is a narrow ridge to the south east. The main fortress wall is built of sun-dried bricks enclosing an area 170 by 130 yards. The gate was on the south eastern corner and protected by one octagonal and one round tower. Inside was a roadway and high wall. Further on was an arched gateway. Beyond the wall was an entrance hall. The hall has an apsidal end.Two more buildings are at and and Buttress walls are divided by narrow vaulted recesses of two stories. and . On another wall were stucco figures in flat relievos of horsemen and a lion. A small room is in the western corner of the terrace . Stein found a Doric column on the second buttress. and .

Section III - Remains of Mural Paintings
In this section Stein focuses on two paintings he found in a narrow passageway. and After removing some bricks Stein discovered four figures and part of a fifth much damaged by weather and insects. He assesses these as the first pre-Mohammedan paintings found in Persia. With great care he managed to remove 12 panels for safe-keeping. Fearing subsequent damage and loss, Stein provides a very lengthy and detailed description of these unique paintings. One he identifies as Rustam, the legendary hero of Sistan. This, he notes, is very similar to a painting from Dandan-oilik and discussed in Ancient Khotan. Shows an artistic rendition of a fragile wall painting.

Section IV - Remains on the Hill-top
This section is a detailed description of the extensive fortifications, cemetery and other buildings remaining on the top of Koh-i-Khwaja. Stein provides a few photos as well. The northern fortress wall The plateau edge is 200 feet above the top of Ghagha-shahr ruins. There is a road up.It leads to a small walled ruin Kok-i-Zal. An area of square, vaulted rooms totaling 50 by 30 yards is enclosed by massive brick walls of large sun-dried bricks designed to protect the area below from attack from above. Ascending further from the dip along the wall for 50 yards one one finds a small mound Stein concluded this was the remains of a defensive tower designed to guard the road. Another 160 yards is another tower just above the end of the Ghagha-shahr wall, below. Moving west along the plateau edge 1/3 mile one finds a ruined fort called Chihil-dukhtaran (40 maidens) on the southern end of a tongue precipice and valley with access to the plateau which the fort was to guard. This fortress wall is of solid bricks in an oblong about 40 by 30 yards. The wall is loop holed. There is a gate in the east face flanked by two small round towers of which one still has its vaulting below the second story. And there are round bastions on the corners. There is a long hall, formerly vaulted, along the inside of the western wall. About 80 yards to the north are the remains of another, square enclosure of rough stones and there is a series of similar rooms 20 more yards north. Elsewhere on the plateau are remains of Buddhist and Muhammadan worship - the cause of the place being deemed sacred. shows graves and shrines. The plateau is visited by thousands of pilgrims each year.
Stein connects this sacred mount with places in the Yasht, the earliest religious Iranian text, the Avesta.
He then devotes space to discuss a local tribe, the Sayed fisher-folk.and .

Chapter XXIX - Ruined Sites within the Oasis of Persian Sistan

Section I - Remains at and near Shahristan
Shahristan is among the oldest sites with remains near the Helmand delta. It is on a north-south detached ridge about a mile long above the plain and The fortified area on the south side is 800 by 250 yards. The massive walls are of sun-dried bricks with towers and bastions. At the northern end where the slope is easier there are two ruined walls. Toward the south there was also a citadel 140 by 80 yards. All the walls were greatly decayed indicating their ancient age. The bits of pottery found there also indicated a very ancient age, predating Muhammadan occupation and even Sasanian times.
Stein next visited ruins at Atish-kadak (the fire temple). These are on the northern end of another ridge some 6 miles from the former site. On top of this ridge there was an enclosure 72 feet square with walls 4 feet thick. Next to it was another enclosure 32 feet square. Inside was a circular tower At the northern end was a more impressive structure. A hall 35 by 27 with walls 5 feet thick rising in places to 20 feet. The name indicates that Zoroastrians practiced their religion here until late middle ages. Stein found more towers in the vicinity. The section ends with the usual 1 pages of descriptions of items found here.

Section II - The Band-i-Sistan and the Ancient Name of the Helmand
This is a large barrage 8 miles from Shahristan at a split in the Helmand delta. Such constructions were necessary to enable irrigation canals from the Helmand. That these barrages are ancient Stein indicates by connecting the modern name, Helmand, with ancient names found in the Avesta.

