The Peoples of the West

from the Weilue 魏略

by Yu Huan 魚豢
A Third Century Chinese Account
Composed between 239 and 265 CE
Quoted in zhuan 30 of the Sanguozhi
Published in 429 CE
Draft English translation
by
John E. Hill
© September, 2004
 
“I was not born knowledgeable,
I am devoted to antiquity and am quick to seek knowledge.”
Kong Qiu 孔丘 (Confucius).
Lunyu, 7, 19.
 

Contents
Preface
Acknowledgements
Introduction
About this Translation
About Fonts and Characters
About the Text
Translator’s Notes
About the Dating and Background of the Text
Background Reading
About Measurements and Administrative Divisions

The Text
Section 1. The Di Tribes
Section 2. The Zilu Tribes
Section 3. The Qiang Tribes
Section 4. The three main overland routes to the Western Regions
Section 5. The Southern Route
Section 6. The Kingdom of Linni (Lumbini)
Section 7. The Kingdom of Juli (the ‘Eastern Division’ of the Kushan Empire)
Section 8. The Kingdom of Panyue (Pandya)
Section 9. The Central Route
Section 10. Previous Misconceptions
Section 11. Da Qin (Roman territory/Rome)
Section 12. Products of Da Qin (Roman territory)
– Product List
Section 13. The Sea Route to Da Qin (Roman territory)
Section 14. Roman Dependencies
Section 15. The Kingdom of Zesan (Azania)
Section 16. The Kingdom of Lüfen (Leukê Komê or modern Al Wajh)
Section 17. The Kingdom of Qielan (Wadi Sirhan)
Section 18. The Kingdom of Xiandu (‘Aynūnah = Leukos Limên?)
Section 19. The Kingdom of Sifu (Petra)
Section 20. The Kingdom of Yuluo (Karak)
Section 21. The Kingdom of Siluo
Section 22. The Far West
Section 23. The New Route of the North
Section 24. The Kingdom of Northern Wuyi (Khujand)
Section 25. The Kingdoms of Liu, Yan, and Yancai (the Alans)
Section 26. The Kingdom of Hude
Section 27. The Kingdom of Jiankun (Khirgiz)
Section 28. The Kingdom of Dingling
Section 29. The Kingdom of Duanren (‘Pygmies’)
Section 30. The Author’s Comments

Abbreviations and Bibliography

Appendices
A. The Main Caravan Routes.
B. The territories of Haixi, Haibei and Haidong.
C. The “Great Seas” and the “Western Sea.”
D. Sea Silk.
E. Wild Silks.
F. Maritime Commerce and Shipping during the Han Period.
G. The Water Cisterns on the Route between Petra and Wadi Sirhan.
H. The Identification of the City of Angu with Ancient Gerrha and Modern Thaj.
I. The Spread of Ideas and Religions along the Trade Routes.
J. Climate and other Changes along the Silk Routes.
K. The Identification of Jibin as Kapisha-Gandhāra.
L. The Introduction of Silk Cultivation to Khotan in the 1st Century CE.
M. The Canals and Roads from the Red Sea to the Nile.
N. Kanishka’s Hostage in History and Legend.

Preface
There are several important Chinese texts relating to the early development of the “Silk
Routes” that have not been translated into English previously. They include the ‘Chapter on
the Western Regions,’ several biographies of the Chinese generals who expanded Chinese
power to the west, from the Hou Hanshu (covering the period from 25-220 CE); and an
important 3rd century geographical and historical text called the Weilue. My aim is to
complete up-to-date, fully annotated translations of these texts, and make them easily
available to all.
•  My first translation, a draft annotated version of The Western Regions According to the Hou
Hanshu, appeared on the Silk Road Seattle website in May of 2002. The response from
readers was beyond any expectations. I was inundated with a wealth of new material,
comments and suggestions from scholars in more than 30 countries. This led to a thorough
revision and updating of my original draft which had been available on this site since July
2003. It is presently being revised once more before being published in book form.
•  I am hoping, by the publication of this draft version of my annotated translation of the
Weilue on the same site to elicit a similar response. This should lead to a more accurate and
useful final document. I intend to add the biographies of several of the Chinese generals who
were instrumental in opening the main “Silk Routes” to the west at a later date.
•  I have included a number of lengthy quotations in the notes because I believe they are of
importance, well-stated, and of particular interest. I have also included some notes sent to me
privately in emails. I have tried to get permission from all these correspondents but have not
heard back from all of them yet. If authors have any objections to my use of their material I
hope they will contact me and I will gladly make adjustments.
•  Some of the longer notes, which may be of more general interest, I have included as
Appendices at the end of the document so they can be easily found and accessed.
•  Publishing this translation will, I hope, make the Weilue of interest to the general reader,
while retaining enough information in the notes to make it useful for specialists. I have tried to
keep the text itself as uncluttered as possible so that it may be comfortably read as a whole.
•  I hope the work will help rekindle interest in the extensive early contacts and exchanges
between East and West, and how they shaped the development of our cultures and our
technologies. May it inspire readers to search for answers to some of the many remaining
mysteries in the text. I also hope that you will share at least some of the great pleasure I have
received while studying this fascinating work.
•  Your comments, corrections, criticisms or suggestions are most welcome and will be taken
into account in future revisions and, if used, full credit will be given. Please contact me directly
– not through the Silk Road Seattle website. I will try to answer any questions – my contact
details are:

John E. Hill
PO Box 467
Cooktown, Qld.,
AUSTRALIA 4895
Email: wynhill@bigpond.com

Acknowledgements
Many, many people have helped to encourage me and with the research that went into this
new draft version of the Weilue.
          I would like to acknowledge the constant help and encouragement I have received from
Jo Wynter, my beloved partner of almost 30 years. Without her untiring patience and constant
help, editing and suggestions, none of my historical work would have ever come to fruition.
          My special thanks go to Professor Fida Hassnain, who originally fired my enthusiasm in
early Indian history, took me to visit many Kushan sites, helped me explore the archaeological
collections held in Kashmir, while conveying the knowledge he had gained through his long
career; Professor Daniel Waugh for his encouragement and for making it possible to publish
my work on the Silk Road Seattle website; and to Professor Victor M. Mair for ongoing advice
and assistance. Others who have generously provided valuable help include:
Nettie K. Adams, Dr. Farhad Assar, Dr. Thomas Bartlett, Professor Christopher I.
Beckwith, Dr. Craig Benjamin, Professor Alison Betts, Professor E. Bruce Brooks,
Professor Felix Chami, Dr. T. Matthew Ciolek, Professor Joe Cribb, Chris M. Dorn’eich,
Professor Étienne de la Vaissière, Aayko Eyma, Professor Richard N. Frye, Professor
M. Gawlikowski, Dan Gibson, Gaston Giulliani, Dr. Irene L. Good, Dr. David T. Graf,
Paul Greenhall, Chris Hopkins, Professor Karl Jettmar, Agnes Korn, Henriette Kress,
Whalen Lai, Valérie Lefebvre-Aladwi, Renzo Lucherini, Pavel Lurje, Thomas K. Mallon-
McCorgray, Dr. Michael Macdonald, Professor Daniel L. McKinley, Raoul Mclaughlin,
Felicitas Maeder, Josef Maier, Samir Masri; Professor Irina Merzliakova, John Moffett,
Khademi Nadooshan, Professor Giorgio Nebbia, Mark Passehl, Prof. E. G.
Pulleyblank, Lic. Paola Raffetta, Fr. Yves Raguin, S.J., Professor Nader Rastegar,
Professor Donald Redford, Joachim K. Rennstich, Janet Rizvi, Peter Rowland, Dr.
Edmund Ryden, Orit Shamir, Michael Schimmelpfennig, Professor Steven Sidebotham,
Professor Nicholas Sims-Williams, Sören Stark, Dr. Sebastian Stride, Dr. Mehmet
Tezcan, Professor E. H. Uphill, Dr. Chunyun WANG, Professor Donald B. Wagner,
Antonia Willis, Richard Wong, and the Editorial staff of Shen-Nong of Integrated
Chinese Medicine Holdings Ltd., in Hong Kong (www.icm.com.hk).
Almost inevitably I will have forgotten some who have helped me along the way – and a few
have asked not to be named. I extend my heartfelt gratitude to you as well.
          Thank you all so very much. I will be forever in your debt. This is your work as well as
mine – there is no way I could have written it without the kind support and assistance I
received from you. I hope you will find it worthy and will be pleased with it. I look forward to
any suggestions you may have to improve it in the future.
 

Introduction
About this Translation
•  This translation has been made from the text of the Weilue as contained in the five volume
Sanguozhi published by the New China Bookstore Publishing House, Beijing, 1975, zhuan
30: 858-863. I have also checked critical passages against other, earlier, editions.
•  As in the modern world, the borders of countries were constantly changing. In addition,
many of the peoples mentioned in the Weilue were nomadic, and regularly moved from place
to place. Peoples of different ethnic backgrounds and even languages were sometimes
grouped together under a common name as “confederations” or “tribes,” which at times can
be quite confusing for the reader.
•  I have divided the text into numbered and headed sections for clarity and ease of use.
Modern place-name equivalents are in rounded brackets after the Chinese names. Many are
well-established and widely accepted. Tentative identifications are indicated with a question
mark, and the evidence is discussed in the notes.
          The modern place-names adopted here sometimes only refer to the general location of
the ancient sites mentioned in the text. Usually I have only given the name of the nearest
modern town, or the main town of an oasis. For example, the oasis of Kashgar (Shule)
contained several towns, as it still does, and these are sometimes referred to individually.
Literal translations of place-names and products have been put within single inverted
commas, such as: ‘Eastern Division.’
          Where needed for clarity, I have added comments and notes in square brackets, eg:
“the three heavenly bodies [the sun, moon, and stars].” Identifications that remain uncertain
are indicated by a question mark.
•  Because the older Wade-Giles system of Romanizing Chinese is still commonly used, you
will find it employed in many of the quotes given in my notes. I have, therefore, included the
Wade-Giles equivalents after the Pinyin for many names and terms in italics and within
square brackets.
•  Chinese characters are omitted from the translation itself to make it easier to read. The
characters for all major place-names and terms are included in the appropriate notes.
•  For those wishing to check the reconstructed ancient pronunciations I highly recommend
first checking Edwin Pulleyblank’s masterful Lexicon of Reconstructed Pronunciation in Early
Middle Chinese, Late Middle Chinese, and Early Mandarin (1991), UBC Press, Vancouver. It
is arranged alphabetically according to the Pinyin renderings and also contains references to
the entries in Karlgren’s earlier, but still useful, Grammata Serica (Recensa). Pulleyblank’s
listing of the reference numbers to the characters in Karlgren’s work probably provides the
most convenient way of finding particular characters in that work, which is notoriously difficult
to access. Some additional reconstructions have been added from the Grand dictionnaire
Ricci de la langue chinoise. 7 volumes. Instituts Ricci (Paris – Taipei). Desclée de Brouwer.
2001, which is abbreviated as GR in the entries.
•  Note that the EMC reconstructions are only reliable back to the time when the Qieyun
dictionary was completed in 601 CE, as Pulleyblank himself notes on p. 20 of his Lexicon.
This means there was a gap of over 300 years between the composition of the Weilue and
the best phonetic reconstructions for the characters we have at our disposal.
          Although the reconstructions back to the late 6th century are frequently useful in
helping to identify place-names, there were undoubtedly significant phonetic changes
between the 3rd century and 6th century CE. Also, there were likely significant differences in
pronunciation between the Chinese soldiers and settlers on the northwest frontiers and the
inhabitants of the capital.

•  The reconstructions of “Archaic Chinese” according to Karlgren’s Grammata Serica, in

which he attempts to reconstruct pronunciations back to the Chou period (up to circa 220
BCE), are also included. These reconstructions of “Archaic Chinese” are indicated by the use
of a preceding asterisk: *. Sometimes Karlgren’s attempts to provide these earlier
reconstructions are of value, but they should be regarded with caution. They are followed by
Karlgren’s “Ancient Chinese,” which are his reconstructions for the period equivalent to
Pulleyblank’s EMC.
          As entries are often difficult to find in Karlgren’s book, I have included his numbering
system preceded by “K”, so a typical entry from his work will look like this: K. 139s *g’ân / γân.
•  Quotes from French authors have been translated into English and usually adapted (e.g. by
changing the French E.F.E.O. romanizations into Pinyin, leaving out unnecessary footnotes
and some of the Chinese characters).
•  Some notes from my earlier translation of the Chapter on the Western Regions According
to the Hou Hanshu (abbreviated here as CWR) are duplicated here to save the reader the
tedious task of checking the notes from one work to the other.
 
About Fonts and Characters
•  I have avoided using Chinese characters in the Text itself. Chinese characters as found in
the Notes will require the enabling of “Unicode” Chinese characters. Most modern computer
programs come with the ability to display Chinese characters but some readers may have to
install or “enable” them in their browser to be able to read the Notes and Appendices properly.
•  Some of the rarer characters may not be available in the fonts on your computer. In this
case, if you are using Windows 2000 or XP format, try to obtain the very extensive “Simsun
(Founder Extended)” font which is available on the (unfortunately very expensive) Microsoft
Office Proofing Tools CD.
          Those with Office XP 2002 or later should be able to install it from your Office XP CD
using the information available at: http://www.i18nwithvb.com/surrogate_ime/background.htm
•  For Mac users I recommend checking the following website for information on Chinese
fonts for the various OS X operating systems:
http://www.yale.edu/chinesemac/pages/os_x.html
  
•  For Linux users I recommend checking the following website for information on Chinese
fonts: http://seba.studentenweb.org/thesis/linux.php     
•  For the balance of the document (including Chinese romanizations and the quotes from
various other languages), I have used “Gentium” font throughout as it elegant, and contains
the greatest number of diacritics (or accents) needed. Best of all, is free and available now in
Linux, Mac and PC formats.
If this font is not on already your computer it can be easily downloaded and installed
from this site, or: information on the Gentium font can be found at:
http://scripts.sil.org/cms/scripts/page.php?site_id=nrsi&item_id=Gentium&_sc=1
Gentium fonts may be downloaded from: http://scripts.sil.org/cms/scripts/page.php?
site_id=nrsi&item_id=Gentium_download&_sc=1 Fortunately, it is free, and not a large
file, and so should not take too long to download and install in your “Fonts” folder.
•  The use of “Gentium” has allowed me accurately represent almost all the diacritics
employed in the quotes. Don’t forget, if you can’t find exactly the diacritic you need already in
the list of characters, you can combine marks from the “Combining Diacritical Marks” heading,
by typing the character you wish to mark and then going to “Symbol” and adding the
appropriate diacritical mark. This works reasonably well in most cases.
          In spite of this, some readers may still experience difficulties. However, I believe that
most of the notes will be meaningful even if the odd character is missed. If readers continue
to have problems, or wish to discuss some point, contact the author directly at:
wynhill@bigpond.com . But please – do not contact the Webmaster.
 
About The Text.
•  The Weilue fills in many gaps in our knowledge of the extensive international contacts and
trade networks at this early period. The kingdom of Wei was one of the ‘Three Kingdoms’
(Wei, Wu and Shu) formed after the disintegration of the Han dynasty in 220 CE. Its capital
remained at Luoyang [Loyang], which had also been the capital of the Later Han dynasty.
“The Wei controlled the north and north-west, being based essentially on the Yellow
River valley with their capital at Loyang ; the Wu in the south and south-east ruled the
Yangtze valley and the two Kuang provinces, while the Shu were based on the
Szechuan basin in the east, but also commanded the hills of Kweichow and part of
Yunnan.” Needham (1978), p. 40.
•  The original text of the Weilue, or “Brief Account of the Wei Dynasty,” by Yu Huan has,
sadly, been lost. Fortunately, this chapter on the xirong, or ‘Peoples of the West’, was quoted
in as extensive footnote to the Sanguozhi by Pei Songzhi, first published in CE 429.
•  Unfortunately, Yu Huan does not mention his sources in the text that has survived. Some of
this new data undoubtedly came to China via traders from Da Qin. Land communications with
the West apparently continued relatively uninterrupted to the northern state of Wei after the
fall of the Han dynasty.
          Wei was the northernmost of the three kingdoms the Han empire had split into and it
controlled access to Dunhuang and the main trade routes to the west. It was also, of course,
the state that Yu Huan lived in. An entry of the “Chronicle of the Three Kingdoms” for the year
222 CE informs us:
“Second month (Mar. 1-29). The Kings of Shan-shan 鄯善, Ch’iu-tsŭ 龜玆 (Kutcha),
and Yü-tien 于闐 (Khotan) each dispatched an envoy to offer tribute. [The Emperor
said in an edict: “‘The Hsi-jung 西戎 came to submit to his arrangements,’ and ‘The Ti-
ch’iang 氐羌 came to seek acknowledgment,’ – these lines are sung in praise in the
Shih 詩 and the Shu 書. Now, the distant barbarian tribes of the Western Regions have
all come to offer submission and allegiance to us. Envoys shall be sent to soothe
them.”] From this time on, the Western Regions maintained contact with China, and the
wu-chi chiao-yü 戊己校尉 was appointed.” Fang (1965), p. 98. [Note: this latter title is
usually transcribed as: “wu- (or mao-)chi hsiao-wei” – which translates as the “Maoji
Commandant” – see Hucker No. 2456 and TWR notes 1.5 and 22.5. Also, “the Ti-
ch’iang” above would be better rendered as “the Ti and the Ch’iang.”] 

•  There is information in the Weilue about the maritime routes to the Roman Empire and it is
quite possible that some, or all, of the new information on the Roman Empire and Parthia
came from foreign sailors. One record of information obtained from such sources (which may
have been available to Yu Huan) is recorded in the Liangshu:
“During the 5th year of the Huangwu period of the reign of Sun Quan [= CE 226] a
merchant of Da Qin, whose name was Qin Lun came to Jiaozhi [Tongking]; the prefect
[taishou] of Jiaozhi, Wu Miao, sent him to Sun Quan [the Wu emperor], who asked him
for a report on his native country and its people. Qinlun prepared a statement and
replied. At the time Zhuke [nephew to Zhuke Liang, alias Kun Ming] chastised Dan
Yang [= Jiang Nan] and they had caught blackish coloured dwarfs. When Qin Lun saw
them he said that in Da Qin these men were rarely seen. Sun Quan then sent male
and female dwarfs, ten of each, in charge of an officer, Liu Xian of Huiji [a district in
Zhejiang], to accompany Qin Lun. Liu Xian died on the road, whereupon Qin Lun
 returned direct to his native country.” Adapted from Hirth (1885), pp. 47-48.
Note: In the name of the “Roman” merchant Qin Lun above, Qin, as is standard Chinese
practice with foreign names, stands for ‘from Da Qin’ or the Roman Empire. The old
pronunciation of the personal name Lun 論 is reconstructed as: K. 470b *li̯ wən / li̯ uĕn or *lwən
/ luən; EMC lwən or lwənh. This, as Renzo Lucherini has kindly pointed out in a private
communication of 23 May, 2004, may well have represented the Greek name of Leon.
•  Yu Huan apparently never left China, but he collected a large amount of information on the
countries to the west of China including Parthia, India, and the Roman Empire, and the
various routes to them. Some of this information had reached China well before Yu Huan’s
time, and can also be found in the sections dealing with the ‘Western Regions’ of the Shiji, the
Hanshu, and the Hou Hanshu.
•  In spite of this repetition of earlier (and sometimes fanciful) information, the Weilue contains
much new, unique, and generally trustworthy material. Most of it dates from the late second
and early third centuries CE. It is this new information that makes the Weilue such a valuable
source. Most of the new information appears to have come from the Later Han dynasty,
before China was to a large extent cut off from the West by civil wars and unrest along its
borders during the late 2nd century CE.
“The time of the Han dynasty, especially the Later Han, was one of the relatively
important scientific periods in Chinese history. There were great advances in
astronomy, improvements in the calendar, an outstanding development in the earth
sciences, and foundations laid for methods of classifying plants and animals; alchemy
flourished, and the first book ever written on the subject appeared (A.D. 142). A
sceptical and rationalist way of thinking developed, particularly about A.D. 80 in the
hands of Wang Chhung [Wang Chong].., while there were two Han princes who also
took part in active intellectual life. One, Tê of Ho-Chien, was a scholar and bibliophile
who preserved the important ‘Artificer’s Record’ section of the Chou Li (Records of the
Rites of Chou), the other was the almost legendary Liu An of Huai-Nan, who gave his
name to the Huai Nan Tzu, a compendium on all the science of the day and one of the
most important monuments of ancient Chinese scientific thought. Indeed, bibliography
as a whole received great stimulus, for the Han period marked the first systematic
development of book lists; compiled by experts in astronomy, medicine, military
science, history, magic and divination, these were incorporated into the Han histories
and list some 700 works written on wooden or bamboo tablets, and on silk. Buddhism
also entered China in Later Han times and the first sutras were translated into Chinese
at the capital, Loyang.
          In technology the Han age was marked by the invention and spread of the use of
paper, by numerous developments in ceramics such as the first glazes and the
introduction of a material that was the forerunner of porcelain, by advances in
architectural techniques such as making decorated bricks and tiles, and by raising the
level of textile technology to a stage not approached by Iran or Europe until centuries
later. A large number of natural products new to China were also imported: alfalfa and
the grape-vine from the west, oranges, lemons, betel nuts and lychees from the south
and south-west. From the west also came improved breeds of horses, and from
Khotan, possibly from Burma too, jade arrived in large quantities. Perhaps the greatest
achievement of the Han people in nautical technology was the cardinal invention of the
axial rudder at least as early as the first century A.D.
          Towards the end of Later Han times, palace revolutions became increasingly
 frequent, and in 184 a farming crisis led to a peasant revolt guided, in this case, by the
‘Yellow Turban’ secret society. Although the revolt was suppressed, it left some of the
army generals in positions of great power, and by 220 the central government found
itself ineffective. The country became divided, and for the next half century remained
fragmented into three independent kingdoms in a state of permanent mutual hostility.”
Needham (1978), pp. 39-40.
•  Along with these great scientific and technological exchanges came new ideas,
philosophies and religions. Foreign ideas and religions spread incredibly quickly via the trade
routes across the whole of Eurasia and much of Africa. The acceleration of information
exchange was unprecedented. Buddhism became established in Central Asia well before the
turn of the millennium and in China during the 1 st century CE.
•  There is also some fairly convincing evidence that Christianity and Judaism had reached
both China and India by the first century CE, and Christianity was definitely well-established in
southern India at least by the second century CE. This rapid spread of religions was to
continue in later centuries with Manichaeism, Nestorian Christianity and Islam. See: Appendix
I: “The Spread of Ideas and Religions Along the Trade Routes.”
 
Translator’s Notes
•  The section on Da Qin (Roman territory) from the Weilue was translated into English, with
excellent notes, by Friedrich Hirth in his pioneering volume, China and the Roman Orient, first
published in 1885. He also included translations of a wide range of other Chinese texts
relating to Da Qin (the Roman Empire) and the Chinese text of each is included, making it an
essential reference, even today.
•  This was followed in 1905 by a translation of the rest of the text of the Weilue into French by
Édouard Chavannes, under the title of, “Les pays d’occident d’après le Wei lio.”
          Chavannes’ translation is accompanied by copious notes in which he clarified
numerous obscurities, and convincingly identified many of the countries and towns mentioned
in the Weilue, especially along the eastern sections of the overland trade routes. These are,
to my knowledge, the only translations of significant portions of the text into European
languages to date.
•  In 1980 I was living in India, beginning to study the history of the Kushan Empire, when
Professor Edwin Pulleyblank very kindly sent me copies of Chavannes’ annotated French
translations of the Weilue and the chapter on the ‘Western Regions’ as well as other key texts
from the Hou Hanshu. I translated Chavannes’ accounts into English as an aid to my studies.
What a wonderful treasure chest of information I discovered there!
          It was soon clear to me, however, that the translations and notes were badly in need of
expansion and updating, and, as there had never been a complete translation of the whole
texts into English, I decided to began teaching myself Chinese so I could study and translate
the original Chinese texts.
          I had not realised when I first started in 1980 what an immense and lengthy, but
rewarding, task the translating and annotating the Weilue would be. It was an audacious
undertaking, as my knowledge of Chinese was (and still is) very limited. It would have been
impossible without the help of many experts and friends and any merits this new translation
might have are due largely to their kind and generous suggestions and advice.
•  An excellent and detailed review of Chavannes’ translation of the Weilue (which,
unfortunately, excludes the section on Da Qin or the Roman Empire) was published by Paul
Pelliot in BEFEO 6 (1906), pp. 361-400, in which he corrected some of the major mistakes
and weaknesses in Chavannes’ pioneering translation. I include here some of his more
important observations and notes:
“Mr. Chavannes always makes use of the edition of the twenty-four historians
published by the library of Tushujicheng in Shanghai from 1888. This edition has the
advantage of being printed clearly in a convenient format and is relatively inexpensive.
It accurately reproduces the Imperial edition published in the 18 th century by order of
Qianlong and which is authoritative in China today. It is just that this edition in moving
characters1, generally correct for the Shiji or the Histories of the Han, and which is at
the same time the first and the only true reflection of the dynastic histories, is quite
careless from the Sanguozhi onwards. Additionally, Mr. Chavannes has had at his
disposal the edition of the Sanguozhi known as the Baorentang (p. 550, n. 2; p. 555, n.
1), but he does not seem to have always checked it for, in at least two cases it is
unlikely that the Baorentang edition gives readings which, in the edition of Shanghai
are clearly printing errors: on p. 522, “自項氐 Zixiang Di” is incorrect for “白項氐
Baixiang Di” and the correct reading is found in the example in the xylographic edition
published by Jiangnanshuju in 18871. It is the same for the 魏卑 Weibi of p. 526 where
Mr. Chavannes clearly sees that it ought to be written 鮮卑 Xianbi and which is, in fact,
correctly written Xianbi in the edition of Jiangnanshuju and, very probably, in that of the
Baorentang. The edition of 1877 that I quote here is, however, far from being
satisfactory itself. In the section on Da Qin that Mr. Chavannes has not translated, it
presents a printing fault which has misled Mr. Hirth and which I have already had the
occasion to note (B.E.F.E.-O., iv, p. 175, n. 3). As to the rest of this section taken from
the Weilue, one will find in the large format edition of 1887 the faulty readings: 祿福
Lufu in place of the 福祿 Fulu of Mr. Chavannes (p. 521); 絶精 Juejing instead of 精絶
Jingjue (p. 556). On the other hand, this same edition of 1887 gives readings or
characters in certain places that one cannot reject a priori: such as the fact that it
always writes Yuedi 月氐 and not Yuezhi 月氏 2, 領 ling in place of 嶺 (although I do not
believe the two characters were used interchangeably); in the name of Yulai (p. 558),
one finds 于 yu in place of its equivalent 於 ; Danduo (p. 526) is written with 柘 to and
not with 拓 tuo [note Chavannes (1905), p. 526, n. 5, writes: “The character 拓 is also
pronounced zhi; but the pronunciation tuo appears preferable when it refers to the
pronunciation of foreign sounds [there is a typographical error here where Chavannes
gives ‘nons’ instead of ‘sons’].” The 皮亢 Pikang of p. 558 is not a priori better than the
皮宂 Pirong given in the edition of 1887. In the enigmatic title that the edition of the
library of Tushujicheng gives in the form of 白疏聞 bosuwen (p. 550), the edition of
1887 agrees with other sources that have 閒 xian in place of 聞 wen. All these
examples show that one cannot safely translate using a single contemporary edition of
the dynastic histories. The first palace edition alone deserves complete trust for the
text adopted under Qianlong, and yet modern criticism can only see there the version
which was followed by the scholars of the 18th century, but not a sufficiently reliable
text that comparison with the editions of the Sung, Yuan, or the Ming would be without
profit3.
          All these editions, from whatever period, have this in common, that they have not
modified the text even if it was clearly in error. Disregarding the copying or printing
faults that they inevitably present in greater or lesser numbers, the differences between
 the editions to which Chinese or European science are able to refer to always provide
various readings furnished by previous printed or manuscript examples, and the
various editors have not chosen between them in the same manner. This prudence,
this respect for the text, is one of the principal merits of Chinese scholarship and it is,
in part, due to this that the dynastic histories have retained such great authority. But, as
a result, commentaries are necessary to establish, whether by comparison between
the dynastic histories or referring to other works of Chinese literature, if a certain
passage is certainly or probably in error, and in which manner it ought to be corrected.
It is principally under the present dynasty, which is the great period of Chinese
exegesis, that this research has been undertaken.”
Page 365, note 1. I have several times, and with others with me, spoken of the
lithographic or photolithographic edition of the twenty-four historians. This is the
edition used here by Mr. Chavannes ; it was published in 1888 and in the
following years in a small format, and, in fact, has been carried out with the use
of mobile metallic characters. The same applies to the corresponding edition of
the Tushujicheng.
Page 366, note 1. This xylographic edition of 1887 does not, however,
reproduce the official edition of the 18th century, but that published under the
Ming by the 汲古閣 Jiguge. It is known that the Jiguge of the 毛 Mao family was
the best publishing house existing during the Ming. There is the catalogue of
what was published there (cf. WYLIE, Notes on Chinese literature p. 60). The
edition of the Jiangnanshuju which appeared in 1887 is in the library of the
École des Langues orientales.
Page 366, note 2. This form 月氐 Yuedi has not been neglected, if one refers to
the remarks of Mr. FRANKE in his Beiträge aus Chinesischen Quellen zur
Kenntnis der Türkvölker und Skythen Zentralasiens (Berlin, 1904), where its
existence prior to the Weishu is disputed : yet there was a printing fault. One
sees that it is a matter of the edition. In reality, I believe that the ancient
manuscripts rarely distinguished between 大 da and 太 dai, 氐 di and 氏 zhi, 祗
zhi and 祇 qi. The unity of the ancient forms of these dual characters has
survived until now in spirit. As for the form 月支 Yuezhi, it should be noted that it
has also served for writing the name of a Korean principality (Sanguozhi, ch. 30,
folio 13).
Page 366, note 3. We have not so much as mentioned ancient Chinese
manuscripts. Meanwhile, exception should be made for those that have been
rediscovered in Japan over the last few years. Among them is a manuscript from
the Tang period giving the 食貨忎 Shihuozhi of the Qian Hanshu of Ban Gu with
commentary by Yan Shigu, that is to say, a portion of the three canonical
histories which have never ceased attracting attention and which, as a
consequence, have been transmitted with the greatest care. Now, on this
chapter alone, there are about a hundred characters different from the usual
text. Cf. on this subject B.E.F.E.-O., ii. 335.
Translated and adapted from Pelliot (1906), pp. 365-367 and nn.
•  Throughout this translation I have relied primarily on the Weilue as quoted in the
Sanguozhi, New China Library 1975 Edition, published by the New China Bookstore
Publishing House, Beijing. This is generally regarded as an authoritative an accurate
rendition, with the added advantage of including punctuations. Occasional small differences
with other editions have been dealt with in the Notes.
•  Paper was a new invention, first recorded in China in the year 105 CE (although recent
research indicates it was probably invented previous to this usually accepted date). Prior to
this books were usually written on bamboo slips or on silk. It is unclear whether Yu Huan had
access to paper or not. The reader should be aware that this chapter has only survived
because it was included as an extensive note to the Sanguozhi. Often it is possible to tell
whether a bamboo slip has been lost because they usually only had a limited number of
characters on them; a page of paper could contain a larger text. Unfortunately, here one
cannot tell for certain, but it does seem possible that one or more bamboo slips were lost
before the chapter was recorded in the Sanguozhi – particularly near the end of Section 10.
•  The notes, which I hope will make the translation more meaningful and accessible for
readers, have proved to be even more difficult and demanding than the translation itself. In
particular, the identification of some of the place-names and products mentioned in the text
are still unresolved; and continue to be vigorously debated.
 
