INNERMOSTASIA - CHAPTERS
Detailed Report on explorations in Central Asia, Kan-su and Eastern
Sir Aurel Stein K. C. I. E.
Oxford, At the Clarendon Press,
Oxford, England 1928
Volume I Text
Table of Contents
Introduction - vii - xx
List of abbreviated titles - xxiii - xxvi
List of 288 Illustrations in Volume I - xxxiii - xxxix
Chapter I - Through Chilas, Darel and Tangir - 1 - 35
Section I - From Kashmir to Chilas - 1
Section II - Chilas and its Past - 7
Section III - On the way to Darel -13
Section IV - Darel Old and New - 20
Section V - Through Lower Darel and Tangir - 30
Chapter II - From Yasin to Kashgar - 36 - 65
Section I - Yasin in History and Geography - 36
Section II - Through Yasin to the Darkot Pass -41
Section III - From the Yarkhun Head-waters to the Tagh-dumbash Pamir - 47
Section IV - In the Valley of Tash-kurghan - 53
Section V - By the Kara-tash River to Kashgar - 58
Chapter III - From Kashgar to the Khotan River - 66 - 97
Section I - Along the Outermost T'ien-shan - 66
Section II - Old Remains and Routes beyond Marl-bashi - 74
Section III - A Hill Range in the Taklamakan - 81
Section IV - Past the Mazar-tagh of Khotan - 90
Chapter IV - From Khotan to Lop - 98 - 155
Section I - Antiques from Khotan Sites - 98
Section II - List of Antiques acquired at Khotan - 101
Section III - Finds at Sites near Domoko - 127
Section IV - The Niya Site revisited - 140
Section V - List of Antiques from Niya Site - 148
Chapter V - On the Way to Lop-nor - 156-179
Section I - Charchan and Yash-shahri - 156
Section II - The Sites of Koyumal and Bash-Koyumal - 163
Section III - Resumed Labors at Miran - 169
Chapter VI - Remains of an Ancient Delta - 180 - 213
Section I - The Ruined fort of L.K. - 180
Section II - List of objects found near or excavated at Fort L.K. - 189
Section III - The Sites of L.L and L.M. - 192
Section IV - List of Antiques excavated or found at the Sites L.L.., and L.R. -
Section V - Across the Ancient Delta of the Kuruk-darya - 204
Chapter VII - The Remains of Ancient Lou-lan - 214 - 280
Section I - Work resumed at and around the Lou-lan Site - 214
Section II - Miscellaneous Objects found at or near Lou-lan Site L.A. - 219
Section III - Relics of an Ancient Burial-ground - 225
Section IV - The Textile Relics of L.C. - 231
Section V - The Decorative Designs of the L. C. Fabrics - 235
Section VI - Miscellaneous Sepulchral Deposits and Descriptive List of Antiques
from L. C. - 245
Section VII - The Ancient Castrum L. E. and the Remains of Mesa L. F. - 259
Section VIII - From the Lou-lan Station to Altmish-bulak - 269
Chapter VIII - The Search for the Ancient Chinese High Road - 281 - 312
Section I - To the Easternmost Outpost of Lou-lan - 281
Section II - The location of the "Town of the Dragon" - 290
Section III - Across the Salt-encrusted Lop Sea-bed - 295
Section IV - The "White Dragon Mounds: - 304
Chapter IX - To the Su-lo Ho Delta- 313 - 342
Section I - By the Eastern Coast of the Cried-up Sea - 313
Section II - The Valley of Besh-toghrak - 321
Section III - An Ancient Terminal Basin - 327
Section IV - The Delta of the So-lo Ho - 333
Section V - Transport Problems of the Ancient Lou-lan Route - 337
Chapter X - To Tun-huang and An-hsi - 343 - 370
Section I - The Limes North-west of Tun-huang - 342
Section II - Tun-huang and the "caves of the Thousand Buddhas' revisited -
Section III - By the Han Limes to An-hsi - 362
Chapter XI - In Search of the Limes to Su-chou - 371 - 403
Section I - The Limes Line North of the Su-lo Ho - 371
Section II - From Ch'iao-wan-ch'eng to Shih-erh-tun - 378
Section III - Hua-hai-tzu and its Limes Remains - 389
Section IV - The Limes traced East of Hua-hai-tzu - 397
Chapter XII - From Su-chou to the Limes of Mao-mei - 404 - 428
Section I - The Limes along the Pei-ta-ho - 404
Section II - Past the Mao-mei Oasis and its Outposts - 409
Section III - List of Antiques from Ruins of Han Limes - 414
Chapter XIII - The Etsin-gol Delta and the Ruins of Khara-khoto - 429 - 506
Section I - The Lower Etsin-gol and its Terminal Basin - 429
Section II - Khara-khoto and its Remains - 435
Section III - Remains outside Khara-khoto - 445
Section IV - The Remains of a Rural Settlement and Marco Polo's "City of
Etzina' - 453
Section V - List of Antiques from Khara-khoto and Neighboring Sites - 462
Chapter XIV - To Kan-chou and the Central Nan-shan - 507 - 521
Section I - A Desert Route towards Kan-chou - 507
Section II - To Nan-kou-ch'eng and the Eastern Head-waters of the Kan-chou
River - 511
Section III - Return from the Nan-shan to Mao-mei - 518
Chapter XV - Across the Pei-shan to Barkul - 522 - 547
Section I - Through the Desert Ranges of the Pei-shan - 522
Section II - Across the Easternmost T'ien-shan - 529
Section III - Past the Karlik-tagh and Barkul - 535
Section IV - Historical Relations between Barkul and Hami - 539
During the third expedition Stein returned to several of the important sites he
had explored during his second expedition. Since he was still editing the
official report of that second expedition well after returning from the third
he was able to incorporate much information from the latter tour into Serindia.
In this report he mentions when that occurs and refers the reader back to the
previous report. From the point of view of the reader interested in
fortifications, I believe, the chapters on Stein's continued search for and
study of the Han wall and the chapter on the fortress at Khara-khoto may be the
most interesting. For the reader interested in exploration perhaps the chapter
on the return to Lou-lan and especially the description of Stein's remarkable
trek across the dry Lop Salt Sea may be more exciting. For those with
antiquarian and historical interests the chapters on the new lost mountain
valley societies that Stein described may be of interest, along with the
revealing photos he made of their inhabitants.
Aurel Stein here describes his plans and the considerations that determined the
dates for his expedition. He gives copious acknowledgment and thanks to the
many officials and others who sanctioned the expedition and provided official
funds. He thanks foe Surveyor General of India for assigning Rai Bahadur Lal
Singh, the eminent senior surveyor, again as his chief assistant. plus two
other experienced surveyors, Muhammed Yakib Khan and Mian Afraz-gul Khan, (of
the saintly Kaka-khel clan) from the Kyber Rifles. The Indian military provided
Naik (corporal) Shamsuddin of the First K. G. O. Sappers and Miners as
engineering assistant. In view of the official financial support it was agreed
that all archeological 'finds' would be provided to the museum in New Delhi.
The expedition traveled nearly 11,000 miles during a period of 2 years and 8
months. Stein was determined also to develop information on the past and
present conditions prevailing along the medieval caravan routes used by
specific travelers including Marco Polo as well as countless military and
commercial travelers. One important consideration for the route and departure
date was the temporary opening of secluded mountain areas in northwest India
normally closed by their local chieftains to European visitors.
Stein provides in this chapter a summary of the content of the expedition, the
routes and locations visited and the principal results obtained.
The results were so voluminous and consisted of so much very specialized
material from soil samples to documents in unknown languages to examples of
high Buddhist fine arts that the full study and description has involved a
large number of noted specialist scholars who have prepared appendices to this
report as well as separate publications. Stein gratefully describes the
contributions of these scholars. We have made photographs of the illustrations
in Innermostasia and
also have provided links to some of them in this text. Also we have links to
some of the maps. A listing of all the map sections is at
Chapter I - Through Chilas, Darel,
Section I - Kashmir to Chilas
For this, his third expedition, Stein, as always, planned on following a new
route through yet unexplored territory. This time he took advantage of a
temporary lull in the typical local fighting between mountain chieftains to
travel along a more westernly route than on the first and second expeditions,
visiting mountain valleys not previously opened to Europeans. For this he
obtained special permission through the Indian government negotiations with
chiefs in Chilas, Darel and Tangir to visit their domains. This also involved
crossing numerous high passes and moving through nearly impassible river
gorges. Again, the timing for the expedition departure was critical, too early
would mean high passes closed by deep snow, but too late would mean river
gorges were swollen with the raging torrents from melted snow.
Stein was especially keen to visit the Darel valley as it had been mentioned by
those medieval Buddhist pilgrims he followed as the route used between the Oxus
and Indus rivers. This and other valleys were occupied by fanatical Dard
tribesmen. Stein notes that more recently one Raja Pakhtun Wali,
descended of the Kushwaqt family and son of Mir Wali, former
ruler of Yasin, had gained control of Tangir and by 1909 had expanded his power
over Darel and adjacent valleys. This 'ruler' was bent on developing better
relations with the British Raj in order to gain support for his own precarious
rule. Stein was quick to take advantage of this unusual situation and propose
to route his expedition through this local chief's domain, that is through
Chilas and Yasin. Darel and Tangir are on the mountain route between the Indus
and Yasin. However, intense diplomatic negotiations and the extensive
preparations for this lengthy expedition prevented Stein from departing Kashmir
until 31 July.
The journey began by boat, down the Jhelam or Vyath - Hydaspes river from
Srinigar to Bandipur on Volur lake. From there the extensive and heavy baggage
was sent with Muhammed Yaqub Khan and Naik Shamsuddin via the Gilgit Transport
Road to Hunza. Stein liked to travel as light as possible and for this section
of his trip such was essential.
On 2 August with Lal Singh and Afraz-gul Stein departed directly toward Chilas
through the deep gorges of the Kishanganga and across the watershed towards the Indus by the Barai pass at
14,250 feet. At that point they had reached Chilas territory. Two more days and
across the Fasat Pass at 15,200 feet brought them to the fort at Chilas (The
Baral and Fasat Passes, Chilas, and Stein's route as far as the Shobat Pass as
described below are all shown on map aias
Stein digresses to discuss the historical events of the Chinese military
campaign in 747 AD in which General Kao hsien-chih brought his army through the
Pamirs and across the Darkot to drive Tibetans out of Yasin and Gilgit. It was
at that time that the Tibetans were forcing the Chinese out of Turkistan and
attempting to gain alliance with the Arabs to the west. The Chinese expedition
was designed to block this effort. Stein describes all this in more detail in
Ancient Khotan. He notes that then and still recently attacks on Chilas
and Gilgit came from the Chitral side (west) until in 1893 the Sikhs invaded
and finally a garrison was permanently established in Chilas Fort. The Chinese
reported the same danger in 749 AD.
Stein continues with detailed discussion of alternate routes over various
passes. each of which has its peculiar difficulties, making reference to
various map sheets in the Indian Survey series. He also discusses the rather
complex political situations among the various tribes who inhabit these
isolated valleys. (All of this area is now part of the problem facing the
Section II - Chilas and Its Past
While these sections on life and the peoples of northwest India (Pakistan) are
tangential to his study of Chinese Turkestan, they are a fascinating look into
society there only a hundred years ago. His simple matter-of-fact descriptions
of the difficulties he overcame in simply moving through narrow gorges and over
snow covered passes and glaciers reveal his personality. Unfortunately time and
space prevents more than a bare summary here. Readers should study his
multi-volume texts. His many illustrations, made with the primitive glass plate
camera of that time, are excellent in showing the people and places.
Stein continues with reference to other medieval notices of the Chilas area. In particular he quotes from Alberuni's
India and his own translation of the Rajat (Kalhana's
Rajatarangini, a Chronicle of the Kings of Kashmir in 2 volumes). Chilas
and the surrounding valleys are occupied by Dard peoples, an ancient race that
has held out for centuries against campaigns by Sikhs and others from the south
and south-east. Stein laments that his late start and lack of time for more
extensive research. The general opinion of the Dard peoples of Chilas is that
they are warlike and engage in constant raiding of their neighbors.
On 8 August he crossed the Baral pass in rain and snow.
On 9 August he crossed the Fasat pass at nearly 15,000 feet elevation on a very
steep track and snow. He passed through Niat village at 7,000 feet. On 10
August he continued down the Niat valley. Near Basha at 5,500 feet the scenery
changed from thick forest to barren rock. Soon he reached a spot about 1000
feet above the Indus. There he was welcomed by the local British agent, Captain
C. T. Daukes. With the captain as guide he visited the fort of Chilas about a
mile from the Indus and garrisoned by a double company of Imperial Service
troops. He notes that Chilas only became part of British Indian's protection in
1893. There is more interesting description of the area in this section.
Section III - On the Way to Darel
Stein was met at Chilas by four agents sent by Raja Pakhtun Wali from Darel to
secure his passage. Stein notes the double role of these worthies and the armed
troops - that is to protect him from the many bandits or would-be bandits and
also to ensure that he stayed on the correct route and didn't bring any British
mischief into Darel. In fact his entrance and passage though this
semi-independent area was a great concession by its war lord. The Raja had
insisted that Stein could not be accompanied by a single person from the Gilgit
On 11 August Stein started again. The first several day's progress included
devious routes through the Hodar valley to the north of the Indus. But first
the route continued through rocky defiles on the Indus itself below Chilas.
to travel down stream on a skin raft made of 6 bullock skins guided by 4
swimming experts. As always, he measured the river flow and found it was 14
miles an hour - as he noted "an exciting journey, but at the same time
refreshing and restful" The raging river was confined between sheer walls
and only 200 -300 yards wide. And by that time it was already 24 feet below
high flood stage. At the location where the river passed the Hodar valley
Stein's 'exciting journey' stopped and the baggage was ferried across. Above
the village there Stein spotted a ruin on a ridge some 300 feet above so of
course climbed to examine this 'fort' - 160 yards by 100 yards in size.
following day Stein began moving through even more narrow gorges in which
animal transport was impossible. So he had the first of the many teams of local porters assigned
to carry the baggage while he proceeded on foot. This situation continued until
he reached Yasin. Near Dar, he found another ruined fort on a spur beside the
track. That evening they camped above Pakora village at 7,600 feet. The next
day they crossed the Unutai-gali pass at 10,510 feet. This
was the entrance to Darel, the Khanbari valley. The mountain slopes were again
covered with thick forest. At the pass the previous porters were relieved by a
crew of Darelis. A half mile on Stein was met for official welcome by
Mehjtarjao Shah Alam, the Pakhtun Wali's nephew and escort of armed troops of
the Wali's personal guard.
Stein had received permission to conduct survey of this unmapped area. Lal
Singh had begun the plane table work at the entrance to the Hodar valley and
continued it through Darel. Stein describes his visit to Darel in detail making
specific mention of the assistance provided by Shah Alam and Pakhtun Wali. The
route led over more passes, including the Phuno-phuno at 13,650 feet and the
Chiyagal pass at 14,000. (All these places are on the same map linked above.)
Section IV - Darel Old and New
In this section Stein again digresses to provide as much historical information
and his current observations of the people as he can. He starts as usual with
the descriptions provided by those trustworthy Chinese pilgrims, Fa-hsien and
Hsuan-tsang. The former passed through in 403 AD. Stein provides and analyzes
his accounts and the T'ang Dynasty annals. He proceeds to narrate his trip
through Darel. The first main stop was at Mankial. As usual with foresight he
had sent the Wali a list of 'old places' he wanted to visit. So he was promptly
provided with 'greybeard' guides to these locations. Without losing time Stein
visited as many 'kot's as possible. He concluded that most pre-dated
Muhammadan times. He describes each in turn in the order visited.
first was Ramal-kot and oval enclosure on a rocky ridge with long axis 100
yards. The second was Zhomi-kot and the third Taronali-kot. He then visited
Bojo-kot, actually a series of fortified houses on a walled terrace.
visiting several walled villages Stein was taken to Raji-kot on a rocky hill
about 500 feet above the river valley, then past Bodo-kot and Gali-kot on spurs
to stop at Gali-kot. On a high ridge above Raji-kot Stein found a massive wall.
Throughout the valley, or rather on the ridges and spurs around it, Stein was
shown many more ruined walls and enclosures. One had walls 16 feet thick.
decided that the Raji-kot for had been the main residence of the ancient rulers
of Darel. Now Raja Pakhtun Wali had decided to build his castle (palace),
called Gumare-kot, in a plain just south of this hill. Stein was greeted there
by the Raja in full state. Stein provides a summary biography of this worthy
Khushwaqt chief who had conquered the region for himself. Unfortunately, as
Stein note, Tangir tribesmen murdered the Raja in 1917 and burned down his new
building forcing his wife and children to flee to Kandia..
Section V - Through Lower Darel and
On 18 August Stein moved from Gumare-kot down the main Darel valley southward
back toward the Indus, passing walled terraces, old irrigation canals, villages
and more forts. (Darel
and Tangir are shown on mapa2as) Near Shaha-khel he
found a ruined fort on a high peak called Lokilo-kot (Red Fort) built of clay -
a rectangle 174 by 115 feet with corner bastions 12 feet square. It commanded
the entrance to the valley. To avoid return to the Indus Stein crossed the
Shardai pass toward Tangir over a very dangerous ascent by a zigzag trail. High
in this pass Stein found another fort 183 by 262 feet with 12 foot square
bastions, designed to guard Darel. After 5 hours climbing they crossed the pass
at 10,050 feet. The descent to the Tangir river was equally difficult and
dangerous. The Raja had another fortress in Tangir at Jaglot.
valleys are at similar elevations between 6,000 and 7,500 feet. Stein describes
his passage of two days up the Tangir valley. This was dangerous territory due
to the fanaticism of the inhabitants plus their unruly and warlike nature. He
observed that the Raja's escort spread out and established flank and lead
guards as a distance. Along the way he continually compared the racial
characteristics of the local population in each valley and noted that those in
Tangir differed considerably from those in Darel. When he had time he would
call together a representative group and perform anthropological measurements
as well as make photographs and write down as much of the local language as
possible. He found that in Tangir there was an active logging enterprise
operated by the Kaka-khel traders whose sancity enabled them to conduct
business in tracts too dangerous for others. The Tangiris lived in small
separated and isolated villages whereas the Darelis lived at least much of the
year in larger centralized communities. The upper part of the Tangir valley was
held by Gabar-khel tribesmen who showed their animosity as Stein passed through
with armed escort. He camped with Mian Shah-zada, a Kaka-khel from Ziarat and
uncle of Afraz-gul, and whose recommendation had helped Stein gain passage. And
during the crossing it was Shah-zada whose influence kept the restless local
tribes at bay.
On 21 August Stein continued through forest upwards to over 10,000 feet
there the climb was above the tree line over steep rock slopes. They reached
the top of the Sheobat pass over 15,000 feet, which is the watershed between
the Indus and Gilgit rivers and the border of Raja Wali's domain. There he was
met by a large team of Gupis porters sent up to take over the baggage. (here
his route leaves the map.)
Chapter II From Yasin to Kashgar
Section I - Yasin in History and Geography
By crossing the Sheobat pass Stein entered the territories of Gupis and Yasin.
Gilgit Agency had been established only in 1877. Stein restricts himself to
general comments on the history of and society in Yasin. He was now in the main
historical route from the Darkot pass to India from the Pamirs. He mentions
some of the Chinese reports. As early as 722 AD. the Chinese sent an army into
Yasin in an effort to block the Tibetans from joining forces with the Arabs to
the west. They were active again in 737 and it was in 747 that the famous
general Kao Hsien-chih brought his army across the Pamirs and over the Darkot
pass. He discussed all this in Ancient Khotan. Further Chinese
expeditions to Yasin took place in 750 and 753. But the defeat of the Chinese
main army by the Arabs in 751 north of the Pamirs resulted in ultimate Chinese
withdrawal from Turkestan. After Chinese withdrawal there have been no reliable
records about Yasin for a 1000 years. Then the Khushwaqts and Katur families
gained control and maintained rulers over Chitral and Mastuj for generations.
describes the topography of Yasin and remarks on its fertility and openness to
sunshine. The region was only held back economically by the incessant warfare
of the ruling clans. Normally Yasin's position and ease of defense provides
security. But when powerful outside forces such as Chinese or Tibetans need to
use the direct route south from the Darkot and Baroghil passes does it get
invaded and overrun.