Section III - The Site of Zahidan and Later Ruins to the North-west
Zahidan was an important ancient and medieval city, the capital of Sistan that was captured by Timur in 1383 AD. It was about 6 miles north west of Shahristan on another ridge. It was abandoned not long after Timur's attack. Stein provides a plan. The best preserved ruin there is the citadel with its massive towers and bastions. I skip Stein's detailed description. He also describes other ruins in the general area. and and and

Section IV - List of Pottery Specimens and other Small Objects from Later Sites in Northern Sistan
This is the usual detailed list of objects found, 4 pages long.

Chapter XXX -In the Desert Delta of Sistan

Section I - Ruins Ancient and Modern
On 19 December Stein continued south. In this section Stein describes ruins south of the Helmand River delta. The area was some times part of Afghanistan. (Actually the 'border' between Afghanistan and Iran was very fluid over the centuries.) He describes many ruins, some of them forts, such as Kundar and another near Hauzdar . and . He found a "Chigini (windmill) and a nearby mansion and At Ramrud he found another fortified village. and . At Kalat-i-gird, 3 miles from Ramrud, he found an almost circular fort with diameter 160 yards and walls 8 feet thick. And near this location he found more structures. West of Kalat-i-gird he saw terraces that reminded him of the Lop Desert. This area was rich in ruins mostly from Muhammadan times.

Section II - Remains of Prehistoric Settlements
In addition to Muhammaden ruins Stein found copious amounts of pre-historic potsherds and other bits and pieces lying on the open ground. In many places there were so thickly distributed that they even protected the clay surface from wind-erosion. He termed this 'civilization' as 'chalcolithic' and considered that it has occupied these mesas for a very long time. He commented that the density of the deposits of human production on the ground was due to the powerful effect of the wind blowing away all the softer and lighter soil over centuries. Thus the pottery and other fragments would represent the remains of centuries of occupation. He remarked that a complete collecting effort would fill many carts ,even railway cars. But he did bring back a significant sample. He photographed another ruin called Burj-i-chakar and which he discovered was the first in a while line of watch towers. He found a section at . Two miles further he found another mesa on which was more pottery and then another outpost Further sout west there was a small fort and . All of these and others Stein describes in detail along with the items found at them, many of which are also shown in the plates in Vol III. In conclusion Stein notes that the geographic location of Sistan between the Indus civilizations to the east, Mesopotamia to the west and the Caucasus - Caspian area to the north makes it a likely location for the intermingling of these societies.

Section III - List of Objects found at Sites of the Southern Helmand Delta
Stein categorizes the artifacts in this 16-page long list both as to the specific site and also as to pre-historic or later and the pre-historic items are divided between grey, buff and red. The items from an historic era are either glazed or unglazed.

Section IV - Ruins of an Ancient Border Line
Stein notes that these ruins were in the area in which he found the prehistoric pottery and bones described in the previous section. So in this section he focuses back on the forts themselves. First noted is Burj-i-chakar and with walls to 25 feet high. It was a two-story square, 60 feet on each side with 10 foot wide towers in the corners. The interior was divided into three chambers. Three miles distant was another ruin and plan 59 again. This one was 48 feet square with round towers in the corners. He found several more numbered xxiii, xxiv and xxv. In total I count about 35 watch towers or forts in Sistan that Stein mentions or describes in detail

Section V - From Sistan to India and London
In early February Stein finished his exploration of Persian Sistan and headed back to India - Baluchistan. He reached Koh-i-Malik Siah and then followed the main caravan route. Along the route were government rest stops with wells. One of the small posts is shown here. Despite this route being relatively well provisioned, Stein passed many dead camels along the way to the rail head.
On 21 February he reached it at Nushki after 15 days ride. He stopped briefly at Quetta and Sibi, then spent a week at Delhi. He also went to Dehra Dun to check on the maps with the Trigonometrical Branch of the Surveyor General of India. These are found in Volume IV (a box) and I have photographed them for this summary. In mid-March Stein returned to Kashmir to find his 182 cases of artifacts save at Srinagar. From there Stein returned again to London in order to continue work on the material gathered during his Second Expedition.

Appendices A to R Together with the index these fill 176 pages.