About the Dating and Background of the Text
Chavannes, in his introduction, convincingly dates the composition of the text:
“The biography of Yu Huan has not been admitted to the canonical histories. Therefore,
we would only be able to guess at the date at which this author wrote if a celebrated
critic of the Tang period, Liu Zhiji 劉知幾, had not left us, in his Shi tong 史通 published
in 710, this very short bit of information:
‘Previously, during the Wei period (220-265), Yu Huan, originally from the capital
(Changan), composed the Weilue without being officially given the job. The
narrative of these events comes to a halt during the reign of Emperor Ming (227-
239). . . . ’
The evidence of Liu Zhiji, dating from a time when the Weilue had not yet disappeared,
cannot be put in doubt. It fixes the composition of the Weilue in the twenty-six years between
CE 239, the end of Emperor Ming’s reign, and 265, the end of the Wei dynasty.” Translated
and adapted from Chavannes (1905), pp. 519-520.
Pelliot adds the following information about the date and status of the text in his review:
“The first question to resolve was to establish clearly in which period the Weilue was
composed. It is known that the author was called 魚豢 Yu Huan, and various
indications support the late testimonies that place him under the Wei (220-265), but Mr.
Chavannes is the first to base this date on a text definitely from the 8th century. This
text is found in the 史通 Shi tong of Liu Zhiji 劉知幾 [661-721], published in 710. Mr.
Chavannes believes that it is unique and decisive. In fact, it is truly the only text that
the Chinese bibliographers quote regarding Yu Huan that is not taken from the
canonical histories. However, the fact of not coming from the official compilations,
regarding material on Chinese history, does not give more authority to a work.
Henceforth we will be able to call upon a text more than a hundred years older, and
more reliable. It is said in the chapter on literature of the dynastic history of the Sui
(581-617) that Yu Huan occupied a post of langzhong 郞中 (“Palace Gentleman”)
under the Wei2.”
 2. Sui shu, 淮南書局 Huainan shuju edition (1871) ch. 33 folio 4 b.
Translated and adapted from Pelliot (1906), p. 362.
 
“After the Tang, the only title that survived, before the complete disappearance of the
work, is the Weilue in 50 chapters, mentioned still in 1225 in the 史畧 Shilue of Gao
Sisun 高似孫. Xin Zhu reports another work of Yu Huan, the Zhongwai guan 中外官, of
which the title has been preserved for us in the Nanqi shu 南齊書 chapter dedicated to
the administration. This was, without doubt, a sort of table of the metropolitan and
provincial functionaries. Yu Huan is this time qualified as a 官儀 guanyi, but there is no
doubt that it refers to the same individual. Here again it is said that Yu Huan lived
under the Wei. As the Nanji shu deals with the years 479-501, and was compiled in the
first half of the 6th century, we have in this passage new evidence, 100 years previous
to the Shui shu, and 200 years before the Shih tong, which allows us to fix the period in
which the Weilue was written in the second third of the 3rd century.” Translated and
adapted from Pelliot (1906), pp. 363-364.
•  Although the Weilue was never classed among the official or ‘canonical’ histories, it has
always been held in the highest regard by Chinese scholars as a unique and precious source
of historical and geographical information. Pelliot notes:
“Tianlue and Weilue are classed among the 雜吏 zashi. Mr. Chavannes translates this
term by “historians of mixed value.” I am not sure that this is the meaning. Wylie (Notes
on Chinese Literature, p. 25) renders zashi by “miscellaneous” and perhaps he was
right. The term tsa “mixed” could apply here to the nature of the subjects dealt with,
which are “various,” and not to the greater or lesser knowledge or talent which the
author would have to prove.” Translated and adapted from Pelliot (1906), p. 362, n. 2.
[Note: The ABC p. 1230, defines 雜吏 as an “unofficial history.”]
•  Since the time of Chavannes and Pelliot, there has been almost a century of scholarship
devoted to various aspects of the text by scholars from many countries.
          Recent archaeological finds, and research on other key texts, notably the Periplus of
the Erythraean Sea, and the chapters on the ‘Western Regions’ in the Shiji, the Hanshu, and
the Hou Hanshu, have helped to throw light on this difficult but important work.
          In spite of all this attention, many place-names in the text remain unidentified (or the
identification is not convincing), and some sections of the routes outlined in the text have
remained unclear. This is especially true of the sections relating to the Roman Empire, and
the sea routes between China and Egypt, where the data available is very sparse.
•  The Weilue contains many place-names which are no longer known in Chinese and which,
if left unidentified, make the trade routes, and much else besides, impossible to decipher.
•  Local place-names frequently change, and the ancient names of places are often long
forgotten. Nor are we certain of the local pronunciations in the second and third centuries CE.
As in English, the Chinese sometimes used descriptive names, such as ‘Salt Lake’, or ‘Red
Valley,’ and, at other times, literal translations of foreign names.
•  The pronunciation of words change over time, as do the pronunciations of the Chinese
characters that have been used to transcribe them. The Han transcriptions of the sounds of
local place-names often amount to little more than rough approximations. Sometimes
syllables were dropped, sometimes the pronunciations were drastically altered, particularly as
certain foreign phonemes did not exist in Chinese. These processes are also common in
English where we find examples such as ‘Roma’ transcribed as ‘Rome’ (single syllable) and
‘Paris with an ‘iss’ sound at the end instead of the French ‘ee’.
•  It was recognised by both Hirth and Chavannes that, although the Weilue was composed
during the Wei dynasty, most of the geographical information it contains, especially that on the
regions to the west of the Tarim Basin, must have been collected at an earlier date.
•  The Weilue includes much of the information on the Roman Empire already recorded in
zhuan 118 of the Hou Hanshu on the “Western Regions.” This information seems to have
been mainly based on the accounts of the Chinese envoy, Gan Ying, who had been sent by
the famous Chinese General Ban Chao, Ban Yong’s father, in CE 97, to the west to gather
information. It has been either paraphrased from the Hou Hanshu itself, or taken from the
same sources. On the other hand, much of the information on Parthia and the Roman Empire
is additional to that included in the Hou Hanshu. It was presumably collected after the report
of the Chinese General Ban Yong to the Emperor in, or just before, CE 125.
•  Gan Ying got as far as the banks of the Persian Gulf but was persuaded not to go further by
the Parthians. He returned to China in 101 CE. Much of this information is duplicated in the
Weilue. Fan Ye, the compiler, who died in CE 445, added a few bits of later material to the
Hou Hanshu (dating up to about CE 170). These include this fascinating passage:
“The king of this country [Da Qin] always wanted to send envoys to the Han, but Anxi
(Parthia), wishing to control the trade in multi-coloured Chinese silks, blocked the route
to prevent [the Romans] getting through [to China].
In the ninth Yanxi year [166 CE], during the reign of Emperor Huan, the king of Da Qin
(the Roman Empire), Andun (Marcus Aurelius Antoninus), sent envoys from beyond
the frontiers through Rinan (Commandery on the central Vietnamese coast), to offer
elephant tusks, rhinoceros horn, and turtle shell. This was the very first time there was
[direct] communication [between the two countries]. The tribute brought was neither
precious nor rare, raising suspicion that the accounts [of the ‘envoys’] might be
exaggerated.” Hou Hanshu, ch. 118. See TWR Section 12.
•  Much of the new information in the Weilue is very specific and quite detailed, giving
distances and directions between cities, and must have been based on actual travel notes.
Who supplied this information is not clear, although the routes described strongly suggest that
they were mainly gathered from Arab, possibly Nabataean, traders.
         However, the many references to Anxi (Parthia) indicate that the information in the
Weilue on the Roman Empire and Parthia must date from before the collapse of the Parthians
and the founding of the Sasanian Empire in 224 CE.
          Indications in the text strongly suggest that the information on Parthia and the Roman
Empire was gathered after the accession of Meredat to the throne of Mesene/Characene (i.e.
sometime after CE 116), and before the fall of Sura on the Euphrates (along with the whole
region between Dura-Europos and Edessa), to the Romans in CE 164-165.
•  Wherever possible, the information in the Weilue has been checked with that of the 1st
century Periplus of the Erythraean Sea. This work can now be confidently dated to between
40 and 70 CE and, most probably, between CE 40 and 50. See: Fussman (1991); Robin
(1991); and Casson (1989): pp. 6-7.
 
Background Reading
To gain background on the period, and especially of the trade between the Roman Empire
and the Orient, the Weilue can be read alongside the following texts:
– the Shiji [Shih-chi] by Sima Qian [Szu-ma Ch’ien], particularly chapter 123 on
Dayuan [Ta-yüan] which covers up to the end of the 2 nd century BCE. See the very
readable translation in: Watson, Burton, 1961, II: 264-289. 
– the chapters on the Western Countries (zhuan 61 and 96) of the Hanshu which
covers the period 125 BCE to 23 CE (translated and amply annotated by Hulsewé and
Loewe in China in Central Asia, 1979). (Covers 125 BCE to CE 23)
– the chapter on the Western Regions (zhuan 118), of the Hou Hanshu (23 CE to late
2nd century, with most of the information on distant countries dating prior to 125 CE):
See the 2nd edition of the annotated translation of my The Western Regions According
to the Hou Hanshu (abbreviated as TWR in this work), which is freely available on the
Silk Road Seattle website:
http://depts.washington.edu/uwch/silkroad/texts/hhshu/hou_han_shu.html
– the chapter on Chinese expansion into the Tarim Basin (zhuan 77) of the Hou
Hanshu (French translation by Chavannes in T’oung pao 7, 1906, pp. 149-234). (1st
and early 2nd century CE). (I hope to make an English translation of these biographies
available soon.)
– the Periplus of the Erythraean Sea (c. 40-50 CE). See the excellent translation (from
H. Frisk 1927 with up-dates and improvements) with the Greek text and extensive
notes by Lionel Casson: The Periplus Maris Erythraei. (1989), Princeton, Princeton
University Press. Also still useful is William H. Schoff’s 1912 translation: The Periplus
of the Erythræan Sea. Wilfred H. Schoff. New York, Longmans, Green, and Co.
Second Edition. Reprint, New Delhi, Oriental Books Reprint Corporation. 1974. The
main text from the earlier translation by Schoff is now fortunately available to all on this
website, although one should still consult both his book and Casson’s more recent one
for their useful commentaries.
– the Naturalis historia by Pliny the Elder - completed, except for finishing touches, in
77 CE. For a full translation see: Natural History. Pliny the Elder (77 CE). Translation
by W. H. S. Jones, Loeb Classical Library, London/Cambridge, Mass. (1961). For a
widely available, readable and useful selection see: Natural History – A Selection. Pliny
the Elder (77 CE). Translated by John F. Healy, London, Penguin Books. (1991).
– the Geography by Ptolemy completed c. 150 CE. The only available English
translation is by Edward Luther Stevenson in Geography of Claudius Ptolemy. Reprint
New York: Dover Publications, Inc. 1991. It is, unfortunately, full of mistakes.
Additionally the Greek names have been Romanized, which often makes identifications
unnecessarily difficult.
I have not included maps with this translation but strongly recommend having a good atlas at
hand while reading the book. Some of the maps available on the “Silk Road Seattle” website
will also prove helpful, especially the detailed map showing the main routes across the Tarim
Basin and the one accompanying the Schoff’s translation of The Periplus of the Erythræan
Sea. Interactive maps and much additional information may be accessed at the following
sites: http://www.fas.harvard.edu/~chgis/chgis_home.html, and
http://www.people.fas.harvard.edu/~chgis/tools/
 
About Measurements and Administrative Divisions
•  The basic units of measurement employed in the Weilue, were the 里 li and the chi 尺. They
have varied considerably through the ages and from district to district. Fortunately,
measurements remained stable over the Qin and Han periods, and an inscribed bronze
standard measure, dated 9 CE, was discovered in 1924 at the Imperial Palace in Beijing. This
has allowed accurate conversions to modern measurements. The Han chi, or ‘foot’ is given as
equalling 0.231 metres (or 9.095 inches).
•  On the basis of this lucky discovery, the li was calculated by Dubs (1955): 160, n. 7 to be
415.8 metres. See also Chapter IV, Appendix I, Standard Weights and Measures of Han
Times, ibid., 276-280.
          In most cases, I have given the exact equivalents to the nearest kilometre based on this
measurement even when the number of li given is obviously an approximation (e.g. 8,000 or
10,000 li). For relatively short distances (less than 100 li), I have calculated to the nearest
tenth of a kilometre.
•  Most, though not all, of the distances between places given in the Weilue – where they can
be checked – prove to have been surprisingly accurate, especially those in areas controlled at
times by the Chinese.
          Sometimes there are mistakes in the distances given in the text. This may be because
it is not clear exactly where the ancient route went between two points (or which of several
alternatives was used). Occasionally, especially over longer distances, the figures given are
obviously widely inaccurate.
•  All other conversions of Chinese measurements are based on the values calculated by Wan
Kuo-ting for the Qin [Ch’in] and Han periods as described by A. F. P. Hulsewé in T’oung pao
Archives, Vol. XLIX, Livre 3, 1961: 206-207.
•  The main administrative divisions of the Han Empire were the 郡 jun “commanderies” (or
‘provinces’) and 王國 wangguo ‘kingdoms.’ These were subdivided into 縣 xian “counties,” 鄕
xsiang “districts,” and 里 li “wards.”
•  Some Chinese words have no exact equivalent in English. One such example is the word
城 cheng, which literally translates as ‘walled town,’ but it was also used for large towns that
were not walled. It is sometimes rendered ‘city,’ but only a handful of the cheng mentioned in
the Weilue would be large enough to be called a ‘city’ in our age of megalopolises. Most of
them were what we would think of as country towns or provincial centres. I have translated
the word simply as ‘town’ and leave it to the reader to add the nuances according to the
context.
          Similarly, 國 guo is used to refer to entities ranging from tiny fiefdoms or even villages
to entire empires and can be translated as ‘kingdom,’ ‘fief,’ ‘nation,’ ‘state,’ ‘country,’ or
‘empire.’ I have translated it as ‘kingdom,’ unless the context specifically indicates otherwise,
as most of these territories seem to have had a hereditary system of rulers at this period.
•  I have used “Western Regions” for the term 西域 xiyu rather than the usual translation as
“Western Countries.” I emphasize that xiyu is sometimes used in the specific sense of the
kingdoms actually controlled by China on the routes to the west of “China Proper” (the “Inner”
region – the land within the Wall), and sometimes in the far more general sense of all
countries to the west of China.
•  The character du 都 is frequently translated as ‘capital’ in spite of the fact that there are
many examples of more than one du existing in a single state. Additionally, Dubs (1938), p.
28, n. 2, says that in “ancient times” it referred to “a large walled city”. I have, therefore
usually translated this word as ‘major centre’ or ‘large town.’ Sometimes the Weilue
designates a town as a wangzhi 王治 which translates literally as the “king’s chief town or
residence” and this is much closer to the concept of ‘capital’ as we use it today, so I have
translated it variously as “the (king’s) capital,” or “the king’s residence.”
          In one place in Section 11, just after referring to the fact that Rome controlled hundreds
of minor kings, the text becomes more specific, using the term wangsuozhi cheng 王所治城
which can be translated as “the king’s centre of administration city,” which I have
rendered: “the king’s administrative capital” – i.e. Rome itself. See note 11.25.
•  Territories referred to as ‘Han,’ ‘Tianzhu’ (India), ‘Anxi’ (Parthia), or ‘Da Qin’ (Rome) were
sometimes used very specifically for the home territory, but often far more loosely for
territories controlled by, or tributary to, the main seat of power. Thus, in the Hou Hanshu, we
find references to the ‘King of Da Qin’ (that is, the king of Rome) and, at the same time,
subject territories such as Egypt, or the ‘Roman Orient,’ are also referred to simply as Da Qin.

THE TEXT OF THE WEILUE

Section 1 – The Di Tribes
The Weilue’s Chapter on the Western Rong (‘Peoples of the West’), 1 says:
The Di2 have their own kings. They originally came from far away.
When the Han opened Yi zhou,3 and established Wudu Commandery,4 they drove back the
people of this race who dispersed, and took refuge in the mountain valleys. Some were at
Fulu,5 and others in the neighbourhood of Qian 6 and Long. 7
They are not an homogenous race. They are said to be the descendants of Panhu. 8 Some
are called the Qing Di (Green Di), others the Bai Di (White Di), and others the Ran Di (Giant
Python Di),9 referring to the class of reptiles in which they are placed.
The people of the Middle Kingdom name them according to the colour of their clothes, 10 but
they call themselves Hezhi.11
Each (of their tribes) has its kings and chiefs who, in most cases, received their lands and
titles from the Middle Kingdom, and are required by it to fulfil their responsibilities, or they are
demoted.
During the Jianan period [CE 196-220], Angui, king of the Xingguo Di (Xing Kingdom Di), 12
and Qianwan, king of the Baixiang Di (White Section Di), 13 each had tribes of more than ten
thousand men.
In the sixteenth year [= CE 211 – but should read 213 CE], they joined forces with Ma Chao
and rebelled.14 After (Ma) Chao had been defeated [in 214], Angui was attacked and killed by
Xiahou Yuan. Qianwan made his way to the southwest and entered Shu (Western Sichuan). 15
His tribes were not able to get away and all submitted.
The (Chinese) government transported those who had taken a questionable stance during
these events, to a separate place in Meiyang (in the Commandery) of Fufeng. 16
These are now the two tribes of Anyi (Peaceful Yi) and Fuyi (Governed Yi). 17 A Military
Protector manages the region.18
As to those who behaved wisely, (the Chinese Government) left them their place on the
borders of Tianshui19 and Nanan.20 These are the ones who are now dependants of
Guangwei Commandery.21
Their customs and language are not like those of the Middle Kingdom, but similar to those of
the Qiang22 and several Hu peoples.23 Each person has a family name, like the family names
of the Middle Kingdom. They prefer blue and deep red clothes.
They are commonly skilled at weaving cloth; they are good farmers; they breed and rear pigs,
oxen, horses, donkeys, and mules.
When a woman marries, she wears a renlu24 that, in the way that it is trimmed and decorated,
sometimes resembles the renlu of the Qiang, and sometimes the tunic of the Middle Kingdom.
All braid their hair.
Many of them know the language of the Middle Kingdom because they have lived in the
Middle Kingdom and mixed among the people. Nevertheless, when they return to their tribes
they naturally speak the Di language.
Their marriage (customs) resemble those of the Qiang.
These are, in fact, the people who were previously called the Western Rong, 1 and now live in
Jie,25 Ji,26 and Huandao.27
Although, at present, the fiefdoms28 are administered by the (Chinese) commanderies and
kingdoms,29 previously they had their own kings and feudal princes living in (now) empty
villages. They also had more than ten thousand (people) living in the region of Wudu, 30 and in
the neighbourhood of Yinping31 and Jie.32
 Section 1 – The Di 氐 Tribes
1.1. By the time of the Han dynasties (206 BCE – 220 CE), the term Xirong or ‘Western Rong’ 西戎, was used in the
very general sense of ‘the tribes to the west of Han.’ The term once applied to a specific people from whom a
number of tribes descended, which helps explain how the more generalised version of the name came about. It will
be useful here to have a look at early Chinese concepts of ethnicity and how they had developed by the period of
the Han dynasty:
            “By the beginning of the first millennium BC an awareness of being “Chinese” had emerged in
northern China. This identity is intimately connected with the growth of early states, especially the Shang
and the Zhou, whose centers lay in the region of what is today Henan 河南, Shanxi 山西, and Shaanxi 陕
西 Provinces. This area, known as the Zhongyuan 中原 (Central Plains) or zhongguo 中國 (central
states), formed a nucleus around which early “Chinese” civilization flourished.
            Texts written down by the mid-first millennium BC suggest that the people living in this area
regarded themselves as “civilized,” a concept that was closely connected to li 禮. Li is commonly translated
as ceremony or rite, as well as etiquette. In reality, the concept of li is complex and implies an entire
system of behaviour and ritual – a kind of customary law. Its principles regulated the lives of the people
living within the zhongguo. Whoever practised li was eligible to be regarded as a member of the Hua-Xia
“We-group.” Thus, it seems entirely reasonable to suggest, as have many (Creel 1970; Eberhard 1982;
Pulleyblank 1983), that the concept of being “Chinese” at this early date was not fixed and static. It was a
 flexible category defined in cultural and political terms, and membership expanded along with the spread
of the early Chinese state.
            Some of the earliest categorizations and perceptions of “other” people in ancient China are found in
the Zhouli 周禮 and the Erya 爾雅, texts that date from the latter half of the first millennium BC. What is
striking in both of these texts is the classification of groups into broad categories that set up oppositional
clusters of cultural traits. The key here is the word “broad” – the categories are defined in such a way that
individuals or groups of people could easily shift in and out of the categories as the boundaries of “Chinese
civilization” expanded and people adopted the trappings of what the “Chinese” perceived as “being
Chinese.”
            The Zhouli (c. third century BC) defines the Chinese world in terms of “inner” (nei 內) and “outer”
(wai 外). According to this view, the world consisted of nine concentric circles surrounding a central zone
known as the king’s domain. In this schema, the six zones closest to the king’s domain were considered
neifu 內府, i.e., outside. They did not belong to the world of the king’s domain and were inhabited by
“outsiders.”
            This basic concept of “inside” vs. “outside” continues as a constant theme in Chinese literature and
is further developed in the Erya, a Han period (second century BC) text. In the Erya the inside / outside
theme is given a stronger geographical basis. The “outsiders” were consequently subdivided into four
groups associated with the four cardinal directions – the “Four Seas.” There were the northern Di 氐, the
Eastern Yi 東夷, the southern Man 蠻, and the western Rong 戎. Each of the Four Seas was further split
into “inside” and “outside.” The nei or inside groups were again dissected into sheng 生 (raw), i.e.,
uncivilized and hence potentially dangerous, and shu 熟 (cooked or ripe), i.e., civilized in the sense that
these groups had been tamed by the influence of Chinese civilization (see Cushman 1970: ch. 2 for a
detailed discussion of this analysis).
            The constant sub-theme running through these texts is politics. The groups are first defined
according to categories of “inside” and “outside.” Those groups on the inside are linked with the growing
authority and strength of the early Chinese state and are beginning to lose their sense of differentiation.
Those groups living at the peripheries of the Chinese state are, in contrast, perceived of as “different” and
are consequently potential threats to the unity of the state.         
            To sum up, while Chinese texts do provide information on how the Chinese texts do provide
information on how the Chinese distinguish between themselves and others, the categories are not fixed or
rigid. Individuals or groups could abandon those cultural traits regarded as “raw” or non-civilized, and
adopt those used by the Chinese. In doing so, the people became “cooked,” i.e., more civilized, and
potentially less ethnically distinct.
            During the Han Dynasty, descriptions begin to appear in texts that still reflected the broader
categories of “inner” and “outer,” and “raw” and “cooked,” but also provided more detailed information
about what these “other,” i.e., non-Chinese, people looked like, how they lived, and what names they were
called.” Peters (2002), pp. 83-85.   
 
          “The Sai 赛 tribes, which appeared in the valleys of the Ili and Chu rivers by the end of the seventh
century B.C. had possibly come from the east. The precursors of the Asii, the Tochari, the Gaisani and the
Sacarauli seem to have been the Rong of the surname Yun 戎姓之允, the Daxia, the Yuzhi 禹知 (Yuzhi
禹氏) and the Suoju 莎車 who appeared in pre-Qin records and books. In 623 B.C., Duke Mu 穆 of Qin 秦
dominated the Western Rong 戎 and opened up territories which extended for 1,000 li [416 km]. This
event possibly caused the Sai 赛 tribes’ westerly migration.
          The Rong 戎 of surname Yun 允, the Daxia 大夏 and the Yuzhi 禹知 (Yuzhi 禹氏) can
respectively be traced back to the Shaohao 少昊, the Taotang 陶唐 and the Youyu 有虞.
          The Shaohao 少昊, which has known [sic] as the state of the surname Yun 允, originally dwelt in the
valley of the Ruo 若 River, then moved to Qiongsang 窮桑 in the north of Lu 鲁. A branch of the
descendants of the Shaohao 少昊 dwelled at Ruo 鄀, and of them, those who moved to Guazhou 瓜州
were called “the villains of the surname Yun 允”. Among the “villains of the surname Yun 允”, some
 moved inwards (the Central Plains) and the others went westwards. Of the latter, those who reached the
valleys of the Ili and Chu rivers became a tribe of the Sai 赛 people, but those who remained to the west of
Hami 哈密 (Kumul) were known as the Wusun 烏孫.” Yu (2000), pp. 1-2.
There are references in Chinese literature to a work or works by Yu Huan, the Weilue 魏略, or “Summary of the Wei
Dynasty,” and the Dianlue 典略, or “Records of the Wei Dynasty.” The originals have long since disappeared and
there are varying accounts of how many chapters each contained or, indeed, whether they were even separate
documents.
The surviving text translated here was quoted by Pei Songzhi in his notes to the Sanguozhi 三國志, or “Memoir of
the Three Kingdoms,” published in 429 CE. It begins thus: “The Weilue’s Chapter on the Western Rong says: . . . . ”
For further information see: Chavannes (1905), p. 519; Pelliot (1906), pp. 362-363; de Crespigny (1970), pp. 75-76.
“According to the Book of History, Yu divided the world into five concentric domains, the outermost of
which was the wilderness domain 荒服. Yen Shih-ku says, “The Jung and Ti [occupied] the wilderness
domain, hence it is said, ‘The four wildernesses.’ It says that it is a wilderness, [where] they suddenly go
and come without any regularity. The Erh-ya says that Ku-chu [in the north], Pei-hu [in the south], Hsi-
wang-mu [a place in the west], and Jih-hsia [in the east] are called the four wildernesses.” Dubs (1938), p.
263, n. 1.
The Rong, Qiang, and Di peoples were known to the Chinese by Shang times [1765-1122 BCE]. ‘Rong’ was originally
used by the Chinese to refer to a people they later described as the ancestors of the Qiang tribes. Considered a
major threat, the Shang mounted large campaigns against them.
The Rong moved into the area to the south of the Yellow River soon after the middle of the 7 th century BCE.
Although the Di are usually considered “a clearly defined national and political grouping,” the name ‘Rong’ was
commonly used in the looser sense of “barbarian” or “bellicose.” By Han times, the meaning of ‘Rong’ had been
expanded to refer generally to all non-Sinitic populations of the West. See: Průšek (1971), pp. 38-42, 210; Molè
(1970), p. 86, n. 61; Dictionnaire Français de la Langue Chinois. 1976: 478, No. 2486. See also the notes on the Di and
Qiang (nn. 1.2 and 1.19).
 
1.2. The Di 氐 [Ti ] people.
Di 氐– K. 590a: *tiər / tiei; EMC: tɛj.
The Bamboo Chronicles contain the first historical mention of the Di referring to the capture of “twenty Di kings”
during an expedition against the Gui or Di peoples by the Zhoujing Ji in the 12 th century BCE.
“It appears that the term Di referred to people originally of Qiang origin settled in the Qin Ling ranges and
further south. As this territory was occupied by the Chinese, the Di became increasingly heavily influenced
by that culture, though still retaining elements of the Qiang. [Chinese records refer to a “White Horse”
(Baima) Qiang and a “White Horse” Di, both on the frontiers of Wudu and Guanghan commanderies in
northern Sichuan. These are probably references to the same people.] See, in particular, HHS 86/76, 2359-
60, and also HHS 87/77, 2898-99.” de Crespigny (1984), p. 470, n. 8. See also: de Crespigny (1977), pp. 6-7, and
n. 8; Rogers (1968), pp. 4-5, and nn. 6, 9; Molè (1970), p. 83, n. 50; Holmgren (1982), p. 116; Wu (1982), pp.
107-108; CICA: 101, n. 178.
      “Earlier History of the Ti People. Of the “Five Races of Barbarians” traditionally associated with the
“Sixteen States,” Fu Chien’s clan was affiliated with the Ti race, a proto-Tibetan group ethnically related to
the Ch’iang 羗, with whom they are associated in the Book of Songs. [Fu Chien (Pinyin: Fu Jian), 338-385,
took power in 357 and subdued all of northern China, founding the Former Ch’in or Qin – one of the ‘Six
Kingdoms’]. There are no clearly identifiable remains of the Ti language in the Chinese sources. 6
Information about the history and distribution of these people is contained in the dynastic histories in
sections specifically devoted to them as well as in separate documents; 7 of the latter, a memorial composed
by Chiang T’ung 江统 (d. 310) is especially informative.
      In pre-Han times the Ti were spread principally in the mountainous country that extended from
 southern Kansu to northern Szechwan, one of their most important seats being at Ch’ou-ch’ih. In 36 A.D. a
considerable number of Ti and Ch’iang were moved into Kuan-chung and Ho-tung, a fact which Chiang
T’ung later deplored on the grounds that the descendants of those people, having grown in numbers and
strength, oppressed the Chinese inhabitants of the areas. Eruptions of the Ch’iang and Ti which began in
107 lasted a decade and were unprecedentedly devastating. Chiang T’ung asserts that during the last years
of the Later Han the barbarians of Yung Province reduced Kuan-chung to desolation, and the Wei-lüeh’s
account of the activities of the Ti in Kansu, Shensi, and Szechwan tends to confirm this. There can be no
doubt that, as Chiang T’ung says, the Ti profited substantially during the rise of the kingdom of Wei from
the rivalry between it and the state of Shu; indeed, the bargaining position which they then enjoyed must
have been of prime significance in the history of their entrenchment in China. One episode which is of
particular interest for the history of the Fu clan was the removal in 236 of on Fu Chien, who then held the
title “King of the Ti of Wu-tu,” with his following of 400 Ti house-holds into Kuang-tu (in Szechwan). This
Ti leader is probably to be identified with the illustrious ancestor after whom Fu Chien’s uncle, Fu Chien,
was named.”
6. Eberhard states that the Ti could be regarded merely as more sinicized Ch’iang, were it not for the
fact that their language is said (in TT) to have differed from that of the Ch’iang; but according to the
Wei-lüeh their language was the same as that of the Ch’iang.
7. WS 101, SS 98, Chou-shu and PS 96. The relevant portion of the Wei-lüeh, preserved in P’ei Sung-
chih’s commentary on SKC, was translated by Chavannes (“Wei Lio”). Its opening section is devoted to
the Ti.
Rogers (1968), pp. 4-5; 80, nn. 6 and 7.
 