Stein notes that most of the inhabitants speak Burushaski, a language distinct
from the other Dardic languages of the Hindukush. The inhabitants call
themselves Wurish. Two main sections of Yasin are Ghizar and Kho.
Section II - Through Yasin to
the Darkot Pass
Stein reports that Ghizar was too far from his route of march but that he did
manage to visit Kho.
On 23 August he descended from the Sheobat pass over the remains of a glacier.
Stein took a short cut to a pass giving direct access to Gupis.
the unexplored pass almost defeated the porters. They spent 8 hours climbing
over huge masses of rocks and boulders. before reaching a narrow gully that was
the pass at 16,000 feet. They were still at 15,000 feet by nightfall when
forced to make camp. The next day they moved down to Gupis through more steep
and narrow passages.
On 24 August they stayed in Gupis. There they met the garrison of Gupis fort
which guards the entrance to the Yasin valley and route leading to Mastuj and
Chitral. It was at this point also that the Chinese army destroyed the bridge
across the Gilgit river to block Tibetan advance.
On 25 August Stein marched the 13 miles from Gupis to Yasin. All this is
discussed in Serindia and Ancient Khotan. By spending a day in
Yasin Stein had time to visit the Governor, Raja Shahid-ul-Ajam, another
On 27 August he proceeded again. Stein notes that it was his desire to see the
pass over which Chinese general Kao Hsien-chih had brought his army in 747 AD.
(see Serindia) , the Darkot pass, which prompted him to climb to the
pass through deep snow early in the season during his second expedition. He
described all that in Desert Cathay. So now with the season being much
later he was determined to repeat the effort.
On 29 August he accomplished his desire at 15,250 feet according to his
barometer, or 15,380 feet according to the Trans-frontier map.
This whole section if a fascinating account of Yasin ancient and modern and its
inhabitants and topography, which, unfortunately, we have to skip .
Section III - From the Yarkhun
Head-waters to the Tagh-dumbash Pamir
In this section Stein narrates the course of his movement from the Darkot pass,
he was met by Captain H.F. D. Stirling of the 57th Wilde's Rifles (who later
died during WWI), with a team of Mastuj porters. From
there Stein was forced by a rock slide closure of the north-eastern glacier
route to take the more difficult north-western descent.
there Stein also had to avoid Afghanistan to which he had obtained permission
in 1906 that was now lacking. Therefore he had to cross laterally across the
Yarkhun and Karmbar rivers above 12, 000 and 14,000 feet toward Hunza.
route also crosses the Chillinji pass at 17,500 feet.
On 30 August a long march took Stein from the Baroghil saddle to Murgach near
the Karambar pass. In these very high mountain valleys he met Wakhis herdsmen
with their flocks on summer grazing grounds, even though they were Afghans. He
passed a small fort built by Wakhis as defense against Kirghiz raiders. He was
now at 14,420 feet elevation passing glaciers that fed rivers to both west and
On 1 September Stein continued down and met a team of 40 Wakhis porters waiting
for him near the Chillinji glacier.
On 2 September they again went up to the Chillinji pass over fresh snow.
climb took 8 hours to reach the 17,500 foot elevation. They continued east to
camp at 15,000 feet. His
illustrations dramatically display the scene on these glaciers.
On 4 September they reached Spandrini in the Chapursan valley.
On 5 September they crossed the Kermin pass at 13,600 feet into the Derdi
there they reached the main road up the Hunza valley to Top-khana 5 miles from
Misgar. Stein had used this route for the first expedition in 1900. At Murkushi
he met Muhammad Yaqub and Naik Shamsuddin who had brought the heavy baggage
there directly from Kashmir via Gilgit and Hunza. Thus Stein's lengthy
excursion of 5 weeks west was purely to enable him to see unexplored valleys.
Once again, rather than repeat a crossing of the Kilik pass used in 1900 he
chose a new route over the Ming-taka pass. ( map3bs shows the Ming-taka, Kilik and
On 7 September they reached the crest at 15,650 feet by mid-day.
crossed 15 passes at elevations from 10,000 to 17,500 over a total distance of
520 miles of which 4/5 was on foot, and without leaving India.
Section IV - In the Valley of
From 8 to 11 September Stein passed through the Tagh-dumbash valley to the
capital of Sarikol, Tash-kurghan. (map 3cs shows the
Sarikol with Tash-kurgan to the north). ( map 3aas
shows detail - location of Tash-kurgan beside the river and most of Sarikol).
He had been through this valley already during both his first and second
expeditions. Here he mostly describes changes he observed. As he was to note
throughout this expedition he found that cultivation had expanded since his
visit in 1906. He connected this with an increase in population since the raids
from Hunza had been stopped by British control. Along the way he searched for
evidence of an ancient legend of a canal "Faryad-ariki". He
reports that his old friend, Rashid Beg, explained that the expansion of
irrigation around Dafdar was due to the reopening of this very canal by the
military Amban at Tash-kurghan. Stein traced its impressive construction over
many miles. Stein refers also to the mountain fortress at Kiz-kurghan, which
was already a ruin when visited by Hsuan-tsang in the 7th century and which
Stein described in Serindia. ( map 3es shows
the Kiz-kurgan fort at the bend in the river). His point was that this ruin
could not have survived even in its dilapidated condition except for the
dryness of the climate, which in turn was the reason agriculture depended on
irrigation using runoff of snow from the higher elevations. He again marvels at
the view from the valley of massive Muztagh-ata far to the north (which he had
tried to climb on the earlier expedition). ( map 2r
shows Muztagh-ata and his earlier routes.) He stopped on 11 September to visit
another ruined fortress known as Bazar-dasht. the northwest wall measured 190
yards but the remains rose only 3-4 feet above ground. Stein also revisited the
fortifications of Tash-kurghan. A detached enclosure was measured at 193 by 83
feet with round towers 10 feet in diameter at three corners. There were gates
remaining in the north and south walls. The walls were only 2-3 feet thick
built of large bricks or stones. Further down the ridge he found another
enclosure 53 by 26 feet. Stein spent 12 September in Tash-kurghan making
arrangements for transport further on.
Section V - By the Kara-tash River to
On 13 September Stein departed Tash-kurghan for Kashgar. Again, he selected a
new and unexplored route, through the Kara-tash valley, which was only open at
certain times of the year. (The route also is in map 2r and map 2w ) Crossing the eastern, outer ridges of Muztagh-ata
on the Chichiklik plateau and Dershat valley, Stein felt he was in the footsteps of his 'patron' Hsuan-tsang,
(in 642 AD) as he had been also in June 1906. See Serindia for more
descriptions. On Chichiklik he again found an enclosure and apparent travelers
way station that conformed to Hsuan-tsang's descriptions. From there Stein
followed his 'patron' through the narrow Tangri-tar rock gorges to Tar-bashi.
On 15 September Stein divided his team at Toile-bulung. (For this expedition he
had the additional capabilities of three Indian surveyors.) He sent Lal Singh
eastward to move directly through Yarkand and Khotan to the main section of the
K'un-lun above Kapa and east to Charchan to extend the previous triangulation
higher in the range further east. He expected to meet Lal Singh again at Miran
4 months later on 15 January.
Stein again expresses his confidence in Lal Singh. "I could place absolute
reliance on my own travel-companion's devoted zeal and energy."
Stein sent the heavy baggage on to Kashgar with Afraz-gul Khan and Naik
Shamsuddin. He then went due north with Muhammad Yaqub Khan to survey a new
route across the Merki pass and
down the Kara-tash valley river. In 1906 he had tried to accomplish this by
sending surveyor Rai Ram Singh but the route was blocked. This time the Kirghiz
reported that favorable climate might result in the gorges remaining open. On
16 September in the morning Stein crossed the Merki (or Buramsal) pass at
15,000 feet by using yaks through the deep snow. ( It is in
map 2ffas ) The descent to the north was more
difficult to reach a Kirghiz camp at 12,180 feet where they remained over
night. With fresh yaks he continued down through the river bed. They found
Kirghiz cultivation at the confluence of the Merki and Kara-tash rivers at
10,000 feet. There they obtained Kirghiz camels at Chimghan.
On 18 September they continued north through more gorges which required 2 days
to advance 20 miles. . Further on they had to make use of 'rafiks' that is
artificial galleries attached to the sides of sheer rock walls. At such points
the camels had to be unloaded and the baggage carried by porters. The ponies
and camels had other difficulties threading their way through bounders in the
river bed. Finally they came to a complete blockage and had to retrace steps.
On 19 September Stein found passage for the camels even more difficult as they
had continually to cross and recross the river. Eventually the 'rafiks' became
impassable for the ponies. They were then attached to ropes and had to swim the
river while attached to the camels. It took the men 2 hours to advance 3/4 mile
while climbing a steep slope to 800 feet above the river and then climbing down
again. But they made it through the river gorge to Saman where the valley
On 20 September they continued through desolate terrain. The last defile was at
Tushkuch where Stein measured the 40-yards-wide river flow at 1,400 cubic feet
per second. This was at low water stage. He estimated that at flood season the
flow would be 3-4 times greater. From there they reached a small but
flourishing village at Altunluk. (map 2exxas shows
Stein's several routes into Kashgar, with Altunluk at the lower right side). He
was met there by Kichik Beg, another old acquaintance, sent from Yangi-hissar
Ya-men to await him.
On 21 September Stein completed a 40-mile ride to Kashgar. Muhammad Yaqub Khan
surveyed the route as Stein rushed ahead. Stein passed a walled village,
Ak-bash town, with walls of 129, 144, and 164 yards length. The walls were of
stamped clay to an average height of 20 feet. Above this rampart there was a
wall of sun-dried bricks 13 x 13 x 2 inches to an additional height of 10 feet.
At the northeast corner the wall was 10 feet wide at top with a 3.5 feet wide
parapet to 7 feet height. Stein was welcomed for the night by Sir George
Macartney at the British Consulate General (note now not simply a
representative of the Government of India.)
Chapter III From Kashgar to the
Section I - Along the Outermost T'ien-shan
Once again Stein had to remain in Kashgar for an extended period in order to
organize his work for the coming year and acquire men and transport. As always
his prior planning was meticulous and effective. He had written ahead and
secured reliable Hassan Akhun as head camel man and Hassan had personally
selected 12 special camels for the expedition. Other of the former team of
Turki assistants had also eagerly come to join up. Only the matter of a Chinese
secretary, interpreter, had Stein been unable to find a suitable person. The
fellow from the first tour was incompetent. The excellent individual from the
second tour, Chiang Ssu-yeh, was too infirm to risk lengthy journeys in desert
and mountains. At length Stein had to settle on one Li Ssu-yeh who left very
much to be desired. Although Stein did note that the translator - interpreter -
had never failed in that department and was totally discreet and trustworthy.
It was just that he was not interested in history, archeology or any of Stein's
purposes that had fully absorbed Chiang's attention. Sir George warned Stein
that the Chinese revolution had brought into power officials who would not be
so friendly and moreover there were revolutionary activities about as well.
On 9 October Stein departed for the longest tour in Turkestan he had yet
attempted. He was anxious to reach Lop, over 600 miles away in time to work
during the winter. But of course there were many sights to see and places to
dig on the way. Naturally the route had to pass through Khotan, which Stein was
eager to visit again anyway. But rather than take the 'highway' directly, which
he had traversed before, he got the idea to take a 'short cut' by going
directly east across an uncharted section of the desert and on finding the
Khotan river follow it due south. He sent the heavy baggage ahead by the high
road to Khotan. And Muhammad Yakub was sent on a road east to Maral-bashi.
Stein with Afraz-gul and Naik Shamsuddin headed north east to explore new
ground along the foothills of the T'ien-shan before reaching Maral-bashi.
section is Stein's account of that portion of his trek. (His route is in
On 18 October he reached Maral-bashi.
Section II - Old Remains and Routes
Stein halted at Maral-bashi to make last minute preparations for the dangerous
move into the full desert. ( map 8aas shows the
region around Maral-bashi.) This section is his narration of that failed effort
to reach the Khotan river directly. And one senses a bit of excuse making for
this failure. It was the remarkably high sand dunes that did it.
prepared - he had brought 6 galvanized iron water tanks and 40 some goat skins
from India, but the desert won anyway. He reduced his party still more and sent
more baggage south to Yarkand. While accomplishing these preparations Stein
took a detour to visit a ruin at Lal-tagh north east of Maral-bashi. He records
considerable detail and his conclusions about the road between Kashgar and
Ak-su. He made a circuit and returned to Maral-bashi.
Section III - A Hill Range in the
On 25 October Stein finally departed Maral-bashi into the desert. He wanted to
reach the Khotan river directly at the hill fortress Mazar-tagh, over looking
the river. He explains his rationale for attempting this exploit. He claims (no
doubt rightly) that he was under no illusions about the difficulty of the
endeavor. He tramped south eastward for 3 days. ( map
8acs shows Stein's route south-east into the desert and his return).
By 31 October he was beginning to recognize the inevitable. Some sand dunes
rose over 300 feet high directly across his route. Some of the extra hired
camels were already breaking down. In one ten hour period they traveled only 11
miles but actually advanced only 7 miles. At that point Stein reckoned he was
still 100 miles from Mazar-tagh. That meant that even at maximum rate of march
it would take two more weeks to reach the river and water. Stein also knew
about the disaster Sven Hedin had survived a few years before. He provides an
the forbidding terrain ahead. Facing the strong likelihood of disaster Stein
turned around and headed back to the Yarkand river near Maral-bashi. Stein
concludes this section with recommendations for future exploration of the area.
And also he supplies more detailed descriptions of the topography. He was
convinced that there had been a very ancient mountain chain or at least ridges
connecting Maral-bashi and Mazar-tagh and wanted to prove it.
Section IV - Past the Mazar-tagh of Khotan
In this section Stein describes his alternate route to Mazar-tagh. He turned
back from his effort to cross the desert directly and reached the Yarkand river
in three days, which flows north along the entire western edge of the desert
between the desert and the enclosing mountains.
From 5 to 8 November Stein rode north-east until he reached the outskirts of
Ak-su cultivation. Once reaching the confluence of the Yarkand with the Khotan
river flowing from the south, Stein turned that way and hurried upstream
(south) on the Yarkand River. ( map 12das and
map 12xx2as show the river confluence) This was
over ground he had surveyed in 1908.
On 17 November Stein and party again surveyed to Mazar-tagh ridge and its
Tibetan fortress. (The
length of the Khotan river requires several map sections -
map 13 bas shows the location of Mazar-tagh.
Description is in Serindia.) There Stein found waiting Muhammad Beg, the
head man at Islamabad, with a team of diggers from the south that Stein had the
foresight to order up previously. Again raiding the extensive refuse dumps
revealed more documents and other artifacts. Stein also found a Buddhist shrine
he had missed in 1908. Stein had to abandon further exploration and rush on to
On 21 November he reached the oasis town to refit for the next phase, a winter
of work at Lou-lan. The section ends with two pages of the usual detailed
descriptions of artifacts recovered around Mazat-tagh.
Chapter IV - From Khotan to Lop
Section I - Antiques from Khotan Sites
This section briefly describes Stein's preparatory activities in Khotan for the
coming winter in the desert. And it contains notes about some of the more
interesting of the artifacts listed in section II. (Khotan is in many of
Stein's maps from all three expeditions and his text also is in the various
reports - we consolidate all this at Khotan..
On 29 November Stein departed Khotan.
Section II - List of Antiques
acquired at Khotan
In this section Stein includes 27 pages of detailed descriptions of items
brought to him at Khotan. As always Stein carefully distinguishes the items he
has personally found from those supplied by other of the wide net of 'treasure
seekers' he encouraged to bring in their finds. In this case much of the
material listed came from Badruddin Khan and Tokhta Akhun.
Section III - Finds at Sites near
The section ends with 8 pages of detailed descriptions of items found at or
near Domoko. ( map 14das shows the relationship of
Domoko and Kadalik to the caravan route between Khotan and Keriya.
On 30 November Stein left the edge of Khotan cultivation. He still had 700
miles of travel to reach the Lop Desert. He was in a rush, but not so much as
to skip a return to Domoko, even though he had explored the oasis in 1901, 1906
and 1908. In this section he describes significant changes he observed, mostly
increased irrigation and cultivation. He notes extensive additional irrigation
around Gulakhma, Ponak and Domoko. He was guided to a new find by his former
guides, Mullah Khwaja, Turdi and Kurban. These eager 'treasure seekers' had
found bundles of manuscripts they knew Stein would want. But in their eagerness
they had obliterated the structure of the shrine itself. All Stein could
recover from the ruin was fragments of frescos - plaster painted in tempera
that he judged the artistic equal of the completed work he had uncovered at
Khadalik in 1906. He judged that the shrine had been
abandoned toward the end of the 8th century. Stein remarks about the new
discovery of a Buddhist shrine that with the deceptive terrain full of
tamarisk-cones he is not surprised and expects still more such ruins will be
The section closes with 8 more pages of detailed descriptions of items all
carefully organized to show from whom or where they were obtained..
Section IV - The Niya Site revisited (many photos are here.)
Stein again spent three days at Keriya, among other
preparations hiring 12 additional camels. Despite his sense of urgency to reach
Lop he could not avoid the pleasure of returning once more to Niya, one of the
locations of his spectacular achievements during the first expedition.
On 8 December he reached the oasis. He carried orders from Tai Ta-lo-yeh, the
Amban at Keriya, which promptly secured for him 40 laborers and a month's
supply of rations to support his onward journey to Charchan. (
map19bas shows the area south and east of Niya
On 13 December he started north to the ruin. ( map
19bxas shows the route ). He noticed that the camel tracks from his visit
in 1906 were still visible. Three days' march brought him back to the Niya
excavated at several sites, marked N xlii and N xliii and wide reconnaissance
located a few more. He describes the Niya ruin area in detail with
illustrations. Besides buildings he found a former orchard and vineyard. With
the much larger team of laborers this time he was able to clear more buildings.
- N iii, xii, N xxxix
On the way he noticed that summer floods had damaged the irrigation system -
yes, too much water could be as devastating as too little. From the point of
view of the archeological explorer, he remarked that centuries hence an
explorer observing the abandoned buildings would not be able to ascertain the
real cause of their abandonment. Stein, always ready to return favors, after
returning from the Niya ruin, put his 40 laborers to work helping to repair the
irrigation barrage and canals paying for this out of his own funds.
always carefully measured the flow of rivers he encountered. With the
assistance of his long time 'factotum' Ibrahim Beg they determined that the
flow would support additional cultivation.
On 18 December Stein departed Tulkuch-kol ruin area, with the temperature
already reaching 42 degrees F. below freezing, headed east for Charchan.
Section V - List of Antiques from
This section is 7 pages of detailed descriptions of items from Niya.
Chapter V - On the Way to Lop-nor
Section I - Charchan and Vash-shahri
On 19 December Stein departed Niya river headed for Lop-nor. He followed his
1901 and 1906 route along the Yar-tungaz river for two days. ( map23cbs shows the first part of this route.) (He
always appeared apologetic for going along a previous route rather than explore
a new one.) In this case it enabled him to note that on sand dunes he could see
the footprints of his camels from those prior trips. From the Yar-tungaz he
crossed the desert directly toward the Endere. He noted that the Endere had
again changed course, a phenomena that he already had pointed out created great
difficulties for establishing continued irrigation systems. Stein refers to the
report of his second expedition - Serindia - for his description of the
region. He did not stop this time at Endere but proceeded directly to Charchan,
where he stopped for two days. ( map 22abs is detail
of Charchan town.) There he received word about a disturbance at Charkhlik
related to the Chinese revolution. He provides an interesting account of the
impact of the revolution on public affairs in Turkestan (not good). In this
case the bandits had captured Charkhlik. ( map26aas
continues the route with side visit to Vash-shahri, and map26bas continues along the river.)
On 31 December Stein was advised to proceed with caution, which he did over the
142 miles to Vash-shahri. This required 7 days ride along the left bank of the
Charchan River. At the edge of Vash-shahri he was stopped by armed local
Muhammadans who believed his group to be more bandits.
was cleared up by Stein's previous associate there, Roze Beg. Roze Beg again
guided Stein's brief exploration of Vash-shahri which he had already explored
in 1906 and described in Serindia. In this report he discusses a few
more remains including coins confirming that the site was occupied into the
12th century. The place was known as Hsin-ch'eng in the Tang records which,
with other records, indicates it was founded around 627-49 AD. Stein also
studied the current situation. He noted extensive improvement, increase in
population, expanded cultivation and prosperity since 1906.
From Roze Stein learned that the bandits had captured, tortured and killed the
Chinese Amban at Charkhlik and taken control of the village. The local
Muhammadans remained neutral and tried to avoid trouble with whichever Chinese
regime had the upper hand. The former Amban had fortunately sent to Urumchi for
help. A strong unit of Tungan troops duly arrived from Kara-shahr. They
secretly entered the village at night and killed off the bandits. This incident
is of interest relating to the history of the Chinese revolution and Central
Asia. But it had another result remarkable to the success of Stein's projects.
Stein learned later that the revolutionary government at Urumchi had sent an
order to the Amban at Charkhlik to prevent Stein's further exploration. This
order was in the Ya-men safe during the incident, unexecuted by the civil
Amban. It remained there since the Chinese regulations prohibited military
officials, such as the commander of the relieving unit, from performing civil
functions. Thus Stein was able to secure all the help with men and camels he
needed at Charkhlik and proceed back into the desert to Lou-lan before this
order was found.
Section II - The Sites of Koyumal and
On 6 January Stein departed Vash-shahri and reached Charkhlik on the 8th. He
found that the 'disturbances' and passage of sizable units of Tungan troops had
depleted the scanty resources of the small town, greatly increasing his
difficulty in hiring workers, camels and supplies for 3 months' work in the
desert. He notes in this section that in Serindia he already described
Charkhlik's history during the era in which Marco Polo passed by. Marco Polo
called the town "City of Lop' and Hsuan-tsang called it Na-fu-po. In the
T'ang annals it was Shik-ch'eng (Stone Town). Stein notes again his
identification of Charkhlik with the Chinese military colony established there
in 77 BC. He had to remain in Charkhlik for 6 days in order to collect the
necessary resources, but used the time also for additional archeological work.
He found remains of a brick ruin 15 feet in diameter near the center of the
ancient walls previously described. This he identified as a stupa. To the
south, outside the cultivated area he found two more ruins. One was called
walls 8 feet thick were made of sun-dried bricks. The walls were about 218
yards long where still existing. In the center was another stupa some 28 feet
square and 14 feet high. The sun-dried bricks were 17 x 9 x 4 inches.
this were two Vihara chapels each about 20 feet long and 9 feet wide. The
remains contained many fragments of plaster and wood. Another shrine was 30
yards to the south. Some documents in Sanskrit and Gupta were recovered.
Stein moved on south about 1.75 miles to another place named Bash-Koyumal. Parts
of its wall remained in segments about 45 feet long made of sun-dried bricks 17
x 9 x 4 inches and 4 feet 9 inches thick. There was also a massive wall 10 feet
thick and 50 feet long near the center.. Fig 107 shows
another stupa 12 feet square and remaining to 9 feet height made of bricks 17 x
9 x 4 inches. More fragments (including a rare one on silk) in Sanskrit
language and Brahmi writing in Gupta style were found. Pieces of stucco relief
included a head of Buddha.
Section III - Resumed Labors at
Miran (many photos are here.)
Stein was relieved and happy when Lal Singh arrived from his 4 months' long
survey in the Kun-lun and the Altin-tagh south of the desert caravan routes.
This was to extend the surveys of 1906 at Kapa further east of Lop-nor as far
as Nan-hu. (The results are in maps 23, 27, and 30)
On 15 January Stein departed Charkhlik despite failing to secure adequate
numbers of workers and amounts of supply. He rushed to Miran which he reached
in 2 days. ( map30acs shows the area between
Charkhlik and Miran.)
On 17 January he found the condition of tiny Miran much improved and expanded
after 7 years. He also found that the water supply of the Miran river exceeded
that of the Charchan river at Charkhlik. Stein was eager to return to the
ruined stupa at Miran which he had excavated and photographed in 1907 as shown
in Serindia. Unfortunately his careful effort at preservation by burying
the frescos was undone and ruined by a Japanese amateur explorer who had
damaged the walls while cutting away some fragments. However other sections of
the walls remained protected by earth which Stein then set about excavating.
Stein photographed these frescos and then began the careful removal he well
knew how to accomplish. In fact special tools for this delicate endeavor had
been developed and created in India at the shops of the 1st Sappers and
was undertaken by Naik Shamsuddin and Afraz-gul and Stein personally. Once
removed the frescos had to be packed in specially built wooden crates. The
effort required 12 days to complete. The results are to be displayed in the new
museum in New Delhi.
Stein was able to incorporate much of the new information he obtained at Miran
in his chapter in Serindia since that report was not completed until
after this third expedition. However, he records here some information about
additional ruins he found at this time. One of these was a site only 1.75 miles
north. This was another stupa remains standing to 5 feet. Stein found enough
objects to date the place to 8th-9th centuries AD. Another site was about 1.25
miles northeast of Mxiv. . It was a tower 17 feet square and 16 feet high made of
sun-dried bricks 18 x 10 x 4.5 inches. Stein was taken to yet another ruin on
18 January, M xv, about
a mile north east of M v. This turned out to be not a stupa but a shrine that
had contained statues. Several life-sized heads were recovered amid the debris
of the collapsed dome and walls.and and
It was at Miran that Stein received a letter from Sir George Macartney
informing him of the edit issued at Urumchi to prevent his archeological and
survey work. In addition he was having great difficulties in hiring the
recalcitrant Loplinks and their camels for the extended work. Very fortunately
at that moment an old friend, Sher Ali Khan, a Pathan merchant traveler
happened by with 40 camels loaded with tea enroute from Tun-huang to Yarkand. He
agreed to exchange camels for those wanting to return to Charkhlik and also to
assist with loading his own camels with the cases bound for Kashgar. Unable to
spare camels Stein had to dispatch Lal Singh for his survey of the Kuruk-darya
north on ponies. The hoped for plan that Stein had made months earlier was for
Lal Singh to meet Abdur-rahim at Tikenlik. Stein also had to abandon his
previous plans for exploration by Muhammad Yaqub Khan around the eastern side
of the Lop Sea and instead send him along the southern shore and then for a
level survey up the Su-lo Ho basin. Stein proposed to excavate remaining sites
at Lou-lan and then search for the ancient Chinese caravan route directly
across the Lop sea from Lou-lan back to Tun-huang.
Eventually Stein managed to collect 30 camels. They had to carry ice for 35 people for a month and food
supplies for a month for all plus food for his own team for another month, plus
of course the equipment, photographic plates, silver, tents, winter clothing
and bedding. Stein records that 'it goes without saying that everyone had to
walk." Thus he was greatly relieved on 31 January to be ready to set out
while seeing off the caravan carrying the carefully packed cases west to
Chapter VI - Remains of an Ancient
Section I - The Ruined Fort of L. K.
In this section Stein describes the topography and vegetation encountered on
his march north from Miran.
On 1 February Stein departed going north with old Tokhta Akhun serving a guide
once more. He had to leave at Miran his Chinese interpreter, Li Ssu-yeh, too
weak to cope with the desert along with two guards and the non-essential
baggage, all of which had to move later on via the caravan route to Tun-huang.
Stein's first objective was a fort found by Tokhta Akhun in 1910. The first
day's travel was to Abdal on the Tarim River. ( map30abs
shows the area around Miran and Abdal.) The following day they crossed the
frozen Tarim River. ( map30das continues the route
to the north-east and then reaches Lou-lan on map29gas. ) After another day they reached Uzun-kol,
where they found fresh water ice to cut and load onto 19 camels. There are many
maps and texts at Lou-lan .
On 4 February Stein was able to see some peaks in the Altun-tagh far to the
south and use them to triangulate his position with his plane table.
On 5 February Stein was able by using field glasses to sight the ruins of fort
L.K. in the distance. They soon noticed another fort west northwest of the
first. Near fort L.K. they picked up stone-age remains on the desert surface
about three miles west of the area in which they had found similar items in
1907. They soon found a Chinese bronze coin. Near the fort they crossed a dry
bed of an ancient river. and, and,
fort was an irregular oblong with longer sides facing northeast and southwest
about 620 feet long and the shorter sides 330 feet long. The corners were
oriented to the cardinal points. The walls were massive but very badly eroded.
But piles of drift sand protected parts of the walls. The surviving sections of
the southwest face clear of sand on the inner side showed the construction
methods. Outside and inside the fort the wind had eroded the ground level to
depths of 25 feet. The wall was built of alternating layers of clay and Toghrak
trunks and branches laid crosswise. The layers were more narrow as they rose
giving the walls a slope inward. It was 32 feet wide at the base. The first
clay was lumps of hard clay 5 feet thick, excavated from near the river. The
second layer of Toghrak was 22 feet wide and 1.5 feet thick with a leveling
layer of tamarisk brush under it. Then came a layer of clay 4.5 feet thick
followed by another of Toghrak timber 15 feet wide and 2 feet thick. Above that
was a clay layer 4 feet deep and then a layer of Toghrak 10 feet wide. The top
of the wall was eroded too much for accurate measurements. But Stein was sure
that the original top layer had been clay with a likely parapet. He presumed
that the wall had originally been coated on both sides with clay. This had
eroded along with the outer portions of the clay layers.
sketch shows that the original wall must have risen to over 21 feet. The layers
of wood increased in depth with each layer upwards while the clay layers
decreased in depth. This method was to reduce 'top heaviness'. The wall was
reinforced by vertical timber posts in pairs inner and outer, connected for a
frame and some 15 feet apart. Stein believed that the design indicated Chinese
engineering. The original gate was in the northeast face about 100 feet from
the eastern corner. The top of the gate was eroded away but there was a
considerable remains of the timber framework. The sides of the gate were
revetted by nine posts on each side set in two massive foundation beams each 22
feet long. A cross beam joined the two near the entrance indicating that the
gate had been 10 feet wide and 10 feet high. The gate was closed near the outer
end by a massive wooden door of two leaves each 5 feet wide. One of these was
on the ground with its boards 3 inches thick secured by stout cross-joints. The
cross beam had sockets that once held the door jambs. The adjoining posts had
holes into which fitted the cross-bar securing the folds when closed. This
facility was similar to the gate found at Kara-dong in 1901.
Inside the walls there was an area near the middle of the northeastern wall
about 130 feet by 100 covered with heavy timber debris . To the south of it
there were remains of two small groups of timber and wattle built quarters. The
timber was cracked by centuries of erosion. No refuse was found in this fort.
Excavation revealed that the original walls were of a Toghrak frame with
vertical wattle packed with tamarisk branches secured to cross beams joining
the posts. The plaster on both sides reached 8 or 9 inches in thickness. A
western room measured 27 by 20 feet. In it was a fine carved double bracket
capital 3 feet long similar to some at Miran and Lou-lan. They are similar to
Gandhara relievos and Persopolitan models. But only a few small objects were
found in the ruin. Stein continues with descriptions of a number of other
ruined buildings. No written remains were found at L.K.
Stein estimated the fort was dated from the same period as Lou-lan, that is to
the 3rd century AD. The fort L.K. is on a straight line between Lou-lan and
Miran, which was at that time the capital of the region and known as Yu-ni.
about 30 miles south of Lou-lan, and was no doubt designed to protect the
important route between those two posts.
Section II - List of Objects found
near or excavated at Fort L. K.
This section of 3 pages describes the objects found there. The wooden double
bracket capital is described in detail.
Section III - The Sites of L. L and
On 7 February Stein went to the small, nearby found by Tokhta Akhun and
designated L.L. It was 3 miles west of L.K. It was also near the dry river bed.
The walls were of stamped clay and layers of tamarisk in an oblong shape with
shorter sides 138 feet and long sides 218 feet in length. The preserved rampart
was 26 feet thick and seven layers of clay, each 16 inches thick, high. Each
tamarisk layer was 6 inches thick. The east wall probably had contained the
gate but was eroded. The east wall projected out 42 feet to make space for a
small inner enclosure about 68 feet long.
From L.L. Stein walked northwest about 3 miles to another site,
contained small wood buildings, the ruins thereof, The first excavated was a
room 25 by 30 feet.
Another room had the much sought for pile of refuse, 2 feet deep. Textiles
including silk and leather were found, along with some documents. Another room
also contained items such as a lacquered casket. Stein describes clearing more
buildings with varying contents. There was a document written in Sogdian script
and others in Cursive Brahmi and Chinese. Stein excavated many buildings in
this rather large area over which they were spread sometimes hundreds of yards
apart. He figured that the survival as ruins of these massively constructed
buildings indicated that it was likely that many more lesser structures located
in between had vanished.
In March 1915 Afraz-gul returned to the area while completing the surveys along
the Kuruk-darya and around Lop during which he found another site not far to
the west that they labeled L.R. (See later chapter). Stein was convinced that
the string of settlements from L.K. to L.R. along a line of 10 miles were the
southern settlements on the Kuruk-darya delta while it still reached Lou-lan
area. The area around L.M. furnished the local agricultural resources needed at
L.K. and together these settlements were on the direct route between Lou-lan
Section IV - List of Antiques
excavated or found at the Sites L. L., L. M. and L. R.
This section of 6 pages is description of the items found at these locations.
Section V Across the Ancient Delta of
On 9 February Stein moved northeast from L. M to return to Lou-lan, that is
station L.A. where he set up his base. Stein had described this area in
Serindia, but did so again in more detail in this section. The route lay
across repeated eroded ditches, Yardangs, and sand dunes. They also crossed the
dry bed of the branches of the Kuruk-darya. Along the way they again found many
stone age implements on the desert floor. They found many Han era items as
On 10 February ,upon reaching the familiar stupa at Lou-lan at night Stein set
a big bonfire to guide the camel caravan that lagged far behind. The section is
completed by two pages of detailed descriptions of items found along the march
Chapter VII - Remains of Ancient
Section I - Work resumed at and around the Lou-lan Site
In this brief section Stein focuses on describing some of the new results,
having already thoroughly analyzed the Lou-lan area in Serindia.
On 11 February Stein began work promptly in the morning. (
map29eas is detail of central Lou-lan.) His first
task was to send most of the camels with Tokhta Akhun to the spring at
Altmish-bulak locate to the north at the foothills of the Kuruk-tagh, to
recover. ( map29cas shows the area around
Altmish-bulak and Lou-lan). He sent Afraz-gul to explore to the north and
northeast searching for signs of the caravan route from Tun-huang. Stein notes
here that he has already incorporated most of the new information obtained into
the text of Serindia, which was completed after his return from this
third expedition. But it was this third expedition from which he obtained much
of importance. In 1906 he could only discern two of the walls running east
north east to west southwest. This time with more opportunity for detailed search he found the
two other walls at right angles. and Fig.
the remains of a large Ya-men-like structure at L.A. ii.
On 12 and 13 February he sent a detachment to search for more ruins. The rest
dug away in the piles of refuse. They found many Chinese documents on wood and
paper. 40 more records were uncovered at L A iii. including one in Kharosthi.
More rubbish heaps revealed many more finds. His descriptions fill several more
Section II - Miscellaneous Objects
found at or near Lou-lan Site L. A.
This section is 5 pages of detailed descriptions of items newly found around
Lou-lan. (Note - illustrations of many relics described in this and the other
sections are published in volume III of Innermostasia)
Section III - Relics of an Ancient
Stein comments that the weather was often 44 degrees below freezing but clear.
He sighted the red ridge of the Kuruk-tagh to the north and even the peaks of
the Kun-lun far to the south.
On 15 February Stein tramped to the northeast to the locations found by
Afraz-gul. Two sick Lopliks were left in camp with Ibrahim Beg and the ice
depot. Stein and his team had to walk, carrying the essential baggage. They
came upon a mesa, shown in , 56
yards long and 32 yards wide at maximum. The mesa had been selected for a
cemetery Stein numbered L.C. Some
graves had become exposed by erosion but the significant ones had to be
immediately found excellent pieces of silk. The graves were all a jumble with
bones mixed with boards and fabrics of all kinds. There were also Chinese
documents. Stein soon realized that these graves in which not a single intact
skeleton was found had actually been the result of a movement of the remains
from some other grave sites. Stein describes the graves themselves, their
content and the local topography in detail. He dated the removal and reburial
to not later than the first part of the 3rd Century AD.
Section IV - The Textile Relics of L.
In this section Stein discusses on detail the textiles found in the graves
discussed in section III. He notes that the variety of fragments is the result
of the Chinese procedure of wrapping the bodies tightly in a mass of used
garments and other fabrics. In this case the largest number of fragments were
of silk, with wool coming second. The numerous examples have given experts a
great trove of different types.
Section V - The Decorative Designs of
the L. C. Fabrics
Stein continues here with descriptions of the individual fragments. They
represent a wealth of different designs and constitute a selection of the
earliest known examples of Chinese silk production. Stein confines himself to
discussion of the main types and groups into which these fabrics may be
classified. That this section fills 10 pages is indicative of the content, even
at the level of generality Stein claims it is.
Section VI - Miscellaneous Sepulchral
Deposits and Descriptive List of Antiques from L. C.
In this section he turns to discussion of the other remains found in these
graves. Many objects are of personal use. There are trays of food. The section
concludes with 13 pages of the usual very detailed descriptions of each
numbered individual item.
Section VII - The Ancient Castrum L.
E. and the Remains on Mesa L. F.
Stein left the graves for later study and marched on to the northeast. His
objective was a fort found earlier during Afraz-gul's reconnaissance. This was
numbered L.E. It was about 19 miles from L.A. . This fort was identical in construction to the Han wall near
Tun-huang, dating it to the second century BC. Its purpose was determined by
its location (or reverse) namely to serve as the first fortified post on the
Lou-lan side of the caravan route there from Tun-huang. The walls had withstood
2000 years of ceaseless wind erosion. (plan 12). The fort was approximately a
rectangle The walls were oriented so that two would lie along the direction of
the prevailing wind. The east and west walls were 450 feet long and the north
and south walls about 400 feet. The main gate was 10 feet wide and in the southern wall. The walls were of
remarkable strong construction. They were built of layers of closely tied fascines of tamarisk
twigs a foot thick alternating with strata of stamped clay 5 to 6 inches thick.
The salt impregnation resulted in the walls being like concrete. In
addition the walls had revetments of longitudinally fixed fascines. The
walls appeared to have been about 12 feet thick but in some places as much as
18 feet thick. In places it still rose to 10 feet in height. Stein judged the
walls as sufficiently strong to resist any Hun raiding party.
On 16 February Stein and team marched on to mesa L.F. on which Afraz-gul had
found some graves. Stein's photo shows this to be an imposing ridge over 100 feet
high amid erosion. Once on top they found a wall 5-6 feet thick of clay slabs
with gate and a cemetery. The enclosure was 200 by 80 feet. The timber gate remained
upright. The refuse in two rooms yielded several documents in Chinese and
Kharosthi. The outpost with such a commanding view in all directions was on the
direct caravan route from Tun-huang.
The first grave contained a body in remarkable state of preservation.
the body and its covering clearly identified it as non-Mongolian, a member of
the Indo-European tribes that inhabited the region prior to the Chinese
arrival. (Such mummies are now displayed in the museum in Urumchi and have
generated a lot of discussion about the first inhabitants of Central Asia.)
describes the body and clothing and accompanying artifacts but could not remove
more graves were then opened to reveal more non-Mongolian bodies.
noted they were members of the indigenous population of Lou-lan. Stein
refers to the Former Han Archives in which the descriptions of the inhabitants
also indicate they were Indo-European. Stein found many artifacts on and around
the mesa including Neolithic tools and Chinese coins.
Stein's survey showed that all these sites were on a straight line to the
northeast from Lou-lan, the route he predicted for the caravans from Tun-huang.
So he planned to begin his own exploration across the Lop sea in search of that
route from station L.E. But first he had to prepare carefully for the dangerous
trek. He and his team returned to L. C. and then to Lou-lan (L.A.)