1.3. Yi zhou 益州 [I-chou]. During the Han a zhou 州 was a territory under Chinese control, but outside China
proper. It was equivalent to a jun 郡, which is usually translated as: ‘commandery.’ Yi zhou was established in 109
BCE by Emperor Wu. Its administrative centre was to the east of modern Puning County, Yunnan (Prefecture and
Province). The Yi zhou of the Han must not be confused with the Yi zhou of the Five Dynasties which corresponds
with Chengdu fu of Sichuan. See Chavannes (1905), p. 521, n. 1. Yi zhou, during the Han, included most of modern
Yunnan and southern Sichuan. It bordered on Jiaozhi (centred near modern Hanoi) to the southeast, and modern
Burma to the southwest.
1.4. Wudu Jun (Wudu Commandery) 武都郡 [Wu-tu chün] was established by Emperor Wu in 111 BCE to the west
and southwest of Changan.
“South of the Wei and the Tao was Wudu commandery, which had been in Yi circuit under the Former Han,
but was in Liang Province of Later Han. The commandery covered the mountainous country of the Min
Shan in present-day southern Gansu and the borders of Sichuan, with the upper course of the Han River
and of the Bailong Jiang. In this region also, both non-Chinese and the Chinese settlements were scattered
and isolated among the mountains at the foothills of Tibet and the base of the great ridge of the Qin Ling.”
de Crespigny (1984), pp.13-14, and the maps on pp. 92, 98.
“Han Wu-ti decreed the establishment of Wu-tu Commandery at Ch’ou-ch’ih in 111 B.C. (HS 28: Ti-li chih
8B. 1a; Dubs, II, 82). Ch’ou-ch’ih was the seat of the “White Horse (Po-ma) tribe, which is described as the
strongest of the Ti tribes. Ti rebellions against the Chinese colonists of that region were recorded for the
years 108 B.C. (Wei-lüeh; Dubs, II, 93; de Groot, Urkunden, II, 198) and 80 B.C. (HS 7; Dubs, II, 163).” Rogers
(1968), p. 81, n. 9.
See also: Note 3.8, and Chavannes (1905), p. 521, n. 2, where he mistakenly gives 118 BCE instead of 111 BCE for the
establishment of Wudu Commandery. He says that its centre was 80 li to the west of Cheng County (Sub-prefecture
of Jie, Gansu Province).
1.5. “Fulu 福祿 was the name of a County (xian) of Jiuquan Commandery (now Su County, Gansu Province).”
Translated from Chavannes (1905), p. 521, n. 3.
“It was not until the time of Emperor Wu, shortly before 100 BC, that the Han established a military and
political presence northwest across the Yellow River and founded the commanderies of Hexi [Ho-hsi] “West
 of the River”. Jiuquan [Chiu-ch’üan], Zhangye [Chang-yeh] and Dunhuang [Dunhuang] were probably
established in 104 and subsequent years, Wuwei [Wu-wei] and Jincheng [Chin-ch’eng] in the half-century
following.
      Under Later Han, the commanderies of Wuwei, Zhangye, Jiuquan and Dunhuang stretched in that order
from southeast to northwest along the present-day Gansu corridor. On the southwest, they were backed by
the Qilian Shan [= Nan Shan] and the mountainous regions of present-day Qinghai [Ch’ing-hai]. To the north
and east they faced the Helan [He-lan] Shan, the Tengger and other deserts on the edge of the Gobi. The
cities and settlements were based on oases, supplied by the snowmelt streams which flow from the high
ground to the south and then disappear into marshes in the desert. As in the Tarim basin of central Asia,
irrigation agriculture was maintained around these cities, and the settled farming economy was sufficient
to provide a frontier defence for the trade and communications of the Silk Road which led through the
Western Regions to India and Rome.” de Crespigny (1984), pp. 7-8.
For a discussion of the dates of the establishment of the four commanderies, see also: Dubs (1944), p. 83 and note
23.1.
1.6. Qian 汧 [Ch’ien].
“The ancient County of Qian was to the south of the present County of Long (Fengxiang Prefecture, Gansu
Province)”. Adapted from Chavannes (1905), p. 521, n. 4. Qian County was approximately 180 km northwest
of Changan (modern Xian) on the Long road.
“(Shensi, S of Lung hsien) LTYT 36a, SLKCYC 4133.2. A prefecture of Fu-feng Commandery, Ssu-li.” Rogers
(1968), p. 310.
1.7. Ancient Long 隴 [Lung] County was to the west of Qian County, and included the strategic Long Pass which is
approximately 215 km northwest of Changan on the Long Road.
“(Shensi, Lung hsien) LTYT 36b, SLKCYC 4137.1. Area of Lung Mountain near the Lung-ch’eng Prefecture
that lay in Lüeh-yang Commandery in Ch’in Province. It was the western of the Four Fortresses of Kuan-
chung.” Rogers (1968), p. 324.
“Lung is the name of a chain of mountains at the sources of the river Wei on the border between Shen-si
and Kan-su”. Molè, (1970), p. 71, n. 13.
“The main communications route east and west along the Wei River was the Long Road, so called from the
mountain by which it passed. It seems most probable that the ancient road generally followed the line of
the modern railway from Xian along the Wei valley, then crossed the watershed to the Tao River, and then
went north to the Yellow River near Lanzhou, where it joined the Silk Road leading northwest into central
Asia. Most of the country of eastern Gansu is rolling loess hills, not a major obstruction to open movement
and manoeuvre, but sufficient to render attractive the silted flood-plains of the major streams, while in the
upper reaches of the Wei and about the Yellow River the terrain is steep enough to make travel away from
the valleys quite difficult.
      Near Long Mountain, however, on the borders of present-day Gansu and Shaanxi, the Wei River runs
through gorges in the loess. In this region the road left the Wei valley and crossed the hill country through
the Long Pass, north of the river. The Long Pass was a fortified barrier, and it appears also that the Long
Road itself was protected as a military highway, with patrols, garrisons, stores and arsenals at intervals
along its course.” de Crespigny (1984), pp. 14-15.
1.8. Panhu 縏瓠 [P’an-hu]. In spite of the popular etymology recounted below, the name Panhu was probably a
transliteration of a Di name.
“At the beginning of Chapter CXVI dedicated to the Barbarians of the South (translated by WYLIE, Rev. de
l’Extr. Orient, 1882, pp. 200-201), the Hou Hanshu has recounted in full the legend of the dog, Panhu, who
married the daughter of Emperor Gaoxin (identified by Sima Qian with Emperor Ku), and who was the
ancestor of the Barbarians of the South. The commentary on the Hou Hanshu quotes, on these remarks, a
passage from the Weilue itself which indicates to us a popular etymology of the name Panhu: ‘Emperor
Gaoxin had an old married woman who lived in the house of the king, and who had an earache. While
 removing the problem, an object was found big as a cocoon. This woman placed it in a gourd (hu), which
she covered with a bowl (pan). In an instant, the object transformed itself into a multicoloured dog. That is
why it is called Panhu.” Translated from Chavannes (1905): 521, n. 6. Note: the reign of the legendary
emperor Gaoxin or Di Ku is traditionally assigned to c. 2436 BCE to 2366 BCE.
1.9. Ran Di 蚺氐 [Jan Ti] – ‘Giant Python’ Di. Williams (1909), p. 403, identifies the word jan as:
“A large serpent found in southern China, described as 50 [Chinese] feet long, which can seize deer
for food; it has long teeth and a bright variegated skin which is cured for covering guitars; it carries its
head close to the ground. . .    the gall is reputed to be useful in curing consumption; this description
doubtless refers to a sort of boa like that reported to be found in Hainan Island.”
This description of a snake up to 50 chi [about 17 metres] is undoubtedly an exaggeration, but the description could
otherwise refer to either of two giant pythons common in Southeast Asia: the Reticulated Python (Python
reticulatus), or the Indian Python (P. molurus).
          The Reticulated Python ranges from southern Burma to Indonesia and the Philippines and is probably the
world’s longest snake, with specimens found up to at least 9.6 metres (31 feet) long, although the Anaconda is
heavier. It has been known to take deer and even, occasionally, children.
“The reticulated python gets its name from the distinctive color and pattern on its scales. According to
Webster’s Third New International Dictionary the word “reticulated” is an adjective defined as “having
lines intercrossed, forming a network.”  It is also known as the regal python (regal is a word that refers to a
king). Its scientific name is Python reticulatus. . . . Reticulated pythons live in tropical forests on the
continent of Asia. Their range extends from Myanmar and India, across Southeast Asia and on many of the
islands of the Philippines and Indonesia. They are at home on the ground, in caves or in trees and they
have adapted to living in towns and cities where they hunt chickens, ducks, rats and domestic cats, dogs
and pigs. Large reticulated pythons have eaten monkeys, wild boar, deer and even people. There are not
many cases of these pythons capturing and eating people, but it has been reported even in recent years. . . .
The largest reticulated python ever found in the wild was reported in 1912 from the island of Celebes (now
known as Sulawesi) in Indonesia.  This snake measured thirty-three feet.” Shwedick (2002).
However, ran, according to GR Vols. III, p. 449 and IV, p. 334, refers to Python molurus bivittatus Schlegel. It adds that
it is edible and that its flesh and bile are used as medicines. This snake, commonly known as the ‘Indian Python’, is
usually less than 4 metres but is sometimes twice that length, or about 8 metres (26 feet).
          I have translated ran as ‘Giant Python,’ as it is impossible to decide which of these two species of python was
intended in the text.
1.10. Early Chinese accounts refer to at least two main groups of Di: the ‘Red Di’ and the ‘White Di’. It has been
suggested that this could be a reference to geographical divisions, with the White Di to the west, and the Red Di to
the south, but it seems:
“quite clear that the term Red Di was not a geographical one and did not distinguish the bearers of it from
Ti living in other places; it was a political and particularly a social term, distinguishing one group of
implacable enemies of Chin from others who were tired of fighting and anxious to come to terms; it was
primarily the distinction between one aristocratic ruling group and another subordinate group.” Průšek
(1971), p. 219.
There seems to be little reason to doubt the early claims, repeated here in our text, that the names refer to the
colour of the clothes worn by various clans or groups. “Thus explaining the terms ‘Green Di’ 青氐, ‘White Di’ 白氐
above.” Translated from Chavannes (1905): 522, n. 2.
1.11. Hezhi 盍稚 [Ho-chih]. Ancient pronunciations:
he – K. 642n: *g’âp / γâp; EMC: γap; he: g’âp / γâp
zhi – EMC: drih. Reconstructed pronunciations of this word are not included in Karlgren.
1.12. Xingguo Di 興國氐 [Hsing-kuo Ti]. Xingguo refers to the ancient state of Xing in the Hebei plain at the foot of
the Taihang Shan, which the Di conquered circa 660 BCE and was settled by the ‘Red’ Di. See: Průšek (1971), pp. 144-
145. GR Vol. II, p. 1196 states that it refers to the ancient prefecture of Hubei which is now known as Yangxin.
1.13. Baixiang Di 白項氐 [Pai-hsiang Ti] literally reads: “The White Section of the Di”. Chavannes (1905), p. 522.
mistakenly gives Zi [Tzu] for the similar Bai [Pai] (which can also be pronounced bo).
“Mr. Chavannes always makes use of the edition of the twenty-four historians published by the library of
Tushujicheng [T’u-shu-chi-ch’eng] in Shanghai from 1888. This edition has the advantage of being printed
clearly in a convenient format and is relatively inexpensive. It accurately reproduces the imperial edition
published in the 18th century by order of Jianlong [Ch’ien-lung] which is authoritative in China today. Only
this edition in moving characters, generally correct for the Shiji [Shih-chi] or the Histories of the Han which
are the first and only true readings of the dynastic histories, is quite negligent from the Sanguozhi [San kuo
chih] onwards. Moreover, Mr. Chavannes has had at his disposal the edition of the Sanguozhi known as the
Barentang [Pao-jen-t’ang](p. 550, n. 2; p. 555, n. 1) but it does not seem to have been always checked, for in at
least two cases it is unlikely that the Barentang edition gives readings which, in the edition of Shanghai are
clearly printing errors: on p. 222, “Zixiang Di” [Tzu-hsiang Ti] is incorrect for Baixiang Di [Pai-hsiang-
ti],and the correct reading is found in the xylographic edition published by Jiangnanshuzhu [Chiang-nan-shu-
chu] in 1887. It is the same for the Weibi [Wei-pi] of p. 526, where M. Chavannes clearly sees that it ought to
be written Xianbi [Hsien-pi] and which is, in fact, correctly written Xianbi [Hsien-pi] in the edition of
Jiangnanshuzhu and, very probably, in that of [the Barentang].” Translated from Pelliot (1906), pp. 365-366.
“Po-ching 百頃 was another name for Ch’ou-ch’ih. It is explained as indicating the area (100 ch’ing) of the
Pond (ch’ih); in late editions (e.g. K’ai-ming: Wei-chih 30. 1006. 2) Wei lüeh wrongly writes Tzu-hsiang 自
項, which Chavannes in his translation (n. 7) fails to correct. Po-na (SKC 30.31a) has Po 白-hsiang.” Rogers
(1968), p. 84, n. 47.
The Chinese text of the Weilue that I have been using (the five volume Sanguozhi published by the New China
Bookstore Publishing House, Beijing, 1975, zhuan 30: 858-863), like the Po-na edition, also has Bai- (or Bo-)xiang 白
項 [Pai- (or Po) hsiang]. This name would appear to be correct and carries the meaning of the ‘White Section.’
          This interpretation seems to be confirmed by the description given earlier in the text of the various sections
(or clans) of the Ti peoples: “Some are called the Qing Di (Green Di), others the Bai Di 白項氐 (White Di), and
others the Ran Di (Giant Python Di), referring to the class of reptiles in which they are placed. The people of the
Middle Kingdom name them according to the colour of their clothes, but they call themselves Hezhi [ 盍稚].” See
also notes 1.11 and 3.8.
1.14. Ma Chao 馬超.
“In chapter 1 of the Wei zhi section of the Sanguozhi, one reads that Ma Chao’s rebellion indeed broke out in
211, but it was only in 213 that he obtained the cooperation of the Di tribes: ‘In the eighteenth Jianan year
(213), . . .   Ma Chao, finding himself at Hanyang (to the south of the County of Qingfu, Xuchou Prefecture,
Sichuan Province), began to make trouble again using the Qiang and the Hu. The king of the Di, Qianwan,
rebelled to join forces with him. (Ma) Chao camped at Xingguo. Xiahou Yuan attacked him.’ In the first
month of the nineteenth year [29 January to 26 February, 214], Xiahou Yuan defeated Qianwan and put him
to flight. He exterminated the people of Xingguo.” Translated from Chavannes (1905), p. 522, n. 3.
          “During the Chien-an period (196-220), according to the Wei-lüeh . . . , the Ti leaders A-kuei 阿貴 of
Lüeh-yang and Ch’ien-wan of Ch’ou-ch’ih each controlled tribes numbering more than a myriad [10,000].
They made common cause with the rebel Ma C’hao 馬超 (176-222; SKC 36. 6b-9a) in 213 (SKC 1, 38a; Wei-
lüeh’s 211 seems to be wrong). After Ma Ch’ao’s defeat in the following year, A-kuei and his followers were
wiped out by Hsia-hou Yüan 夏侯淵 (d. 219; SKC 9, 3b-7b), while Ch’ien-wan went southwest into Shu,
where his tribes all submitted to Han authority. The latter deported those whose conduct had been
equivocal, placing them in Mei-yang; at the time of the composition of the Wei-lüeh these were under the
jurisdiction of a Protecting Army (hu-chün). Those who had “conducted themselves wisely” were left in
Nan-an and T’ien-shui commanderies.” Rogers (1968), p. 82, n.12.
“By 214, however, Ma Chao was finally driven into exile in Sichuan, and in the same year Han Sui and his
allies, who included the Shaodang Qiang, were decisively defeated at the Changli River in Hanyang,
identified with the present-day Hulu River, north of Tianshui. In the winter of 214, Xiahou Yuan followed
that success with a campaign against Song Jian. Song Jian died, his capital at Fuhan was captured, and all
 his officials were killed. With this victory, the whole territory east of the Yellow River was in the control of
Cao Cao.” de Crespigny (1984), p. 165.
1.15. For a more detailed description of these events see: Haloun (1949-50), pp. 126-130; de Crespigny (1984), pp.
162-172. Shu 蜀 = modern Western Sichuan.
“Shu designates the western part of presentday Ssu-ch’uan province, the Red Basin around Ch’eng-tu, or
the commandery of that name. . . . ” CICA, p. 220, n. 829
“(Szechwan) A state of high antiquity traditionally thought to stem from the enfeoffment of a cadet line of
the descendants of Ti-kao as Marquis of Shu. Used specifically as an area designation for the part of
Szechwan centering on Ch’eng-tu and generally for the whole Szechwan basin.” Rogers (1968), p. 331.
“6. (Hist. Geog.) Shu : a. Anc. name of the region corresponding to the central part of 四川 of modern Szu-
ch’uan (Sichuan). b. A state annexed by the state of 秦 Ch’in [Qin] during the 戰國 [403-222 CE] era of the
Warring States and corresponding to the plain of 成都 Ch’eng-tu (Chengdu), in 四川 modern Szu-
ch’uan. Against this state a military operation was organised by 孝公 Hsiao Kung [Xiao Gong] (381-338
BCE) of the state of 秦 Ch’in, which prepared the way for the great military expeditions of the 秦 Ch’in
and 漠 Han empires. c. chün4 or commandery, under the 秦 Ch’in and Han, corresponding to the anc.
State.” GR Vol. V, No. 9891.
1.16. Meiyang 美陽.
“Meiyang was a county (xian) [縣] of the Commandery of you Fufeng (that is to say, the Fufeng to the right,
or west, of the capital). It is to the southwest of the present District of Wugong (Jian County, Shensi
Province).” Translated from Chavannes (1905): 522, n. 4. Meiyang was approximately 100 km northwest of
Changan. See also: Rogers (1968), p. 82, n. 12.
1.17. Anyi 安夷 [An-i] literally, ‘Peaceful Yi’ and Fuyi 撫夷 [Fu-i] literally, ‘Governed Yi’.
“The I, “barbarians of the east” settled in southern Shang-tung, northern Kiang-su and northern An-hui,
had generally faded before the Chou epoch and the last remnants disappeared in the IIIrd century B.C.
(Eberhard, Kultur und Siedlung, p. 385ff). As is evident here, the name can only have the very general
meaning of “barbarians”.” Molè (1970), p. 86, n. 61.
1.18. Hujun 護蕈 [Hu-chün] = Military Protector.
“(1) HAN: Military Protector, briefly from A.D. 1, an officer on the staff of the Defender-in-chief (ta ssu-ma),
one of the eminent Three Dukes (san kung); rank apparently 2,000 bushels, but functions not clear; not
continued in Later Han. . . .  HB: commissioner over the army.” Hucker, No. 2775.
Professor Dubs translates hu-chün as ‘The Protector of the Army’. See: de Crespigny (1967), p. 12. I think it is quite
clear that the title refers here to a military officer supervising previously pacified tribes. I have, therefore, used
Hucker’s suggestion of ‘Military Protector.’
1.19. Tianshui 天水 [T’ien-shui].
“The centre of Tianshui Commandery was to the southwest of the present Tong County (Gongchang
Prefecture, Gansu Province).” Translated from Chavannes (1905), p. 523, n. 1.
“During the Han period, under the arrangements of Emperor Wu, the commandery of Tianshui was
established to control the Wei valley immediately west of Long Mountain. . . . ” It was renamed Han-yang in
AD 74. de Crespigny (1984), pp. 13, 70.
“The commandery of Tianshui (later Hanyang), in the upper Wei valley, appears to have been regarded as
marginal territory between East and West.” de Crespigny (1984), p. 57.
1.20. Nanan 南安 [Nan-an].
Chavannes (1905), p. 523, n. 2, places Nanan County in Jianwei Commandery in present Sichuan. This must be a
mistake. The reference here is undoubtedly to Nanan Commandery that was established in CE 188 to the west of
Hanyang Commandery. See the map in de Crespigny (1984), p. 148.
“. . .    the city of Nan-an (near Kung-ch’ang). . . . ” Kung-ch’ang is to the south of Lan-chou, near Min in the
lower valley of the T’ao shui (western Kan-su). Molè (1970), pp. 126, n. 236, and 86, n. 59.
1.21. Guangwei 廣魏 [Kuang-wei] Commandery.
“The text reads thus: 今之廣平魏郡所守是也 [jinzhi guangpingwei jun suoshoush ye]. But further on, one
finds the name 廣魏郡 Guangweijun, which proves that the word 平 [ping] is superfluous here and, in
fact, the introduction of this parasite character renders the text unintelligible. Meanwhile, the name of the
Guangwei Commandery is itself very puzzling, for it does not figure in the geographical chapters of the
Jinshu (chaps. XIV and XV) and, consequently, is not found in Li Zhaoluo’s dictionary of historical
geography which is solely based on the canonical historians. By good luck, the geographical chapters of the
Jinshu were made the object of a study by Bi Yuan 畢沅 published in 1781 under the title of 晉書地理志
新補正 “The Geographical Treatise of the Jinshu Recently Completed and Corrected”. It is in this work
that we finally find the desired solution. Indeed, we read there (chap. I, p. 4 b of the reprint made in the
Jingxuntangcong shu 竞訓堂从叢書:” Emperor Wu (= Cao Cao, who lived from 155 to 220), of the Wei
dynasty, established Guangwei 廣魏 Commandery. Under the Jin, during the Taishi period (265-274), this
name was changed for the first time to Lüeyang 略陽.”
          The name of Lüeyang Commandery is, in fact, raised again in the Jinshu (chap. XIV, p. 15, b) with a
commentary confirming that the name of Lüeyang was formerly Guangwei. Li Zhaoluo’s dictionary,
furthermore, informs us that the administrative centre of Lüeyang Commandery (ancient Guangwei), was
90 li [37 km] to the northeast of the present Qinan 秦安 County (Qin 秦 Prefecture, Gansu Province).
Thus, the site occupied by Guangwei Commandery is resolved. – We remark, incidentally, that the use Yu
Huan made of the name of Guangwei, which was only in use from about 220 to 265, is in perfect agreement
with the approximate date assigned by Liu Zhishi to the work written by this author.” Translated from
Chavannes (1905), p. 523, n. 3.
1.22. The Qiang 羌 [Ch’iang]. See note 3.1.
1.23. The Hu 胡 peoples. Hu is a rather vague term used for northern and western peoples of non-Sinitic origin,
usually, but not exclusively, for those of Caucasian appearance. It was commonly used for people of Persian,
Sogdian, and Turkish origin, Xianbi, Indians, Kushans and even, occasionally, for the Xiongnu (who, however, are
usually clearly differentiated from the Hu).
          Pulleyblank (1991), p. 126, gives as definitions of hu 胡: “dewlap; interrogative adverb, why?, how?; general
name for horse-riding nomads (Han); for Iranians from Central Asia (Tang); foreign, western.”
          On this subject I am greatly indebted to Thomas Bartlett of Latrobe University in Melbourne, Australia who, in
a private communication dated 14th April 2002, wrote:
“Etymologically, the graph “hu(2)” means literally “old flesh.” According to the Shuo-wen dictionary, this
character originally meant the dewlap – loose skin that hangs down from the neck of an ox. Later the word
was used to refer to certain central Asian peoples, perhaps because the heavy beards of such men may have
seemed to Chinese to resemble the flesh hanging down from the necks of oxen (dewlaps). Such central
Asian peoples were usually cattle-raising pastoralists, so the classifying epithet “Hu” may have pejoratively
implied a fundamental affinity between the people and their animals. Many ancient Chinese names for
foreign peoples had such uncomplimentary implications.
          Later, “Hu” in China became a general term for “northern barbarians,” whether they were of central
or east Asian origin. This usage apparently has less to do with beardeness than with non-Chineseness.
Thus, for example, Han period historians refer to a certain Tungusic people as “eastern Hu.” In the 17th
and 18th centuries, one purpose of the literary inquisition carried out in China by the Manchus, a Tungusic
people not known for heavy beards, was to remove derogatory remarks to “Hu” in extant Chinese literary
works.”
Hulsewé’s translation of hu as “nomad” (CICA: 80, n. 71) cannot be justified, as his own translation from the Hanshu
on the state of Xiye [Hsi-yeh] shows. Ibid. p. 101. To say that the people of a ‘land of nomads’ are ‘different from the
nomads’ is meaningless.
          I have translated the word hu here as “Westerner” with considerable hesitation, but feel that this, at least,
closely represents the way it is used in the Hou Hanshu and the Weilue.
1.24. The ‘renlu’ 衽露 [jen-lu].
Ren refers to the front of a garment, a skirt, or the lapel or flap in front of a coat which is buttoned under the right
arm, and lu means to bless, or blessed. The ‘renlu’, therefore, is likely to have been similar to the pang-gdan, or
striped apron, which Tibetan women, to this day, wear from the time they are married.
1.25. Jie 街 [Chieh].  
“The locality of Jie is mentioned further on as adjoining the ancient Commandery of Wudu and the District of
Yinjin. It is, therefore, probably necessary to read 邽 in place of 街 and to identify this place with the District of
Shanggui 上 邽 which was a dependency of the Commandery of Longxi, and which was to the southwest of the
present Secondary Prefecture of Qin 秦, in Gansu Province.” Translated from Chavannes (1905), p. 524, n. 2. See also
note 1.32.
1.26. Ji 冀 [Chi].
“Ji was a District of the Commandery of Tianshui. This is nowadays the District of Fuqiang (Gongchang
Prefecture, Gansu Province).” Translated from Chavannes (1905), p. 524, n. 3.
Ji was also the name of the capital of Tianshui Commandery (renamed Hanyang Commandery in CE 74), and was
west of the Long Pass on the main route to the northwest from Changan. See the map in: de Crespigny (1984), p. 92.
1.27. Huandao 獂道 [Huan-tao].
“The District of Huandao (the correct orthography is 豲道) was in the Commandery of Tianshui. It was to
the northeast of the present District of Longxi 陇西隴 (Gongchang Prefecture, Gansu Province).”
Translated from Chavannes (1905), p. 524, n. 4.
1.28. The fiefdoms du 都 [tu]. The word is used here meaning a fiefdom. Although it often refers to a large city or
capital, it can also mean a fief granted to a prince. See, for example, Williams (1909), p. 846; GR Vol. VI, No. 11668,
4b.
1.29. The ‘Commanderies and Kingdoms’ shaoguo 郡國 [shao-kuo].
“. . .    this term designates the principal administrative divisions of the Chinese who had divided all their
territory into a certain number of Commanderies and Kingdoms.” Translated from Chavannes (1905), p.
524, n. 5. Also, see the discussion in: de Crespigny (1984), pp. 1-3.
1.30. Wudu 武都 [Wu-tu]. See note 1.4.
1.31. Yinping 陰平 [Yin-p’ing] 陰 平
“The District of Yinping was to the northwest of the present District of Wen (Jie Prefecture, Gansu
Province.” Translated from Chavannes (1905), p. 525, n. 2.
1.32. Chavannes interpreted this passage somewhat differently than I have, thus:
“Regarding those who live (in the region of) Jie 街, Ji 冀, and Huandao 獂道, although they are now under
Chinese administration, they have, nevertheless, preserved their kings and chiefs who live in their
territory and among their tribes. Besides, in the ancient region of Wudu, in the region of Yinping 陰平
and Jie 街, there are some tribes numbering more than 10,000 men.” Translated from Chavannes (1905),
pp. 524-525. See also n. 1.25.                                                                                                                                           
Section 2 – The Zilu Tribes
The Zilu1 originated among the Xiongnu.2 Zi is the name the Xiongnu used for slaves.
Formerly, in the Jianwu period [CE 25-55], the Xiongnu became weak and dispersed. Their
slaves fled and hid in the region of Jincheng, 3 Wuwei,4 and Jiuquan (Commanderies),5 north
to Hei Shui (‘Black River’) and Xi He (‘Western River’). 6
Wandering from east to west, they care for their flocks by leading them in search of water and
pasture. They make raids on the territory of Liangzhou. 7 Their tribes have progressively
increased until they number several tens of thousands of men.
They are not the same as the eastern tribes, who are the Xianbi. 8 They are not all of one
race. There are Dahu,9 Dingling,10 and also quite a large number of Qiang living among them.
This is because they were originally slaves of the Xiongnu.
During the period (at the end) of the Han and (the beginning) of the Wei [circa 220 CE], one of
their great chiefs was named Tantuo.11 After he died, some great chiefs, descendants of his,
were living to the south near the frontier of Lingju (Prefecture) in Guangwei (Commandery). 12
There was Tugui13 who came (to invade our territory), and rebelled several times. He was
killed by (the Prefect of) Liangzhou. Now there is (the chief named) Shaoti. 14
Sometimes these tribes come to submit, sometimes they withdraw in hiding. They often make
trouble on the routes to the western provinces. 15
 Section 2 – The Zilu Tribes
2.1. The Zilu 貲虜 [Tzu-lu]. 
zi 貲 – ‘ransom,’ ‘property,’ ‘valuables’ – not in K; EMC: tsia̭ /tsi
lu 虜 – ‘capture’, ‘captive’. K. 69e *lo / luo; EMC: lɔ’
The Zilu gradually grew into a powerful state centred around (Lake) Koko Nor. They were later known to the
Chinese as the Tuyuhun [T’u-yü-hun], and to the Tibetans as the ‘A-zha.
          After many years of warfare, they were decisively defeated by the Tibetans in 663, and never recovered their
independence. Some of them fled to the Chinese, others remained, and were gradually absorbed by the Tibetans.
See: Molè (1970), pp. 2, 30, and 73, n. 22.
2.2. Xiongnu 匈奴 [Hsiung-nu].
xiong 匈 – ‘breast’, ‘heart’. K. 1183d *χįung / χįwong; EMC: xuawŋ.
nu 奴 – ‘slave’, ‘dependents’, ‘wife and children’. K. 94l *no / nuo; EMC: nɔ
“As the Hsiung-nu empire continued to expand throughout Mongolia, Sinkiang and parts of Manchuria, the
Han rulers were forced to greater efforts to defend themselves and garrison the silk route. During the early
Han the throne had become so weakened by internal revolts that a takeover by the Hsiung-nu would have
been possible, but the Shan-yü may have avoided the temptation in the belief that conquest would only
 serve to alienate them completely and that permanent control would be impossible. The Shan-yü may not
have been quite so astute, but all the alien powers that have gained the Chinese throne were either forced
to retreat or were eventually assimilated.
          Toward the end of the Han, the loosely organized Hsiung-nu empire fell apart through over-extension
and internal dissensions. The northern faction broke away to appear later in the Orkhon-Selenga region,
while the southern faction fled further southward. The northerners were shortly faced with the emergence
of the Hsien-pi, who became powerful contenders for control of the northern steppes. For a time they held
their own, but eventually they were forced to retreat westward. The defeat of the northern Hsiung-nu and
their flight to Ili, the land of the Wu-sun – whom they defeated - marks the beginning of the western
Hunnic empire and the pillaging of Europe under Attila in the fifth century AD.” Bowles (1977), p. 260.