Section VIII - From the Lou-lan
Station to Altmish-bulak
Upon his return to L.A. Stein was relieved to find that Lal Singh had arrived
from his survey route of over 150 miles of the Kuruk-tagh and Kuruk-darya. And
the camels had returned from Astin-bulak. Lal Singh had met Abdurramim in
Tikenlik (see later for location south of Singer) as Stein had requested months
earlier. This desert hunter was a master camelman as well, owning 5 magnificent
camels that Stein always admired. Lal Singh had found four more cemeteries in
the desert before joining Tokta Akhun at Astin-bulak. With all the teams now
joined, except Muhammad away doing a level survey up the Su-lo Ho, Stein was
ready to move on. First, however, the camels and men needed at least a brief
rest at the spring at Altmish-bulak. Stein sent some camels and the Lopliks
back to Miran along with cases containing the recent finds. All this plus the
baggage previously left at Miran had to be transported on directly to Tun-huang
under Ibrahim Beg's supervision. Stein ordered that the caravan from Miran
would met his team at the wells of Kum-kuduk.
By 18 February, with all plans safely underway, Stein and his main crew went
north-west picking up more coins and bronze items along the way past the stupa.
On 19 February Stein sent Abdurramim with the camels on directly toward
Altmish-bulak while he detoured with Lal Singh and Afraz-gul to inspect one of
the cemeteries Lal Singh had found. A few miles further they found a ruined
wood structure. This was on the direct line between Lou-lan and the Kuruk-tagh.
A few miles further they found an underground dwelling on a ridge above the
surrounding terrain. The room contained three coffins. Lal Singh's discovered
cemetery lay another mile to the northwest. Here they found four coffins
containing the usual mix of personal items, food and also Chinese silk.
time running out Stein's examination was rushed. Even so they did not make it
to Abdurramim's camp until after dark. Thanks to Abdurrahim's going out looking
for them, they reached the camp in good order. After another day's march they
reached the welcome spring at Altmish-bulak. They found plenty of good ice,
timber to melt it with, and extensive reeds and tamarisks for the camels. They
spent a few days resting and refreshing for the coming ardurous crossing of the
Lop sea. The section ends with a 3 page description of artifacts found during
Chapter VIII - The Search for the
Ancient Chinese High Road
Section I - To the Easternmost Outpost of Lou-lan
From 21 to 24 February Stein's party remained at Altmish-bulak. (
map29ccs ) The coming exploration would be even more dramatic and dangerous
than his remarkable journey south across the heart of the Taklamakan to the
Keriya River and his long trek through the high mountains and western Tibetan
plateau in search of the sources of the rivers flowing to Khotan during his
second expedition. Meanwhile Stein was busy planning further operations
including a return to map the Kuruk-tagh the following winter. Stein divided
the teams in order to accomplish more. His group included Afraz-gul,
Shamsuddin, Tokhta Akhun and another Loplik hunter plus camelmen. They
had 20 camels, eight for ice, four for fuel, and eight for baggage and
provisions for the long trip to Tun-huang. Lal Singh and his three assistants
had Abdurrahim and his five remarkable camels. Lal Singh was assigned to survey
around the north and northeastern edges of the Lop Sea and the sections of the
Kuruk-tagh to the east, while Stein crossed the dry salt sea more directly.
They were to meet at Kum-kuduk wells on the Tun-huang caravan route south of
the Lop Sea. This was also the place Ibrahim Beg was to bring the other caravan
from Miran. Stein expected to complete the crossing in 10 days, but would need
to carry not only ice but also the fuel to melt it and cook meals. The camels
probably could not survive more than 10 - 12 days. He would have to navigate by
compass and dead reckoning. He noted that success depended as much on 'good
fortune' as on his careful planning. (These routes show on the maps linked
On 25 February they departed Altmish-bulak early. Lal Singh and Abdurrahim went
east while Stein went south southeast towards L. F. He
describes the terrain. They continued on 26 February through mesas, ravines and
ridges. They reached the mesa Afraz-gul had previously visited, L. I. bypassing
for the time the graves on L.Q. They were forced to halt there in order for
Hassan Akhun to resole the camels and rescue one that was foundering. Stein
used the time to explore the three mesas L.I. He found evidence of periodic
occupation by indigenous herders. He then explored the vicinity looking for
evidence of the ancient caravan route. A few miles northeast he found another
mesa exactly on the same line as the previously explored sites northeast of
Lou-lan with the remains of a much eroded watch tower, designated L.J. He noted
that the route to the northeast was by no means direct to Tun-huang but had the
advantage of skirting the worst of the dry salt sea.
Stein describes the procedure for sewing bull hide directly onto the skin of
the camel's feet which were continually cut by the sharp salt cakes. Another
unpleasant task was to administer a dose of rape seed liquid by pouring it down
the camel's nose. These required a team of men to hold the camel down and took
all together half the night.
Section II - The Location of the
'Town of the Dragon'
On 27 February Stein and team got an early start toward mesa L.J (which is on
map32aas ). From the top of L.J. they could see
another mesa in the distance at a bearing of N.60 degrees E. so set out for it.
Only a mile further on Tursun Akhun, a camelman, saw Chinese coins on the
surface. Stein picked up 211 of these coins lying in clumps along a line over a
distance of 30 yards and only 3-4 feet wide. (on same map, which also shows Lal Singh's route further north)
What a remarkable find. Stein assumed these must have fallen out of a bag or
box carried by a Chinese caravan at night and left in situ for over a thousand
years. Apparently this line was slightly off the standard caravan route or the
coins surely would have been found at some later date. Further search by Nail
Shamsuddin found a batch of bronze arrow heads of Han period type. The location
and alignment of this remarkable find assured Stein he was indeed on 'the right
track'. It also greatly raised the morale of the Turki followers and their
belief that Stein not only knew what he was doing but was guided by benign
spirits. They continued on for ten miles moving just south of rows of mesas.
After another 18 miles they reached the end of this group of mesas.
Stein again digresses with extended discussion of ancient Chinese records and
memoirs. The topic is the identification of a 'Town of the Dragon' prominently
mentioned in those accounts. Stein discussed the same topic in Serindia.
Another issue is that the ancient Kuruk-darya likely flowed east past Lou-lan
and into the sea at its northern corner during centuries prior to its turning
south to flow toward the southwest corner. But in any event Stein was sure that
the ancient Chinese accounts describing a "Town of the Dragon"
referred not to an actual town but to a physical area that might appear in myth
as a town. Stein guessed that this unusual line of mesas constituted the real
'town of the dragon.
Section III - Across the
Salt-encrusted Lop Sea-bed
On 28 February they again started out at daybreak.
(continuing on map32bas ) Stein had to choose the
direction of march. He headed east toward a high mesa. Again, they were lucky
to find a Chinese Wu-chu coin within half a mile. On reaching the mesa another
unusual discovery took place. One of the men saw three more Wu-chu coins.
Beyond that were various Chinese implements and tools of iron or copper plus
more coins. All
these were confirmed as Han era items. But Stein was at a loss to explain why
the objects were there. Stein decided to continue eastward through a series of
salt encrusted ridges. The ground soon turned to sharp salt cakes that again
hurt the camels' feet. They had to cross many lines of salt-encrusted hills.
These forces Stein to alter the direction to N 20 degrees E by South 20 degrees
west in order to skirt this trying line of hills. He turned northeast. (He
notes at this point that a year later Afraz-gul from the same spot turned
southwest and found to open ground within 2 miles. A few miles further they
again found better ground so turned back to N 55 degrees E. Again, after a few
more miles he turned East. By evening they reached the edge of the flat dry
salt sea. Again the night was taken with resoling the poor camels.
On 1 March they could see the real dry salt sea before them. They turned S 94
degrees East. They continued across very rough salt cakes. That night one camel
failed to continue and had to be shot.
On 2 March they moved on toward the southeast. Once across the salt sea they
encountered again rows of ridges and hills. Proceeding for another half mile to
the south southwest they again came upon a Chinese Wu-chu coin.
Amazing luck. They they found a glass bead and further on
another coin. Niaz Pawan, one of the two Lopliks, noticed human footprints.
These led to yet another coin. Then the footprints led to those of two more men
plus a pony and donkey. Stein could not imagine the cause. But that evening
Mahmud, one of the camelmen who had been with Lal Singh, recalled a story he
heard about a Chinese merchant being robbed by three Khotanese who had fled
with his donkey and pony. The mystery was solved. But the fate of these bandits
was only imagined, namely that they would surely perish before going too much
further into the Lop salt sea.
Section IV - The 'White Dragon
Once again Stein digresses into a historical topic. The question is to locate
the "White Dragon Mounds" so frequently mentioned in the Ancient
Chinese annals. Stein contends that this formation was the line of mesas they
passed on 28 February as described above. He quotes and analyzes the account
left by Li Tao-yuan and those of various other ancient authors. He adds to this
information on the topography obtained by Afraz-gul while conducting his
assigned survey starting with the mesa on the west side of the sea the
Chapter IX - To the Su-lo Ho Delta
Section I - By the Eastern Coast of the Dried-up Sea
Stein and party continued on to the south southeast in hopes of
avoiding more of the salt terrain that damaged the camels. They soon turned
south southwest along the edge of the sea. They found two cairns whose purpose
they could not discern. After crossing another inlet of the salt sea they again
found the track of the robbers so knew they were on the right route to
Kulm-kudak. ( map32fas shows the routes Stein and
Lal Singh took to their meeting and the route the baggage took from Miran.)
They covered 21 miles that day.
On 4 March they saw the coming of the buran season. They continued through
varied terrain. Then suddenly found the clear line of the Han caravans etched
deeply into the salt forming a straight line some 20 feet wide. Evidently the
heavy traffic of caravans and carts had crushed the salt blocks into a
relatively smooth roadway. They continued to find small objects that indicated
they were on the right track of the Chinese caravans. They covered 22 miles
that day. (the routes are shown on map32xx1as and
continue shifting southward on map32xx1bs )
On 5 March they continued for a while and then took a brief break to rest all
before crossing the next salt sea bay. The stop gave them time for explore the
area and soon turned up some pottery and iron items.
On 6 March they started again. They found tracks of wild camels, the Chinese
caravans and of humans. Later on they even found the remains of Lal Singh's
camp over the night of 4 March. From the top of a 100 foot high mesa Stein
could see the line of mesas near Kum-kuduk. They turned to travel South 150 degrees East. That day they
reached the much sought for caravan route. They had been following the track of Lal Singh's cyclometer
wheel and it led them to reunion at his established camp site. (
map 32xx4as )
Stein and Lal Singh spent the following day studying the latter's plane table
survey and report of his journey around the northern and eastern sides of the
Lop sea. His route is shown on map 32. The men spent the day watering the
On 7 March the caravan with baggage from Miran arrived under Ibrahim Beg's
control. Turdi the Dak carrier also brought Stein's usual huge postal mail bag.
Among the letters was one from Sir George Macartney confirming that the British
Embassy in Peking had arranged for the Chinese government to rescind all orders
against Stein's work and replaced them with the opposite, to support him. Lal
Singh kept busy as usual with surveys into the high sand dunes.
Section II - The Valley of
On 10 March Stein released the camels from Miran and Charchan with their
owners. Tokhta Akhun and Niaz departed for home at that time. Turdi carried
Stein's mail back to Khotan. The next project was to survey the Lou-lan route
beyond the end of the Su-lo Ho eastward toward Tun-huang. (the route is on map35bas ) Stein wanted to survey the north - right
bank - of the river valley as far as Besh-toghrak. For
that purpose also Stein had sent Muhammad Yaqub from Miran to perform a level
up the river valley. Stein sent Lal Singh and Afraz-gul directly to the
northeast from Kum-kuduk. Stein took the heavy baggage by the caravan route as
far as Yantak-kuduk. He reported some results already in Serindia. At
that point Stein left the camels with baggage to continue along the caravan
route while he moved north northeast. He soon found that subsoil water was
plentiful. He continues his narrative with detailed descriptions of the
terrain, soil and vegetation.
On 13 March he was joined at the wells of Besh-toghrak by Lal Singh and
Afraz-gul thus completing a survey along both sides of the Su-lo Ho basin.
Meanwhile Muhammad Yaqub's level survey up the basin showed Stein that the fall
of the land, 4.2 feet per mile, was in agreement with his guess that the Su-lo
Ho had earlier flowed clear to the Tarim marshes. All this effort was expended
for geographical reasons, to show the relationship of the Su-lo Ho to the Tarim
at and around Kum-kuduk.
Section III - An Ancient Terminal
On 14 March, after a day's stop at Besh-toghrak, ( on map 35aas ). Stein and company set out toward the
Tun-huang section of the Han wall. He was sure that the ancient Chinese caravan route to Lou-lan
followed this route. This section contains Stein's description of his team's
survey work during the next few days up to March 16. He tasked Muhammad Yaqub
to run a level in order to prove that the Su-lo Ho could have reached the Tarim
as noted in previous section and then complete another level. Stein and Lal
Singh both made detailed plane table surveys. The section is only 6 pages long
but full of details.
Section IV - The Delta of the So-lo
Stein continues with his narrative and analysis of the topography of the lower
So-lo Ho. He
devoted considerable time and effort to this subject during both expeditions
and completed detailed surveys by repeated passages of himself, Afraz-gul and
Lal Singh. This is 4 more pages of details. The reason for all this effort is
that the question was one of the then current controversial topics being
discussed by the various explorers - like the role of Lou-lan - the course of
the Kuruk-darya - and the routes west from Tun-huang.
Section V - Transport Problems of the
Ancient Lou-lan Route
In this section Stein summarizes the information collected from his extensive
surveys to focus on what they reveal about the difficulties the Chinese must
have had in moving caravans between Tun-huang and Lou-lan. In this section he
marshals all his knowledge of military logistics as well as the evidence he has
personally found by crossing the area in question. He describes the conditions
along a practical route from east to west - or rather from southeast to
northwest. He believes that the supply and maintenance of Chinese garrisons at
Lou-lan and beyond was an enormous undertaking. He starts with the locations on
the Han wall that he found in 1907 - that is the magazine (depot) building T
xviii east of the ancient Yu-men. His discussion on the importance of this
building is in Serindia. He discusses the features of the first two day's march
from T xviii which lead along the wall to its extreme western end at T iva near
Toghrak-bulak. For this distance water and forage was easily available from the
Su-lo Ho. and the terrain was hard and passable. But for the next two days the
situation was worse. During this march the route crossed the 'Three Ridges
Sands'; and the ancient lake bed and ridges of drift sand before arriving near
present Besh-toghrak. But water was probably still available. From near
Besh-toghrak where there may have been another 'granary' the route would lead
along the northern side of the valley. Water was available from wells and there
was vegetation for grazing over a distance of 80 miles, or four days march. At
that point one would look for the 'Sha-hsi' well. At this point the
salt- encrusted bed of the ancient sea is near the foot of the cliffs that mark
the old shore line. Stein doubts that there was much if any vegetation beyond
that point. But to that point water and forage should have been not too
difficult to supply. But at that point difficulties became serious. There was
no water or vegetation over a distance of 125 miles. At best this might be
shortened by 12 miles by going over a longer stretch of the difficult salt
block covered terrain. So there was a problem for supply of water and forage.
Compared to today, however, at that time there was the agricultural station at
Lou-lan. While today the first oasis is 140 miles further northwest on the
Tarim. Moreover, at that ancient time the Kuruk-darya brought water and
vegetation eastward at least as far as L. I. Still the requirement was for men
and animals and carts to travel for at least 5 days without water, fuel or
forage. Stein states that this problem was greater than any modern military has
faced. He notes that there are no ancient Chinese reports or records describing
in detail how this problem was solved. Yet, he has shown that it was solved.
Thus he must offer 'conjectures.
Stein continues by noting that some system of depots and supply bases must have
been organized. He refers to one Chinese text and considers that it indicates
that supplies were sent east from Lou-lan to some intermediate point at which
the caravans could obtain them. But this point would be 50 miles east of
Lou-lan station L.I. And it was 25 more miles west from L.I. to the center of
Lou-lan at L.A. For the first two long marches from Besh-toghrak the water and
forage was probably supplied from that point. He notes also that while water
and forage could be obtained there the food and supplies for the men had to
come from Tun-huang, over 220 miles further east.
He compares the route to Lou-lan to the still current caravan route between
An-hsi and Hami, a road through the dry Pei-shan hills. For the first 9 of its
11 days no local supply can be obtained except for limited water. Yet it is
clear that the Chinese military does use this route and in 1877 conducted a
major military campaign along it. Then they brought at least 40,000 troops to
Hami by stages from Su-chou to An-hsi and then in small units to Hami. He notes
that from AD 73 on that was the main route - line of communication - to the
Tarim. Comparison of conditions on the An-hsi - Hami road with those from
Tun-huang to Lou-lan reveals great difference. Ten day's march across the
Pei-shan bring the traveler to the outskirts of a fertile oasis well fitted to
serve as a bridge head. On the Lou-lan route ten days of march would still
leave the traveler 125 miles away from the nearest water and 150 miles from the
Lou-lan cultivated area. Plus, Lou-lan could never provide as much supplies as
Hami. Stein suggests that only large caravans of camels could carry sufficient
water and forage. But camels cannot be used in the desert from May to August.
Stein refers to the R. E. Field Service Pocket-book by Col Scott-Moncrieff for
data on the supply problems when using horses, donkeys and ox drawn carts.
(This is similar to the use by authors today when estimating the campaigns of
Alexander the Great.)
The calculation is that more than half of the load possibly borne by a mule or
horse drawn cart is absorbed by the water and fodder alone needed for the
animals. And the remaining load would suffice for four passengers with minimum
baggage or for the water and rations required by four mounted men. The daily
ration for a horse, mule or ox is 8 gallons weighing about 80 pounds. Then add
8 pounds for fodder per day. Assume a two-horse drawn cart had to cover 125
miles in 5 days with four halts at which water and food had to be carried from
the start point. This would mean 640 pounds of water plus 120 pounds for the
cask and 64 pounds for fodder. This adds up to 824 pounds out of the possible
useful load of 1,344 pounds. Stein considers that use of camels in winter would
reduce the problem. And he imagines the possible use of camel drawn carts.
His final assessment; "It is hard to form an adequate conception of the
enormous scale of the supply and transport arrangements which such enterprises
along the Lou-lan route must have called for, or of the extent of human
suffering which these terrible desert wastes must have witnessed. But since the
substantial correctness of the contemporary record left by the "Herodotus
of China" is not subject to doubt, we must recognize in this conquest of
all the formidable difficulties of the desert route one more proof of that
wonderful power of organization which likewise enabled Chinese leaders to
triumph over nature's greatest obstacles in other regions and other
Chapter X - To Tun-huang
Section I - The Limes Line North-west of Tun-huang
In this section Stein describes his work along a section of the
Han wall from a point east of the section northwest
of Tun-huang that he explored in 1907 to a point more directly north of the
town by the southern edge of the Kara-nor lake. (Here are three special maps -
On 17 March Stein began again March to locate and excavate the towers and
sections of the Han Dynasty wall he had found in 1907 during his second
expedition. Then he had explored mostly the wall to its end northwest of
Tun-huang and a section to the north-east. His schedule then forced him to
leave gaps in his explorations. Since he already published in Serindia
much detail on the history of the wall he does not repeat that here. He
does includ some information provided by M. Maspero from translations of the
Chinese documents found in 1907. ( map35abs and
On 17 and 18 March he moved from Toghrak-bulak to the tower T xviii that he
identified already as a supply magazine for the garrisons on the wall. He made
new photos of towers Tiva built on a clay terrace overlooking the depression at
which the wall terminated. . Fig
the reed-covered marsh basin in which the Su-lo Ho terminates as seen south of
Tiva and shows the beginning of the wall.
tower T vii. Figures and
stacks of reed fascines near watch towers T xi and Txiii. Fig
tower T xiii and remains of rooms adjacent. (same tower is in fig 180 of
Serindia). Having excavated at the towers toward the western end of the
wall in 1907, Stein focused on the towers next to the Kara-nor lake further
east starting with tower T xiid. Starting with that tower Stein moved along the
southern shore of Khara-nor. Were ever the Chinese engineers could count on the
lake or a wide marsh-bed to prevent raiders they economized on wall building
and spaced the towers only to provide distant observation. The wall could be
traced for 10 miles. directly east from Tower xxii c to tower xxiii b.
About the middle of the length of the lake shore there was a prominent ridge
which reduced the width of the lake but also provided a high plateau. There
were three watch towers T xxii d to f within
a distance of 2 miles.
Tower Txxii d is on
eroded clay about 80 feet above the marsh to the west. The tower was 16 feet
square and even ruined remained 9 feet high. It was built of sun dried brocks
14 x 7.5 inches and 4 inches thick. There were thin layers os straw between
every fourth course of bricks. Adjacent to it were the remains of three rooms
one of which likely was used as a stove to heat the others. The tower provided
an usual mass of refuse from which Stein obtained Chinese documents including
one dated 16 December 47 AD. signed by Tsung-min and Shou-kuan, the former was
from theP'o-hu of the barrier, likely the name of a section of the wall
Another 3/4 mile east was another ridge, 90 feet high, on which was tower Txxii
c. It was 14 feet square and 9 feet high on which was the remains of a 6 foot
square room. The tower was built of sun-dried bricks 14 x 7 x 4 inches with
reed layers between every 5th course. It commanded a wide view across and along
the lake - towers T xxiiic and Txxiiie being visible. Again the refuse yielded
Chinese documents on wood. One document was instructions on the fire signals
Less than a mile to the north-east of T xxiie was a belt of erosion terraces
curving round from T xxiii with another tower Txxiif. This one was similar but
16 feet square at the base with a guard room at a height of 8 feet being 7 feet
square. .There was an entrance through a narrow passage in the south-east
corner. More refuse revealed more Chinese records including a part of a
calendar for the year 13 B.C. The tower T xxiif was on a terrace that provided
a full view of the lake shore. There was no wall for the next 5 miles eastward
along the lake shore. But the gap between T xxii f and T xxiii b was not
completely unguarded. There were two small towers, T xxiii and T xxiiia at the
end of the plateau adjacent to the lake..
On 20 - 22 March the group of towers T xxiii b to g and the wall connecting
stretches of ridges was found to be marshy.
On 21 March Stein moved camp to a spring-feed pool at Chien-ch'uan-tzu.
For this section Stein found a ancient road sunken 3 or 4 feet below the
adjacent ground level. It was obviously the result of much traffic passing by
over centuries. He discusses the history of the area and the various names that
Chien-ch'uan-tzu had in Chinese records over the centuries. Those records
describe the 'brackish spring' in practically the same condition as Stein found
About 2.5 miles ENE of Tower T xxiiia is another tower T xxiiib on an eroded
ridge 50 feet high. The line of the wall here is much eroded coming from T
xxiiic to the east. T xxiiib is 16 feet square and 13-14 feet high built of
bricks 14 x 7 x 4 inches. Fragments of tapestry and rug were found here. Almost
a mille and half due east of T xxiiib is a 90-foot high mesa on which is tower
Txxiiic. There was much pottery in the refuse. The wall here extended through
the depression and right up the side of the mesa. It was then extended round
the end of the mesa. T xxiii c is a well preserved tower built of layers of
stamped and 14.5 feet square at its base. It remained intact to a height of 15
feet. The eastern face of the tower showed foot-holes flanked on either side by
smaller holes intended to afford a grip for persons climbing to the top. On the
northern side there adjoined a room about 13 feet square with walls built of
brick and about 1 foot, 8 inches thick. The wall facing to the north stood to 8
feet height but that to the west was much broker and that to the east
practically destroyed from the prevailing east wind.
From the northwest corner of this tower the Han wall was 8 feet thick and
turned to the south-west toward a knoll as high as that of the tower but
steeper. The knoll was a natural defense on which the wall had a gap of 30
feet. Past this the wall was built of bricks of 14 x 7 x 4 inches and was 3
feet thick. The wall continued down the far side over a steep slope for 27 more
feet. This Stein noted was the only section of the wall that he found built of
brick masonry. Beyond that stretch the wall was again built of reed fascines
and clay toward the south-west for a further 90 feet, then turned
west-north-west toward tower T xxiiib.
The area by the tower on the high knoll was covered with much pottery debris
showing lengthy occupation. From the usual refuse Stein found Chinese records
on wood and various small objects including two bronze arrow heads. Tower T
xxiii c had a fragment of a calendar dated year 4 B.C.
A mile and half north of this tower was another, T xxiii e outside the line of
the wall itself. It guarded the lake shore which could not be seen from tower T
xxiii c. But boggy ground prevented Stein from reaching T xxiii e. That tower
served an important purpose in covering an angle section of wall. While the
guard on T xxiii c could not see the close area under the edge of the mesa he
could see very widely at a long distance to the north-east to where the
Su-lo-Ho joined by branches of the Tang-ho entered Khara-nor.
From tower T xxiii c the decayed wall went south-east towards a 100 foot high
mesa at a distance of a mile. Over this distance the wall was nothing but an
earthen mound. On that mesa was a completely decayed tower T xxiii d. There was
much pottery debris all around. These were later dated to Tang and Sung eras.
At tower T xxiii d the wall turned ENE for another mile to another clay ridge
on which at 35 feet above the adjacent terrain there was tower T xxiii f
was built of lumps of clay with layers of thin Toghrak branches between them.
It was 14 feet square at the base and remained to a height of 16 feet. The
ridge had been widened by a built up area of a clay platform. About 6 feet from
the tower was a well, 3 feet in diameter. Again, there was much refuse all
around from which Stein found more Chinese documents on wood and other various
From tower T xxiii f the wall turned to the south-east for another 7 miles, but
at less than a mile was tower T xxiii g on a 30 foot high terrace. . Station T
xxiii g was a tower with a chamber 7 feet square between thick walls of bricks
but broken down to only 5 feet high. The entrance was on the south-at corner.
The Han wall passed at a distance of ab out 20 feet to the north. Only a small
amount of refuse was found there. A mile further the wall passed another
isolated clay terrace about 15 feet high that formerly had held another tower,
but only pottery remained. The wall continued for another 2 miles across
depressions full of reeds. The wall was again found on another hell - with
Tower T xxiii h . This
one had a base of 16 feet square and was built of bricks 14 x 7 x 4 inches to a
height of 11 feet. The upper portion had a guard room 8 feet square. Over the
next mile Stein found three more watch towers T xxiii i, j, and k along the
wall. These were all built like T xxiii h. But Stein did not have time to
excavate those three towers on 22 March. From tower T xxiii l Stein had to move
rapidly onward due to shortage of water. But T xxiii l was a brick tower of the
same dimensions and appearance as the others. The guard room was 6 feet square.
entered by a narrow passage on the south . It was full of refuse to 4 feet deep. Stein recovered over 2
dozen Chinese documents on wood. Among these were instructions on signal
fires.. Two posts were named as Wei-hu Chih-k'ou. A refuse heap was also found
outside in which there were an additional 3 dozen records.
Stein continued on and reached decayed towers T xxiii m and n. each 3/4 mile
apart. The wall continued toward the south-east. Stein noted the towers were
not in a straight line, which he guessed might be due to enabling visibility of
each tower if they were not right behind each other. Beyond these the wall
disappeared in marshy terrain. Further on was tower T xxiii o on the end of a
low terrace. Its bricks remained to a height of 15 feet. Again refuse yielded
more Chinese documents. The wall continued over gravel past low mounds
remaining of towers T xxiii p and Txxiii r to the east to tower T xxiii s. But
Tower T xxiii q remained to a height of 12 feet. T xxiii s was 17 feet high and
both were of usual brick masonry - the former had reeds between every 2 courses
and the latter had them between every 5 courses.
Stein continued to Tower T xxiii t to the east-south-east about a mile and half
away. The direct route was blocked by a sheet of water and a bog.
On 23 March Stein moved toward the south toward Tun-huang.
On 24 March they continued next to marshes and found two more towers. One was
named Txxiii u and measured 29 feet square at the base. It was likely meant as
a refuge place for local farmers. It was old but not part of the Han wall
defenses. The wall probably continued east from T xxiii t to join the section
Stein had explored in 1907 to T xxx. The second, smaller, tower was even more
recent. Stein continued on to the high walls of Tun-huang.
Section II - Tun-huang and the 'Caves
of the Thousand Buddhas' revisited
Stein could not remain in Tun-huang more than a few days despite the urgent need for rest for his
team and camels because he had to hurry on to explore the desert north of
Su-chou before the summer heat made that too difficult. (His discoveries at the
Caves of the Thousand Buddhas are fully described in Serindia and
Ruins of Desert Cathay). ( map 38fas ) He found that his former friend, Wang Ta-lao-yeh, had been
replaced as magistrate at the Hsein-kuan Ya-men. The new official was much less
helpful. But the new military commandant was Shuang Ta-jen, who had been
Stein's helpful official at Chia-yu kuan in 1907. Stein's hired Chinese
interpreter, Li Ssu-yeh, proved more and more worthless.
Section III - By the Han Limes to
In this section Stein returns to the Han wall north of Tun-huang at a point he
had found it north-east of the town in 1907 and explores it eastward toward
On 8 April Stein departed Tun-huang toward An-hsi with the objective of
locating more of the Han wall than he had found in 1907. Meanwhile Lal Singh
was surveying toward the south into the Nan-shan eastward and then north to
An-hsi. Muhammed Yaqub was to survey down the Tang-ho to its confluence with
the Su-lo Ho and then turn east along the right (north) bank to An-hsi. The
main load of heavy baggage went directly to An-hsi while Stein moved north to
find the wall. He planned to start at tower T xxxv where he had stopped in
1907. (map38bcs shows wall near An-hsi with towers from xxvi to xxxviik and
Stein had to avoid treacherous terrain in several locations created by
underground drainage from the Nan-shan emerging to create salt-bogs.
On 10 April he continued hurrying to reach water. Finally he found the evidence
of the wall, in a straight line of low ridge only 3-4 feet high but clearly man
made. But they had to hurry on north to reach the Su-lo Ho by nightfall. They
reached to river to camp and replenish their water supply. The next morning
Stein turned back south-west to find the wall. They found the low but straight
mound when practically falling on it. It was about 4 feet high but 32 feet
wide. They found a mound that was all that remained of a tower designated T
xxxvii a. The construction here was inferior to that Stein found further west.
But there were no signs of ruined watch towers further west. Turning east Stein
followed the wall and found that it rose to a height of 9 and 12 feet in
alternating layers of clay and brush wood. There were sections that evidenced
efforts to burn the wall. A mile east they found another decayed mound
designated T xxxvii b. At another half mile they found another tower Txxxvii c
built of stamped clay. This one had a remaining square base 20 feet but most of
the northern and eastern faces were decayed. The remaining height reached 14
feet. The wall continued to the east with a height of 6 to 8 feet composed of
layers of brushwood. The next watch tower was reached after 2 more miles. This
one, T xxxvii d was only a mound, but the wall made its usual semi-circle to
the north around it at a distance of 50 yards. Further the wall continued in a
direction of S 97 degrees E. They found another decayed tower T xxxvii e and
then stopped for the night in order to find the camp that had been set up by
On 12 April Stein returned to the line of the wall, south of the river. They
found another tower, Txxxvii f that had a height of 18 feet and was 18 feet
square at the base. Its construction was unusual, being formed of an outer
casing of stamped clay with an interior of natural clay. Wooden beams had been
fixed into square holes cut into the natural clay and then the stamped clay had
been built up around these beams. There was evidence of signal fires having
been set on the top. No significant refuse was found. Its location puzzled
Stein. He could not find a clear trace to the south-west toward tower Txxxvii
e. Stein searched all around and finally found the Han wall again a mile and
half to the east-south-east of T xxxvii f. But at this location he found two
lines of wall about 90 yards apart. These then united a half mile to the
southeast at T xxxvii h, a decayed tower. Stein decided that the two segments
had been built to rectify the defense near a pre-existing dike that was evident
joining them. He considered this more evidence of the hasty and relatively poor
construction methods employed for this section of the wall.
From tower T xxxvii h the wall continued for a mile and half to T xxxvii i.
This tower was also nothing but a mound of clay. But 30 yards west of it and
within the wall here was the remains of a small cell in better condition. This
was 6 feet 3 inches square with masonry walls to a height of 2-3 feet. The
bricks were 9 x 6 x 4 inches. This appeared to be a small shrine.
at T xxxvii i the line of the wall took a sharp turn to the north-east. After
another mile there was another clay mound remanent of a tower. Then the line
continued in the same direction for another mile and quarter to another ruin T
xxxvii k. The wall rose to a height of 6 to 8 feet in places. From T xxxvii k
the wall turned due east toward a large tower some mile away. This one appeared
new, but might have been built from an earlier tower. Along this stretch the
wall was again different. It was two narrow walls of earth and fascines about 6
feet apart with the space between filled with loose earth. All this section had
been decayed by moisture. By this time Stein had reached cultivated ground at
Erh-kung. Further east they could find only a short section of decayed wall
amid the cultivated area so abandoned the search and proceeded to An-hsi.
Further on they found again the series of towers Txxxviii a to c
they had found near An-hsi in 1907. ( tower T xxxviii a, shown in and
described in Serindia.)
At this point Stein had completed his project to fill the gaps in his
exploration of the Han wall between its western end at An-hsi and along the
south bank of the Su-lo Ho. He had used An-hsi already twice and again made it
a storage depot for his antiques and baggage. Lal Singh also arrived from the
Nan-shan to the south. Two days later Muhammed Yaqub arrived from his survey
along the north bank of the Su-lo Ho. ( map 38cas
and map 38das
and map 38eas
show details of tower locations.)
Chapter XI - In Search of the
Limes to Su-chou
Section I - The Limes Line North of the Su-lo Ho
From 14 to 17 April Stein remained at An-hsi from April to tend to preparations
for further explorations and accomplish correspondence. He notes that An-hsi
despite its small development has been since the first century AD the key
location for departure from the Su-chou area to the northwest to Hami and
beyond to the Tarim. It is still the end of the main Chinese main highway along
that route. (See Serindia.) Stein's next objective was to find the Han
wall to the east of An-hsi and the north of Su-chou a section he did not have
time to explore in 1907. He wanted to find the Han wall all the way to the east
where it would generally meet with the much later Ming Great Wall. Further, he
wanted to explore the Etsin-gol river northward to Mongolia and see the lost
city at Kara-khoto, recently discovered by the Russian explorer, Colonel
Kozlov. And he had to accomplish both before the heat of summer would make such
(Stein was always eagerly competing with the many other European explorers who
were criss crossing Central Asia at that time.) ( map
shows Stein's route on the main road from An-hsi to
Hami he used during the 2nd expedition). ( map 40eas
shows Stein's route east as described in the
following several chapters.)
He left all possible baggage at the An-hsi Ya-men with Ibrahim Beg. At An-hsi
Stein obtained the services of Zahid Beg, a Mongol interpreter, who knew Turki
as the locals along the Etsin-gol were Mongol nomads. He sent Muhammed Yaqub on
directly by the main road with other baggage to Su-chou.
On 18 April Stein and the others departed An-hsi by crossing the Su-lo Ho to
follow the right bank between the river and the slopes of the Pei-shan
mountains. By the second day they reached the location where the Su-lo Ho
passes through a defile near Hsiao-wan. Stein had passed this same spot but on
the left bank in 1907. He noted then that at this location the Han wall crossed
the river and continued east on the north bank. In 1907 he had found sections
of the wall to the east along the north bank. Now he wanted to fill the gaps in
exploration. ( map 40cbs
and map 40cas
show the Su-lo Ho east of An-hsi.
The defile through which the Su-lo Ho forces its way west is between offshoots
of the Pei-shan to the north and outer ridges of the Nan-shan to the south. At
this point the main high road to the south of the river crosses a ridge 200
feet above the river. There, not surprisingly, it is guarded by two Chinese
fort towers. Now Stein found a tower T xl a near the river and almost opposite
the two modern towers on the other side. It is a small walled enclosure about 19.5 feet square inside
with a tower 8.5 feet high in the north-east corner. It is built of the usual
bricks 13 x 7 x 4 inches. The walls are strengthened by later construction.
About 2 more miles to the east he found another tower, T xl b. And a third
tower, T xl c was on the top of a high hill projecting from the Pei-shan
overlooking the river from the north. About half way to T xl b Stein found a
section of the wall itself at a bearing of S 100 degrees E near the defile. It
was about 34 feet wide and 8-9 feet high. Along the north side there was a
ditch from which the wall earth had been dug. Stein reached tower T xl b
short distance north of the wall and on higher ground. It was exactly like the
towers near Tun-huang. It was 20 feet square at the base on 26 feet high. It
was built of solid layers of stamped clay 6 inches thick. From there Stein
climbed the 300 foot high spur north of the defile. From there he had an
extensive view of the entire defile and the valley to the east as far as the
large fort at Bulungir. To the north he could see the Pei-shan hills and to the
west the gravel plain toward An-hsi. The tower T xl c on the summit was in a
perfect location to provide observation. It was built of bricks with layers of
tamarisk brushwood after each 3 courses. It was 23 feet square at base and
remained 13 feet high. The south and southwest sides had collapsed due to the
slope itself subsiding. On the east face one course of bricks was laid
vertically between two courses horizontally. There was an observation post 4
feet wide on top in which Stein found Han era Chinese documents.
Returning down the spur Stein and team found the wall again to the east of T xl
it was constructed of layers of gravelly earth and tamarisk brushwood. It
followed the slope of the hills that line the north side of the defile only 200
yards from the river bed. There the wall was commanded by the crest of the
hills that rose 100 to 150 feet above it. This indicates that this section of
wall was intended not for military defense but only to secure greater safety to
border police. Where the wall descended closer to the river its brushwood
layers had completely decayed. Searching back west Stein found the wall formed
by the gravel mound straight to tower xl a and 20 feet high.
On 20 April Stein crossed the river to the left bank. The Su-lo Ho was about 4
yards wide and 3 feet deep as the flood season had not yet begun. he measured
flow at 1,600 cubic feet per second. On the river's left bank below camp 122
they found the mouth of a small Nullah from the Wan-shan-tzu spur occupied by a
ruined temple. These spurs form an ideal site for defensive fortification,
especially before the Han wall was extended west to Tun-huang. This location
may have been a site for the "Jade Gate' prior to its move west and then
much later back to Chia-yu-kuan. Stein insists that the Wan-shan-tzu defile is
the ideal location for a defense and frontier gate. He notes that the Chinese
established a large garrison at Bulungir or Pu-lung-chi 10 miles to the east.
Returning to the line to the east on the right bank in less than a half mile
they found the wall again where it passes between two low ridges at the
south-eastern end of the spur previously visited and described. They could
trace it clearly for 120 yards as a double embankment. The southern wall was 24
feet wide and 10 feet high, the northern wall was less wide and only 5-6 feet
high. They were about 44 feet apart. Further on the wall was lost in soft loess
near the river bed. A few miles further on they found the remains of another
tower, Txli a on a wide terrace. Stein had seen this one in 1907 from the south
side of the river and visited by Lal Singh that year. It was built of layers of
stamped clay and was 20 feet square at base. The northern face had fallen and
the rest split in two. With this evidence Stein was assured that the Han wall
was built close to the Su-lo Ho. But no trace of the wall was found in the
thick vegetation. However, they did find another tower high above a bend in the
river. This was numbered Txli b. It was of layers of stamped clay and still
rose to 29 feet on a base of 20 feet square. By use of ropes and the foot holes
built into it one of the men reached the top where he found a wooden spoon and
some other fragments. There was plenty of Han pottery around the base. There
was an enclosure some 27 feet along the northern side and had joined the tower
on the west side.
Going another 2 miles east across bare clay they found another terrace which
had been converted into a tower Txli c by simply cutting and digging out the
clay on four sides. This also was 20 feet square and 21 feet high. There were
many Han potsherds all around. The 10 feet high wall itself passed around the
west, north and east sides of this natural clay tower at a distance of 32 - 36
On 21 April Stein and party continued east. Another tower, Txli d,
found only another mile and half east. This one of stamped clay was only 30
yards from the river bank. Wind erosion had cut down the northern foot. The
remaining tower still had a height of 28 feet. After another mile and half
across the wind-eroded clay plain between the river and ridge they found
another rocky ridge jutting out toward the river. On a small hillock some 30-40
feet high there were remains of another watch tower, but only the base of about
2-3 feet high remained. But the refuse did contain Han dated remains. After
another mile they came to tower T xli f. This was on the top of a detached hill
about 150 feet above the river level. This one had a loop-holed parapet but was
clearly Han era, although enlarged to the east, south and west by additional
masonry. In this the bricks were set vertically in Kan-su style The original
tower was solid with bricks of 15 x 10 x 4 inches and reed layers between the
courses at 3 feet 6 inch intervals. The base was 24 feet square. The added
later masonry of bricks 14 x 6.5 x 3 inches increased the size to 32 feet
square. The height was 32 feet. There were the usual foot-holes on the southern
face. The tower commands an excellent view along the river to east and west. A
careful search all around found rubbish about 20 feet down the slope and 2 feet
thick. Many remains were retrieved as shown in Stein's included listing. There
was a small enclosure (T xli g) at the foot of the hill. But the refuse in it
showed it was modern.