“The Hiung-nu headed a powerful alliance of stock-raising tribes in the late 3 rd – the early 2nd century B.C.
and dominated the eastern part of Central Asia during two centuries, laying the foundations for the
emergence of tribal alliances there in the Middle Ages. The military-political events of Hiung-nu rule are
well-known from written sources but the origin of the Hiung-nu themselves and the early stages of their
history remain obscure to this day. It is difficult to “picture and expound consistently” all those stages, as
Ssu-ma Ch’ien, a contemporary of the Hiung-nu, pointed out in his time. Having rounded up all the
information about that tribe, the great historiographer of old, remarked only that “the Hiung-nu
descended from Shun-wei, a scion of the Hsia rulers’ family.” The evidence of written sources alone is not
sufficient to resolve the above question.” Minyaev (1985), p. 69.
“Hsiung-nu is the designation for the nomad tribes living to the north of China ; they have often, but by no
means certainly, been identified with the Huns ; see Sinor (1963), p. 263 (cf. p. 220) for the literature on the
point ; cf. also Pulleyblank (1963), p. 39, for further identifications.” CICA, p. 71, n. 4.
2.3. Jincheng 金城 [Chin-ch’eng] Commandery was to the east of the salt (Lake) Koko Nor and had its centre to the
northwest of the prefectural capital of Lanzhou (Gansu Province). See: Chavannes (1905), p. 525, n. 7.
“It was not until the time of Emperor Wu, shortly before 100 BC, that the Han established a military and
political presence northwest across the Yellow River and founded the commanderies of Hexi “West of the
River.” Jiuquan, Zhangye, and Dunhuang were probably established in 104 and subsequent years, Wuwei
and Jincheng in the half-century following.” de Crespigny (1984), p. 7.
“It is not certain where the Silk Road from China crossed the Yellow River during Han times, but it was
surely in the region of present-day Lanzhou; and this frontier place of early Han became the base for
expansion to the west of the river. The commandery of Jincheng was not formally established until 81 BC,
but administrative and political control had been maintained for a generation before that time, based
notably upon the garrison city of Lianju, on the Datong River, about a hundred kilometres northwest of
present-day Lanzhou. . . .
          The establishment of Jincheng commandery, therefore, served two purposes: firstly as the base for
the communications line across the Yellow River which led north through Wuwei and into central Asia;
second as an area for colonisation by the Chinese, north of the Yellow River and among the Xining valley.”
de Crespigny (1984), pp. 11-13.
“Around 60 B.C., the Han extended the name Chin-ch’êng to cover the whole region inhabited by the
Ch’iang between Huang-ho and the Kuku-nor and set up a protectorate there centred upon present-day
Lan-chou.” Molè (1970), p. 92, n. 88.
2.4. Wuwei 武威 [Wu-wei] Commandery was situated to the west of the Tengger Desert, its centre at what is now
modern Wuwei (seat of Liangzhou prefecture since Tang times), in Gansu Province.
“For the dates of the establishment of the four commanderies of the north-west, see Hulsewé (1957), pp. 6-
7, and RHA I pp. 59-60, where it is concluded, tentatively that (i) Chiu-ch’üan and Chang-i were established
in 104, (ii) Tun-huang was established shortly afterwards, at least before 91 B.C.; (iii) Wu-wei was probably
set up between 81 and 67, although minor administrative units had existed there previously. See also
Chang Ch’un-shu (1967), p. 748 : Chiu-ch’üan 111 B.C.; Chang-i between 111 and 109 B.C.; Tun-huang
between 101 and 94 B.C. ; Wu-wei c. 7 B.C.” CICA, p. 75, n. 40.
2.5. Jiuquan 酒泉 [Chiu-ch’üan] Commandery was centred where modern Jiaquan is now, just to the southeast of
Jiayuguan [Chia-yü-kuan]. Situated on the main road to the west, it also protected the approaches to the strategic
Etsin Gol delta to the northeast. See CICA, p. 75, n. 40 in note 2.4.
2.6. The Hei Shui 黑水 [Hei Shui] literally ‘Black River;’ and the Xi He 西河 [Hsi Ho] literally ‘Western River.’
“It appears the Hei shui can be identified with the Dang He, or the Shazhou [Dunhuang] river. Cf. Sima Qian,
French trans., bk. I, p. 126, n. 2. The term ‘Hei shui’ designating a river, and not an administrative district. It
must be the same with the term ‘Xi He’ which cannot apply here to the Commandery of Xi He straddling
the Huang He in the north of Shanxi and Shenxi. I therefore consider the Xi He in our text as being the
western branch of the great loop of the Huang He. The domain of the Zilu is thus bordered by Shazhou
[Dunhuang] to the west, and the Helan Shan massif to the east.” Translated from Chavannes (1905), p. 525,
nn. 5 and 6.
While I agree with Chavannes that these are references to rivers, not administrative areas, I do not agree with his
identifications. I believe it is far more likely that they refer to the upper and lower reaches of the Ruo Shui or Etsin
Gol (Etsin River), which flows north into two lakes, the Sogo Nur (Sokho Nōr) and the Gaxun Nur (Gashun Nōr).
          The well-watered valley of the Ruo Shui led directly southwest from the homelands of the Xiongnu would
have provided the shortest and most practicable escape route for their slaves to reach Chinese-controlled territory.
          The river which flows from the Loulang Nanshan and Lenglong ranges across the ‘Gansu corridor,’ through
ancient Jiuquan [Chiu-ch’üan] Commandery to the north and west of Zhangye [Chang-yeh ; formerly known as Kan-
chou], to join the river at the entrance to the Etsin Gol valley, is known to the Chinese as the Hei He.
          Still, today, it is called the Hei He on some maps as far as the Heli Shan range where it changes its name to the
Ruo Shui. This is undoubtedly the Hei Shui of this text (the characters shui 水 and he 河 are often used
interchangeably for river).
“The Edsin Gol springs from two sources, in the Kan Chou and Hsü Chou oases at the foot of the Nan Shan.
After watering several minor oases, of which the chief is Chin-t’a (the Golden Pagoda), they unite near
Mao-mei, off the southeastern extremity of the Pei Shan. Thence they flow somewhat east of north, as the
Hei Shui or Black Water, crossing the indeterminate borders of Kan-su and entering Inner Mongolia. The
Hei Shui then separates into the Eastern and Western Edsin Gol, which, after reaching Outer Mongolia, end
in two communicating lakes or meres, Gashun Nor and Sokho Nor.” Lattimore (1929), p. 205.
Further north, the Ruo Shui splits into the Xi He (‘Western River’ – also known as Mörün-gol or Ar-gol), and the
Dong He (‘Eastern River’ – also known as Ümne-gol or Iké-gol), before flowing into the two lakes. The former, the Xi
He, is almost certainly the same river as the Xi He of the Weilue. I should note here, however, another possibility: Xi
He [Hsi Ho] is also used to refer to “the eastern vertical leg of the great bend of the Yellow River.” Rogers (1968), p.
84, n. 44.
The importance of this region, at the junction of the main ‘Silk Route’ to the west, and the easiest, and most direct
route to central Mongolia, is hard to overestimate:
“Nature, by affording water and grazing over a continuous line of some two hundred miles, has at all times
provided in the valley of the Etsin-gol an exceptionally easy route for raids and invasions from the Altai
region, that true home of the Mongols and other great nomadic races, towards the line of the westernmost
oases of Kan-su. These, extending along the foot of the Nan-shan, constitute the great natural highway
between China and innermost Asia. Wide belts of desert and barren hill-ranges stretch both to the west and
the east of the Etsin-gol. These belts, very difficult for any large bodies of men to cross, hardy nomads
though they may be, help to protect this important ‘corridor’ for trade and military operations against
serious attack from the north. But the valley of the Etsin-gol stands open, like a gate inviting invasion. . . . It
will suffice to point out that those who since the first Chinese advance under the Emperor Wu-ti into ‘Ho-
hsi’ were concerned with the safeguarding of this indispensable passage land between China and Central
Asia, were not likely to ignore or neglect the advantage that a cultivated area, well to the north of the great
highway and yet easily capable of support from the side of both Su-chou and Kan-chou, would necessarily
present for the purpose of a barrier whereby to close that gate against inroads, or as an advanced base for
offensive movements against nomadic hosts.” Stein (1928): Vol. I, pp. 409-410.
“It is here that the route of invasion from the Mongolian steppes cuts through the ancient border
line drawn by the Chinese when they first occupied the passage land to the north of the Nan-shan. The
ruined forts of imposing size and evident antiquity which we found here on both sides of the river were, no
doubt, intended to guard the gateway for invasion here presented. One fort built with clay walls of
exceptional strength looked an exact counterpart of the ancient frontier post of the “Jade Gate” as located
by me seven years before on the Limes in the desert west of Tun-huang.
            As we moved down by the Etsin-gol from that last outlying Chinese settlement we found the sandy
bed of the river nearly a mile wide in places but absolutely dry at the time. Only at rare intervals could
water be obtained from wells dug in deep hollows below the low rocky spur thrown out by the Pei-shan
and then spreads out in a delta extending for some 110 miles to the north before it terminates in a line of
brackish lakes and marshes.
            The conditions brought about here by a succession of low-water seasons furnished a striking
illustration of the appearance which the Lou-lan delta may have presented before the Kuruk-darya had
finally dried up. Where river beds lined by narrow belts of jungle had been left dry for long years, we found
many of the wild poplar trees already dead or dying. The wide stretches of ground separating the several
beds showed but scanty scrub or else were absolutely bare. No wonder we heard sad complaints in the
scattered camps of the two-hundred-odd families of Torgut Mongols which are established in the Etsin-gol
delta, and about the increasing difficulties caused by inadequate grazing. Yet this extensive riverine tract,
limited as are its resources, must always have been of importance for those, whether armed hosts or
traders, who would make the long journey from the heart of Mongolia in the north to the oases of Kansu.
The line of watch towers of later construction met at intervals afforded proof that this route into Mongolia
had been frequented and guarded during late medieval times.
            The analogy thus presented with the ancient Lou-lan delta impressed me even more when I
proceeded to examine the ruins of Khara-khoto, the “Black Town,” which Colonel Kozloff, the
distinguished Russian explorer, had been the first to visit in 1908–09. There remained no doubt for me then
that it was identical with Marco Polo’s “City of Etzina.” Of this we are told in the great Venetian traveler’s
narrative that it lay a twelve day’s ride from the city of Kan-chou, “toward the north on the verge of the
desert; it belongs to the Province of Tangut.” All travellers bound for Karakorum, the old capital of the
Mongols, had here to lay in victuals in order to cross the great “desert which extends forty days’ journey to
the north and on which you meet with no habitation nor baiting place.”
            The position thus indicated was found to correspond exactly to that of Khara-khoto, and the
identification was completely borne out by the antiquarian evidence brought to light at the ruined site.
This soon showed me that through the walled town may have suffered considerably, as local Mongol
tradition asserts, when Chingiz Khan with his Mongols first invaded Kansu from this side about A.D. 1226,
yet it continued to be inhabited down to Marco Polo’s time and at least partially even later, down to the
fifteenth century. This was certainly the case with the agricultural settlement for which it had served as a
local center, and of which we discovered extensive remains in the desert to the east and northeast. But the
town itself must have seen its most flourishing times under the Tangut of Hsi-hsia rule from the beginning
of the eleventh century down to the Mongol conquest. . . .
            There was much to support the belief that the final abandonment of Khara-khoto was brought about
by difficulties of irrigation. The dry river bed which passes close to the ruined town passes some seven
miles away to the east of the nearest branch still reached by the summer floods. The old canals we traced
leading to the abandoned farms eastward are removed considerably farther. It was impossible definitely to
determine whether this failure of irrigation had been brought about by a reduction in the volume of the
Etsin-gol’s water or had been caused by a change in the river course at canal head with which the
settlement had for some reason been unable to cope. Anyhow, there seemed good reason to believe that
the water supply now reaching the delta during a few summer months would no longer suffice to assure
adequate irrigation for the once cultivated area. Even at Mao-mei oasis, over 150 miles farther up the river,
and with conditions far more favourable for the maintenance of canals, serious trouble had been
experienced for years past in securing an adequate supply of water early enough in the season. Hence,
much of the once cultivated area had been abandoned.” Stein (1931), pp. 188-191.
“In one part of this region, now known by the Mongol name of Edsin Gol, the Ruo Shui flows past the
present city of Jiuquan for more than three hundred kilometres into the desert. Nowadays, the Edsin Gol
provides little more than brackish water and salt pans, but in the time of Han these marshes were fertile,
and abundant with wild life. The whole river system then provided a salient of arable land stretching into
the heart of the desert.
          This territory, called Juyan by the Han Chinese, was maintained and garrisoned by the empire from
 the time of Emperor Wu till the last century of Later Han. Militarily, the outpost of the Great Wall was
important for two reasons: as a supply point for the garrisons in the northwest and, perhaps more
significant, as a means to deny this prosperous region to the northern nomads. Left undefended, Juyan
would have provided an ideal route for attack against the Chinese commanderies of the corridor itself.” de
Crespigny (1984), p. 9.
2.7. Liangzhou 涼州 [Liang-chou].
“The Liang province of Later Han was divided in two by the Yellow River, flowing eastwards from the
Tibetan massif and then north towards the desert land of the Ordos. In this region, unlike other territories,
the Yellow River was of only minor importance as a communications route: its valley and its waters
provide some opportunity for travel upstream or down, but river transport is generally practicable only
during the high water of summer, and there were, in any case, few places of interest or value along the
stream. On the contrary, in fact, in the time of Qin and at the beginning of the Former Han the Yellow River
served as the frontier line of the empire, and during later centuries it was a barrier to overcome for
communication between China and central Asia.” de Crespigny (1984), p. 7.
2.8. The Xianbi or Xianbei 鮮卑 [Hsien-pi or Hsien-pei] people.
“The Hsien-pi, who took over control of Mongolia after the fall of the Hsiung-nu state, had emerged as a
powerful tribal union as early as the first century B.C. The main clan of the Hsien-pi had set up their
nomadic camp in south-east Mongolia and lived along the middle course of the Liao-ho river. A large
number of Hsien-pi now settled in central Mongolia and over 100,000 Hsiung-nu families, who had settled
there earlier, adopted their tribal name. T’an-shih-huai, leader of the Hsien-pi tribal union, in A.D. 155
established the Hsien-pi state, which rapidly became one of the most powerful empires of its day, as
powerful as the previous Hsiung-nu Empire. The Han court considered that the Hsien-pi’s horses were
swifter and their weapons sharper than those of the Hsiung-nu, and the Hsien-pi, too, managed to acquire
good-quality iron from the border regions of China. Their political centre, the headquarters of T’an-shih-
huai, was in the south-east near the Darkhan mountains but was later moved to the former shan-yü’s
headquarters in the Khangay mountains.
          Between A.D. 155 and 166, T’an-shih-huai conducted a series of major military campaigns that led to
the extension of Hsien-pi power over the Great Steppe as far as southern Siberia and from Ussuri to the
Caspian Sea. Until the third decade of the third century A.D. the Hsien-pi was the leading power in Central
Asia.” Ishjamts (1994), pp. 155-156.
“The place of origin of the Hsien-pi as a political force was in the Khingan range area of the upper Amur
basin, in a region inhabited later by speakers of Tungusic languages; for this reason it was assumed that
they were also Tungusic, but more recent research links them with the Mongols. . . .
          Having defeated the northern Hsiung-nu in AD 166, the Hsien-pi shifted to the Orkhon-Selenga basin
just west of the Amur in northern Outer Mongolia. There they formed the nucleus of what became within a
few centuries the Mongol empire. Just as the control of Sinkiang shifted from Indo-European speakers to
the Turks with the defeat of the Yüechih-Tochari between 174 and 161 BC, so the control of Mongolia
switched from the Turks to the Mongols.” Bowles (1977), pp. 260-261
“The histories are agreed that the manners and customs of the Xianbi were very close to those of the
Wuhuan (HHS 90/80, 2985, SGZ 30, 836, commentary quoting the Wei shu of Wang Shen; Schreiber, “Hsien-
pi”, 147 ff. and 162-163. . . .
          The Xianbi are said to have taken their name from the mountain called Xianbi, now identified as a
peak of the Great Xingan range, west of the Horqin/Khorqin West Wing Centre banner in Kirin (Gezu jianshi,
46). It is equally possible, however, that the mountain took its name from the tribe. . . . ” de Crespigny
(1984), p. 524, n. 12.
          “Initially, the socio-political institutions of the Hsien-pei centred around Hsien-pei clan and tribal
alliances, but the Hsien-pei ultimately sought to create a nomadic “state on horseback.” As the Northern
Hsiung-nu retreated, relations between the Hsien-pei and other nomadic groups became more volatile,
with the Hsien-pei staging frequent raids to secure essential goods and expand their grazing areas and
power. “[The Hsien-pei] looted along the [Han] border on the south, resisted the Ting-ling on the north,
repulsed the Fu-yü on the east, attacked the Wu-sun on the west, and occupied all the old territories of the
Hsiung-nu, all of which spanned [approximately] four thousand li [1,663 km] east to west and seven
thousand li [2,911 km] south to north.”39
          Hsien-pei power solidified under the leadership of Tan-shih-huai (d. 181 A.D.), and the Han dynasty
viewed their growing strength as a threat. During the reign of Emperor Huan-ti (r. 147-167 A.D.), the court
responded by ordering Lieutenant General Chang Huan, who was responsible for the supervision of the
Southern Hsiung-nu, to attack the Hsien-pei, but he failed to subdue them. The court thereupon sent an
envoy with a seal and sash to confer the title of prince on Tan-shih-huai and propose a marriage alliance
with him. Tan-shih-huai rejected the offer, and the border intrusions and lootings worsened. 40
          . . . .  As a powerful nomadic force north of China during the decline of the Han, the Hsien-pei became
deeply involved in the struggle for power in China as that dynasty disintegrated. Hsien-pei’s relations with
Ts’ao Ts’ao, the founder of the Wei dynasty (220-264 A.D.) of the Three Kingdoms period (222-280 A.D.),
suggests that the pattern of interaction between the nomad and agriculturalist courts changed very little
during he middle of the third century. The Hsien-pei leader, K’o-pi-neng, initially allied himself with Ts’ao
Ts’ao and aided him in pacifying a rebellion led by T’ien Yin in present-day Hopei. He then joined with the
Wu-huan when they revolted against Ts’ao. K’o-pi-neng’s forces were defeated by Ts’ao’s, and he was
forced to retreat north of the Great Wall. However, he soon sent tribute to Ts’ao’s Wei dynasty in northern
China and sought to establish peace. The Hsien-pei leader’s desire to normalize relations suggests that he
was faced with the same economic difficulties that the Hsiung-nu had suffered. After he and his people
withdrew from close proximity to arable territory toward more remote areas, it was difficult to acquire
agricultural products, and during this time, the Hsien-pei were not strong enough to breech and loot the
powerful Wei borders. Consequently, K’o-pi-neng was forced to negotiate peace with the Wei. The Wei
court, eager to gain the Hsien-pei as allies, granted K’o-pi-neng the title Fu-I Wang (“Prince of Upright
Subordination”), thereby suggesting his subordinate status to the dynasty. K’o-pi-neng accepted the title
and took advantage of renewed ties with the Wei to lead three thousand horsemen in driving twenty
thousand horses and oxen to the border markets for exchange.” 42
39. [Hou Han shu] 90, chüan 80, “Account of the Hsien-pei,” 9b.
40. San-kuo chih, “Book of Wei,” 30, “Account of the Hsien-pei,” 6a.
41. Ibid., “Biography of K’o-pi-neng,” 8b-9a.
42. Ibid., 8a.
Jagchid and Symons (1989), pp. 34-37.
          “In transcriptions of the Later Han period we begin to find cases of *kh- and we find *th also in the
transcriptions of the Chin-shu. These my perhaps reflect increasing penetration and admixture with the
Eastern Hu, that is the Hsien-pi and Wu-yüan 烏桓 (or 丸) [should be read Wu-huan (or –wan)] M. ‧ou-
h̑wan  <  *‧aĥ- ĥwan = Avar, who probably spoke a Mongolian type of language. It was the Hsien-pi who
became dominant on the steppe after the collapse of the Hsiung-nu empire in the second century A.D.”
Pulleyblank (1963), II, p. 242.
“The Wu-huan and the Hsien-pei people, who created the most powerful nomadic states after the Hsiung-
nu decline, also traded their livestock and furs: “In the twenty-fifth year [of Chien-wu, 4 AD], Hao-tan, the
leader of the Wu-huan on the west of the Liao [River] and others . . . admired [our] culture. They led their
people to the court and presented their tribute, male and female slaves, cattle and horses, bows and the
furs of tigers, leopards and sables.” And again, “The Hsien-pei are a branch of the Eastern Barbarians
[Tung-hu]. . . . Their animals, which are different than those of the Middle Kingdom, are wild horses, great
horned goats, and chiao-tuan cattle. The bow made from horns is commonly known as the chiao-tuan bow.
Besides, there are sables, na [= seals – presumably the seals from Lake Baikal], and ermines. Their skin and
hair are tender and soft and they are known as the best furs under heaven.” Jagchid and Symons (1989), p.
167. From the Hou Hanshu, zhuan 80, “Account of the Wuhuan”, 5a.
“When the ruler of the Northern Hsiung-nu was beaten by Chinese forces in 91 and fled in an unknown
direction, a new people, the Hsien-pi, took the opportunity to migrate, and settled on his territories. The
remaining Hsiung-nu clans, which numbered more than 100,000 yurts, began to call themselves Hsien-pi,
and from that time on the Hsien-pi began to gather strength.
 According to the Chinese chronicles, the Hsien-pi originated in a land of forests and high mountains near
the basin of the River Amur. Their language and customs are described as similar to those of the Wu-huan,
except that before a wedding they first shaved their heads, then held a large assembly on the river during
the last month of spring; they feasted, and once the feasting was over, celebrated the marriage. Wild birds
and beasts not found in the Middle Kingdom of China lived in the territories of the Hsien-pi, who made
bows out of horns. There were also sables, foxes and squirrels with soft fur, from which fur coats renowned
for their beauty were made in the Celestial Kingdom. The breeding of cattle, sheep, goats and horses by the
Hsien-pi is also mentioned and they are said frequently to rustle each other’s herds of livestock and horses.
          The Hsien-pi were described by one of the Chinese emperor’s councillors in 117 as follows:
After the Hsiung-nu fled, the Hsien-pi, who took over their former territories, grew in strength. They have
hundreds of thousands of warriors, they are remarkable for their physical strength, and are more quick-
witted than the Hsiung-nu. It should also be noted that, as a result of lack of discipline at the guard-posts
on the line of fortifications, there are many ways of evading the embargo, which robbers use to obtain fine
metal and iron of good quality. The Chinese get in [through these gaps] and become the main counsellors
of the Hsien-pi, and so they acquire keener weapons and faster horses than the Huns.
During the reign of the Han emperor Huang-ti (146-168), an energetic leader named T’an-shih-huai
appeared among the Hsien-pi. He subjected the elders to his authority, introduced laws, gathered large
forces and defeated the Northern Hsiung-nu around 155.
All the elders of the eastern and western nomadic communities submitted to him. As a result of this he
looted the lands along the line of fortifications, repulsed the Ting-ling in the north, made the Fo-yü
kingdom retreat in the east, attacked the Wu-sun in the west, and took possession of all the former Hsiung-
nu territories, which extended for more than 14,000 li [5,821 km] to the east and the west, were intersected
by mountains and rivers, and had large numbers of fresh and salt-water lakes.
Thus the territories of the Hsien-pi extended as far as those settled by the Wu-sun in the Ili basin in the
west, while in the north they adjoined those of the Ting-ling alliance of tribes which occupied the Altai
mountains, the basins of the upper and middle Yenisey and the areas adjoining and to the west of Lake
Baikal.” Kyzlasov (1996), pp. 318-319.
Apparently, the “edition of the twenty-four historians” published by the library of Tushujicheng [T’u-shu-chi-
ch’eng] in Shanghai from 1888 of the Weilue, which Chavannes used, has mistakenly used Weibi 魏卑 here instead
of 鮮卑 Xianbi, as Chavannes himself recognised, (1905), p. 526 n. 1. See also the discussion in Pelliot (1906), pp.
365-366, quoted in note 1.13.
For an excellent French translation of the major texts referring to the Xianbi, see: Mullie (1969), pp. 24-51.

2.9. The Dahu 大胡 [Ta-hu] people.

da, dai 大 = ‘great,’ ‘big.’ K. 317a *d’âd / d’âi; EMC da’, dajh, dah 
hu 胡 = ‘why?’ or ‘how?’ – name for non-Chinese peoples to the west of China. K. 49a’  *g’o / γuo; EMC ɣɔ
As the name does not seem to be verified in any other text, it is possible that Dahu 大胡 [Ta-hu] was a simple
copyist’s error for Donghu 東胡 [Tung-hu] or ‘Eastern Hu.’
“東胡 tung1 hu2 (Hist.) Tung-hu or Eastern Hu : anc. name of the Tungus, a horse-breeding people from
southeastern Mongolia and the basin of the 遼 Liao, and who, at the time of the 漠 Han dynasty (206 BCE –
220 CE) were dominated by the 匈奴 Xiongnu. The 漠 Han dynasty (206 BCE – 220 CE) tried to put an end
to this domination, but also to control the Donghu’s territories.” GR 11836, p. 347.
However, the Weishu states that the Xianbi are the remnants of the Donghu – see Mullie (1969), p. 41. So, it is rather
odd to find the Weilue saying that there are Dahu (= Donghu?), as well as Dingling and Qiang, living among the
Xianbi.
          It seems more likely that the Weilue was referring here to another group of people and the obvious similarity
between the name Dahu and that of the Dahur or Daghur people mentioned in the following quotes suggest that
they may be one and the same people. The present-day Chinese name for this Mongol people is Dawoer 達斡爾
[Ta-wo-erh].
          There is, on the other hand, a very long gap between the time of the Weilue and when the first accounts of
these people came to the notice of Western scholars, and so the identification remains uncertain.
“Daghur, also called DAGUR, DAHUR, or DAUR. Mongol people living in the Heilungkiang Province of China.
Their language, formerly thought to be Tungistic of a mixture of Mongolian and Tungus, is now known to
be an archaic Mongolian dialect preserving features found in 13 th-century documents. Their own name is
Daghur; the Manchu form is Dahur; the Russian form Daur occurs in the name of the Daur mountain range.
          Russian settlers in the 17th century found the Daghur well established in eastern Transbaikalia and
the Amur region. . . . Their chief occupations are agriculture, logging, hunting, stock raising, and horse
breeding. The clan system prevails. The religion is shamanistic, although some of them are adherents of
Tibetan Buddhism.” NEB III, p. 343.
“The Tungus tribes today are divided loosely on a regional and linguistic basis into two groups: the
northern Evenk and Even (Lamut) of Siberia and the upper Amur basin; and the southern Nanay (Goldi),
Ulchi, Oroch (including Udege), Orok, Negidal and Solon. To these some would add the Dahur, who are
generally classed with the Mongols. In a special category are the Manchu and ancient tribes of the lower
Amur, some of whom appear later in history as the Koreans and as part of the ancestral population of
Japan.
          The origin of the Tungus is closely related to the east coast Neolithic province which spread
southward into the Yangtze basin and northward into the Tung-Pei (‘East–North’) – the basins of the Liao
and Amur. The Amur or Hei-lung (‘Black Dragon’) basin forms the northern third of the Chinese Neolithic
culture area of the second millennium. With the commencement of the Chinese Bronze Age Tung–Pei
became the land of the Tung or ‘eastern’ barbarians [Donghu or Tung-hu]. The term Manchuria has been
applied only since the emergence of the highly sinicized Manchus, successors to the earlier Jurchen Tungus
who ruled China as the Jin or Chin dynasty.
          The earliest Chinese reference to the inhabitants of Tung-Pei is to the Su–Shen of the second
millennium BC, the ancestors apparently of the I-lou. By the time of the Han dynasty the I-lou were located
in eastern Manchuria. They seem to have combined cattle, horse and pig-breeding with millet and wheat
agriculture. They lived in semi-subterranean houses with corridor entrances, wore hemp and shredded
tree-bast garments and enjoyed falcon hunting. The economy is similar to that of other sedentary Tungus
tribes of northern Korea of approximately the same time.” Bowles (1977), pp. 282-283.
2.10. The Dingling 丁令 [Ting-ling] people
“Nor. . .    should the detail, recorded by Yü Huan [in the Weilue], be overlooked that in the early part of the
third century old men of K’ang-chü still told of their journeys – 10,000 li in extent – beyond the kingdom of
the Yen-ts’ai to the kingdom of the Dwarfs, in other words, to the country of the Lapps. . . . Further, they
[the Xiongnu] were in contact, north of the T’ien Shan, with the Wu-sun, on the river Ili, and with the Ting-
ling on the Irtish.” Teggart (1939), pp. 204, 212, and n. 48.
“A description of the Ting-ling (probably a Turkish people) given in the Wei-lüeh was translated by
Chavannes (“Wei Lio,” pp. 560 ff.). The accuracy of the Wei-lüeh account [especially that there were two
different northern peoples named ‘Ting-ling’] has been challenged by Mori Masao in his two articles and
upheld by Uchida Gimpū. Maenchen-Helfen concludes that from the third century B.C. to the third century
A.D. the Ting-ling occupied the territory from Lake Baikal to slightly beyond the Yenissei. Immediately
following the Fei River debacle of 383, Ti Pin rebelled against Ch’in . . .   , and his successors ruled a quasi-
independent nation until 392, when it was wiped out by the Later Yen (CS 9.9b, TCTC 108.1b).” Rogers
(1968), p. 231, n. 274.
“On the Dingling people, who lived to the north of the Xiongnu, in the general region of Lake Baikal, there
is little recorded. Most of the material has been collected by Wang Jih-Wei “A Brief History of the Ting-ling
People”. Pulleyblank, “The Chinese and their Neighbours”, 445, identifies them as a proto-Turkish people.”
de Crespigny (1984), p. 510, n. 25.
2.11. Tantuo 檀拓 [T’an-t’o]: a great chief of the Zilu [Tzu-lu]. “The character tuo is also pronounced zhi; but the
pronunciation tuo appears preferable when transcribing foreign names.” Translated from Chavannes (1905), p. 526,
n. 5.
2.12. Lingju 令居 [Ling-chü] Prefecture in Guangwei 廣魏 [Kuang-wei] Commandery. For Guangwei Commandery,
see note 1.21.
2.13. Tugui 禿瑰 [T’u-kuei]: a Zilu Chief.
2.14. Shaoti 劭提 [Shao-t’i]: a Zilu Chief.
2.15. Xizhou 西州 [Hsi-chou] Chavannes, in his translation of the Weilue, translated this term by the
“districts of the western wards,” Chavannes (1905), p. 526. This is the sense it should be understood
here. It should not be confused with the later use of the same name. The whole territory of Turfan
was reoccupied by the Chinese c. 640 and turned into the district of Xizhou. Later again it was
specifically used to refer to the town of Yarkhoto, about 20 li (8 km) west of Turfan. See: Chavannes
(1900), pp. 6, 8, 357; Stein (1928), pp. 578.
Section 3 – The Qiang Tribes1
From Dunhuang in the Western Regions2 to the Chuo Qiang (‘Disobedient Qiang’)3 in the
Nan Shan (‘Southern Mountains’),4 and several thousand li west to the Congling (the
Pamirs),5 are the remnants of the Yuezhi6 and the Congzi (‘Brown Onion’),7 the Baima
(‘White Horse’),8 and the Huangniu Qiang (‘Yellow Ox’ Qiang). 9
Each of these peoples has its’ own chief. They are bordered to the north by various kingdoms.
Neither the distance (from China), nor the extent (of their territories), is known.
It is rumoured that the Huangniu Qiang (‘Yellow Ox’ Qiang) are of a separate race, and are
born after a pregnancy of (only) six months. To the south, they border on the Baima Qiang
(‘White Horse’ Qiang).
 Section 3 – The Qiang 羌 Tribes
3.1. The Qiang 羌 [Ch’iang] tribes. Qiang is a general term referring to the tribes living to the southwest of the
‘Gansu corridor,’ in the area of present-day Qinghai province, Shenxi, Shu and Han. See: Molè (1970), p. 75, n. 25;
CICA: 80, n. 69.
          The Qiang have been commonly referred to as ‘Tibetans,’ which is misleading. They appear in the literature
many centuries before a ‘Tibetan’ state had emerged. While it is true that many Tibetans are descended from Qiang
tribes, they were only one of many peoples who contributed to the genetic and cultural inheritance of modern
Tibetans. For a detailed discussion of the various Qiang groups during the Han, see de Crespigny (1984), pp. 54-75.
“Collectively the tribal confederacies and petty principalities were referred to as the 150 Ch’iang (Chiang)
tribes. The ideograph means simply ‘sheep-raisers,’ and their land was called the ‘grass country’ (ts’ao-ti).
White stone-worshipping Ch’iang, who claim to be the pastoralists of Chinese history, still survive near Li-
Fan on the edge of the plateau [between Kansu and Burma].” Bowles (1977), p. 257.
“In my “Die Bedeuttung der Na-khi für die Erforschung der tibetischen Kultur” (Hummel 1960), p. 308, I
have set the presence of the Ch’iang in the Küke-noor region and in A-mdo around 2000 B.C., and the
beginning of a southward migration of the Miao (akin to the Ch’iang), possibly in connection with the
arrival of ox-breeders from the Eurasian steppe-belt, at the close of the 3 rd century. Another possible
explanation for this movement of people is offered by the so-called Pontic Migration, the last offshoots of
which reached the Küke-noor area before the middle of the 1 st century B.C. The presence of the Ch’iang
(which the Chinese believe to be the descendents of the Miao) in this region would then have to be fixed
accordingly. By and large, this would be in agreement with the annals of the Han period. Concerning the
Indo-European influences in Tibet see M. Walter and C.I. Beckwith (1997) “Some Indo-European Elements
in Early Tibetan Culture”. Hummel (2000), p. 64, n. 19.
“The early histories describe conquest and pressure by the Chinese against the western frontier peoples,
and HHS 87 states that in the time of Ch’in and Han the territory of the Ch’iang lay west of the region of
modern Lanzhou.
          The greater part of Ch’iang territory remained forever beyond the frontiers of Han, so that much of
the geographical description is inevitably vague. . . . But the Ch’iang tribespeople with whom the Chinese
had greatest contact were living in the east of the great salt lake, the Kokonor, along the upper reaches of
the Yellow River and its tributaries, the modern provinces of Tsinghai, Kansu and parts of Shensi. From this
point of view, though the term Ch’iang is sometimes rendered as “Tibetan,” the ascription is not
particularly helpful. The Ch’iang who dwelt on the frontiers of Han can be traced as distant ancestors to
the peoples of modern Tibet, but they were not then closely associated with that territory, and there is
clear implication that they had a long history in the northwestern region of China.
          In discussing Chinese dealings with the Ch’iang during the Han dynasty, the official histories make
some distinction between the Western and the Eastern Ch’iang: the Western Ch’iang were those of the
frontier valleys and hill country, the Eastern Ch’iang inhabited the lower ground and loessland of the
present-day provinces Kansu and Shensi. The distinction is not always clearly maintained: some tribes
either emigrated or were forcibly resettled from the west to the east, and the records do not indicate how
many of the Ch’iang people were formerly settled under Chinese control east of the Yellow River. The
earliest references to the Ch’iang describe them as inhabitants of the frontier region in the west.
          This territory of the Ch’iang is bounded on the north by the Nan Shan, or Richthofen range, along the
Kansu corridor, and on the south by the Min Shin, a ridge of the great Tsin Ling divide. The climate of the
region is cold and dry. . . . ” de Crespigny (1977), pp. 4-5.

“In these accounts [in the Hanshu and the Hou Hanshu], the Qiang barbarians of the Han period were
identified with the San Miao, who were banished to the lands of the west by the legendary Emperor Shun.
The name Qiang is related to the ancient clan-name Jiang and the history of these tribes is identified with
that of the Rong and Di barbarians of the west during the time of Zhou and Qin. The early histories describe
conquest and pressure by the Chinese against the western frontier peoples, and Hou Hanshu 87/77 states
that in the time of Qin and Han the territory of the Qiang lay west of the region of modern Lanzhou. . . .
          According to the tradition of the Zuo zhuan, the Qiang-Rong people of the Zhou period had been
farmers in the region of modern Gansu, and there is archaeological evidence for some farming and painted-
pottery settlements even in the upper reaches of the Yellow River.” de Crespigny (1984), p. 55-58.
“The geographical area covered by the Zhang-zhung confederacy, which comprised north-eastern
Tibet, and above all the ethnic links with the Ch’iang, should naturally induce us to shift the focus of our
linguistic comparisons towards the northern border regions of the Sino-Tibetan settlements, rather than to
the western Himalaya.17 This would also solve some problems raised by Stein (1951, “Mi-ñag et Si-hia”), for
example the fact that in Tibetan texts mu (in the forme rmu [dmu, smu]) appears to be a typical indicator of
the Zhang-zhung religion, as a more specific term for the country of Zhang-zhung, but at the same time
rmu is also used to indicate the Mo-so (or Na-khi) who once populated north-eastern Tibet, and were
beyond doubt akin to the Ch’iang. The Ch’iang in turn call themselves rma [rme, rmi]. In fact, rme means
‘man’ and ‘tribe’ in the Si-hia language. Probably no connection exists between the meaning of Zhang-
zhung-smar [smra and dmar] and rma [rma, rme] or rmu [dmu, smu], even if these ancient words are
occasionally mixed up or used one for the other by the Tibetans. It is, however, possible that an identity
exists between rmu or rma [rme, rmi] = ‘man’ and dmu [mu, rmu] = ‘sky’ in Zhang-zhung, or mu [ma] used by
the Ch’iang and mo in Si-hia. This view is supported by an investigation of the origination myths and of the
lists of divine ancestors of northern Mi-nyag, located around the Küke-noor, which was anciently part of
the reign of Si-hia, annihilated in the 14th century. These legends are reminiscent of the myths of ‘O[d]-
de[lde]-spu[r]-rgyal as ancestor of the Central Tibetan royal family, equally of north-eastern Tibetan
provenance.”
17My views on the eastern-Tibetan origin of the Tibetan tribes, and hence of their language,
seem to be shared by D.L. Snellgrove: “. . . it would seem certain that the various waves of people who
occupied Tibet, speaking early styles of Tibetan, came from the east, pressing ever further westward.
They certainly penetrated at an early period deep into the Himalayan Range to the south, as is proved
by the survival of ancient oral traditions, still intoned largely uncomprehendingly by priests of the
people now usually referred to as Gurungs and Tamangs, who live mainly on the southern side of the
main range almost the whole length of present-day Nepal. Is it therefore conceivable that those early
 Tibetan speakers did not also press westward up to the main river valley of the Tsangpo (Brahmaputra)
and so reach the land of Zhang-zhung? It is also significant that Tibetan dialects are still spoken far to
the west of the boundaries of modern Tibet, not only throughout Ladakh, but also in Gilgit and
Baltistan, now controlled by the Pakistan Government.” (D.L. Snellgrove, 1987, Indo-Tibetan Buddhism, p.
392).
Hummel (2000), pp. 9 and 63, n. 17.