Further to the east Stein found remains of the wall again. But it soon
disappeared likely due to wind erosion. Again, further yet, they did find
another section of decayed wall further north a mile away from the river. And
there they also found another tower, T xli h. But there was not much in or
around it. ( map 42aas shows a section of the Han
wall west of Mao-mei.)
Section II - From Ch'iao-wan-ch'eng
Before even reaching tower T xli h Stein saw the ruined walls of a small
fortified town destroyed in the Tungan rebellion. He had noticed it in the
distance while passing on the high road south of the river. The first name for
it given him was Ch'iao-wan-ch'eng but later he also learned a name -
P'eng-chia'chuang. Either way it was an impressive fortification even though
now ruined. It had massive walls of stamped clay forming a rectangle of 380
yards by 135 yards. The southern wall was within 100 yards of the river. There
were large vaulted gates of brickwork in both south and north walls and both
were protected by square outworks also entered by gates.
There were broad straight streets between the gates. In town center they were
crossed by another main street. The many well built houses were nor roof-less.
Some monks lived in one house. There was also a decayed temple containing some
statues. The gates had ornamental pavilions above and these were also above the
corner towers. No locals could inform Stein about the date at which the
fortress had been built, but Stein guessed from its location far west of the
Ming wall that it was from the later Manchu era. This location enabled it to
guard the camel caravan route across the Pei-shan from Hami to the river. But
caravans using carts could not use that route but had to pass through An-hsi.
Outside the fortified town there remained the ruins of suburbs which had once
supported the garrison. It also had its wall, but one much inferior to the
fortress. Once the town was destroyed caravan route shifted to reach the river
further east. There was a large temple in ruins about 150 yards from the
northern gate. A further 100 yards north Stein found the remains of the Han
wall as it passed east to west in a straight line on the bare gravel to a
height of 3-4 feet. Next to it there were two parallel walls with a course
between them. A local guide explained to Stein that this was a training ground
for mounted soldiers practicing musket fire.
Further along the wall Stein found the point at which the caravan route to Hami
crossed, about 350 yards northwest of the fortress. There he saw a row of five
small stupas north of the wall and another group of three south of the wall.
Nearby was a larger modern memorial temple shrine occupied by Tibetan Lamas.
Stein remarks that he continually drew attention to the Chinese practice of
establishing shrines at places where trade routes crossed a wall. The spot were
the caravan route to Hami was exactly half way between towers Txli h and Txli
i, each a mile west or east respectively. At those distances, Stein guesses
that originally there would have been a tower with gate at the crossing point.
He presumed that the garrison for the fort was established in the early 18th
century. He presumed that this location was selected as the best to which
supplies could be delivered from Su-chou.
Stein analyzes the tactical and topographic reasons for placing the Han wall
and the later fortress north of the Su-lo Ho rather than south of it where one
might consider the river itself as part of the defense. He notes that the river
all along this stretch is easily fordable throughout and thus no barrier, but
from north of the line of the Han wall there is nothing but desert ridges of
the Pei-shan with no water for many miles. Thus the Chinese engineers wanted to
guard the river's water itself from enemy use rather than rely on the river as
a defense. In cases of both the Han wall and the Manchu fortress the
fortification system was part of the expansionary policy pushing Chinese power
west into Central Asia. Furthermore along this eastern part of the river basin
there was cultivated ground right up to the southern side of the river, while
further west there was no such ground but only desert with Tun-huang located
much further from the river.
On 22 April Stein resumed exploration to the east. Another mile on he found
another watch post with walls 3 feet thick of bricks 14 x 9 x 3 inches. More
relics were retrieved there. The low mound of the wall continued for more miles
eastward. There was another tower, Txli j at a mile and another, Txli jj after
a further mile. There and further eastward wind erosion had nearly destroyed
both wall and towers. After four miles they found another ruined remains of a
watch post T xli k. Another mile of mound (wall) was passed to reach tower, T
kli l. At this one Stein found several small brick stupas Beyond this tower the
wall was again lost on the hard clay surface. They found a tower, Txli m
further south of the expected line but determined that it was a much later
construction. Stein and team turned south to regain the river where it makes is
90 degree bend from north to west. There they found a tower T xli n of stamped
clay 12 feet square where the caravans halt at a village called Ma-ku-t'an.
From there they turned back northeast to reach another tower on the Han wall
line. This was T xli o. It
was constructed of stamped clay on a base 32 feet square. It was about 50 yards
north of the Han wall. Another tower T xli p was seen to the northeast on
Stein followed the wall line to the southeast from T xli o. They found sections
as high as 5-6 feet in three places a quarter mile apart. After 3 miles they
reached tower Txli r built of bricks 14 x 9 x 7 inches. Next to it was a guard
room only 6 feet square.
From there Stein continued on across soft ground liable for flooding. After
another mile and half they found a low rocky ridge on which were towers T xlii
a to d above the village, Shih-erh-tun. Stein had briefly visited the area in
1907. The team halted at the village for a rest day. Stein found the remains of
the Han wall continued along a succession of low narrow ridges of the Pei-shan
range overlooking the Su-lo river valley. About 300 yards along the ridges he
found w T xlii a. It was on a small rocky hill about a furlong south of the
rampart itself and at a height of 50 feet above the plain. It commanded like
towers T xlii b to d stretched eastwards. The tower was of stamped clay with
thin layers of reeds between each layer. It was in great decay and was only 10
For another 3/4 mile the wall went towards another rocky hill where they was a
completely decayed clay mount about 12 e high that was tower T xlii b. They
found lots of Han pottery there and a Wu-chu coin. From tower T xlii b which
was directly behind the remains of the wall, the wall was traced clearly for
another mile across stony ground but was then lost eastwards in a belt of sandy
soil. Layers of brushwood were exposed along the sides of the wall. The rampart
nowhere more than 4 - 5 feet high was about 14 feet wide on top. This section
of Han wall was guarded by two towers, T xlii c and T xlii d. Tower t xlii c
was made of stamped clay with reinforcing layers of reeds. It was 20 feet
square plus later additions and 14 feet high. The west face additions had
fallen away. And directly south of the wall was the remains of a small fort..
Stein makes special note of the two towers being so close together, deciding
that the view from T xlii d was blocked. Tower T xlii d had been repeatedly
repaired and added to. ( in
which it is behind the small fort). It was 33 feet square at base counting the
additions. It was 13 feet high. There were 5 small, new P'ao-tas along the
ridge to the east. On a rocky terrace some 30 feet below T xlii d was the small
fort enclosure. This
was 58 feet from east to west inside and 46 feet across. The walls were 18 feet
high excluding the parapet of bricks measuring 12 x 8 x 4 inches and probably
of later construction. A large section of the west wall was fallen. There were
no remains inside the fort. Its massive walls and location showed its
antiquity. . The
plan shows the route leading from the village toward T xli o
toward Hami passes close below this small fort and so does the route that leads
to Ch'iao-wan-ch'eng. The
village no doubt was the last cultivated spot before the caravan route to Hami
crossed the desert the location of a defensive fort to serve as a 'gate' is
logical. The fort was known as "Hsiao-fang-p'an or the small
protective camp." Moreover, there is a typical small shrine at the
junction of the two caravan routes near the wall.
Section III - Hua-hai-tzu and Its
On 24 April Stein continued east from Shih-erh-tun. Only a half mile on from
tower T xlii d the wall was lost on low ground in reeds. They continued in the
same direction towards T xlii e a tower visible from the village and came again
to the wall making a straight line of reed fascines. Near tower T xlii that was
about 2 miles further the wall rose to a height of 6 feet. From there on east
it remained clearly visible for 12 more miles. Across this distance the wall
was built of alternate layers of reed fascines and stamped clay. . Along that
stretch they found towers T xlii e to j all of similar original construction
and also showing later additions and repairs. The caravan route to Su-chou is
only a mile south of the wall. They were built of stamped clay and are from 22
to 28 feet square at the base. The heights vary between 18 and 25 feet
including brick parapets of later construction. Each tower now is on the
north-western or north-eastern corner of a later walled enclosure (tower T xlii
f . The
enclosures are also built of stamped clay 3.25 to 4 feet thick but less solid
than the towers. They form enclosures of 60 - 62 feet. They showed wind erosion
on their western sides. Little refuse was found in these but there were Han
potsherds plus later pottery.
Beyond T xlii j the wall could be seen for about another mile but then
disappeared in a depression. From there they turned southeast to reach a well.
Across the depression and on another high ridge they found a large conical
tower about 200 feet above the marsh. This was built of layers of clay
reinforced with large trunks of poplar and was 33 feet square at the base. At
12 feet the top had a small look-out platform. The tower had a very fine view
to the north and northwest.
On 25 April Stein sent the baggage on to Ying-p'an, a small walled oasis town,
and rode with Lal Singh and a few helpers to the north to search for the wall.
They found so much that it took two days to collect it all. They found the wall
across a wide plateau trending from west-north-west to east-south-east. In
parts it was eroded by wind. But there were many segments that still rose
nearly intact to height of 6-7 feet. It was built of alternate layers of
stamped clay and fascines, each 8 to 10 inches thick. The clay layers were very
hard. The wall was about 5 feet wide at the top. With the wall itself so well
preserved Stein was glad to find also other remains. The first tower, Txliii a
was decayed to a mound but had many potsherds around and from a refuse heap
they found wooden slips inscribed in Chinese. and much else. Two records
mentioned Chinese convicts exiled to frontier service.
Another half mile to the south-east they found the remains of a potter's kiln.
Another half mile on they found tower T xliii b. It was 11 feet high, built of bricks 15 x 8 x 5 inches and was
16 feet square at the base. It had been enlarged to 20 feet wide but the outer
brick casing had fallen away revealing the original white plaster on the wall.
The next two towers, T xliii c and d were only mounds. At T xliii d there was a
row of eight low mounds as spacing of 30 - 50 yards. Another mile further there
was another tower T xliii e now only a mound about 15 feet high and 22 yards
across of layers of clay and brushwood. Many Han era potsherds were found.
Another mile they came to the mound of tower T xliii f were the wall turned to
due east. The wall could be seen for another mile as a clay bank 3-4 feet high.
Then the line of the wall disappeared for 3 miles where they found a low mound
with tower T xliii g only 5 feet high. A coin, arrow head and a few other items
were found. They could not see any wall further on but found tower T xliii h on
a hill about 30 feet high and 50 feet in diameter. A section of wall was there.
From there then found tower T xliii j further east.
On 26 and 27 April Stein continued to excavate along the wall. At tower Txliii
h they found 16 more Chinese records and other items. Two documents were dated
39 BC and 13 AD. They continued to towers Txliii g and T xliii i about a mile
apart with T xliii h between them. But they could find no trace of the wall
itself by the two western ones. At T xliii i on a hill the wall appeared. Nine
more Chinese records were found. Beyond that tower the wall again was 5-6 feet
Another mile on a mound south of the wall was the remains of tower T xliii j.
Only the east wall remained. But 24 Chinese documents were recovered.
The next tower to the east, Txliii k, was a half mile further. Originally of
stamped clay and reeds, it had decayed to a mound But more relics and documents
were found there. Some dates were 56 BC, 21 BC, 86 BC, 48 BC, and 89 AD. At
that tower the wall was in good repair and its line changed direction to
east-north-east. The next tower, Txliii l was 24 feet high. It
also had been repaired and added to later. There was also a later enclosure.
This wall was formed of stamped clay 4 feet thick and on the south face
remained at 10- feet high. As so often the west face was completely eroded.
While built during the Han dynasty as part of the entire wall, this post
remained in use into medieval times. Northwest of T xliii l at a distance of 40
yards there were the foundations of four small buildings constructed against
the wall. They were 13 feet north to south and at intervals of 18 feet. Perhaps
these were shelters for the guards. Beyond tower T xliii l the wall was seen to
the east for half a mile.
Section IV- The Limes traced East of
On 28 April Stein left camp at the springs of Hsiao-ch'uan-tzu headed
east-south-east along the caravan route to Su-chou.
On 29 April Stein and Lal Singh searched again for the Han wall. Lal Singh
found it about 8 miles north and Stein found it with towers to the northeast.
The northernmost tower was T xliv a about 5 miles north of their camp. Then
they found the wall about 5 miles further north amid a mass of tamarisk cones. The
wall was only a low mound in drift sand but was 9 feet wide and still 4 feet
high. It was composed only of bundles of wood. The wall disappeared again
toward the east and west under sand dunes. It appeared again several more miles
to the west where it was 10 feet high and 6.5 feet wide
also it was built only of bundles of tamarisk wood. They found no more towers
along that section. Stein believed the construction was due to lack of water.
And subsequent guards would have been stationed at posts further south. To the
east they later managed to trace the wall for 45 miles clear to the Pei-ta-ho
river. Leaving the wall to return south for four miles they found another mound
about 70 feet long and 35 wide. There was an enclosure some 94 yards square
built of brushwood bundles (Txliv e, ) .
Returning to discuss tower Txliv a Stein notes again that it was the northern
most tower in this section and was distinct from the wall. It was 32 feet
square at the base and 14 feet high, built on a low plateau of gravel. It was
built of layers of clay 6-7 inches thick and brushwood layers between. There
was a small structure on the east side. There was much refuse to clear in which
they found 7 Chinese records on wood and other items including a coin and
bronze arrow head. Less than 3 miles to the south-east they found tower Txliv
was 21 feet high and 27 feet square. There was little wind erosion. The top of
the tower was covered with straw and refuse. Two Chinese tablets were found.
There were two dozen Chinese documents in the refuse. Dates found were for 62
and 112 AD. Names of individual officers and of units were given.
Continuing south east for another mile and half they found the decayed post
with foundations of a clay tower T xliv c about 16 feet square with a room next
to it. The southern wall alone of the room survived to a height of 3 feet of
bricks 15 x 8 x 4 inches. The refuse contained three Chinese records on wood
and a bronze arrow head. A fourth watch tower, T xliv d was seen about 3 miles
to the south east. It was of bricks in decent preservation 16 feet square and
14 feet high. There was a guard room on top with walls 3 feet high. Ten more
wood Chinese records were found.
They found another tower, T xliv f about 16 feet square and 8 feet high about
half way between T xliv c and the well at Ko-ta-ch'uan-tzu. It was built of
bricks 10 x 7 x 4 inches - that is different from the usual size and possibly
of later date. It was out of the straight line between towers T xliv a to d.
Clearly the four watch towers were built during the time of the Han wall but
they were located rather far from the line of the wall as found by Stein. Stein
guessed that either these were to protect the route between Su-chou and the
wall or were built so far back due to lack of water along the wall itself.
Stein was unable to trace the wall between the section described here and a
section further east by towers Txlv a to h near the Pei-t-ho river. At that
point Stein deputized Lal Singh to try to find traces of the wall further east.
On 1 May he had to rush south to accomplish business in Su-chou to prepare for
the expedition to Khara-khoto. Along this route he passed many great towers,
but these were a part of the Ming Great Wall complex. He passed through the
Great Wall at Huang-ts'ao-ying that he had visited in 1907. He illustrated a
Great Wall tower at
Although some 15 centuries separates this tower from the Han wall the
construction is practically identical. It still has the footholes for
climbers.Very interesting view of the great fort and the Ming wall on both
sides can be seen using Google Maps.
Chapter XII - From Su-chou to the
Limes of Mao-mei
Section I - The Limes along the Pei-ta-ho
Stein returned to his same quarters of 1907, the pavilions of the temple at
Chiu-ch'uan (Fountain of Wine" in Su-chou. Since the territory along the
Etsin-gol north of Mao-mei was controlled by local Mongol headmen, Stein
obtained permission and a request from the Tao-tai at Su-chou. Sir George
Macartney from Kashgar had obtained authority from the Chinese Foreign Office
in Peking for this which was sent to their Tao-t'ai, Chou Wu-hsueh, along with
a considerable quantity of silver drawn on Stein's account at Kashgar. He had
also to collect provisions for two months of work far in the desert along with
camel and donkey transport. He stayed in Su-chou for 6 days accomplishing all
this. While he was doing this Lal Singh accomplished more survey work toward
the Nan-shan closely west of Su-chou. ( map 43das
shows Su-chou and the Ming Great Wall north of town)
He notes that the Etsin-gol contains the combined water flow of the Pai-ta-Ho
and Kan-chou rivers that unite at Mao-mei and then flows north into the desert
on the Mongolian border. Mao-mei is the last significant oasis but the river
divided into several channels and enables enough vegetation in various places
along its course for Mongol nomads to establish grazing grounds. He had hired
"Malum', his Mongol interpreter from Kara-shahr already thinking ahead for
the need of his services.
In 1907 he had surveyed sections of the area between Su-chou and the Su-lo Ho.
This time he again split the caravan. Lal Singh went east at first to the
Kan-chou river and then survey north to where that river breaks through the
hills to Mao-mei. Stein went along the Pei-ta-ho river past Chin-t'a oasis in
order to search again for the Han wall.
On 10 May they departed Su-chou. Stein passed again through the Great Wall and
then crossed the hills as above Yeh-mao-wan where there was another line of
watch towers. ( map 43eas
shows the entire area south and south-east of
On 12 May north of Chin-t'a they passed through orchards and cultivated fields.
At that point Malum obtained information from locals about a 'old wall' in the
desert. Therefor Stein let Muhammad Yakub take the baggage on directly to
Mao-mei while he rode off to explore for this new site. He soon reached the
outer ridges of the Pei-shan hills. (map 42abs shows
some towers ) Sure enough, he found remains of the Han wall in a mound that
rose to 9 feet high and 8 feet width. In this section slabs of stone had been
laid as revetments along the earthen rampart. But the necessity to turn west to
Mao-mei prevented Stein from extensive exploration along the wall. He returned
to do this in September. He did follow the wall far enough east to find tower T
xlvi a on a rocky hill 30 yards south of the wall. This one was built of
stamped clay and brushwood 16 feet square and still 9 feet high. There was much
Han pottery all around. Continuing east for a mile he found tower T xlvi b on
another hill - it was but two heaps of stone. Another mile and half along the
mound that was the wall on another rocky ridge 60 feet high they found tower T
xlvi c composed of stone walls. Rubbish there reveled more Chinese documents
and a bronze arrow head. Tower T xlvi d was found on another ridge a mile
further on. This one was a decayed mound 18 by 7 feet and only 4 feet high. The
remaining bricks were 17 x 8 x 4 inches. Beyond these ridges the wall mound
only 4 to 8 feet high crossed an open gravel Sai to the south-east. Riding
another 3.5 miles Stein found traces of a watch post Txlvi f with much Han
pottery. Post Txlvi g was a mound another mile east. But it was adjoined by an
enclosure 57 feet by 79 feet., Those walls also were only mounds of layers of
gravel and brushwood. Further on was tower Txl h. This was built of bricks 13 x
7.5 x 4 inches with reeds between the courses. It had collapsed but had
measured 16 feet square. There also were quarters with brick walls 24 by 16
feet. 14 Chinese records and other items were recovered. One record listed men
on fatigue detail, and another specified what to do if bandits were
encountered. One was dated 69 AD. Continuing east Stein found three more towers
at mile intervals. T xlvi i was the same as Txlvi h, but had been repaired to a
height of 12 feet. The next tower, Txlvi k, was near the road between Chin-t'a
and Mao-mei and appeared modern. Stein saw that the Han wall trace continued
east but had to quit in order to reach Mao-mei by nightfall. However he at
least caught sight of another more modern tower T xlvi l further east. Stein
followed the cart track then for 12 miles and crossed the Kan-chou river to
reach the walls of Mao-mei.