“Our next stop was the homeland of a small ethnic minority of some 100,000 people known as the Qiang,
who live north of Chengdu in Sichuan Province. The villages of the Qiang resemble fortifications, with
slender watchtowers that rise as high as 13 stories, or roughly 30 meters. From a distance the towers look
like factory smoke-stacks. They are usually located at the most strategic places, on cliffs or precipices with
the farthest view. The abundance of these towers, which today are used mostly for grain storage, attests to
a darker period in Qiang history.” Wong (1984), p. 105.
3.2. Dunhuang 燉煌 [Tunhuang]. Often written in later works as Dunhuang 敦煌. 
“The nearest oasis [to the ‘Caves of the Thousand Buddhas’] is Tunhwang (Blazing Beacon), which is marked
on many maps as Shachow (City of Sands). The latter name is appropriate to a town standing among
towering sandhills, and the former is equally suitable, for at a short distance from Tunhwang there are
several of the desert landmarks called tun by the Chinese. These old erections were used to convey
messages by fire-signal across desert spaces, hence the name Blazing Beacon for the town and tower placed
at this strategic point. When Shachow was destroyed the new town was built on the old site and the
ancient name of Tunhwang, which dates from the Han dynasty (206 B.C. – A.D. 220), was revived. The
locality is one which figures prominently in Chinese history by reason of its geographical position, for it
stands at the point where the oldest trade-route connecting China with the West is crossed by the road
which leads from India through Lhasa toward Mongolia and Southern Siberia.” Cable and French (1943): 41.
See also CICA: 75, n. 40 in note 2.4.
          “It does not seem possible, either in hsieh-sheng series or poetic rhymes or transcriptions, to
distinguish separately *-l (Sino-Tibetan –r) and *-n words. In transcriptions we find the same characters
used for both, thus 安敦 M. ·an-tuən = Anton(inus), but 安息 M. ·an-si̯ək = Aršak and 敦煌 M. tuən-
h̑waŋ = Sogdian δrw”n, Greek θρόανα [Throaua]. This means that the two phonemes must have coalesced at
an early period.” Pulleyblank (1963), II, p. 228.
3.3. The Chuo or Er Qiang 婼羌 – [Ch’o or Erh Ch’iang] – Literally: ‘The Unconquered’ or ‘Disobedient’ Qiang.
The Chuo 婼 (or Er) Qiang were the first people the Chinese met on the ancient Southern Route after leaving
Dunhuang, on the way to Shanshan. Several scholars have discussed the various pronunciations of the name. See,
for example: Chavannes (1905), p. 526, n. 8; CICA, p. 80, n. 70.
          Paolo Daffinà (1982), pp. 313-314, makes the point that historically the name would have been read êrh rather
than the more common ch’o: “…all ancient commentators (Fu Ch’ien, Mêng K’ang, Su Lin) are unanimous in stating
that as a name of one of the Qiang tribes the character must be read either 兒 êrh < *ńźie̯ < ńi̯ĕg (873a), or according
to the fan-ch’ieh 兒遮 êrh + chê < ńźi̯a < *ńi̯ăg (982a + 804d).” Pulleyblank’s EMC gives: chuò [ch’o] 婼 trhiak; and,
for er [erh] 兒: ɲiə̆ / ɲi.
          Personally, I doubt whether the Chinese word is an attempt to transcribe a local name but is more likely to
refer to its meaning of ‘disobedient,’ ‘disobliging,’ or ‘intractable.’ This seems probable as the word was used,
originally, to refer to any of a very wide-ranging group of Qiang tribes in the Southern Mountains who were not yet
under the control of, or recently, and tentatively, subject to China. It would have distinguished the various Qiang
tribes, stretching in a wide arc from south of Dunhuang through to the Pamirs, from the tribes further east, who
were mostly under Han control. They were often as a direct threat to China and the trade routes, explaining why
the name was meant in the sense of ‘intractable,’ ‘unruly,’ or ‘unpacified.’ The name seems to be used in this
context in the Hanshu (see, for example, CICA, pp. 80, 96, 97, 103), as well as here, in the Weilue.
“The greater part of the Qiang territory remained forever beyond the frontiers of Han, so that much of the
geographical description is inevitably vague. There are references to the Fa or “Distant” Qiang, who appear
to have inhabited the higher ground of the Tibetan massif, and the Account of the Western Regions in
 Hanshu tells of the Er Qiang who lived south of the Silk Road in the Tsaidam Basin. But the Qiang people
with whom the Chinese had greatest contact were living to the east of the great salt lake, the Koko Nor,
along the upper reaches of the Yellow River and its tributaries, the modern provinces of Qinghai, Gansu
and parts of Shaanxi.” de Crespigny (1984), p. 56.
Towards the end of the Former Han at least some of the Chuo Qiang tribes were forced to ally themselves with
China:
“Before Han secured the Ho-hsi area, the Western Regions had served as the meeting ground for the
Ch’iang and Hsiung-nu. As Wang Shun and Liu Hsin pointed out in 6 B.C., Wu-ti had established the frontier
commanderies of Tun-huang, Chiu-ch’üan, and Chang-i with the specific aim of separating the Ch’o-
Ch’iang from the Hsiung-nu, thereby “cutting off the right arm” of the latter. The Ch’o-Ch’iang were a
powerful Ch’iang tribe, described as the first state southwest of the Yang barrier on the route to the west
(in the mountains southeast of Lop Nor). By the middle of the first century A.D. they had dwindled to
insignificance, with a registered population of only 1,750 individuals. But in the early years of the Han
dynasty, they had been active throughout an extremely large area in the Western Regions, stretching along
the K’un-lun Mountains from the neighbourhood of Tun-huang in the east to the Pamir in the west. The
king of the Ch’o-Ch’iang bore the unique title ch’ü-Hu-lai, “the king who had abandoned the Hsiung-nu and
made over to the Han empire.” This suggests that the Ch’o-Ch’iang must have been forced to switch sides
after Han expansion to the northwest. After their submission the Ch’o-Ch’iang not only joined the Han side
to fight against the Hsiung-nu, but also occasionally took part in punitive campaigns against other Ch’iang
tribes.” Yü (1986), pp. 424-425.
          “Setting out from the Yang barrier the state nearest to Han is that one of the Ch’iang [tribes that is
termed] Ch’o. Its king is entitled Ch’ü Hu lai (abandoner of the nomads who made over to the King).” CICA,
p. 80.
This submission to the Chinese did not last long:
          “In addition, T’ang-tou the Ch’ü-hu-lai-wang king (abandoner of the nomads who make over to the
king) lay close to the Red Water Ch’iang of the Great Tribes,667 and was several times subjected to raiding.
Finding the situation intolerable, he reported a state of emergency to the protector general, but Tan Ch’in,
the protector general [who held this post between 4 and 13 CE] did not bring him relief or help at the right
time. T’ang-tou was in a grave and urgent situation ; angry with [Tan] Ch’in he went east to seek [the
means of] defence from the Yü-men barrier, where he was not admitted. He took his wife and children and
over 1000 of his people and fled to surrender to the Hsiung-nu. The Hsiung-nu received him and sent an
envoy [to Han] with a letter describing the state of affairs.”
667 The text reads: 國比大種赤水羌 (Ssu-ma Kuang, TCTC 35, p. 1137, under pen-shih 2 = A.D. 2, adopts a
different reading). The Han texts know no Ch’ih Shui, “Red River”; we have been unable to locate it [but see note
22.2 where Ch’ih Shui is shown to refer to the Kāshgar-daryā]. We have not been able either to find a formal
distinction between “greater” and “lesser” tribes, but it is perhaps not without significance that Fu Ch’ien says of a
certain name “this is the name of the Ch’iang of the Lesser Tribes” (HSPC 69.8b).” CICA, p.191 and note 667.
3.4. The Nanshan 南山 [Nan shan], literally, ‘Southern Mountains’.
“Nan Shan . . . the general term applied to a vast mountain range, actually a complex of ranges, in
Northwest China. The ranges lie between the Tsaidam Basin to the southwest and the plateau of
northwestern Kansu Province (sheng) to the north. The Nan Shan consists of a complex system of ranges
with a predominantly northwest to southeast axis. . . . The ranges are for the most part about 13,000-16,000
ft (4,000-5,000 m) high; however individual peaks often exceed 20,000 ft and the highest peak reaches
20,820 ft (6,346 m). The ranges are higher and more complex in the west, to the south of Tun-huang and
Yü-men (both in Kansu), where, in spite of the aridity of the climate, many peaks are covered with snow
and glaciers. The eastern section of the mountains is somewhat lower, and only a few high peaks have a
permanent snow cover. Among the ranges are a number of large intermontane depressions and fault
basins. The largest of these is the depression in which lies Koko Nor (lake).” NEB, VII, p. 183. See also: de
Crespigny (1977), p. 5.
3.5. Congling [Ts’ung-ling], literally the “Onion Range,” refers to the mountain ranges at the southwestern and
western end of the Tarim basin = the Pamir Mountains.
“At last Semyonov [the famous Russian explorer and collector, in 1857] reached the top of a pass so high
that the mountains now appeared as an undulating plain, dotted with green lakes only partly covered by
ice. It was the highest point they had reached – well above 15,000 feet. The expedition now descended on
the south of the main range, crossing the alpine meadows thickly strewn with blue and white gentians,
pale blue ranunculi and white and golden buttercups. The explorers also found broad glades covered with
the golden heads of an unclassified species of onion, shortly to be named after Semyonov (Allium semenovi).
Semyonov later learned that onions were so widespread here that the Chinese had given this part of the
Tien Shan the name Tsun lin, or “onion mountains”.” St. George, et al. (1974), p. 153.
“The Ts’ung-ling or Onion Range, so called because of the alleged growth of wild onions there, has long
been identified with the Pamirs, see, e.g. E. Chavannes (1907), p. 168.” CICA, p. 72, n. 8.
“The Ts’ung, or ‘Onion’ range, called also the Belurtagh mountains, including the Karakorum, and forming
together the connecting links between the more northern T’een-shan [T’ien-shan] and the Kwun-lun [Kun-
lun] mountains on the north of Thibet.” Legge (1886), p. 20, n. 2.
3.6. Yuezhi 月氏 [Yüeh-chih]
“Ta yüeh-chih 大月氏 or 氐, GSR 317a, 306a and 867a or 590a : *d’âd / d’ai - ngįwǎt / ngįwɐt, *d̑įěg/ źięg
or tiər / tiei. In view of the fact that our text [Hanshu 96A] further on mentions Yüeh-chih as the name of
this people when they were still living in the present-day Kansu area and that it calls this remnant that
stayed there after the main group had migrated, the hsiao, i.e. “Little’ or “Lesser” Yüeh-chih, it seems likely
that the word ta, meaning “great”, does not belong to the name, as in the case of Ta Yüan and Hsiao Yüan.
          For the Yüeh-chih see Haloun (1937) and Pulleyblank (1966), (1968), (1970), and cf. Pelliot (1929), pp.
150-151. Pulleyblank (1963), p. 92 (cf. ibid., p. 106 and Pulleyblank (1966), p. 17), is inclined to accept the
identification (already suggested by Marquart, Eranšahr, 1901, p. 206) of the Yüeh-chih with “the ’Іάτιoι on
the north side of the upper Yaxartes in Ptolemy”, but this refuted by Daffiná (1967), p. 45, note 5.
Maenchen-Helfen (1945), p. 77 and p. 80, note 110, believes Yüeh-chih to be a transcription and
etymologization of “Kusha”, the Moon people.” CICA, p. 119, n. 276.
“The Little Yuezhi were descended from those of the Yuezhi people who had taken refuge in the Qilian
ranges at the beginning of the Former Han period, when the Yuezhi were attacked by the great Xiongnu
leader Modun and their main force was driven west into central Asia. In later Han times, they evidently
numbered some nine thousand fighting men, their chief centres of population being in the Xining valley
and the territory of Lianju in Wuwei [Wu-wei], with a few groups further north in Zhangye. See HHS 87/77,
2899.” de Crespigny (1984), p. 478, n. 15.
“What may have been a crucial formative influence on the proto-Tibetans was the migration of the people
known in Chinese sources as the Hsiao- (or “Little”-) Yüeh-chih, a branch of the Ta- (or “Great”-) Yüeh-chih.
After defeat by the Hsiung-nu in the second century B.C., the Ta Yüeh-chih migrated to Bactria , and are
generally identified with the Tokharians, who according to Greek sources invaded and conquered Bactria
at just that time. Those among them who were unable to make the trip moved instead into the Nan Shan
area, where they mixed with the Ch’iang tribes, and became like them in customs and language. 7
Unfortunately, we know nothing substantial about the customs of the early Tokharians, and cannot guess
what sorts of practices and beliefs they may have introduced.”
7HHS, 87:2899. See B. Watson, Records of the Grand Historian of China, Translated from the Shih Chi of Ssu-
ma Ch’ien (1961) 2:163, 264, 267-268, for a translation of the famous account of the fall of the Ta
Yüeh-chih. It is my opinion that the Chinese name Ta Yüeh-chih was etymologized by the ancient
Chinese to give a convenient name to those who had settled in the Nan Shan. If the Greek
transcription Thagouroi (i.e., T’a-gur if converted to a Chinese-style notation) – for a people
thought to be in the area of the Nan Shan – is indeed a reflection of the name of these “Lesser”
Tokharians, one could not object to the vowel of the initial Ta-. A form gar lies behind the T’ang-
period Yüeh according to the earliest phonetic transcriptions of Chinese, the T’ang-period Tibetan-
script works. The final -chih may be either a Central Asian ending, as thought by some scholars, or
the Chinese word (the same character, pronounced in all other cases shih) meaning “clan” or
 “family”.
Beckwith (1987), p. 6, and n. 7.
“Hsia-hou’s lieutenant, Chang Ho crossed the Huang-ho [in late CE 217] and reached the territory of “Little
Huang-chung” to the east of Köke-nōr, the seat of the Yüeh-chih tribe which had been the prime movers of
the rebellion.” Haloun (1949-50), p. 128.
Huang is the name of a river in Gansu, a tributary of the Datong and Huang He near Xinan fu. Formerly a portion of
the department was called Huang zhou and Huang zhong. See Williams (1909), p. 370, and Couvreur (1890), p. 526.
3.7. Congzi 葱茈 [Ts’ung-tzu]. Literally, ‘Brown Onion.’
3.8. Baima 白馬 [Pai-ma] or ‘White Horse’ Qiang. These are the same people who are also described as the Poma Di,
the most powerful of the Di tribes. They are variously referred to as either Qiang or Di. Their seat at Zhouchi was
made the centre of Wudu Commandery in 111 BCE. See note 1.4. The Baima Di still survive in their ancient home in
northwestern Sichuan, near the border with Gansu and Qinghai:
“Deep inside the Min Shan, home of the giant panda, we visited a little-known tribe [of about 10,000 people]
sometimes referred to by outsiders as the White Horse Tibetans. . . . The name derives from the White
Horse Valley, one of the areas they inhabit.
          The tribe calls itself the Di people – a name that appears in ancient Chinese histories. Yet all written
records of the Di end around the year A.D. 420, more than 15 centuries ago. Though the Di have no written
language, they enjoy a colorful oral history. . . . ” Wong (1984), p. 305; and note on p. 288.
It seems probable that the “White Horse Valley” mentioned above is the original home of the White Horse Qiang or
Di. This valley is on the upper reaches of the Min Xiang (river), which flows south from the Min Shan (mountains)
near the town of Zhangla [Chang-la]: 32.50° N, 103.40° E.
3.9. The Huangniu Qiang 黄牛羌 [Huang-niu Ch’iang] or ‘Yellow Ox’ Qiang. Note that Yu Huan, while
reporting this folk tale, is careful to say only that it “is rumoured that” (傳聞 – chuanwen) the
Huangniu Qiang are born after a six month pregnancy.
Section 4 – The three main overland routes to the Western Regions
It was at the beginning of the Han that the routes were opened leading to the kingdoms of
Xiyu (‘The Western Regions’ – the countries of the Tarim Basin and adjoining areas). 1
At this time the kingdoms numbered thirty-six. Later they split into more than fifty. From the
Jianwu period [CE 25-55] to our time, they have torn each other to pieces, and destroyed one
another, and now they number twenty.2
There were previously two roads, but now there are three 3, which go to the Western Regions
from Dunhuang and the Yumen guan (‘Jade Gate Frontier Post’) 4 :
[1] Heading west from the Yumen (‘Jade Gate’) Frontier Post, and passing through (the
territory of) the Chuo Qiang (‘Disobedient Qiang’), 5 one turns west to pass over the
Congling (the Pamirs),6 and through the Xuandu (the ‘Hanging Passages’ in northern
Hunza),7 to enter (the territory of) the Da Yuezhi (Kushans). 8 – This is the Southern
Route.9
[2] Heading west from the Yumen (‘Jade Gate’) Frontier Post, leaving the Dadu jing
(The Protector General’s Well),10 turning around the northern end of the Sanlongsha
(‘Three Sand Ridges’),10 one passes by the Julu cang (‘Depot Dwellings’). 12 Then, on
leaving the Shaxi jing (‘West-of-the-Sand Well’), 13 and turning northwest, passing by
the Longdui (‘Dragon Dunes’),14 one arrives at ancient Loulan15 and, turning west,
 goes to Qiuci (Kucha),16 and on to the Congling (Pamir) mountains. – This is the
Central Route.17
[3] Heading northwest from Yumen (‘Jade Gate’) Frontier Post, passing through
Hengkeng (‘East-West Gully’ = the Bēsh-toghrak Valley), 18 one avoids the Sanlongsha
(‘Three Sand Ridges’)11 as well as the Longdui (‘Dragon Dunes’),14 and emerges to
the north of Wuchuan (‘Five Boats’)19 and arrives in the territory of Jushi at Gaochang
(47 km SE of Turfan),20 which is the residence of the Mao (Wu) and Ji Colonel (in
charge of the agricultural garrisons). 21 Then it turns to the west and rejoins the Central
Route to Qiuci (Kucha). This is the New Route.22 [Note that there is also a ‘New Route
of the North’ outlined below in Section 10].
Previous historians have already described the products of the Western Territories in detail;
therefore, I will now be brief.
 Section 4 – The three main overland routes to the Western Regions
4.1. Xiyu 西域 [Hsi-yü] translates literally as “the Western Regions.” This term is used sometimes to refer all the
countries to the west of China, but often refers specifically to the countries of the Tarim Basin, as in this passage
from the Hanshu, chap. 96A:
“Communications with the Western Regions started only in the time of Emperor Hsiao Wu. Originally there
had been thirty-six states, but afterwards these were gradually divided into more than fifty. These all lie to
the west of the Hsiung-nu and south of the Wu-sun. To the north and south there are great mountains, and
a river flows through the middle. The distance from east to west extends for more than 6000 li and from
north to south more than 1000 li.
          On the east the area adjoins Han [territory], being blocked by the Yü-men and the Yang barriers. On
the west it is confined by the Ts’ung-ling. Its southern mountains emerge in the east in Chin-ch’eng
[commandery] and are linked with the Nan-shan of the Han. Its river has two sources, of which one rises in
the Ts’ung-ling and the other in Yü-t’ien. Yü-t’ien lies at the foot of the southern mountains, and its river
runs northward to join the river that comes from the Ts’ung-ling. Eastward it flows into the P’u-ch’ang Sea.
. . . ” CICA, pp. 71-72.
This confirms that China began communicating with the Western Regions during the reign of Xiaowu or Wudi
[Hsiao Wu or Wu Ti], who reigned 140-87 BCE.
          As the translators make plain in their notes (ibid., p. 72, notes 8-12), Jincheng [Chin-ch’eng] Commandery was in
the region of modern Lanzhou [Lanchou] in southeast Gansu [Kansu]; Nanshan [Nan-shan] referred to the Zhongnan
shan [Chung-nan shan]; south of Xian [Hsi-an], the Congling [Ts’ung-ling] refers to the Pamirs; Yutian [Yü-t’ien] is
Khotan; and the Puchang [P’u-ch’ang] Sea is Lob-nor.
          Although the distance given from east to west – more than 6,000 li [= 2,496 km] is too long, and the distance
from north to south – more than 1,000 li [= 416 km] is too short, there can be no doubt whatsoever that Xiyu [Hsi-yü]
– ‘the Western Regions’ – here refers to the region of the Tarim Basin.
4.2. Hulsewé and Lowe (CICA: 71, n. 2) add:
“The Chinese commentators attempt, with some difficulty, to relate the figure 36 to states actually
mentioned in the text. The figure was probably chosen for a symbolical significance or mythical
connotation, it is in fact a “pseudo number” ; see Liu Shih-p’ei (1928), 8. 6a-9a, Katō (1952), p. 432 and Ise
(1968), pp. 21-37. For the figure of 36 commanderies of which the Ch’in empire was alleged to have been
formed, see Kurihara (1960), p. 76-81, and Kamada (1962), p. 74f. For a reference to the “36 states beyond
the seas” mentioned in the Huai-nan-tzu, ch. 4, see Erkes (1917), p. 65, and Haloun (1926), p. 135.”
The Hou Hanshu, chap 118, says:
“In the period of Emperor Wu [140-87 BCE], the Western Regions were under the control of the Interior
[China]. They numbered thirty-six kingdoms. The Imperial Government established a Colonel (in charge of)
Envoys there to direct and protect these countries. Emperor Xuan [73-49 BCE] changed this title [in 59 BCE] to
 Protector General. Emperor Yuan [40-33 BCE] installed both a Mao and a Ji Colonel to take charge of the
agricultural garrisons on the frontier of the king of Nearer Jushi (Turfan).
           During the time of Emperor Ai [6 BCE-1 CE] and Emperor Ping [1-5 CE], the principalities of the
Western Regions split up and formed fifty-five kingdoms.” From: TWR by John Hill.
The exact number of kingdoms is of little relevance. The stark figures indicate the continuous grouping and
regrouping of these kingdoms throughout the Han period. This was undoubtedly due to the struggle to control the
massive increase in East-West trade and its profits, both by the local kingdoms themselves, and by the major
regional powers of the time, China, the Xiongnu, and the Kushans.
4.3. The routes to the West.
The Weilue actually describes four overland routes to the west after leaving China (not three, as stated in the text):
1. The ‘New Route of the North’ (sometimes referred to as the ‘New Southern Route,’ which travelled to the
north of the Tianshan range, and three routes that crossed the Tarim Basin:
2. The ‘Southern Route’ which left Dunhuang, and went south of Lop Nor to Khotan and Yarkand, via Hunza
and Gilgit, and on to northwestern India or Jibin (Kapisha–Gandhāra).
3. The ‘Central Route’ (called the ‘Northern Route’ in the two Han histories) which headed from Dunhuang
to Loulan north of Lop Nor, and on to the region of Korla where it met up with the route coming from
Turfan and headed on, along the southern foot of the Tianshan range to Kashgar
4. The ‘New Route’ which turned north before reaching Loulan and headed directly to the Turfan oasis.
From Turfan it turned west, where it met up with the ‘Central Route’ at Kucha.
For details of the various caravan routes see Appendix A. For the maritime route, see Appendix F.
4.4. Yumen Guan 玉門關 [Yü-men kuan] or ‘Jade Gate Frontier Post.’ Sir Aurel Stein seems to have definitively
located the Yumen Guan or ‘Jade Gate’ frontier-post or ‘barrier’ of the Han period about 85 km west of Dunhuang:
“From the very beginning, when the western frontier of the empire was extended to the region of Tun-
huang, we find the two ‘barriers’ of Yü-mên and Yang always mentioned in close conjunction by the Annals
of both Han dynasties. There can be no doubt that the frontier troops stationed there were meant to offer
mutual support. We have seen above that the ‘Yang barrier’ must be located at the present Nan-hu, and
that the Jade Gate was certainly situated to the north-west of it and on the line defended by the main wall
and watch-stations of the Lime.” Stein (1921), pp. 695-696.
“In connexion with the documentary evidence from the site of T. xiv it only remains for me to point out
that its identification with the ‘Jade Gate’ headquarters is consistent with our knowledge of the other
localities that we find mentioned besides Yü-men in the records of this site. As regards Ta-chien-tu (or Chien-
tu), named in Doc. Nos. 304, 307, 356, I have already had occasion to show that it must in all probability be
identified with the westernmost section of the Limes, and it was controlled from Yü-men. As the first-
named document, No. 304, dates from 96 B.C. and the last from A. D. 15, we see that this administrative
connexion was maintained for more than a century. The watch-station Ta fu, named in T. xiv. iii. 64, Doc.,
No. 309, is not mentioned elsewhere and cannot be located. We are in a better position as regards the local
name P’ing-wang, Doc., Nos. 313, 314, 377, repeatedly met with also in records from other sites as designation
of a watch-station, a company, or a ‘barrier’. From an examination further on of these records, more
definite than those from T. xiv, we shall see that the name was probably borne by the section of the wall
extending from T. xxii. c, on the Khara-nör, to T. xiv. a, where it joined the ‘Yü-men barrier’.
          If we review the conclusions arrived at by the detailed scrutiny of the records from T. xiv, join with
them what our preceding survey of the natural advantages of the site and its topographical setting has
shown us, it is easy to realize that, for the period covered by the extant dated remains of the Limes, the
position of the ‘Jade Gate, the chief frontier station through which all traffic westwards by the ancient Lou-
lan route had to pass, may now safely be fixed at T. xiv. Well withdrawn behind the defensive line of wall
and watch-towers, and protected besides from direct attack by impassable marshes to the north-west and
south-east, the position was admirably adapted to serve as a point d’appui for the posts along the most
advanced section of the Limes. It was equally well chosen as a head station for controlling traffic by the
route which, from the reign of the Emperor Wu-ti down to the close of the Former Han dynasty, was
 certainly the main link between the Chinese empire and its Central-Asian ‘sphere of influence’. Fortunately
an archaeological discovery made on this ground permits us to clinch the argument, while at the same time
illustrating once again the accuracy of Chinese historical records.” Stein (1921), Vol. II, p. 691.
“For discussions regarding the location of the Yü-men and Yang-kuan, which were located at the western
terminal of the Han defence lines at the northwestern tip of present-day Kansu province, see Hulsewé
(1957), p. 7, Lao Kan (1959), pp. 375-382 and (1960 : 1), pp. 40-52 (abstracted in RBS 6, no. 101), Lo Che-wen
(1964) and Ch’en Meng-chia (1965); see also Chavannes (1902), p. 67, note 2, and Chavannes (1913), p. vi.”
CICA, p. 71, n. 7.
4.5. The Chuo Qiang 婼羌 [Ch’o Ch’iang]. The ‘Unruly’ or ‘Disobedient Qiang’. See note 3.3.
4.6. The Congling 葱嶺 [Ts’ung-ling] or Pamirs. See note 3.5.
4.7. Xuandu 縣度 [Hsün-tu] – the terrifying, narrow and dangerous ‘Hanging Passages’ or rafiks in the Hunza
valley.
It is significant that Xuandu is never listed as a guo (= ‘kingdom’ or ‘country’) in the literature. The name translates
literally as ‘Hanging Passages,’ and it has long been recognised that it refers to the terrifying hanging pathways,
locally known as rafiks, which are so characteristic of the route through the Hunza valley to Gilgit. See, for example,
Chavannes (1905), p. 529, n. 5. For details, see Appendix A, under the subheading: South to India over the ‘Hanging
Passages.’
4.8. The Da Yuezhi 大月氏 [Ta Yüeh-chih] or the Kushans. See note 5.16.
4.9. The Southern Route. Refer to Appendix A, under the subheading: “The Southern Route.”
4.10. The Dadujing 都護井 [Ta-tu ching], literally: ‘The Protector General’s Well’.
“With this point [the location of the ‘Three Ridges’, see next note] of the Wei lio’s itinerary once determined,
it is possible to fix upon the probable locations also for the ‘well of the Protector-General’, which precedes
it, and ‘the Chü-lu granary’, which follows it. In the former, I think, we can safely recognize the depression
within the westernmost angle of the Limes wall, guarded by the watchtowers T. iv. a, b (Map No. 74. B. 3;
also Plate 33), which, owing to plentiful grazing and to fresh water obtainable in springs and wells, would
offer a very convenient halting-space for caravans following the protected border line. My description of
this locality further on will show that I found here traces of what seems to have been a large entrenched
camp, probably dating back to the time when the route and the line of the wall guarding it up to this point
were first established. It is certain that there is no place on the route between the Jade Gate (T. xiv) and
Bēsh-toghrak which could offer similar advantages for a half-way halting-place.” Stein (1921), Vol. II, pp.
556-557.
“Starting from Yü-mên kuan, the famous ‘Jade Gate’, the position of which in Han times near the ruined
fort T. xiv of the Limes west of Tun-huang is established without doubt [see Map No. 35. D. 4], the ‘route of
the centre’ followed the Limes westwards, just as the present caravan track does, to its extreme end near
the watch-towers T. iv. a. b. There I place the ‘well of the Protector General’.” Stein (1928), Vol. I, p. 308.
4.11. The Sanlongsha 三隴沙 [San-lung-sha], literally: ‘The Three Sand Mounds.’
“A clear indication is next supplied in the statement that the route ‘turns back at the northern extremity of
the San-lung (‘Three Ridges’)[desert of] sand.’ With the knowledge of the ground which my explorations
have furnished, I feel assured that by this must be meant that part of the route which lies immediately to
the east of Bēsh-toghrak and has been described above as the end of its second section. A reference to Map
No. 74 and 70 will show that the route, where it passes through the old terminal basin of the Su-lo Ho, in
some places skirts round, and in others actually crosses, the northernmost off-shoots of the high range of
dunes which flanks the Bēsh-toghrak valley all along on the south. This range represents, as it were, only
the foot-hills of successive ranges of drift-sand which extend upwards to the great gravel glacis of the high
Anambar-ula portion of the Āltin-tāgh (Map No. 75. A-D. 1) and further west find their continuation in the
area of high sand dunes spoken of by the Lopliks as Kum-tāgh (the ‘Sand Mountains’), south of the dried-up
Lop lake-bed.
          Looking at the general map, we see quite clearly that the northern extremity of this great desert is
 formed by the low sand ridges which jut out like promontories into the above-mentioned basin and are
crossed by the route. There we can safely locate ‘the northern extremity of the [desert of] sand of the
Three Ridges’. This identification is further supported by the change of bearing which the Wei lio’s account
implies, where it says that the route ‘turns back (revient)’ at this point. As we follow the route on the map,
we see plainly that, after leaving the western end of the Limes, it runs almost straight to the north-west
until it crosses the dune promontory nearest to Bēsh-toghrak, whence it strikes a south-westerly course in
conformity with the general direction of the Bēsh-toghrak valley.
          The agreement between the wording of the text and the topographical configuration is so close that I
am tempted to connect the name of San-lung, ‘the Three Ridges’, with the fact of the route actually crossing
three distinct offshoots or promontories of the high sand ridge on the south. The maps show these plainly
stretching across the line of route between Camp 153 (Map No. 74. A. 3) and the head of the Bēsh-toghrak
Valley (No. 70. D., 4). This feature of the ground must have impressed itself all the more on the Chinese
wayfarers of old because it is only at this point that real dunes have to be traversed on the Lou-lan route.
The dunes of the three offshoots referred to did not appear to me to rise anywhere above 40 feet or so, and
would certainly be practicable for Chinese cart traffic such as I have seen elsewhere. Yet I know only too
well from experience how troublesome an obstacle they are bound to present to weary men and beasts
engaged upon such a desert journey.” Stein (1921), Vol. II, p. 556.
“As regards the ‘Three Ridges Sands’, the evidence furnished by the actual configuration of the ground, by
the reference to the route which there ‘turns back’, and by the very name, makes it practically certain that
we have to place them at the northern extremity of a belt of high sand dunes crossed by the present
caravan route to the east of Bēsh-toghrak. It is at or near the last-named important halting-place that I
consider that the ‘Chü-lu Granary’ was probably situated.” Stein (1928), Vol. I, p. 308.
4.12. Julucang 居盧倉 [Chü-lu ts’ang], literally: ‘The Depot Dwellings’.
It is most unlikely that Julu is meant to be a name as Daffinà (1982), p. 331, points out. The character ju means
‘residence,’ or ‘to dwell;’ and lu means ‘cottage,’ ‘cabin,’ ‘hut,’ ‘house,’ or ‘inn;’ cang means ‘granary,’ ‘depot,’
‘magazine,’ ‘storehouse.’
          It is likely that there would have been several buildings near such a strategic granary and depot. There would
undoubtedly have been guards, and possibly an inn of some sort. I have translated the term as the ‘depot dwellings’
but it could equally be rendered as the ‘Granary Inn’ or even the ‘granary residence and inn.’
“The ‘Chü-lu granary’, which is likely to have been one of the early ‘resting stations’ established soon after
the Lou-lan route was first opened, may with some probability be located at Bēsh-toghrak. There is no
other site likely to have offered such advantages as this place, which nowadays, too, has more grazing than
any other west of the Limes. Being just beyond a difficult stage of the route, Bēsh-toghrak would be
particularly suited for an advanced base of supply. But I could trace no remains to give archaeological
support to the identification, and considering the character of the ground, with subsoil water near the
surface and a good deal of shōr in the soil, no structures of mud bricks or mere clay would have had much
chance of leaving visible traces here after many centuries of abandonment. No one who, like myself, has
seen the wretched mud hovels which serve as Chinese ‘inns’ and guards’ quarters on the desert route from
An-hsi to Hāmi, the modern pendant of the Lou-lan route, could feel any doubt about their complete
disappearance in the course of a thousand years or even less after they were abandoned. And yet they
somehow suffice for a traffic which at times may not be much less than that seen by the Lou-lan route in its
heyday.” Stein (1921), Vol. II, p. 557. See also: Stein (1928), Vol. I, p. 308.   
“. . . .  Wang Kuo-wei, in his Preface to the Liu-sha chui chien, KTCL 17.6a., believed this granary to have been
situated to the West of the sand dunes of the White Dragon Mounds Pai-lung tui, but Enoki (1963), p. 146f.,
referring to Chavannes, (1905), pp. 529-531, has demonstrated that this granary was located to the East of
the White Dragon Mounds, between these dunes and the San-lung sha Desert. . . . ” CICA, p. 156, n. 460
4.13. The Shaxi jing 沙西井 [Sha-hsi ching], literally: “The West-of-the-Sand Well.”
“As regards the position of the Sha-hsi well, we are furnished with a very helpful indication by the
statement that the route there turned to the north-west. This, read in the light which my explorations of
1914 have thrown upon the line followed by the ancient Lou-lan route, takes us clearly somewhere near the
point where it turns the last south-western offshoot of the low Kuruk-tāgh range, overlooking the Bēsh-
 toghrak valley from the north. This point approximately corresponds to 91° 32 ’ long. 40° 23’ lat. in Map No.
67. From there the line of the ancient route, as I have traced it, makes a sharp turn to the north-west and
follows this bearing, along the shore of the dried-up Lop sea, till it reaches the point where its salt-
encrusted bed and the ‘White Dragon Mounds’ flanking it are traversed. It was within about twelve miles to
the north-east from this point that, when tracking in 1914 the line of the ancient route in the opposite
direction, I came upon the first living vegetation at the foot of the clay cliffs lining the eastern inlet of the
ancient sea-bed, north-west of Kum-kuduk. Three miles or so further on we succeeded in digging a well on
a strip of ground where the soil became sandy. Though the water proved too salt even for the camels, its
presence suggests that in early times, when desiccation had not yet proceeded so far, a ‘resting station’
with drinkable water, corresponding to the ‘Sha-hsi well’ of the Wei lio, might have existed somewhere near
this place at the western end of the Bēsh-toghrak valley.” Stein (1921), Vol. II, p. 557. See also: Stein (1928),
Vol. I, pp. 308-309.
“Is it possible that a characteristic feature of the ground here suggested the designation Sha-hsi 沙西 for
this station? It may literally be interpreted to mean ‘the west [end] of the sand’. The place mentioned in the
text is certainly the last westwards to which the light drift-sand covering the bottom of the Bēsh-toghrak
valley extends. Beyond it no drift-sand is met with now on the ancient route until the vicinity of the Lou-
lan site is reached.” Ibid., p. 557, note 20.
4.14. Longdui 龍堆 [Lung-tui], literally, ‘The Dragon Dunes.’ The Weilue mentions only the Longdui, dropping the Bo
(‘white’) of the name Bo Longdui as given in the Hanshu and certain other texts. See CICA, pp. 89, n. 108; 190 (but note
that on ibid., p. 202, they are translated as, simply, the “Dragon Mounds.”).
“It is called “Dragon-shaped Dunes” or “White Dragon-shaped Dunes” 白龍堆, a part of the desert where
the dunes of white sand are stretched out regularly, looking like dragons. Here, in fact, is what the
commentator Meng Kang 孟康 says in the third century of our era (in the Hanshu, chap. XCIV, b, p. 7 b):
‘The Dragon-shaped Dunes have the appearance of the body of an earthen dragon without a head,
but with a tail. The tallest are two to three zhang high [15.2 feet, or 4.63-6.93 m]; the lower ones are
more than one zhang [7.6 feet, or 2.31 m]. All are turned towards the northeast and look alike.”
According to Xu 徐松 Song (Hanshu Xiyu zhuan bu zhu 漠書西域傳補注 1829, chap. II, p. 27 a), the
Longdui, or desert of the Dragon-shaped Dunes, is the part of the Gobi desert of which one crosses the
northern extremity when going through Shisan jianfang 十三間房 [note – Chavannes incorrectly wrote
this place name as: 三十間房] on the route from Hami to Turfan. The locality of Shisan jianfang is shown
about 350 li to the east of Pizhan 闢展 [modern Shanshan or Piqan] on the map of the territory of Turfan
in the Xin jiang shi lue 新疆識畧 of Song Yun 松筠 (1821). This work, describing the present route from
Hami 咍密 to Turfan, says (chap. I, p. 8b of the little edition published in Shanghai in 1894):
“From Hami, walking towards the west, you turn and go towards the north to cross the Wukeke 烏
克克 [Bogdo Ula] Mountains. You go between two mountains to avoid the dangers of the windy
Gobi 風戈壁.” In the margin the author adds: “To the south of these mountains the windy Gobi is
found. It extends for several thousand li in all directions. This is what is called the (bitter) sand
desert of Gahun 噧順沙噧 . These are the White Dragon-shaped Dunes of Antiquity 卽古之白
龍堆也.”
 If you glance at map 62 of the Stieler Atlas (1902 edition), you will verify that the highway from Hami to
Turfan, in fact, describes a circular arc to pass the mountainous northern region. It thus avoids the Valley
of the Devil (Teufelsthal) or, more exactly, the Valley of the Demons, which is situated on the most
dangerous, but most direct route going from Hami to Turfan.
It is this more southerly route which the Chinese ambassador Wang Yande 王延德 [Wang Yen-to] took in
981 A.D. to make the journey from Hami to Turfan (Sungshi, chap. CCCCXC, p. 4 b; cf. STAN. JULIEN,
Mélanges de géographie asiatique, pp. 91-92):
“Setting off from Hami 伊州 (Yizhou), this traveller then went through Yidu 益都 [I-tu], then he
 went through Nazhi 納職 [Na-chih]. . . .  This town is the closest place to Yumen guan [Yü-men
kuan] which is to the southeast of the desert of the extremely evil demons 城在大患鬼魅磧之
東南玉門關管甚近.*
In this region there is neither water nor pasture. (Wang Yande) set out taking some roasted grain with him.
After three days he arrived at the relay of Bifeng (‘Shelter from the wind’) at the outlet to the Valley of the
Demons 至鬼谷口避風驛. Conforming to the rule of this country, he made a sacrifice and offered an
invitation to the gods to stop the wind, and the wind then ceased. After eight days in all, he arrived at the
Zetian (‘Field Fertilising’) Temple 澤田寺. (The king of) Gaochang 高昌, learning of the arrival of the
ambassador, sent some men to meet him. He then went through a place called Baozhuang 寶莊 (this must
be Pizhan), then through Lukchung 六種 (Luzhong), and then arrived at Gaochang 高昌, which is none
other than the district of Xi 西州 (Yar-khoto, 20 li to the west of Turfan).”
* [According to Pelliot (1906), p. 369, this sentence should read, “This town is to be found to the
southeast of the desert of the extremely evil demons; it is very close to Yumen guan.” Therefore
the desert of the demons is between Nazhi and Hami, not between Yumen guan and Nazhi as this
paraphrase by Chavannes indicates].
It was necessary to reunite here all these texts in order to show that the route followed in antiquity, to go
from Hami to Turfan, crossed the northern extremity of the great desert which extends to the south as far
as Yumen guan (near Dunhuang or Shazhou). The term Bolongdui (the White Dragon-shaped Dunes) applies
in fact to an immense region. This is why the Hanshu can tell us that, “straight to the west of Tunhuang
(Shazhou), outside the passes (Yumen guan and Yang guan) is the Bolongdui sand desert and there is
Puchang Lake (Lop Nor) 正西關外有白龍堆 蒲昌海.” It does not necessarily follow that the route
which crossed the Bolongdui headed straight to the west from Shazhou to Hami (or, to be more exact in
speaking of the ancient route, “to the west of Hami”. . . . ), then to Turfan since, between these two
localities, you cross the northern extremity of the Bolongdui. We are going to show in the following note
that it is this second route which must be the route called ‘Central’ in the Weilue.” Translated from
Chavannes (1905), p. 529, n. 7. Also see note 4.9.
“I believe that the designation of Po-lung-tui, ‘the White Dragon Mounds’, was applied by the Chinese, from
the time of the first opening of the route, to that particular portion where it skirts and then crosses the
extreme north-eastern extension of the dried-up salt bed of the ancient Lop-nōr. There strings of salt-
coated clay terraces, all undoubtedly carved out by wind-erosion from what was the lake bottom of an
earlier geological period, run parallel to each other in the direction from east-north-east to west-south-
west, and extend for a considerable distance along both the western and eastern shores of the ancient salt-
encrusted lake bed. Their fantastic and yet curiously uniform shapes would readily suggest to Chinese eyes
the form of ‘a dragon in earth which was without a head but had a tail. The highest rise to two or three
chang (twenty or thirty feet); the lowest to over one chang (over ten feet). All of them are turned towards
the north-east and resemble each other.’ Thus a Chinese commentator of the Former Han Annals, writing
in the third century A.D., accurately and graphically describes them.
          These belts of salt-impregnated ‘Mesas’ form the most striking feature of the dismal ground crossed
by the last two marches but one of the ancient route before it reached the extreme eastern limit of the Lou-
lan area which once possessed water and vegetation. This explains why the Wei lio, where it describes ‘the
route of the Centre’ which led direct from Tun-huang towards Kuchā, places the Lung-tui or ‘Mound in the
shape of Dragons’ immediately before the station of ‘the ancient Lou-lan’. My explorations of 1914 have
proved that on the line followed by the old Han route, the Wei lio’s ‘route of the Centre’, there was, for a
distance of over 120 miles [193 km], a stretch of ground to be crossed which in Han times was already a
waterless desert of salt, bare clay, or gravel. This forbidding waste lay between the line of wells still
available in the long-extended depression which connects the terminal Su-lo Ho drainage with the
easternmost end of the ancient salt lake-bed of Lop-nōr and the furthest point reached by the Kuruk-daryā,
the river branch which is now quite dry, but then stretched its delta to the northern settlements of Lou-
lan, including the ruined station of ‘Lou-lan’.
          It was for crossing this absolutely barren desert without water or vegetation that the Chinese
missions required provision to be made, from the nearest part of inhabited Lou-lan, for guides and for the
 carriage of water and supplies to meet them near the ‘White Dragon Mounds’. Even with the help thus
provided, it remains somewhat of a problem how those ancient Chinese organizers of transport succeeded
in maintaining traffic, including the movement of large bodies of men, over so great a stretch of ground
devoid of all resources and presenting formidable natural obstacles. In any case, the passage from the
Annals plainly shows to what tribulation the use of the ancient route north of the dried-up Lop sea-bed by
large Chinese convoys &c., must have exposed the Lop population, semi-nomadic as it was.” Stein (1921),
Vol. I, pp. 341-342. See also: Stein (1928), Vol. I, pp. 3089-310; CICA, p. 89, n. 108
4.15. Loulan 僂籣 [Lou-lan]
“The evidence thus afforded by the Ch’ien Han shu enables us to feel certain that, from about 77 B.C.
onwards, the capital of the ‘kingdom’ corresponding to the medieval and modern Lop was situated in the
present Charkhlik tract. It also supplies the definite date when the name of the territory was changed from
the original Lou-lan to Shan-shan. There is nothing in the record of the Annals to suggest that this change in
the official Chinese designation was prompted or accompanied by any change in the position of the capital.
...
          In reality the ‘ancient Lou-lan’, which the Wei lio mentions on its ‘route of the Centre’, is identical with
the ruined ‘site of Lou-lan’ to the north of the Lop-nōr, but yet within the Lop region. The exploration of
these ruins, first discovered by Dr. Hedin in 1900, has convinced me by conclusive archaeological evidence
that the ‘route of the Centre’, which the Wei lio’s author knew about the middle of the third century A.D.,
passed this site, and that it was not abandoned until about the middle of the fourth century. Documentary
evidence obtained at the site, and discussed in Chapter XI, shows that the Chinese military station
represented by those ruins was actually called Lou-lan in local Chinese records of the third and fourth
centuries.
          This proves that the Wei lio and the source used in Li Tao-yüan’s commentary on the Shui ching were
right in giving the name of Lou-lan to the Chinese military colony which guarded the route along the north
side of Lop-nōr in their own time. But the continued application of the archaic name Lou-lan to this
particular locality cannot be accepted as proof that the capital of the whole Lop tract or Lou-lan, as the
Chinese called it down to 77 B.C., must also necessarily have stood there. It is simple enough to assume that
the Chinese retained in use or revived the antiquated designation of Lou-lan for that part of Lop through
which the most direct route westwards from Tun-huang led, and which to them was consequently of
special importance, while for the capital of the territory situated to the south of Lop-nōr and the terminal
Tārim the new official designation of Shan-shan took root.” Stein (1921), Vol. I, pp. 343-344.
“The original name of the state of Shan-shan was Lou-lan. The seat of the king’s government is in the town
of Wu-ni, and it is distant 1600 li [666 km] from the Yang barrier and 6100 li [2538 km] from Ch’ang-an.
There are 1570 households, 14100 individuals with 2912 persons able to bear arms. . . .  To the north-west it
is a distance of 1785 li [743 km] to the seat of the protector general [at Wu-lei = modern Yangisar, 350 li or
146 km east of Kucha]. It is 1365 li [568 km] to the state of [Mo]-shan, and to the north-west it is 1890 li [786
km] to Chü-shih [Karakhoto or Kao-ch’ang near Turfan].
          The land is sandy and salt, and there are few cultivated fields. The state hopes to obtain [the produce
of] cultivated fields and looks to neighbouring states for field-crops. It produces jade and there is an
abundance of rushes, tamarisk, the balsam poplar, and white grass [which was said to be used as an arrow
poison in Yarkand]. In company with their flocks and herds the inhabitants go in search of water and
pasture, and there are asses, horses and a large number of camels. [The inhabitants] are capable of making
military weapons in the same way as the Ch’o [‘Unconquered’] of the Ch’iang tribes. . . .
          Shan-shan is situated on the Han communication routes; to the west it is connected with Chieh-mo
[near modern Cherchen] at a distance of 720 li [300 km]. . . . ” CICA, pp. 81-85, 92
“Shan-shan 鄯善 [205a? and] 205a : [di̯an / źi̯än] – dźan / źi̯än. . . .   Shan-shan was the name adopted
when the state had come under Chinese domination in 77 B.C. The name has been identified as being the
origin of Cherchen or Charchan by Hamilton (1958), p. 121 ; see also Pulleyblank (1963), p. 109. . . . ”. CICA, p.
81, n. 77. See also the discussion in Giles (1930-1932), pp. 827-830.
[Note: Pulleyblank gives the EMC for Shanshan as: dʑian’ or dʑianh + dʑian’ or dʑianh]
“In antiquity Lop Nor was a large salt lake at the hub of communications between the Gansu corridor and
the Tarim basin, but changes in the course of rivers have caused it to dry up and become a salt marsh. The
 famous Lou-lan site lies on the north-west bank of the Lop Nor marsh, where the Kongque river now flows
into the marsh. In the first century B.C. it was the capital of the state of Shan-shan. The graves found near
by used to be considered as graves of the Western Han period, but recent Chinese excavations have yielded
material from the seventh to the first century B.C.” Ma and Wan (1994), p. 211.
4.16. Qiuci 龜玆 [Ch’iu-tz’u] = Kucha. Qiuci unquestionably refers to the Kucha oasis. See, for example: CICA, p. 163,
note 506. It has frequently been transcribed as Qizi [Ch’iu-tzu] (as in the previous reference), but this is incorrect,
the last character is properly ci [tz’u] in this name. See: Daffinà (1982), 331; DFLC, p.1026; Pelliot (1920), pp. 181 and
nn. 1-3; 182 and n. 1; and, especially, the detailed study of the name in Pelliot (1923), pp. 126-128 and nn.  It has long
been the most populous and productive oasis state in the Tarim Basin. The Hanshu gives a total population of 81,317
for the oasis, with 21,076 people able to bear arms. CICA, p. 163.     
          “One MS. [of the Tarikh-i-Rashidi] reads Kus and others Kusan. Both names were used for the same
place, as also Kos, Kucha, Kujar, etc., and all appear to stand for the modern Kuchar of the Turki-speaking
inhabitants, and Kuché of the Chinese. An earlier Chinese name, however, was Ku-sien.” Elias (1895), p. 124,
n. 1.
“The country of Ch’iu-tz’u is 170 li̲ south of the Po-shan (“White Mountains”), and 6.700 li̲ west of Ch’ang-
an. . . .  The walled city which is its capital is five or six li̲ square. In its penal laws, a murderer is executed,
and a robber has one arm and one leg cut off. For its military and civil administrative taxes, they measure
the land in order to assess the levies. Those who hold no fields remit in silver specie. Marriages, funerals,
customs, and products are about the same as in Yen-ch’i [Karashahr], but one difference is the climate,
which is here somewhat warmer. It also produces delicate felt, deerskin rugs, cymbals, a great deal of “salty
green,” orpiment, and exotic cosmetics, as well as good horses, wild oxen, and the like. . . .  Three hundred
li̲ to the south there is a great stream, which flows east and is called the Chi-wu River; this is the Yellow
River [a common Chinese misconception – it was actually the Tarim River].” Chou shu 50.13b–14a; covering
the years 557-581 CE. From Miller (1959), p. 10.
“Far more favourable conditions prevail in the T’ien-shan north of the territory of Kuchā. Agricultural
settlements of some size are to be found among the foot-hills. . . ; mines of copper, lead and iron attest
valuable mineral resources ; the presence of conifer forests at the head of several of the valleys draining
the southern slopes affords striking evidence of the effect that atmospheric moisture, carried across the
range from the north, has produced, by clothing the higher slopes with more abundant vegetation and
thus favouring grazing. More important still is the fact that north of the watershed there extends along
this portion of the main chain of the T’ien-shan a series of wide lateral valleys – those of Yulduz and of the
Tekes and Kunges rivers – which provide not only rich grazing grounds but also, in their lower portions,
large areas suitable for cultivation. We know that in Han times these fertile hill tracts were included in the
territory of the powerful Wu-sun nation. . . .
          Channels for profitable trade between these attractive valleys and the oases included in the ancient
kingdom of Kuchā are provided by a number of passes. Of these the Muz-art pass, situated on the flank of
the great Tengri-khān massif, at an elevation of about 11,400 feet. . .   , is the westernmost and best known.
Others lead from the head-waters of the Kuchā and Bugur rivers to the plateau-like top portion of the Great
Yulduz. All of them, though closed by snow during part of the winter and early spring, are practicable with
laden animals during the rest of the year. These routes provide adequate openings for the trade which is
the natural outcome of the abundance of natural products on both sides of the range. Yet owing to their
height, and the narrowness of the valleys by which they debouch southwards, they are far easier to defend
against nomadic inroads and domination than the corresponding routes from the north into the territories
of Karashahr, Turfān, and Hāmi, all farther to the east.” Stein (1928), Vol. II, p. 805.
“Kuchā must always have been a considerable trade nucleus upon the great Central-Asian high road which
passed through it. . . .  The importance of the main oasis in this respect, apart from its local resources, is
sufficiently indicated by the fact that it lies about half-way between Kāshgar in the west and Turfān in the
east; or, if we consider the times when the ancient Chinese ‘route of the centre’ was in use, between
Kāshgar and Lou-lan.
          . . . . During this [Han] period, when the region north of the T’ien-shan was still independent of
Chinese control, there was an additional advantage in placing the administrative centre near Kuchā : it was
easy to watch from this point the several routes leading down from the north, by which barbarian inroads
 might threaten the main line of communication of Chinese trade and military operations. Finally it should
be remembered that the riverine belts of the Tārīm and Khotan-daryā provide the shortest practicable line
of access from the great northern high road to Khotan and the other oases south of the Taklamakān, as well
as to those of Yārkand and Lop in the south-west and south-east.” Ibid., pp. 805-806.
            “The Oasis of Kuchar, the town with its wide lanes of poplars, and the active population make a
better impression than the rest of the Hsin-chian Oasis [sic] south of the Celestial Mountains. The oasis
occupies a very favourable position which from ancient times made it an important economical and
political center. It is situated at the point where the Muzart and the Kuchar rivers emerge from the T’ien
Shan Mountains and direct their course toward the Tarim Basin.
            Kuchar is by far the most important trading center along the caravan route from Kashgar to
Urumchi. The abundant supply of glacier water, carried down by the two rivers of the oasis, and the
numerous springs of subsoil water make the oasis an ideal place for cultivation and an important fruit
growing center of the whole Province. From early times, Kuchar was known as an emporium of Chinese
trade, with nomad tribes occupying the higher grazing valleys of the Tekes and Kunges in the central T’ien
Shan and the steppe country of Jungaria. Nowadays, the oasis conducts an extensive trade with Kalmuck,
Torgut, and Ölöt tribes, occupying the higher mountain valleys of the T’ien Shan and the steppe country
round Karashahr. Kuchar is connected with Khotan by a desert route following the bed of the Keriya River,
and receives by that route its share of the Khotan trade.” Roerich (1931), pp. 98-99.
          “Kucha was the largest of the 36 kingdoms of the Western Regions noted in the second century BC by
the first of the Silk Road travellers, the Chinese emissary Zhang Qian. In AD 91 Kucha surrendered to
General Ban Chao, whose wide-ranging Central Asian campaigns against the Xiongnu brought 50 kingdoms
under the suzerainty of the emperor. By the fourth century, the Kuchean Kingdom of Guici was an
important centre of Central Asian trade and Indo-European culture. Subsidiary trade routes running north
to Junggar and south across the Taklamakan Desert (along the Khotan River) to Khotan intersected with
the Silk Road at Kucha.” Bonavia (1988), p. 154.
“The extensive ruins of this ancient capital of the Kingdom of Guici [the ‘City of Subashi’] lie 20 kilometres
(12 miles) north of Kucha. They are divided into two parts by the Kucha River, which in flood cuts access to
the northern section. The city dates from the fourth century and includes towers, halls, monasteries,
dagobas and houses. The ruins of the large Zhaoguli Temple date from the fifth century. A recently
excavated tomb revealed a corpse with a square skull, confirming Xuan Zang’s claim that, in Guici, ‘the
children born of common parents have their heads flattened by the pressure of a wooden board’. The city
was abandoned or destroyed in the 12th century.” Bonavia (1988), p. 159.
 