Section II - Past the Mao-mei Oasis
and its Outposts
On 14 May Stein, with the assistance of the Hsien-kuan of Mao-mei, Mr Chou
Hua-nan, managed to obtain camels and men. With the heat increasing, Stein had
to move rapidly down the Etsin-gol. Stein notes his expectation that Mao-mei
would prove to have been a significant oasis supporting a section of the Han
lines. He presumed therefor that the wall would have crossed the Etsin-gol
north of the oasis. He remarks here again that the valley of the Etsin-gol must
have always been the major invasion route from Mongolia south into Kan-su for
the centuries prior to Ghengis Khan's invasion against the Tangut's in 1225 AD
. Long before the Mongols, the Yueh-chih, Hsiung-nu (Huns) and Uighur Turks and
others had passed that way. Thus, he reasoned, the oasis must have been
involved in defense of Kan-su since at least 121 - 115 BC. Malum extracted from
locals information about an ancient fort town was located toward the lower edge
of the oasis.
On 15 May Stein rode toward Suang-ch'eng-tzu across the cultivated area. (map42bbs
shows the area around Mao-mei) At that village he
camped. A mile north he found the ruined town. The walls of solid stamped clay
were 16 feet thick and enclosed an area 300 yards wide on north and south faces
and 400 on east and west. There were breaches on the east and south walls from
wind erosion. The walls were 25 feet high. The enclosure was empty but revealed
many Han relics and pottery but none from later eras. Another quarter mile
north he found another, smaller fort 96 feet square inside with 21 feet thick
walls to 30 feet in height. The 10 foot wide gate was protected by another
massive wall 17 feet thick . This was extended around another enclosure by a
lesser wall 10 feet thick and 12 feet high to the parapet. This outer wall had
bricks set vertically, showing its later construction period. Stein decided
that these forts had been garrisoned by troops to protect the Han wall.
On 16 May he returned and crossed the river to the west - left -bank. The river
was over a mile wide indicating its volume during annual floods but held only a
few pools in May. He could not find traces of the wall near the river, but
eventually found a tower T xlviii a on a spur 80 feet above the terrain and
there could detect the line of the wall. It was only a low mound on a bearing
of N. 40 degrees E. but turned at the tower to N. 58 degrees E. The tower was
broken to a height of only 9 feet but was 24 feet square at the base. The
bricks were 14 x 8 x 5 inches. The next tower on a steep ridge, Txlviii b, was
4 miles away. and
a solid tower of stamped clay 20 feet square at the base and tapering to 24
feet high. There the wall line changed direction to N. 83 degrees E. toward the
left bank of a river. The wall had crossed the river at that point but Stein
could not spare time then to look for it on the east - right - bank. He
accomplished that task later (map45aas
shows much of the Etsin-gol north of Mao-mai)
Stein saw that either in Han era or later the governments had expanded the
defenses at the Etsin-gol and made them more powerful than the simple Han wall.
about 4 miles further from T xlviii b he found a massive fort at T xlviii c. It
was similar to the "Jade Gate' at Txiv. It was 32 feet square inside walls
of stamped clay that were 20 feet thick and over 30 feet high. The relics found
were from ancient and medieval eras. A mile north of Ta-wan there was a large
walled enclosure about 220 yards square, T xlviii e ,
close to the left bank of the river. The walls were of stamped clay 18 feet
thick and 18 feet high with large square bastions in the corners. It there were
several ruined buildings inside including a temple. A mile southeast of this
ruin on the river right bank was T xlviii d called by the Mongols
Taralinginduruljin. This fort measured 250 by 185 feet with stamped clay walls
12 feet thick and 25 feet high in places. It had one large square bastion on
the southwest corner and a smaller one part way on the western wall. The gate
in the eastern wall was protected by two massive flanking towers and an outer
enclosure. Then there was a much larger enclosure 700 by 500 feet long to the
east and north (see diagram. Those walls were only 5-6 feet thick. They had
towers in the corner and along the length. Inside the inner fort were two small
buildings. There was much pottery, but difficult to date exactly. Stein guessed
the fort might date originally to pre-T'ang and then post Tibetan eras around
750 AD. From there Stein rode back cross the Etsin-gol to a Mongol encampment.
Two miles further on they found another watch tower , T xlviii f, on a low
ridge commanding a far view of the river plain. I was the same as the others,
20 feet square and tapering to 22 feet high. The bricks were 14 x 8 x 5 inches.
.The forts back at Ta-wan were visible from this tower. It must have been a
forward look out post. Stein noted that on the opposite side of the river he
could see another fort called Ulan-duruljin. He continued north across a bare
Section III - List of Antiques from
Ruins of Han Limes
This section is 12 pages of detailed descriptions of the articles found at each
of the towers and other ruins described in this chapter. Many of them are
illustrated in volume III of Innermostasia.
Chapter XIII - The Etsin-gol Delta
and the Ruins of Khara-khoto
Section I - The Lower Etsin-gol and its Terminal Basin
In this section Stein describes the geography, topography and vegetation along
his route from Mao-mei to Khara-khoto. He also includes some history and
discusses the presence of Mongol encampments that he visited. He mentions also
that he was preceded there by several Russian explorers, including Colonel
Kozlov, who removed a very large quantity of Buddhist art and artifacts to St.
For two days Stein rode north along the left and then the right banks of the
river. ( map 45bas
shows the continuation of the Etsin-gol north and
the location of Khara-khoto to the east.) There were patches of vegetation. At
one he found an isolated tower (fig 232). It was brick - 14 x 8 x 5 inches and
27 feet square at the base with later, additional masonry of vertically set
bricks 16 x 6 x 3 inches. They passed another tower near this one. Another one
was of similar size of bricks 14 x 8 x 5 inches, 16 feet square at base and 20
feet high. These were over 17 miles north of the towers Txlviii f and
Ulan-duruljin. Stein did not number these towers. They passed another tower
further north near Bayin-bogdo hill. They continued riding on 19 and 20 May through the jungle on
the river right bank. They passed three more unnumbered towers and reached
another fort called Bahan-durwuljin near the river bank. This was an enclosure
45 feet square inside walls 11 feet thick of bricks 18 x 9 x 7 inches. Crossing
the river again they found Mongol camps. They passed another, smaller, fort and
another watch tower of similar bricks.
On 21 to 23 May they covered 40 miles through more desolate terrain along the
river. As always Stein describes the terrain, topography and vegetation in
On 24 and 25 May Stein stopped at the camping grounds of the local Torgut
Mongol "Beili" at Dashoba whose assistance was critical for Stein to
hire men and camels to work at Khara-khoto. (
shows Khara-khoto and Dashoba) The Mongol nomads,
not used to any serious work were reluctant even with the generous wages Stein
offered. Stein had the foresight to bring strong influence from the Tao-t'ai at
Su-chou, which helped. Stein described and photographed the camp and the chief.
The Mongols described the serious situation that already had resulted from
greatly reduced summer floods. Lal Singh's survey of the river delta and its
two small terminal lakes showed the significant reduction in water levels. All
this shows on his maps 44 and 45. Stein compares the situation with that which
had occurred at Lou-lan centuries before.
Section II -
Khara-khoto and its Remains
On 26 May Stein had recruited a dozen local Mongols as laborers along with
extra camels. They all started for Khara-khoto. About 2.5 miles from the river
they found the first fort. It had walls 12 feet thick and 24 feet high
enclosing an area of 49 feet square. The solid bricks were 14 x 8 x 6 inches
laid with reeds between every sixth course. They found no remains. They
continued south east across the gravel on which was the usual potsherds
including fine glazed ware from Sung era. They next found the fort called
Aduna-kora. . It
was built of two walled enclosures one inside the other. Both were built of
stamped clay with the inner walls 20 feet thick and the outer walls 12 feet
thick. In both the north and west walls were very badly eroded due to wind and
rain. . The
inner fort was about 83 yards square and the outer walls enclosed an area 220
yards east to west by 180 yards north-south. The inner gate was in the center
of its southern wall. The outer gate was on its east side and protected by a
bastion and court some 40 feet square. Stein found no structural remains. Sung
era potsherds were found along with 5 Chinese copper coins, 4 from T'ang era
and one from 990-1004 AD. Since the ruin of Khara-khoto was another 10 miles
east, Stein concluded that this fort nearer the river was used as a convenient
way-station for caravans proceeding along the river that didn't want to stop at
Khara-khoto. Beyond the for the ground became more sandy. They sighted the high
walls of the city while still 8 miles or so from it. Stein remarks that this
sight was the most impressive that he had ever seen in the desert. Besides the
walls there was a very large stupa on a big bastion.
soon saw also a Muhammadan tomb near the southwest corner
Stein used this structure for storage and pitched his tent beside it. The men
used the inside of a bastion guarding the western town gate for shelter. There
was a huge sand drift against the west wall .
Stein describes how impressive the huge fortress city was in the middle of
total desolation. He
noted the dry bed of a river than skirted the walls. As usual, Stein promptly
sent Afraz-gul off to search the desert to the northeast. He had obtained
information from a local Mongol that there was another ruins in that direction.
He sent the remaining camels and ponies back to graze at Dzusulun-tsakha. The
camels then would be employed in relays bringing water. Two days later he sent
Lal Singh out again to survey up the dried-up branch of the river and the
across the Morun-gol.
On 27 May Stein began excavation within and adjacent to the walls. He quickly
found the Mongol nomads were not that interested in digging. This effort
required 8 days. Plus Afraz-gul reported that indeed he had found another ruin.
Stein records that the most striking aspect of Khara-khoto was the immense
walls. (And even now photos available on the Internet show that this is so).
The area measured 460 yards on the north side and 381 yards on the west. He
assessed it at twice the size of Lou-lan but less than a half of So-yang-ch'eng
near Ch'iao-tzu. The walls were constructed of stamped clay reinforces by a
wooden frame of big rafters. They were 38 feet thick at base with an inward
slope to a width of 12 feet at a height of 30 feet from the ground level. The
wall was wider at the north western corner where the large stupa was located.
In places a parapet one foot thick with loopholes rose for 5-6 feet. There were
ramps leading up to the top of the walls near the gates and at the north
western and south eastern corners. There were gates 18 feet wide in the western
and eastern walls protected by rectangular outworks built as massively as he
main walls. and
addition the walls were defended by large circular bastions at the four corners
and by rectangular bastions along the sides, - 4 each on the western and
eastern sides and 6 and 5 respectively on the north and south sides. The
largest bastions were 47 feet wide. The walls were also opened at two points by
passages of a later date. Stein did not establish the purpose for these late
breaches, but guessed they might have been the work of 'treasure seekers'. The
incessant wind had piled up sand against the outside of the western and
northern walls. Near a bastion the sand pile reached the height of the wall
where they destroyed the parapet and cut into the wall itself. There was some
similar result of wind driven sand on the inside of the eastern and southern
walls. Inside, Stein had difficulty tracing the foundations of former
buildings. He found that they had been built of stamped clay and wood but not
thick. He found relatively few artifacts considering the size of the ruin. But
among them were 230 Chinese and 57 Tibetan (Hsi-hsia and Tungut) documents.
There also were Uighur and Turkish documents.
Later analysis by Mr. Maspero found dates within the period of Yuan - Mongol-
dynasty between AD 1290 and 1366. (Chingis Khan conquered the Tangut kingdom in
1227 and the Mongol dynasty fell to the Ming dynasty in 1368.) Of 17 Chinese
coins found in or around the city 13 show dates between 1008 and 1161 and 3 are
from the T'ang era. The Sung dynasty and absence of Hsi-hsia coins indicated to
Stein that the majority of trade at the time involved the Sung.
Among other small items there were many of glazed pottery dated by experts to
Sung, Yuan and even Ming eras.
Stein considered that the western side of the town was occupied mostly by
shrines. Near the northern wall he found a cella 32 by 50 feet with walls 1.5
feet thick built of sun-dried bricks 12 x 5 x 2 inches set on edge, in which
there was considerable debris. His team cleared a pile of refuse 4 feet deep.
Inside he recognized the base for a large statue and several alcoves. He found
a coin there dated 1068-78 AD. There were many pieces of gilt stucco that
originally comprised the statue. He describes many pieces of stucco depicting
parts of human anatomy and costume. There were also fragments of silk banners
and faience including pieces of decoration from the former roof.
Stein found another shrine on a built up platform of stamped clay 82 by 63 feet
located on the main east west street. Stairs led up from the east. The shrine
chapel measured 12 by 17 feet. Inside he recovered various items in Chinese and
Tibetan. Some 70 yards to the south of K K i ii, was a row of three small
stupas and two shrines. In the south eastern corner Stein found remains of
several buildings. One of these revealed some Muhammadan documents. Dr. Laufer
assessed one fragment of paper as the 'oldest paper money now in existence'.
Stein next described the group of 4 stupas on the top of the walls in the north
west corner. and
largest one is on a platform 18 feet square and reaches a height of 30 feet but
originally was much higher. These stupas were built of bricks set on edge.
Other stupas remain only in total ruin. He found also models of stupas used as
Section III - Remains outside
Stein describes the results of his survey outside the city walls starting with
a group of stupas near the north-west corner. All had been broken into by
'treasure seekers'. The tallest rose to 20 feet height.
also had many small votive stupa models plus clay tablets showing Buddha seated
on a lotus. Stein notes an interesting fact, that the tablets here and relievos
and small statues were mass produced from moulds. Clearing the debris resulted
in finds of packets of Hsi-hsia texts, totaling over 100 pages plus about 50
Tibetan leaves, and many fragments. ( plate
18 ). Another mound, only 10 feet high, some 100 yards from the north east
corner contained more votive stupas and numerous leaves in Hsi-hsia or Chinese.
He considered another structure even more interesting - a ruin marked K K ii
from which Colonel Kozlov extracted his huge collection in 1908. (Now being put
on line in St. Petersburg). Its location was several miles west of the city
was a brick platform 28 feet square and 7 feet high made of bricks 12 x 5 x 3
inches. Stein remarked at the poor condition previous explorers had left the
scene insuring that many valuable items would be damaged or destroyed. He hoped
that this unnamed explorer had at least made photo of the structures and
drawings prior to their destruction. The dome that had covered and
proprotectedcted the contents had been destroyed. Shapir, a Mongol who had also
accompanied Col. Kozlov, explained to Stein what the site had looked like and
what its contents had been prior to the Russian's exploitation. Stein estimated
that the height of the missing dome would have sufficed to house large statues,
and indeed found a huge stucco head in the debris. Shapir stated that the
formerly intact shrine had no entrance. Despite the large quantity of materials
taken by Col. Kozlov, Stein found much more remaining. Stein comments that
photos accompanying Kozlov's article do show the extent of the structure before
it was destroyed. It was built on a three story base with projecting cornices
and an circular drum and above that a cylindrical dome. Stein further comments
that the material is receiving public discussion by Russian scholars. (Now
widely published by the IDP.)
Stein draws attention first to the numerous tests in Hsi-hsia he recovered -
over 1100 hundred of which 300 printed in Hsi-hsia and 59 in Chinese. This
sharply contrasts with the preponderance of Chinese texts throughout the town,
leading Stein to guess that Chinese was the language of daily life and commerce
while Tibetan was reserved for religious texts. He remarks further than most of
the texts, both written and printed, and in Chinese or Hsi-hsia were oblong
books more common in use during Sung times. There were fewer roll type texts
typical of T'ang era. He deplored having found many documents clearly cut in
two by hoe or pickax. But there were more small fragments apparently collected
as a religious custom.
There were also many artistic relics. Stein's laconic (ironic?) assessment
deserves quotation. "The remains of artistic or technical interest
recovered from the wreckage were, as the Descriptive List shows, numerous
enough. But after the account given above of the conditions in which they were
found, it cannot cause surprise that almost all have badly suffered, whether at
the time when the shrine was cleared - and demolished - or subsequently through
exposure. Nevertheless a brief review of them will be useful if only to show
how much it is to be hoped that the large and valuable haul of antiques which
Colonel Kozlov's expedition carried away from this ruin may yet obtain that
adequate study and publication which it deserves."
He continues to describe the mass of architectural remains - sculptural
fragments in stucco of images of all sizes from colossal statues to mere
figurines that must have originally filled the interior. These included many
large parts of bodies from fingers to heads. He found fragments of mail which
he ascribed to a statue of a Lokapala and also a demon's face. There were heads
of Buddhas and Bodhisattvas. There were animal figures as well. There were
numerous tempera fragments from wall and ceiling decorations. The remains of
silk paintings had suffered badly. He found many block printed designs of
Buddhist divinities and other sacred objects. He considered the examples of
block illustrations significant for the study of Chinese wood engraving.
There also were many pen and ink drawings. Again he hopes for publication by
the Russians. He completes his distressed description of 'waste' left behind by
the Russian with comments on silk fabrics.
He lacked the specifically datable materials that he hoped the Russians would
produce. But meanwhile presumes the remains dated from the quarter century
prior to the Mongol conquest of 1226. At the other extreme, he notes that the
Hsi-hsia script was only invented by the Tangut ruler Li Yuan-hao about 1032.
Next he shifts to discuss other ruined structures such as a small brick
platform 12 feet square and a low mound south of K.K. that measured 19.5 by
21.5 feet with walls 1.5 feet thick of bricks 12 x 6 x 3 inches. There was a
domed building and
30 yards to the south west of the south-west corner bastion . Its original
height was nearly 23 feet but part of its dome had collapsed. The domed hall
measured 18.5 feet square, and there was a vaulted porch. As the walls rise
they shift to an octagon and then to a circular drum and high dome. This
building is entirely western and clearly a Muhammadan tomb of 'gumbaz'. The
design is Saracenic. The interior was empty. Stein notes that Marco Polo wrote
that Islam was practiced in the region. Whatever its date, Stein believes it
must be the oldest Muhammadan building remaining in western most China.
Another 3/4 mile north east of the north eastern corner there was a small
stupa, less "Tibetan' looking than the others. . Tis
base measured 11 feet square and had a tapering dome 15 feet high. Stein found
the usual evidence of 'treasure seekers'. Further east there was another mound
that contained a temple. The walls, 19 by 22 feet only 2 feet high at most
nevertheless had tempera frescos some of which Stein managed to extract. There
was the usual image platform 12 feet 3 inches by 11 feet 6 inches. But only
fragments of the former statues remained.
Section IV - The Remains of a Rural
Settlement and Marco Polo's 'City of Etzina'
Afraz-gul returned from his successful reconnaissance in two days. He found
remains of buildings stretched over a distance of 6 miles. Coins indicated it
dated to the same era.
On 1 June Stein ventured to the location while noticing canals along the way.
He found many large and small dwellings amid extensive evidence of ancient
cultivation. He decided it had been a Chinese area. He found 8 Sung era coins
and one T'ang coin. Stein visited and described more and more ruins of homes of
agricultural people over a wide area. There were many fragments of fine pottery
which he describes and illustrates in volume III. Another 17 coins were found
throughout the area, 11 of Sung (1017-22 to 1086-94). He again quotes Marco
Polo to prove that Khara-khoto is the Etezina that Marco found. He continues
with mention that Sir Henry Yule the great publisher, editor and commentator of
Marco's book, had also looked for Yetzina mentioned in 1226 along the river
north of Kan-chou. Stein goes to rather great lengths over several pages to
shoot down any objections to this identification of Khara-khoto.
Stein then turns to the issue of when Khara-khoto was abandoned and why.
Several pages of discussion follow the basic conclusion reached once again is
lack of water due to shifting of the river bed. In addition the volume of water
itself has been reduced since medieval times.
On 5 June, with rapidly increasing heat beating on everyone, Stein was glad to
finish at Khara-khoto. Lal Singh returned from his survey of the Etsin gol
north. ( map44aas
) Stein sent Muhammad Yaqub with the camels
north-east to graze over the summer in a higher region, the Kungurche hills.
shows this area next to Mongolia. ) He also hoped
Muhammad Yaqub could survey to the north but Mongol border guards prevented
that. Stein moved the camp back to Tsondul on the Ikhe-gol where the Mongols
had their main grazing area and from which he had obtained their reluctant
assistance. Hasan Akhun had been in charge of the camels as usual. This clever
fellow discovered another walled enclosure 200 yards square in the desert as he
was bringing the camels south to Mao-mei. Muhammad Yaqub duly entered its plan
on his plane table. Hasan also brought Stein about 200 Tibetan documents and 20
leaves of Mongolian script, plus other remains. .
Section V - List of Antiques from
Khara-khoto and Neighboring Sites
This section is 40 pages of detailed descriptions of the relics Stein brought
back from in and around Khara-khoto.
Chapter XIV - To Kan-chou and the
Section I - A Desert Route towards Kan-chou
On 8 June Stein departed Dzusulun-tsakha southward toward Kan-chou with the
objective of getting into the high Nan-shan again for summer work. The Torgut
Mongols were not interested in the extra work of carrying his baggage through
the hot weather, especially on the route Stein wanted to use through unexplored
desert. Of course he didn't want to use the route along the Etsin-gol he had
followed when moving north, but to explore the desert east of the river to look
for Han wall sections there. The Mongols finally agreed for high wages to carry
his team as far south as Kao-t'ai but only at night thus defeating his
surveying opportunities. From there they agreed since the area traversed would
be higher and hopefully cooler to move during day light, but only early in
morning and late afternoon. Stein was always facing his need to obtain local
workers for whatever projects he had in mind.
From 8 to 15 June they went by night marches with their intendant straying of
animals, loss of loads and such. And even the halts during the day were subject
to intense heat and powerful sand storms. They passed the ruined forts at
Arun-takhai and Tara-lingin examined during the previous march north. He noted
tower T xlviii b where the Han wall hits the left bank of the river. He found a
series of 5 ruined watch towers in a line northeast from the right bank near
the fort Ulan-duruljin on a rising ridge He visited the southern most of these,
Tower T xlviii g, built of bricks 14 x 8 x 6 inches with layers of reeds
between each third course. He lacked more time to search for a connection of
the wall toward Mao-mei nor further toward the north-east as the Mongols who
owned the camels and ponies would not hear of it.
After another 5 nights and over 100 miles they reached Kao-t'ai by the Kan-chou
river. Stein duly marked the route on his maps 42, 45 and 46. They stopped to
rest for 2 days in Kao-t'ai where Stein as usual pitched his tent by a large
temple outside the city gate. From there he had to hire a new transport team
using carts. Stein went directly along the main road to Kan-chou while Lal
Singh took a longer survey route through the foothills along the right bank of
the Kan-chou river. Lal Singh found a ruined town at Lo-t'o-ch'eng with walls
of stamped clay 10 feet thick and of length over a mile east to west and 1,430
yards from north to south. (Once a sizable town). The nearly empty interior was
divided by a cross wall. There were bastions along the walls and at the
corners. The gates in the eastern and northern walls were protected by
outworks. He found several coins from 1644-63 and 1851-62.
On 23 and 24 June Stein marched on from Kao-t'ai to Kan-chou over ground he
described in Serindia. He stayed at the same quarters he had used in
1907, a temple outside the south-western corner of the city walls. (
shows Kan-chou and its oasis to the south.)
(Unfortunately we don't have illustrations of the walls of Kan-chou or Su-chou)
He remained in the town for 10 days making arrangements for his survey again
into the Nan-shan. As usual he accomplished a lot of correspondence with India
and Europe and remarkably was busy checking proofs of Serindia as well,
into which he inserted his recent observations. He hoped to extend the survey
accomplished in 1907 described in Serindia and shown on the maps
published then. He faced the same difficulty as before, namely the strong
reluctance of the Chinese to venture into those forbidding mountains full of
all sorts of physical and human dangers. While the peasants were fearful of
their own danger, the officials were fearful of their responsibilities if
something should go wrong. Luckily for Stein one of his old friends from former
expeditions in 1907 showed up in the person of General Ch'ai Hung-shan the new
military governor, T'i-t'ai, of Kan-chou. Obviously that set matters straight
for Stein. Stein also enjoyed visiting with the Belgian missionaries and
learned much from Father Van Eecke.
Section II - To Nan-kou-ch'eng and
the Eastern Headwaters of the Kan-chou River
Stein mentions his desire to visit the walled village,
Nan-kou-ch'eng and the cave shrines at Ma-ti'ssu due to
information obtained from Professor L. de Loczy. This town was on his proposed
route into the Richthofen range. He sent Lal Singh ahead by a different route
to survey the area. On the way Stein noted that the extensive cultivation all
around Kan-chou was dependent on irrigation from the river rather than any
precipitation. Yet further south and at a higher elevation he observed that the
cultivation had no irrigation but was enabled by rain and snow. He was acutely
aware of the change in total climatic conditions. The village was located in
front of the steep rise of the snow-covered Richthofen range, which provided a
lovely backdrop. He compared the scene to the Italian Alps. In the picturesque
village Stein visited the noted Lung-chiao-miao temple that contained a huge
clay statue of a seated Buddha and standing Bodhisattva and 8 bronze Arhats.
the village Stein went to Ma-ti'ssu monastery through bucolic scenery that he describes with special vigor and
pleasure. There he met with dozens of red-robed Lamas. He delighted in viewing
the elaborate polychromatic decoration of the Tibetan Buddhist style.
were also several stupas and extensive cave shrines and large Chinese style
temples to be admired. He climbed to visit the three stories of caves and
described them in detail. He was given dated records that established the
foundation in 1427 and 1565.
On 8 July Lal Singh arrived, having determined the location where the Kan-chou
river exited the mountains.
On 9 July they moved on south-eastwards to Hung-shui where Stein again found
quarters in a temple garden. ( map46bas
shows Hung-shui and Stein't routes into the
village had its usual Chinese military garrison whose commander appeared eager
to obey the order from General Ch'ai Hung-shan but the owners of the required
pones were not so accommodating. Stein spent three days arguing while the
military commander did his part by threatening, before 17 ponies with their
owners were obtained at a price double the official rate. At that Stein sent
Li-Ssu-yeh (who anyway was useless) and Naik Shamsuddin (whose excellent
engineering skills would not be needed) back with other spare workers to
Kan-chou. (It turned out that with Stein's serious accident Shamsuddin would
have been of very much assistance). ( map43gas
shows part of Stein's routes in the mountains.)
On 13 July Stein, Lal Singh and Afraz-gul set out into the high mountains. They
passed another garrison at Yung-k'ou soon reaching 8,800 feet elevation. The
next day they crossed the water shed of the eastern tributary of the Kan-chou.
Above Pien-tung-k'ou they passed another military outpost. Evidently there was
sufficient reason to consider a danger from Tangut raids. They soon passed
Tangut camps with their large herds of yaks and flocks of sheep. They crossed
the O'po'ling'tzu pass at 12,680 feet. Stein was back in his favorite element -
snow covered mountains. He could see far across the plateau of the O'po'ho
tributary of the Kan-chou. Over the next range would be the Ta-t'ung-ho, a
tributary of the Huang-ho and thus drainage to the Pacific Ocean. There was yet
another fort at O-po, elevation 11,000 feet. At that point Stein faced a
passive mutiny of the Chinese pony-men (similar to those he experienced in the
Kun-lun). Stein managed to get the fort's commander to prevail on the
'mutineers', after an additional compensation in silver. The officer assigned a
NCO to accompany Stein and prevent further difficulties. The section ends with
description of unusual gifts Stein received from the monks at Ma-ti-ssu.
Section III - Return from the
Nan-shan to Mao-mei
In this short section Stein describes his rapid march out of the mountains and
then north and west from Kan-chou back to Mao-mei.
On 16 July he started back. After riding only 16 miles his Badakhshi stallion
reared and plunged and then fell backwards right on top of him. The damage to
his left thigh was serious, muscles were torn, but at least no bone was broken.
He was suffering greatly from the severe injury to his leg. He gamely tried to
walk with support from his surveyors but finally had to quit. He had to be
carried on to the camp by the men in a camp chair. He had to remain immobile in
camp for two weeks but Lal Singh managed to complete quite a bit of the survey
through the mountains until the Chinese workers simply refused to go further.
Lal Singh continued surveying for weeks into August. But Stein finally had to
be carried down on a pony litter to Kan-chou. He stayed in Kan-chou for 10 more
days under care of Belgian missionaries, Fathers Van Eecke and De Smidt. It was
in Kan-chou that he learned about the outbreak of World War One. Lal Singh
reported in and promptly returned to the Nan-shan to survey westward to
complete their coverage of the Kan-chou drainage.
On 22 August Stein departed Kan-chou to Mao-mei traveling along the right bank
of the Kan-chou river. ( map43cas
shows part of this area.) He insisted on riding,
which did no good for his damaged leg. He passed the decayed part of the Ming
Great Wall. That section of the wall continues to the famous fortress at
Chia-yu-kuan. Stein noted that the decay seen in the medieval Ming wall was a
marked contrast to the relative preservation of the much older Han wall that
was built in far worse terrain and climatic conditions. Passing Cheng-i he
noted many towers and defensive positions on hills and ridges. At Mao-mei he
was reunited with both Lal Singh and Muhammad Yaqub..
Chapter XV - Across the Pei-shan to
Section I - Through the Desert Ranges of the Pei-shan
In this section Stein describes his bold trek across the dry Pei-shan hills.
descriptions are vivid, both in detail about the flora and terrain and about
the interaction of the participants. The reader always can get a look into
Stein himself and his thoughts and methods by reading his narratives carefully.
There were a few dramatic moments but nothing really significant by way of
discoveries of ancient civilizations in this bleak region.
Always looking for fresh regions to explore and potential ancient caravan
routes to uncover, Stein decided to avoid the previous route from An-hsi to
Hami. Instead he would move north west from Mao-mei across the desolate desert
like Pei-shan on a direction more or less parallel with the An-shi - Hami route
but significantly further north east of it. His route would strike for the
eastern end of the Karlik-tagh, itself an eastern extension of the T'ien-shan.
From there he would cross and move westward on the northern rather than
southern slope of the mountain chain to visit Barkul and Guchen before
recrossing the mountains back south to Turfan. During his earlier stops at
Mao-mei Stein undertook preliminary oral reconnaissance by quizzing all the
locals he could. The response was meager. His effort to hire Mongol guides
failed. Instead he managed to hire two Chinese who claimed to have traveled
along the caravan routes and were willing to return north. (See below for the
results.) There would be no resources along the route so Stein made special
care about food and water. To
reduce the party and also spare the weak Chinese 'interpreter' Li Ssu-yeh,
Stein sent him with Naik Shamsuddin and two Turki followers back to Su-chou and
from there to An-hsi to join Ibrahim Beg and from there to take all the heavy
baggage by the main road from there to Turfan for a rendezvous in October.
Always trying to expand the survey, Stein proposed that when possible he with
Muhammad Yaqub and Afraz-gul would go along one route while Lal-Singh took
another more or less parallel.
On 2 September they set out but the flooded river required half a day to cross.
The next day they crossed the line previously found of the Han wall west of
Mao-mei. They found two additional towers, Txlv a and T xlv b. made of stamped
clay and layers of tamarisk. The following day they found towers T xlv c , d,
and e on the same line. Stein was suffering from his damaged leg while confined
to a liter on a camel. At the last village he persuaded a gracious home owner
to part with some wood to construct a liter that could be carried on a pony.
This sufficed to carry him for the next two months across desert and mountains.
On 5 September they finally got away from civilization into the Pei-shan desert
hills. Immediately they found yet two more towers, T xlv f and g. Soon after,
the 'guide' lost his way for the first time. From then on Stein relied more on
his own plane table.
On 6 September they continued to ascend a valley and passed two more wells.
They reached the crest at 5,700 feet where they lost any remains of a former
track. But soon they did see the line generated by Lal Singh's cyclometer and
were able to follow it to an old cart track. Eventually they found Lal Singh,
intrepid as always and perfectly at home with the discomforts of desert and
mountain, in his camp near old coal mines.
Next they followed the cart track over 25 miles to a well at Nan-ch'uan.
On 9 September they crossed another ridge at 5,300 feet.
They rode on during 10 and 11 September before stopping at another spring,
Lo-t'o-ch'uan, on 12 September.
On 13 September they rode northwest. By that time the 'guides' were thoroughly
On 14 September they continued and crossed another water shed at 7,000 feet. in
the Ma-tsun-shan range.
On 15 September at Tsagan-gulu they met a group of 6 Mongol families
with their sheep and cattle enroute to their winter grazing area. These nomads
provided much information on the topography of the region, clearing up the lack
of information from the Chinese 'guides'. While Stein's party remained over
night a large camel caravan crossed their past taking rice and flour from
An-hsi north east into Mongolia at Uliassutai. So Stein knew he was at least
near routes still in use. Stein obtained milk and sheep from the Mongols
On 16 September he continued. They turned due west in an open valley. They had
to halt for the night and await the camels. Next morning the 'guides' found the
oasis at Ming-shui only a mile away. ( map40bas
shows the route at Ming-shui.)
were in a valley but at 6,660 feet elevation. By the well they found a mud
enclosure and on a low hill the remains of a Chinese shrine. They rested a day.
in-shui was on the route between Su-chou via Shih-erh-tun to Hami. Therefore
Stein detached Muhammad Yaqub with one of the Chinese 'guides' to survey that
route past Hami to Tash-bulak and Khotun-tam. From Hami he then was to continue
to Shona-nor and on to the Turfan where he would meet Stein's party. Stein felt
that Muhammad would be safe following a well marked caravan route.
Section II - Across the Easternmost
On 20 September Stein started from Ming-shui across the eastern end of the
T'ien-shan beyond the Karlik-tagh toward Bai.
Continuing on 21 September over the northern edge of the Pei-shan for 25 miles
they found a spring in a depression 2 miles long.
On 22 September they continued north westward to reach the foothills of the
Moving up through narrow valleys they found fresh water springs. But at this
point the hired 'guide' admitted he had lost the way and turned frantic from
the dishonor. Searching all around for signs of a caravan route they spent much
of a day fruitlessly going in circles. Fortunately they at least were at a fine
spring from which to replenish the water supply. The 'guide' ran off during the
night, but was found the next day. ( map37bas
shows much of the route.)
On 25 September Stein continued northwest, crossing the ridge at 6,000 feet.
way down was not so easy, full of boulders between rock walls. Stein was
concerned that the narrow gorge might not have an exit further down through
which the camels could pass. And Stein was still being carried in a liter.
But on 26 September Ismail, one of the Yarkandi team, climbed to the top of the
side ridge and reported he could see their objective in the distance. On
exiting the gorge indeed the high peaks of the Karlik-tagh were visible in the
distance, covered with snow. After
a few more miles riding they succeeded in reaching Bai, a hamlet of only about
50 families. ( map37aas
shows the final approach to Bai.) This phase of the
trek had taken 4 weeks across a desert. The oasis was irrigated solely by water
exiting from springs fed from higher up in the mountains. Afraz-gul made a
sketch of a small fort at Bai meant to protect the area from intruders from
Mongolia to the north east. The
fort measured about 320 feet on each face and had gates in the middle of the
north and south walls. There were rectangular bastions at each corner, The
walls were of stamped clay and 8 feet thick. Stein concluded from this
excursion that the more northern route just explored could never have supported
significant military operations. He decided that the well worn route between
An-hsi and Hami was the route - and it still is today as a road map of Chinese
Turkestan will show.
Section III - Past the Karlik-tagh
On 28 September Stein left Bai moving west along the northern slopes of the
T'ien-shan to Pei-ting near Guchen. ( map37has
shows part of this route.)
there he crossed through snow to the Turfan depression. His plan, again, was to
accomplish archeological work in the desert during the winter. West of the pass
that connects Barkul and Hami Stein followed the high road connecting Hami with
the souther part of Dzungaria. From Bai they went to Aturuk north of the
Karlik-tagh. North of the range the terrain was much more favored by water from
streams off the glaciers on top of the Karlik-tagh. The result was extensive
grazing of sheep and ponies. They stopped at the village of a local headman. They then turned
west and the following night again stopped with a local Turki chieftain. These
nomads were in process of moving their flocks down and north to lower
elevations for the winter. ( map34cas
On 30 September they crossed the watershed between lakes Tur-kol and Barkul.
Stein noted that the range was heavily forested between elevations of about
7,500 and 9,000 feet in contrast to the southern side on which there are no
trees or vegetation, just desert. They entered the valley of lake Barkul, which
extends 100 miles east to west, and 30 miles wide north-south.
On 2 October Stein noted a distinct ethnic change in the local population as he
rode from Shor-bulak to Narin-kur. He had passed through the territory of nomad
Turkis with their felt tents, who were subjects of the Wang of Hami. They
entered a region cultivated by Chinese settlers, but no flocks of sheep or
ponies, all the way to Barkul. Stein notes that both valleys, Tur-kol and
Barkul are formed because on its eastern extension the T'ien-shan splits from a
single massive range into two with the southern one rising in places to 14,000
feet and the northern ridge reaching 11,000 feet in places. Barkul valley
itself is at or above 5,000 feet and Tur-kol is over 6,000 feet. Stein notes
the attractiveness of these well watered and lush basins to nomads over the
centuries from which they could and did mount raids and campaigns across the
mountains into the Tarim and eastward into Kan-su. For this reason occupation
of Hami and conduct of trade along the southern side of the T'ien-shan required
also control of the Barkul valley.
Section IV - Historical Relations
between Barkul and Hami
Stein concludes volume I with another digression into history. He again refers
to the Former Han Archives for information about the first Chinese expeditions
into Central Asia. In this case he notes that these archives do not mention a
connection between these two towns on either side of the low mountain range.
Chinese expansion during two centuries after Han Wu-ti in 121 BC required and
resulted in their pushing the Hsiun-nu (Huns) out of the Tarim basin and north
across the T'ien-shan mountains. But the Huns remained in power on that steppe.
From there they continued to mount raids southward. The Chinese could not
secure Hami. This threat was the reason the Han wall north of Tun-huang was
built. The Chinese lost control of the routes due to strong Hun attacks and the
weakness of the last of the Former Han dynasty. But in 73 AD the new Later Han
Emperor Ming again asserted power and sent general Pan Ch'ao back into the
Tarim basin. This time the Chinese were able to stage operations from Hami. At
this time then there are more references in the Later Han archives to Barkul.
The record of Chinese discussion and decisions in 72 AD include recommendations
to defeat the Hun tribes around Barkul. The offensive in 73 followed this
advice. Three Chinese armies advanced on different routes into the steppe
across the Pei-shan and north of the Karlik-tagh. A garrison was then
established at Hami. In 74 AD another Chinese army marched from Tun-huang
attacked to Lake Barkul and then on west into Turfan and Guchen. If the Huns
could be pushed back and prevented from advancing through the area east of the
T'ien-shan then the small oasis societies throughout the Tarim (that is around
the Taklamakan) could be controlled. Chinese control over their critical line
of communications through Hami continued to be threatened. In 77 AD the Chinese
had to withdraw from Hami which fell to the Huns again. It was regained in 90
AD when the Huns were defeated but was lost again in 107 AD. A Chinese campaign
to retake Hami in 119 was soundly defeated leaving the entire route east to
Kan-su open. The records contain an address from 123 AD by Chang Tang, Governor
of Tun-huang to the Emperor urging another offensive. In 123 another army was
sent to establish a military colony at Luchun in the Turfan basin. From there
the Chinese again defeated the Hu-yen king in 126 AD. But by 131 AD they again
were struggling to maintain a garrison at Hami. A Hun campaign in 135 again
defeated a Chinese force. The Chinese were then victorious in 137. The Huns
came right back and won in 151 AD. From then on internal disintegration of the
Later Han dynastic power inside China led to their losses or even much effort
against the Huns around the T'ien-shan.
Stein continues with comparisons between the topographic descriptions in the
Later Han records and his own observations of the terrain in and north and
south of the eastern part of the T'ien-shan.
Stein next discusses Chinese efforts in the area during Sui and especially
T'ang dynasties. By that time the Huns had long gone but there were powerful
nomads, Turki tribes and Uighurs. In 630 AD the Chinese defeated the Northern
Turks, In 640 the Chinese mounted an offensive from Hami and conquered Turfan.
They won a final victory in 658-9. Chinese control lasted several centuries
until the era of Tibetan expansion and then Uighur return. The Chinese did not
regain Hami until the 17th century. Then lost it during the Tungun rebellion
and finally regained the whole Tarim basin after 1874.
Stein continues his narration and
analysis in Volume II which describes his
work in Guchen and then south across the T'ien-shan, months of work in Turfan
basin, cemeteries uncovered at Astana, further exploration back east across the
Kuruk-tagh and along the Konche-darya back to Lop, then the journeyed from
Korla to Kucha and on to Kashgar. From there Stein sent Lal Singh with the
large train of camels directly to India. Stein then entered the Russian Pamirs
and spent weeks exploring southward to the Oxus river Wakan corridor which he
could not cross into Afghanistan. Then he went to Samarkand and from there
crossed into Iran - Khorasan and Sistan exploring southeast Persia and finally
back to India.