4.17. The Central Route.
The ‘Central Route’ (known to the two Han histories as the “Northern Route”) left from Yumen Guan (‘Jade Gate’),
west of Dunhuang, and headed via Loulan to the north of Lop Nor on to Kucha where it met up with the ‘New Route’
and headed on to Kashgar. For further details see Appendix A, under the subheading: “The Central Route.”
4.18. Hengkeng 橫坑 [Heng-k’eng] – literally: ‘East-West Gully’ = the present Bēsh-toghrak Valley. Seeing as
Hengkeng translates literally as the ‘East-West Gully,’ it undoubtedly refers to the Bēsh-toghrak valley which has
always provided the only practicable corridor of communication between Dunhuang and Loulan, to the north of
Lop Nor, and between Dunhuang and Charklik.
“The ground through which the route leads from Achchik-kuduk [= ‘Bitter Well” – east of Lop Nor on the
route from Charklik to Dunhuang] to beyond Bēsh-toghrak, for a total marching distance of over 80 miles,
bears the unmistakable impress of a great desert valley, flanked by the Kuruk-tāgh on the north and the
sand-buried glacis of the Āltin-tāgh on the south.” Stein (1921), Vol. II, p. 550.
          “The northern route leading to Lou-lan must have remained the main line of communication from
Tun-huang westwards during the first centuries after Christ. But when the Later Han Annals mention the
route leading to Shan-shan, they do not give any detail regarding it except that it started from the barrier
of Yü-mên, the ‘Jade Gate.” Fortunately we fare better in the case of the record which the Wei lio, composed
between A.D. 239-65, furnishes regarding the three routes used from Tun-huang to the ‘Western Countries’
 during the ‘Epoch of the Three Kingdoms’. I have already had occasion, when dealing with the historical
topography of the Lou-lan Site, to discuss the interesting information which this text supplies, and which
M. Chavannes’ translation and full commentary have rendered conveniently accessible. I quoted there the
whole of the important passage, and shown that the ‘central route’ of the Wei lio is identical with our Lou-
lan route, passing from the Jade Gate through the Bēsh-toghrak valley to the ancient Lop lake-bed, and
across it to the extreme north-east end of the once habitable Lou-lan area.” Ibid, p. 555.
4.19. Wuchuan 五船 [literally: ‘Five Boats’].
The Hanshu says:
“During the reign-period Yüan-shih [1-5 A.D.] there was a new route in the further royal kingdom of Chü-
shih. This led to the Yü-men barrier from north of Wu-ch’üan, and the journey was comparatively shorter.
Hsü Pu, the Wu and Chi colonel, wanted to open up this route for use, so as to reduce the distance by half
and to avoid the obstacle of the White Dragon Mounds. Ku-kou, king of the further state of Chü-shih,
realised that because of [the passage of] the road he would be obliged to make provisions available [for Han
travellers] and in his heart thought that this would not be expedient. In addition, his lands were rather
close to those of the southern general of the Hsiung-nu. . . .  [Ku-kou was finally beheaded by the Chinese
for disobedience].” CICA, pp. 189-190, 192.
“Wu-ch’uan 五船, lit. “Five Boats”, GSR 58a and 229e: ngo / nguo - di̯wan / dźi̯wan. Hsü Sung thinks these
might be five flat topped hills with steep sides on the Hsiao Nan lu (Lesser Southern Route?) which neither
Chavannes (1905), p. 533, note 1, nor we have been able to locate.” CICA, pp. 189, n. 658.
“That we are so far unable to identify the intermediate locality of Wu-ch’uan (literally meaning ‘the five
boats’) and that of Hêng-k’êng, which the Wei lio mentions in addition, is not to be wondered at, since that
portion of the easternmost Kuruk-tāgh which lies west of the route from Tun-huang to Hāmi, and which
‘the new route of the north’ must have crossed, has up to the present remained practically unexplored.”
Stein (1921), Vol. II, p. 706.
4.20. Gaochang 高昌 [Kao-ch’ang] – the southernmost town of the Turfan oasis; 47 km southeast of the modern
town of Turfan.
“The name Kao-ch’ang goes back to Han times. As Kao-ch’ang-pi, or “Wall of Kao-ch’ang”, it was the
designation of a Chinese military colony which first existed for a short time in 59 B.C. Its aim was to
strengthen Chinese influence in the territory of the king of Anterior Chü-shih, whose capital was at Yar, to
the north-west of Turfan. Kao-ch’ang grew enough in importance to become in 327 the seat of a chün
(“commandery” in Chavannes’ terminology), which was created by the irregular dynasty of the Anterior
Liang of Kan-su.” Pelliot (1959), pp. 162-163.
Gaochang was later chosen by the Uighurs as the capital of their kingdom which flourished from about 850 to 1250
CE. During the Han period it seems to have been mainly a garrison town for Chinese troops.
          There were two kingdoms of Jushsi [Chü-shih], one centred in the Turfan oasis, which the Chinese called
‘Nearer Jushi’, while the other which was over the mountain range to the north, near modern Jimasa, the Chinese
called ‘Further Jushi’. See: CICA, pp. 76, n. 49, 183, nn. 618, 621; Daffinà (1982), p. 312.
“Liu-jong in the east of the Turfan basin was the long-term central location for the garrison troops and
farming colonies [t’un-t’ien]. These troops were under the command of an officer of higher rank called the
Wu-chi commandant. The site of his headquarters was called Kao-ch’ang-pi, this being the origin for the
use of the name Kao-ch’ang (Qocho) for the Turfan basin as a whole. During the latter days of the Eastern
Han, the Wu-chi commandant had become the highest commanding officer of the Western Regions
garrisons, comparable to the secretary-general, who held authority over the inhabitants of the Western
Regions; duties were divided between the two.” Ma and Wan (1994), p. 240.
For details on the early history and structure of the kingdoms of Jushi, see: Shimazaki (1969), pp. 27-81.
4.21. The Mao (Wu) and Ji Colonel(s) 戊己校尉 = Colonel(s) or Commandant(s) in charge of the military
agricultural garrisons.
          Maoji xiaowei 戊己校尉 GR, Vol. VI, p. 655 states that this title refers to the Commandant responsible for the
military garrisons at Jushi [Chü-shih] or Gaozhang [Kao-ch’ang] – in the Turfan oasis – during the Han dynasty. It also
mentions that the first character, 戊 – wu or mou [‘fifth Heavenly Stem’] was pronounced mao until it was
changed by an Imperial edict during the Five Dynasty period (907-960 CE). I have, therefore, given preference to the
original form of mao (reconstruction: *mug) in my translation.
          The text specifically states here that there were two xiaowei here (“maoji er xiaowei”), but it is unclear whether
it means there were two maoji xiaowei, or one Mao and one Ji ‘Colonels.’ I have chosen the latter interpretation as the
most likely. Certainly, between 74 and 78 CE, there was both a Mao (Wu) and a Ji [‘sixth Heavenly Stem’] Colonel.
However: “In 89 only the wu colonel and his regiment were re-established, to be abolished again in 107 CE.” CICA:
79, note 3. Both the Mou and the Ji Colonels are stated later on in the text (see Section 2) to be stationed “within the
walls of Gaochang.”
          Dubs and de Crespigny (1967), p. 65 translate the title as ‘The Wu-and-chi Colonel.’ Hucker (No. 7740) says,
“HAN: Commandant of the Centre (?), rank = 600 bushels, from 48 B.C. the designation of some commanders of
military garrisons in Central Asia; the title seems to reflect the Taoist concepts that the celestial symbols wu and chi
represent the center (chang), but the relevance of this explanation is questionable. . . . ”
          It is difficult to determine what exactly is meant by the terms mao and ji, or maoji, here. Mao (or wu) 戊
usually refers to the fifth of the ‘Ten Heavenly Stems’ and, by extension, can mean ‘the fifth’ or ‘the central’. See GR
No. 12341. Ji 己, on the other hand, refers to the sixth of the ‘Ten Heavenly Stems’ and, by extension, ‘the sixth’. It
also can carry the meaning of ‘to direct;’ ‘to govern;’ ‘to moderate;’ ‘to restrain.’ ‘Moderation;’ ‘moderate.’
          So, it is possible that the two titles together – maoji – traditionally referred to the earth and may have been a
reference to their function of organizing the military agricultural colonies.
          On the other hand, Maoji xiaowei may have implied something like: ‘Central Governing Commandant.’ ‘Nearer
Jushi’ refers to the kingdom or state centred in the Turfan oasis or, sometimes, to the tribe which controlled it.
          Tuntian 屯田 = ‘agricultural garrisons,’ or ‘military agricultural colonies’ were set up to provide for the needs
of diplomatic and trade missions as well as Chinese officials and troops stationed in distant regions. For details on
their establishment and functioning during the Han, see: Hucker (1985) No. 7409; Stein (1921), pp. 740-745; de
Crespigny (1984), pp. 62-67 and 471, n. 17.
“For the wu-chi hsiao-wei, see Lao Kan (1959), pp. 485-496. Lao’s conclusion is that the latter post was
established in 48 B.C. and filled by one officer down to the end of the Former Han, although as early as 31
B.C. a chi regiment was split off from his command and placed elsewhere under the command of a ssu-ma
major. After the renewed penetration of the Chinese into Central Asia in 74 A.D., there were two officers, a
wu and a chi colonel, down to 78 A.D. In 89 only the wu colonel and his regiment were reestablished, to be
abolished again in 107 A.D. See also, Chavannes, (1907), p. 153, note 2, and Ise (1968), pp. 9-14.” CICA: 79,
note 63.
“In 48 B.C. an additional office, that of the wu-chi colonel (wu-chi hsiao-wei), was established at Turfan.
Though the title suggests a post of a military nature, the duties of the office revolved mainly around
financial and logistical matters, especially those related to the management of the agricultural garrisons
(t’un-t’ien), and the general provisioning of food and services for Han forces. At an earlier date, there had
been a post of colonel of agricultural garrisons (t’un-t’ien hsiao-wei) attached to the protector-general. The
office of the wu-chi colonel was in all likelihood a reorganization of that of the colonel of agricultural
garrisons, with expanded functions. Apart from the regular responsibility for the supervision of the
agricultural garrisons, we find wu-chi colonels engaged in a number of other activities: a colonel by the
name of Hsü P’u took charge of road construction around A.D. 3; another named Tiao Hu arrested, in A.D.
10, the king of a Turfan statelet (in Jimasa) who had refused to provide a Chinese diplomatic mission to the
Western Regions with the required supplies of food and service; and a third such officer, Kuo Ch’in, led an
army to attack Karashahr in A.D. 16.” Yü (1986), p. 412.
“In 48 B.C. the Han stationed a wu-chi hsiao-wei (colonel of Wu and Chi) in the Anterior Chü-shih city of
Kocho (Kao-ch’ang), 30 km south-east of Turfan in Xinjiang. The wu-chi hsiao-wei’s main responsibilities
were, first, to command the Han troops from the central plains of China and, second, to make the soldiers
work the agricultural colonies which provided food for the Han troops garrisoned in the Western Regions
and for the Han diplomatic envoys passing through the area. During the period of Yüan-shih (A.D. 1–5), the
wu-chi hsiao-wei Hsü P’u-yü opened the ‘New Northern Route’, which greatly shortened the journey from
the Jade Gate in the Dunhuang limes to the territory of the Posterior Chü-shih. With the opening of this
direct route via Hami (Qomul), Turfan was destined to become even more important to the Chinese than
before.
           In the first century A.D., control of Turfan constantly changed hands between the Han and the
Hsiung-nu. From 73 onwards, and especially after 89 when General Pan Ch’ao of the Eastern (Later) Han
(24-220) brought the Tarim basin back under Han control, the Chü-shih were once again under Han
jurisdiction. Following the conquest of Chü-shih, the Eastern Han re-established, after an interval of some
60 years, the offices of Protector-General and wu-chi hsiao-wei. Increasing numbers of Han garrison troops
were stationed in the area and the newly opened up territory was expanded. Pan Chao’s [sic – should read
Pan Ch’ao’s] son Pan Yung, who was appointed chang-shih of the Western Regions in 123, stationed his
troops at Lukchun (T’ien Ti), an important site located in the centre of Turfan, not far east of Kocho.
Gradually, the Han Chinese from the central plains of China and the Ho Hsi corridor intermingled with the
Chü-shih natives. During the Wei dynasty (220-265), founded by the House of Ts’ao, and the Western Chin
dynasty (265-316), the kingdom of Chü-shih was basically loyal to China thanks to the implementation of a
continuous policy of ‘control by reconciliation’ through the wu-chi hsiao-wei. The so-called ‘Kocho soldiers’
of the Wei and Chin dynasties may have been a local army made up of Chü-shih natives and immigrant Han
Chinese.” Zhang (1996), p. 304.
In 74 CE, a Ji Xiaowei (a ‘Ji Colonel’) was stationed in the town of Chinbu (near Jimasa), on the territory of the tribe
of the Further King of Juwei, and a Wu Xiaowei (a ‘Wu Colonel’) was stationed in Lukchun (near Turfan), a
dependency of the King of the Nearer Jushi. See: The ‘Biography of Keng Kung’ in Hou Hanshu, chap. XLIX, p. 6 b;
Chavannes (1907), pp. 225-226.
4.22. The ‘New Route.’ Refer to Appendix A: subsection (c) “The New Route.”

Section 5 – The Southern Route

The Southern Route1 heads west to:
• the kingdom of Qiemo (Cherchen),2 the kingdom of Xiaoyuan (‘Little Yuan’ – 3
marches south of Qiezhi),3 the kingdom of Jingjue (Niya),4 the kingdom of Loulan
(north of Lop Nor),5 which are all dependencies of Shanshan (Lop Nor and
surrounds).6
• the kingdom of Ronglu (4 marches south of Jingjue or Niya), 7 the kingdom of Hanmi
(Keriya),8 the kingdom of Qule (south of Keriya),9 and the kingdom of Pikang (modern
Pishan or Guma),10 which are all dependencies of Yutian (Khotan). 11
• the kingdom of Jibin (Gandhāra- Kapisha), 12 the kingdom of Daxia (Bactria),13 the
kingdom of Gaofu (Kabul),14 and the kingdom of Tianzhu (Northern India),15 which are
all dependencies of the Da Yuezhi (Kushans). 16
 Section 5 – The Southern Route
5.1. The Southern Route. See the notes in Appendix A in subsection: (a) “The Southern Route.”
5.2. Qiemo 且末 [Ch’ieh-mo] = modern Charchan or Cherchen. There has been some confusion about the Chinese
name as Chavannes (1907), p. 156, and later Stein (1921), 296 ff., gave the wrong romanization for the first character
(Chavannes, using the French EFEO romanization system gave tsiu, and Stein used the Wade-Giles chü). In fact, the
character is correctly represented by qie in Pinyin and ch’ieh in Wade-Giles. There has never been any serious
dispute about its identification with modern Charchan – see for example, Stein (1921), p. 295, CICA, p. 92, n. 125;
although Pulleyblank (1963), p. 109, following Hamilton 1958, p. 121, suggests it was Shanshan (see note 1.13).
          Charchan is strategically located at the junction of the main route from Dunhuang to Khotan, and the route
which goes south through the mountains, around the southern shore of Koko Nor, and on to Xining, and China. A
branch from this second route goes south from Kharakhoto to Lhasa.
          An ancient trail ran from Xining via Koko Nor and the southern turn-off towards Lhasa at Kharakhoto, and
then on to Charchan in the Tarim Basin. Chinese travellers at times attempted to make use of this route to avoid the
horrors of the desert journey between Dunhuang and Loulan, south of Lop Nor. West of Koko Nor the trail goes
through barren country, with little fodder and was inhabited by a hostile Qiang tribe (or tribes) referred to in the
Chinese texts as the Chuo (literally, the ‘Unruly’ or ‘Unsubdued’) Qiang:
“Charchand is reported to lie at a month’s distance from Khoten by a road which leads all the way along the
foot of a mountain range (the so-called Kue-lun of Chinese and European geographers), and between it and
the great Desert of Takla-Makān or Gobi. No roads are known to lead across this range further East than
that from Poloo, which brings the traveler over to the Pangong Lake in Western Tibet; but there is a road
leading eastward into China, which, however, was not used by the Chinese when they were in possession of
the country.” Shaw (1871), p. 37.

“We had to go down the other side of the mountain in a cloud of dust, and then, once more, we were on a
blistered, yellow table-land. It was bordered with abrupt, eroded mountains on which nothing grew. The
great trail from Dulan to Lhasa by Barun wound through here and I even thought I saw traces of the
plough. Yes, I was right. There were field shapes, a wall, an earthen roof. We were at Kharakhoto [west of
Koko Nor at the junction of the trails from Cherchen to Xining and the road south to Lhasa; about two days’
march to the east of Dzun].” Maillart (1937), p. 101.
5.3. Xiao Yuan (or Wan) 小宛 [Hsiao-yüan] was, according to the Hanshu (CICA, p. 92), three days’ march south of
Jushi [present-day Qiemo or Cherchen]: “It lies secluded to the south and is not situated on the route.” It was
bordered on the east by the 婼羌 Chuo [‘Unruly’ or ‘Unsubdued’ – the first character is sometimes transcribed as
er] Qiang. The Hou Hanshu described it is as a small place with just over 1,000 inhabitants that was later annexed by
Shanshan. 
          The name Xiao Yuan (literally, ‘Little Yuan’) is evocative of Da Yuan (‘Great Yuan’), or Ferghana. Brough
suggests that it might have been the home of the smaller group of Yuezhi who settled among the Qiang in the
mountains to the south of the main trade route when the largest group, the Da (‘Great’) Yuezhi – fled to the west
after their defeat by the Xiongnu about 162 BCE. See Benjamin (2003), p. 1.
          They became known to the Chinese as the Xiao (‘Lesser’) Yuezhi. Brough (1965), pp. 592-593; CICA, pp. 93 and
n. 130; 121.
There are two main possibilities for the location of Xiao Yuan:
          1. Stein suggested that if one travelled south from Qiemo, and then southwest, it must have been located near
modern Atqan [Ajiang], about 110 km from Qiemo. Atqan controlled the route running southwest along the
northern foot of the mountain range:
“As to the still smaller ‘kingdom of Little Wan’ or Hsiao-yüan, which lay about three days’ journey to the
south of Chü-mo, and of which a brief account is given in the succeeding notice of the Hsi yü chuan [of the
Hou Hanshu], it is certain that it must be identified with the small settlements of cultivators and herdsmen
which are scattered along the foot of the K’un-lun south and south-west of Charchan, from Achchan to the
debouchure of the Mölcha and Endere Rivers (see Maps Nos 43, 47). To judge from the distance indicated,
the ‘capital’ of this tract, the ‘city’ of Yü-ling, may be placed about Dalai-kurghan, as suggested by Dr.
Herrman. The population recorded for Hsiao-yüan, 150 families, throws light on the modest resources of
this hill tract. It is correctly described as ‘lying out of the way of the high road’ and adjoining on the east of
the territory of the nomadic Jo Ch’iang, who held the high plateaus south of the Altin-tagh, including
Tsaidam.” Stein (1921), p. 296.
          2. However, it seems more likely that the route headed south from Qiemo and then east, up the Cherchen
River gorges, Xiao Yuan must have been near modern Tura [T’u-la], about 125 km [77 miles] from Qiemo or,
perhaps, Bash Mulghun [Bashi Maergong, W-G: Pa-shih-ma-erh-kung], about 22 km further east. Tura and Bash
Mulghun control a valley of rich grasslands, easily-defended and guarding the junction of two important routes. 
          The first of these branches formed an alternative to the main southern “Silk Route” from Dunhuang to Khotan
and is still in use today. It headed west from Lanzhou via Xining and Koko Nor (= Qinghai Hu = Kökenagur or ‘Blue
Sea’ – Bailey (1985), p. 80) past Dzun (or Zongjiafangzi) – where a road then branched south towards Lhasa, and
across the Qaidam [Tsaidam] marshes through Bash Mulghun and Tura to Qiemo.
          The second branch forms the main route to the relatively fertile valleys of Central Tibet. It heads almost
directly south from Bash Mulghun about a thousand kilometres to Xigaze [Shigatse], presently Tibet’s second-
largest city.
          Maillart (1937), pp. 171-175, describes the journey from Bash Mulghum to Cherchen as taking four days;
indicating that the journey from Qiemo to Tura, which is about 22 km shorter than to Bash Mulghum, could be
easily covered by well-rested travellers from Qiemo in three days – exactly the time indicated in the Hanshu.
          Recently discovered evidence indicates the early use of the route through the Qaidam towards Koko Nor and
on to Lanzhou via Xining. The following article was downloaded from: http://news.xinhuanet.com/english/ 2002-
17/03/content_468165.htm on 8 July 2002. 
 
Byzantine Gold Coin Unearthed in Qinghai
Xinhuanet 2002–07–03 14:12:27
DULAN (QINGHAI), July 3 (Xinhuanet) – A Byzantine gold coin recently unearthed in Dulan in northwest
China's Qinghai Province, may shed new light on the history of East-West trade routes.
          Xu Xinguo, head of the Qinghai Cultural Relics and Archeology Institution, said that the coin
excavated from a tomb in Xiangride Township in Dulan County was made during the reign of Theodosius II
(408-450 AD.).
          The tomb was for an ethnic Tubo who lived in the Northern Dynasties (386-550 AD). This is the second
ancient Roman gold coin unearthed in Dulan.
          As sites where coins are found usually indicate the trade and traffic routes, Xu says that archeologists
should think again about the east end of the “Silk Road.”
          A widely accepted theory is that the road entered the Xinjiang Uygur Autonomous Region through
present-day Lanzhou and the Gansu Corridor.
          But Xu said that a number of recent archeological findings from Tubo tombs including this coin had
shifted people's attention to Dulan County deep in the Qaidam Basin. 
          He believed that the Dulan region occupied a very important position for East-West traffic during the
early and middle fifth century. And the route from Xining to Xinjiang through the Qaidam Basin, slightly to
the south, may be equally important, he said.  
          Before sea routes opened between the East and the West, the Silk Road was the land corridor linking
China with Central and Western Asia to the eastern shore of the Mediterranean between 100 BC and 800
AD.
          Experts said that the 2.36 gm coin, with a diameter of 14.5 mm, may have been used as an ornament.
The town of Xiangride mentioned in the article is about 175 km southwest of (Lake) Koko Nor (Qinghai Hu=
Kökenagur or ‘Blue Sea’ – Bailey (1985), p. 80), or 50 km southeast of Dzun (or Zongjiafangzi), on the main road
between Xining and Golmud. The people referred to in the article as Tubo are more accurately described as Qiang.
Tubo (or, more correctly, Tufan) refers to “Tibetans,” who had not yet formed a national identity at this period. For
further information on the recent finds of Byzantine gold coins in China see: Lin (2003).
5.4. The kingdom of Jingjue 精絶 [Ching-chüeh] = Niya. For the identification of Jingjue as the ancient site of Niya
see: Stein (1921), p. 219.
          Stein had difficulties with this identification because of a mistaken distance given in the Hanshu and then
repeated in the Shuijing. The Hanshu states that Jingjue is 2,000 li (832 km) west of Qiemo which has been “long since
identified with Calmad(ana), in the area of modern Cherchen or Charchan.” CICA, p. 92, n. 125. This is clearly a gross
overestimate as the actual distance is only about 250 km.
          The location of ancient Jingjue (Niya) is made certain by the fact the Hanshu gives a distance of 460 li (191 km)
from Qiemo west to Wumi (= Jumi of the Hou Hanshu) which can be confidently identified with the oasis of Keriya,
see note 3.1. As I measure it on modern maps, it is approximately 92 km in a straight line across the desert from
Keriya to modern Minfeng or Niya Bazar, and then about 100 km north, along the course of the Niya He (Niya
River), to the ancient site described by Stein which included the ruins of houses and a stupa. See Stein (1921), Vol. 5,
Map No. 37 – Niya. See also: Enoki (1963), pp. 143 and 159; CICA: 93, n. 132.
“Now this definite mention of Chü-mo or Charchan as a territory with which the ruler of the ancient oasis
represented by the Niya Site stood in close relation, necessarily forces the question as to the identity of his
own ‘kingdom’ upon our attention. Since it is clearly proved by these little tablets that the ancient oasis
possessed its own ruling family, I do not hesitate to identify the site as the chief place of the territory of
Ching-chüeh . . . which the Chinese historical records from Han to T’ang times place to the west of Chü-mo.
In the Former Han Annals ‘the kingdom of Ching-chüeh’ is described as situated to the west of Chü-mo at a
distance of two thousand li. Its western neighbour was the kingdom of Yü-mi at a distance of 460 li. Since
 the latter territory must certainly be identified with the Chira-Keriya tract, we are thus led to place Ching-
chüeh on the Niya River in spite of the greatly exaggerated distance indicated between Chü-mo and Ching-
chüeh. The capital of the kingdom is named ‘the city of Ching-chüeh.’ But the limited size of the ‘kingdom’
is sufficiently proved by the estimates of its population, ‘480 families, comprising 3,360 persons, with 500
trained troops.’
          No details are given about Ching-chüeh by the Later Han Annals, which merely mention it along with
Shan-shan and Chü-mo on the route from Yü-men to Khotan. Ching-chüeh figures similarly in the list of
territories which the Wei lio, composed between A.D. 239-65, enumerates along the ‘southern route’ leading
westwards from Lop-nōr to Khotan. But here we have in addition the distinct statement that Ching-chüeh
along with Chü-mo and Hsiao-wan, another small territory which lay to the south of Chü-mo and evidently
corresponds to the hill settlements between Kapa and Achchan, was dependent upon Shan-shan or Lou-lan,
the territory adjoining Lop-nōr. The statement has its special interest for the identification of Ching-chüeh
with the territory of which the Niya Site may be assumed to have been the chief place. On the one hand, it
dates from the period immediately preceding the time when we assume the site to have been abandoned.
On the other, it helps to explain why among the Chinese documents excavated in 1901 there was the cover,
N. xv. 345, of an edict emanating from the ‘king of Shan-shan’, and why the records of N. xxiv discussed
below include two covers bearing the seal-impression of the commander of Shan-shan.” Stein (1921), p.
219.
          “Niya, with its inhabited surface of 45 sq km, was one of the largest city oases on the Southern Silk
Road in those times. While Stein had already identified over forty ancient structures during the course of
his three excavations, a Sino-Japanese team was able, from 1993 on, to lay open thirty more buildings as
well as two graveyards. As in other oases only half-timbered buildings have withstood the winds and
sandstorms, while the more numerous clay buildings have long since crumbled away. A cautious estimate
of the population figures arrives at about 800 to 1,000 families and perhaps 100 monks. According to the
Han Chronicle, the local prince in the oasis also had kept an army of about 800 men.” Baumer (2000), p. 100.
5.5. The kingdom of Loulan 僂籣 [Lou-lan] = Lop Nor and surrounds. The site plan included at the end of Volume III
of Serindia by Aurel Stein (1921), Plan No. 23 shows that the external stamped clay walls of Loulan were about 4,050
ft or 1,235 metres long and were almost square in form.
          “Loulan is first mentioned in 176 BC, in a letter of the Xiongnu ruler addressed to the Chinese
emperor Wen Ti, in which the Hun leader praises the victory of his commander-in-chief over the Yüeh-
chih and the fact that he had subjugated Loulan as well as twenty-eight other kingdoms. The mention of
Loulan is a clear indication of the significance of that town in those times. In 126 BC, Zhang Qian, the
famous travelling ambassador of the Han emperor Wu Di, concisely yet revealingly describes Loulan: “The
areas of Loulan and Gu-Shi have a walled city and walled suburbs; they are situated on a salt marsh.
Therefore, the city played a military role even then.
          But it seems that Loulan had abused its strategic position on the Middle Silk Road to raid and plunder
Chinese trade caravans, leading to the first Chinese military reprisal under General Cao Po Nu in 108 BC.
Thus, the weak King of Loulan had to send one of his sons to the Chinese court as a hostage and also put a
second son at the disposal of the Huns.
          When a new prince had to be chosen in the year 92 BC, it was the prince who had been educated by
the Xiongnu who was to occupy the throne, as his unfortunate brother had been castrated at the imperial
court. But, because the new king recommenced the plunder of Chinese traders and also informed the
Xiongnu of Chinese troop movements, China sent off a second punitive expedition in the year 77 BC. The
commanding Chinese general, Fu Gia Dsi (Fu Jiezi), captured the treacherous king by an underhand trick,
had him beheaded, and then sent the head to the imperial court as proof of having done his duty. Thus
ended Loulan’s history as a more-or-less autonomous kingdom, for the Chinese no longer installed the new
prince in Loulan, but rather in southern Shan Shan (today’s Kargilik, or Ruoqiang), in this way removing
him from Xiongnu influence. Loulan, however, remained an important garrison city on the Middle Silk
Road and until about AD 330 was part of a chain of forts and watch-towers securing the stretch from
Dunhuang to Korla, as an extension of the Great Wall.
          From the old, dried-up mulberry trees discovered by Stein, one may conclude that at one time
sericulture was practised in Loulan and silk was manufactured. However, as with the rest of the Tarim
Basin, Shan Shan once again fell under the control of the Huns in the beginning of the 1 st century AD when
 the Han Dynasty was shaken by a severe crisis. But although General Pan Ch’ao successfully re-established
Chinese supremacy in around AD 75 (as already mentioned) it is not possible to determine the strength and
duration of Chinese military presence in Loulan after AD 124 and during the following 140 years.
          On the other hand, on the basis of numerous documents dating back to the time between AD 264 and
AD 330, we are well informed in respect of Loulan’s last golden age. It began with the establishment of a
military colony of one thousand men by General So Man about AD 260. . . .
          Loulan’s revival as a garrison city falls in the era of the Western Jin Dynasty (AD 265-316) which,
under the leadership of its energetic first emperor, Wu Ti, once again brought the “Western Countries” –
today’s Xinjiang – under their sovereignty. However, the small garrison city rapidly seems to have fallen
into oblivion, for the last dated document of AD 330 was still written in the name of the last Western Jin
emperor whose rule ended in 316. Hence, in the year 330 Loulan already must have been isolated from the
central government for fourteen years. As mentioned elsewhere, Loulan was abandoned in that year
because of hydrological changes in the Lop Nor area. The military garrison was transferred some fifty
kilometres further south, to Haitou, designated by Stein as L.K. Maybe the successful attack on Shan Shan
in 335 by the Former Liang (317-376) of Gansu also influenced the decision to move the garrison from L.A.
to L.K. But the fort of Yingpan, situated to the north-west, remained occupied right into the time of the
Tang Dynasty.” Baumer (2000), pp. 133-135.
          “Concerning the many registered names of places it is worth mentioning that Miran was called “Fort
of the small Nob” and Karglik “Fort of the large Nob”. The latter is considered to be the capital of the Shan-
Shan principality, to which Loulan probably belonged as from 77 BC.” Baumer (2000), p. 115.
“The evidence thus afforded by the Ch’ien Han shu enables us to feel certain that, from about 77 B.C.
onwards, the capital of the ‘kingdom’ corresponding to the medieval and modern Lop was situated in the
present Charkhlik tract. It also supplies the definite date when the name of the territory was changed from
the original Lou-lan to Shan-shan. There is nothing in the record of the Annals to suggest that this change in
the official Chinese designation was prompted or accompanied by any change in the position of the capital.
...
          In reality the ‘ancient Lou-lan’, which the Wei lio mentions on its ‘route of the Centre’, is identical with
the ruined ‘site of Lou-lan’ to the north of the Lop-nōr, but yet within the Lop region. The exploration of
these ruins, first discovered by Dr. Hedin in 1900, has convinced me by conclusive archaeological evidence
that the ‘route of the Centre’, which the Wei lio’s author knew about the middle of the third century A.D.,
passed this site, and that it was not abandoned until about the middle of the fourth century. Documentary
evidence obtained at the site, and discussed in Chapter XI, shows that the Chinese military station
represented by those ruins was actually called Lou-lan in local Chinese records of the third and fourth
centuries.
          This proves that the Wei lio and the source used in Li Tao yüan’s commentary on the Shui ching were
right in giving the name of Lou-lan to the Chinese military colony which guarded the route along the north
side of Lop-nōr in their own time. But the continued application of the archaic name Lou-lan to this
particular locality cannot be accepted as proof that the capital of the whole Lop tract or Lou-lan, as the
Chinese called it down to 77 B.C., must also necessarily have stood there. It is simple enough to assume that
the Chinese retained in use or revived the antiquated designation of Lou-lan for that part of Lop through
which the most direct route westwards from Tun-huang led, and which to them was consequently of
special importance, while for the capital of the territory situated to the south of Lop-nōr and the terminal
Tārim the new official designation of Shan-shan took root.” Stein (1921). Vol. I, pp. 343-344.
5.6. The kingdom of Shanshan 鄯善 [Shan-shan] included as dependencies the “kingdoms” of Loulan and all the
region around Lop Nor (‘Lop Lake’) and along the southern route including the oases of Miran and Ruoqiang
(Charklik) and Qiemo (Cherchen), all the way west to Niya (Jingjue) and south to Xiaoyuan. The text here informs us
that the kingdom, at this time, extended over 800 km from Loulan west to Jingjue.
          Its capital during early Han times is called Yüni 扜泥. It is often incorrectly transcribed as Wuni in the Hanshu
– see CICA, p. 81-82 and n. 77. It probably referred to the region of modern Ruoqiang or the Charklik oasis, to the
southwest of the died-up bed of Lop Nor. It is sometimes referred to as the kingdom of Krorän.
          “Lou-lan is the Kror’iṃna or Krorayina of the Kharoṣṭhī-documents ; it was originally, it seems, the
name of the whole country and known as such to the Chinese – although they may have been ignorant of
its position – since 176 B.C., when the Hsiung-nu ruler Mao-tun informed emperor Wen of his conquest of
 this and of other states (HS 94 A.10b, Urkunden I, p. 76). In a more restricted sense, Lou-lan continued to
refer to the town of Kror’iṃna, i.e. the area designated LA by Stein (1921), vol. I, pp. 414-415 : see also Enoki
(1963), p. 147.” CICA, p. 81, n. 77. For the use of NW Prakrit in Kroraina (and in Kucha and Karashahr), see
Bailey (1985), pp. 4-5.

Shanshan controlled both the main ‘Southern’ and ‘Central’ routes to the west of Dunhuang:
          “From the mid-third to the mid-fifth century the kingdom of Shan-shan maintained its control over
the southern route of the Tarim, leading from Dunhuang to Khotan, and incorporating the smaller
kingdoms and principalities of Ch’ieh-mo (Calmadana, Cherchen), Hsiao Yüan and Ching-chüeh (Niya,
Cad’ota). At the height of its power, Shan-shan seems to have been composed of a series of rājas or rāyas
(districts) administered by rājadarāgas or rājadareyas nominated by the king. Ching-chüeh was listed, for
example, among the rājas, retaining the original ruler.” Zhang (1996), pp. 288-289.
The kingdom of Shanshan also included the important strategic community, Loulan, located near the northwest
corner of Lop Nor, which, at that time was near the outflow of the Tarim River. Loulan provided an invaluable
supply point on the difficult but important desert ‘Central’ route from Dunhuang to Korla. 
          This has caused considerable confusion about where the “capital” lay. I tend to agree with Stein, Baumer and
Yu Taishan that the seat of government was always in the fertile Charklik oasis:
“On the location of the royal government of Shanshan, there have been two main theories. The first
suggests that Wuni was situated southwest of Lob Nor, around present Ruoqiang 婼羌 county. The second
suggests that Wuni lay northwest of Lob Nor, around the ruins of Loulan (Kroraimna, Krorayina). In
addition, it has been suggested that Shanshan had established its capital at Kroraimna when the name of
the state was Loulan, and later moved its capital south of Lob Nor. In my opinion, Shanshan (i.e. Loulan)
never moved its capital and the seat of the royal government had always been southwest of Lob Nor.” Yu
(1998), p. 197 – and see the whole of his Appendix 2, “On the Location of Capital of the State of Shanshan,”
ibid., pp. 197-211 for his detailed presentation for this scenario.
          “The town of Wuni was not situated northwest of Lob Nor, but was situated in the present Ruoqiang
country (Qarkilik), on the south bank of the Charchen River, by the northern foothills of the Altyn Tagh,
southwest of Lob Nor.” Ibid., p. 201. See also note 5.5.
Both CICA and Taishan Yu have given the wrong romanization for the first character of the name of the capital (the
modern Pinyin for 扜泥 should read yu not wu), hence the following detail:
扜 yu [yü] GR 13088 [64:3] “1. To make a hand sign; 2. to pull to oneself (the string of a bow).” Couvrier (p.
345) gives: “to make a hand sign. To take.” This character is, unfortunately, not listed in either Pulleyblank
or Karlgren.

泥 ni [85:5] EMC: nεj or nεjh; K. 563d * niər.

I believe the original should be accepted, but, as several alternatives are presented in CICA, I have included them
here, for the reader’s consideration:
           “Wu-ni, however, has given rise to considerable discussion because of the uncertainties
surrounding the word here transcribed as wu, viz. 扜. According to Yen Shih-ku, it is pronounced ·o· / ·ua,
and this view is repeated in T’ai-p’ing yü-lan 792.5a (it is not clear whether this passage belongs to the
original Hua-lin p’ien-lüeh of 524, or whether it was copied from a later – T’ang or Sung – manuscript of the
Han-shu only around 983; see Tjan (1949), pp. 60-61; Pulleyblank (1963), p. 89, calls it “an anonymous gloss”,
but the chances are that it is Yen Shih-ku’s remark).
          Secondly, although wu 扜 is included in the dictionary Shuo-wen chieh-tzu of A.D. 100 (see Shuo-wen
chieh-tzu ku-lin 5505a) and even in the earlier wordlist Fang-yen, compiled before A.D. 18, if we follow the
emendation by Tai Chen (1937) in his Fang-yen su-cheng, p. 295, in the Shuo-wen it is not written 扜 but 㧍.
Still more curious, however, is the fact that it does not seem to occur in Han inscriptions or in pre-Han
literature, i.e. it is not found in Uchino’s index to the Li shih (1966) nor in Grammata Serica Recensa.
According to its rare occurrences in Han literature, assembled in T. Moroashi’s Dai Kan-wa jiten, vol. V, p.
 103, no. 11799, wu 扜 seems only to occur in these few placenames in HS 96!
          Thirdly, there are variant readings, where wu 扜 is replaced by 扞 (K. 139q : g’ân / γân), or 拘, K.
108p : ki̯u / ki̯u, or ku / kəu, or g’i̯u / g’i̯u). These variants occur in some editions of SC 123 (Shao-hsing ed.
123.1b. Palace ed. of 1739, 123.3b. This reading has not been adopted by Takigawa, SC 123.7, who writes 扜
without further explanation), cq. HHS Mem. 78.6bff., both not regarding the city of Wu-ni, but the country
of Wu-ni (HS 96.16a ; see note 138).
          Now either Wu-mi or Chü-mi (not ni !) may be correct for the completely different country (see
below), but, as regards the capital of Lou-lan – Shanshan, it would seem that 扞泥 Han-ni is right,
supported as it is by the reading 驩泥 (GSR 158.l : χwân / χuân) in the Hou Han chi by Yüan Hung (328-
375), for this agrees with the word occurring in the kharoṣṭhī inscriptions : kuhani (or kvhani), meaning
“capital” (Enoki (1963), p. 129-135 as well as Enoki (1961) and Enoki (1967), cf. Brough (1965) ), Pulleyblank
(1963), p. 89 reconstructs the “Old Chinese” pronunciation of Wu-ni as ·wāĥ-ne(δ) and believes it
“unnecessary … to adopt the reading 扞 … The variant 驩泥 *hwan-nei seems closer to ·waĥ than to
*ganh as an attempt to render the hypothetical original behind khvani or kuhani”.” CICA, pp. 81-82, n. 77.
The name of the kingdom Loulan was changed by the Chinese to Shanshan in 77 BCE. See: Chavannes (1905), p. 537,
n. 2; Brough, (1965), p. 592; Molè (1970) p. 116, n. 183 [note that the date of the name change in CICA to Shan-shan in
AD 77 is incorrect]; CICA, p. 81, n. 77.
          Pulleyblank (1963), p. 109, suggests that Shanshan was plausibly identified by Hamilton (1958, p. 121): “with
modern Charchan < *Jarjan. The name Shan-shan appears as a substitute for the earlier Lou-lan in the first century
BCE. If the foreign original had indeed palatials at this period, we must suppose that the Chinese palatials were
already beginning to develop, perhaps in and intermediate stage *di̯. There are too many uncertainties, however,
for this to provide a firm argument.”
          Hamilton’s argument does seem to be overridden by the argument that it must refer to the largest oasis in the
region – that of modern Charklik:
“There can be no doubt that by Shan-shan is here meant [that is, in the Weilue] the present Lop tract with
its main oasis of Charkhlik.” Stein (1921), p. 328. See also: Giles (1930-1932), p. 830; Part 4, note 15; Pelliot
(1963), p. 770.
“The northern and southern routes came together again on the eastern rim of the Tarim near the great salt
marsh of Lopnur in the kingdom of Krorän (Kroraina, Loulan) before continuing into the world of the
ethnic Chinese.” Mallory and Mair (2000), pp. 64-65.
“The statelet that formed about the great salt marshes of Lopnur was known in Chinese sources originally
as Loulan and then later was called Shanshan . . . when the territory came under Chinese dominion in 77
BC. The name Loulan reflects an attempt to render in Chinese what we find in later Indian (Kharos̱t̲hī)
documents from the region as Krora’ina or Krorayina (now Krorän). . . .  As for the name ‘Shanshan’, this
has been seen as a precursor to the name ‘Chärchän’, where some of the most spectacular mummies have
been recovered. This is hardly unexpected as the region possesses immense deserts laden with salt that
both early Chinese guidebooks and modern explorers have described in detail. A 1st-century BC document
informs us that from the Chinese outpost at Dunhuang to Krorän there was a desert that stretched for 500
li [208 km] in which there was neither water nor grass.” Ibid. p. 81.
          “During the Han period the population of Krorän is given as 1,570 households, with 14,100 people of
whom 2,912 could bear arms. The agricultural potential of the region is described as limited, its soils being
too sandy and salty, and food crops had to be brought in from neighbouring states. There was an important
nomadic component in the region where asses, horses and many camels are reckoned. Other products were
jade, rushes, tamarisk and balsam poplar.” Ibid, p. 85.
          “Another prominent site to see some excavation is the town of Krorän which was excavated by Sven
Hedin, Aurel Stein and, most recently, by Hou Can. Its stamped earth walls still stand up to 4 m (13 ft) in
height and ran about 330 m (1,083 ft) on each side. It housed clusters of temples, the government central
offie, residential quarters and what has been dismissed as a ‘slum’. Within 5 km (3 miles) of the town Hou
Can uncovered seven tombs. Among the burials was a middle-aged woman who had a child placed over her
head, reminiscent of the ‘Scream Baby’ excavated at Zaghunluq.” Ibid., p. 165.
 “The next town [to the east, after Quemo/Cherchen], Ruoqiang (Charkhlik) is no bigger but is nevertheless
the most important in the vast region encompassing the salt seabed of the dried-up Lop Nor. In the first
century BC it formed part of the Kingdom of Loulan, which was later to change its name to ‘Shanshan.’ At
Ruoqiang the road divides, one branch heading north to Korla, the other taking a more southerly road than
the original Silk road, crossing into Qinghai Province and then turning northeast to Dunhuang. East of
Ruoqiang lies another archaeological site, Miran, which Stein visited in 1906. In the 1970s Chinese
archaeologists found a Han-Dynasty system of irrigation canals here. To the south lie the Altun Mountains,
where a large nature reserve has been established. It was here in the 1880s that the Russian explorer,
Nikolai Prejewalski, discovered the only existing species of the original horse, which was named Equus
prezawalski. Extinct in the wild, the species is now only bred in zoos.” Bonavia (1988), p. 192.
          “Krorän was included in the lists of conquests carried out by the Xiongnu leader Modu in 176 BC and,
with the westward expansion of the Han, it found itself in the middle of two warring empires. . . .  The king
saw that it was impossible to navigate between two such masters and tilted his hand towards the Han, who
took advantage of the situation, assassinating one king and beheading another until they had installed
someone they could trust, and in 77 BC the name of Shanshan. Although ostensibly under Han control, as
late as AD 25 it was recorded that Krorän was still in league with the Xiongnu.
          Understandably, during the floruit of the Silk Road, Krorän was a place of great strategic importance.
About AD 119 Ban Yong, the son of General Ban Chao, recommended that the Chinese governor be sent to
Krorän with 500 men to establish a Chinese colony. . . .  It was intended that that this colony dominate all
approaches to Dunhuang, the main Chinese outpost in the west, by way of both the northern and southern
routes and it was also intended to check any Xiongnu incursions. In order to provide the colony with a
secure agricultural basis, major irrigation works were required and a much later account depicts how
attempts to place a barrage across a river in order to dam the water for irrigation were thwarted by an
intractable river throwing itself against and over the barrage. First the governor tried prayers and
sacrifices to get the water to recede but when these failed he sent his troops in to assault the waters with
swords, spears and arrows, and the river, apparently cowed, dropped its water level and supplied the
desired irrigation channels.
          The administrative capital of Krorän was discovered and investigated by Sven Hedin…, Aurel Stein
and Hou Can. The 429 documents found in these investigations provide contemporary evidence of the
running of the Chinese colony in the 3rd century AD and the approximate date of its abandonment in the
4th century (the most recent document dates to c. 330). In addition to the documents written in Chinese
there were tablets in Prākrit, a north Indian language, which also contained traces of the language of the
native inhabitants. . . .  Stein could only speculate that physical changes robbing the region of adequate
water supplies led to its current deserted state and Hou Can supports this theory with documentary
evidence, indicating pressure on water resources and the need to build a reservoir upstream. The Chinese
abandoned the territory and did not attempt to resettle it during their reconquest of the Tarim in the 7th
century. Krorän apparently went out with a whimper rather than a bang: in his excavations of the home of
one of the leaders of Kroränian society Stein discovered thick layers of sheep dung that preceded its total
abandonment – animals had been stalled in rooms where nobility had once dwelt.” Mallory and Mair
(2000), pp. 86-87.
          “More recent analysis suggests that the Cherchen burials were made about 1000 BC, whilst the
Loulan graveyard bodies seem to have been buried as early as 2000 BC. Though the site is now barren, salty,
sandy and windswept, the rings of dried tree-trunks surrounding the graveyard, the bundles of ephedra
twigs in the graves, and the arrows and baskets all point to a different environment thousands of years ago,
enabling a semi-settled life. When the Chinese of the Han dynasty first set out into Central Asia, Loulan was
still an important caravan stop with water and food in abundance. A disastrous flood in about AD 330
destroyed the town, and the Lop lake gradually dried up into salt flats, although, out of custom, many
travellers still passed through the remains of Loulan on the northern route that offered no shelter or
sustenance. From this time onwards the southern route was safer, though longer.
          The mummies found at Loulan and Cherchen were strikingly European-looking. They had high-
bridged noses, substantial beards, deep, round eye-sockets and fairish or reddish hair. They were tall, if
fully grown, and wore clothing of furs, woven cloth, often in an interesting plaid pattern, leather and
felt. . . . ” Wood (2002), pp. 61 and 63.
Professor Richard Frye (personal communication, 7 July 2003) cautions:
 “If the mummies in fact are to be dated very early (= pre 1000 BC) then it is possible that the ancestors of
the Hunzakut, Burushaki speaking people, were the mummies. This supposes a pre-Indo-European
population extending from the Basques through Rhaetians to the Himalayas etc. (not one people but
various pre-Indo-European speakers). On the other hand, if it can be shown that the ancestors of the
Tokharians were the same as (or related to) the Guti and Hittites, then Victor [Mair] may be right. It all
hangs on the date of the migration of the Tokharians to Gansu from the west. They are hardly the
indigenous Indo-Europeans as Narain thought.”
Perhaps DNA analysis of the mummies will be able to give us some more definite answers to these questions.
          It is of interest to note that a recent study (reported in Nature Science Update and downloaded from
http://www.nature.com/nsu/nsu_pf/030616/030616-15.html on 7 July, 2003) has found that peoples related to the
Basques may have been widespread before the later invasions of Indo-European speaking peoples:
“Goldstein's team collected DNA samples from more than 1,700 men living in towns across England, Ireland,
Scotland and Wales. They took a further 400 DNA samples from continental Europeans, including Germans
and Basques. Only men whose paternal grandfathers had dwelt within 20 miles of their current home were
eligible.
          The Y chromosomes of men from Wales and Ireland resemble those of the Basques. Some believe that
the Basques, from the border of France and Spain, are the original Europeans.”
One hopes that this research is done soon before more damage is done to these important sites by looters. The
following newsbrief emphasizes the need for urgent protection of the sites because of the inherent interest for the
region’s history. It was downloaded from: http://www.archaeology.org/magazine.php?
page=0305/newsbriefs/silkroad on 7 July 2003:
NEWSBRIEFS Volume 56 Number 3, May/June 2003

SILK ROAD THEFT

In the remote Lop Nur desert of northwest China, ancient tombs have been ransacked for the second time
in two years. A team of archaeologists on an expedition to the area reportedly encountered the tomb
robbers and followed their trail back to a previously unknown mausoleum from the Loulan Kingdom, an
important stop along the Silk Road, that flourished more than two millennia ago. Inside the 90-foot-high
domed mausoleum were high-quality silks, colored coffins, and an extraordinary mural depicting
geometric patterns and a gold and a silver camel fighting each other, all of which were damaged by the
looters. Mummies were desecrated and scattered bones thrown from the tombs. Although it is still too
early to be certain, the quality of the grave goods and the rarity of the funerary architecture suggest that
the mausoleum may be royal–or even belong to one of the Loulan kings, whose tombs have never been
found. While investigations into the robbery continue, the local heritage administration now faces the
tremendous challenge of preserving these unique tombs, which are clearly a popular target for looters.
Because of the size of the area in which the tombs are located (they are spread across 25 acres) and the
sparse population of this region of China, this will be an extremely difficult task. – JARRETT A. LOBELL
5.7. The kingdom of Ronglu 戎盧 [Jung-lu] (4 marches south of Jingjue / Niya). Ronglu 戎盧 [Jung-lu] was,
according to the Hanshu, south of Jingjue (Niya), and adjoined Qule (south of Keriya) to the east, and the Chuo
(‘Unruly’ or ‘Unsubdued’) Qiang tribes to the south. It was said to “lie secluded to the south and is not situated on
the route.” It was, perhaps, near modern Atqan – see note 5.3.
5.8. The kingdom of Hanmi 扞彌 [Han-mi] = Keriya. Unfortunately, the name of this kingdom is recorded with a
wide variety of characters used for the first syllable. 
          The Hou Hanshu gives Jumi 拘彌 [Chü-mi] = Keriya. It is described as Yumi 扜彌 [Yu-mi] in the Hanshu.
Unfortunately, this is incorrectly rendered as Wu-mi in CICA p. 94, although the extensive note (p. 94 n. 138)
contains much useful information about the various forms of the name and the geography of the region.  
          The Weilue and the Tangshu both refer to it as Hanmi 扞彌, which Chavannes thinks is the preferable reading,
but there are several other variants in other texts. See the note in Chavannes (1905) p. 538, n. 1 for a discussion of
the various forms of this name. Pulleyblank (1963), pp. 88-89, discusses its derivation and notes that it probably
relates to the Kema in the Kharoṣṭhī documents from Loulan, as well as the Kan 坎, City of the Tang and the Gan 紺,
Prefecture of the 10th century.
          Stein (1907), p. 467, suggested that Hanmi represented the whole area between Chira and Keriya. This was
undoubtedly due to his faulty estimation of the li as being approximately one fifth of a mile, or about 322 metres.
The Han li is now known to have been 415.8 metres, as I have discussed in my Introduction under ‘Measurements.’
          Both the Hanshu, and the Hou Hanshu place this kingdom 390 li [162 km] east of Khotan. This is almost exactly
the distance between Khotan and Keriya on modern maps. Besides, the Keriya oasis is the only place in the region
capable of supporting the populations indicated in the two Han histories (20,040 are mentioned in the Hanshu
account, which does seem a bit high as it only lists 3,340 households and 3,540 persons able to bear arms).
“. . . .  This kingdom must be identified with the territory of Keriya.; the Xiyu shuidaoji (Chap. II, p. 7a), which
follows this identification, remarks meanwhile, that the present town of Keriya is to the west of the river,
while the Tang zhou places the capital of the kingdom of Jumi to the east of the river; but it is clear that the
displacement of the town is not a serious objection.” Translated from Chavannes (1900), p. 128, n. 1. See
also, Chavannes (1907), p. 167, n. 8; CICA, pp. 94, n. 138, and 95-96.
There is confusion caused by the present Chinese name for Keriya where Yutian is written with the identical
characters as for the ancient name for Khotan, only dates back to the late 19th century. See CICA, p. 96, n. 4.1.
The 16th century Tarikh-i-Rashidi has a description of this whole region:
“To the east and south of Káshghar and Khotan are deserts, which consist of nothing except heaps of
shifting sands, impenetrable jungles, waste lands and salt-deserts. In ancient times there were large towns
in these [wastes], and the names of two of them have been preserved, namely Lob and Katak ; but of the
rest no name or trace remains: all are buried under the sand. Hunters, who go there after wild camels,
relate that sometimes the foundations of cities are visible, and that they have recognised noble buildings
such as castles, minarets, mosques and colleges, but when thy returned a short time afterwards, no trace of
these was to be found ; for the sand had again overwhelmed them. On such a scale were these cities of
which, nowadays neither name nor vestige remains! In a word, the habitable districts of Káshghar and
Khotan lie along the western skirts of these mountains. On the frontier of Káshghar is the district of Artuj ; 1
from there to the confines of Khotan, at Kariyá and Jariyá, 2 is one month’s journey. But as for the breadth
of fertility of the cultivated region (from the foot of the western range to the eastward) by travelling
quickly one can leave all cultivation behind in a day or two. On the banks of every stream that comes down
from that range, corn is sown and the land is cultivated.”
1 This place is often mentioned in the Tarikh-i-Rashidi. It is, nowadays, a favourite summer resort of the
townspeople of Kashghar. According to the late Mr. R. B. Shaw the proper spelling is Artush. (J. R. G. S.,
1876, p. 282.)
2 Usually Kiria and Chiria or Chira [= modern Keriya and Qira]. Both exist to the present day, the former a
town of some size.”
Elias (1895), p. 295, and nn. 1-2.
5.9. The kingdom of Qule 渠勒 [Ch’ü-le] (to the south of Keriya). Qule 渠勒 [Ch’ü-le] was probably situated along
the ancient route that led south from Keriya into northern Tibet, near modern Pulu, at the foot of the mountains.
“However, our old Chinese sources do not fail us altogether about the geography of this region ; for the
small territory of Ch’ü-lê. . .   , which the Former Han Annals note to the south of Yü-mi, can be safely
identified with the present submontaine tract known as Tāgh and comprising, as mentioned above, the
various small settlements from the Keriya River to those on the river of Chira. Of Yü-mi I have made it
certain as I believe, that it comprised the whole of the oases between Chira and Keriya, and the Tāgh
subdivision lies, as Maps Nos. 28, 32 show, exactly to the south of these. Ch’ü-lê is described as a very small
territory with only 310 families. We have no means of fixing the position of its ‘capital . . . the city of Keen-
too’.” Stein (1921), pp. 1322-1323.
The Hanshu says:
“The seat of the king’s government is at the town of Chien-tu, and it is distant by 9950 li [4,139 km] from
Ch’ang-an. There are 310 households, 2170 individuals with 300 persons able to bear arms. To the north-
east it is a distance of 3852 li [1,602 km] to the seat of the protector general. It adjoins Jung-lu in the east,
the [land of the] Ch’iang [tribes who are termed] Ch’o in the west and Wu-mi in the north.” CICA, p. 96.
5.10. The kingdom of Pikang 皮亢 [P’i-k’ang] = modern Pishan or Guma. The Hou Hanshu gives: Pishan 皮山 [P’i-
shan] = modern Pishan or Guma.
          The Hanshu Chap. 96 A states that Pishan is 380 li (158 km) west of Khotan, which is as close as I can measure
the distance between the two towns on modern maps. See CICA, p. 97, and n. 152.
          See also Chavannes (1900) p. 125 and (1907) 174, n. 1, and the discussions by Aurel Stein (1907) pp. 99-103, and
(1921a) p. 86 where he says: “. . . . the identity of P’i-shan with the modern Guma is certain.”
5.11. The kingdom of Yutian 于寘 [Yü-t’ien] = Khotan. The identification of the ancient name of Yutian with the
Khotan oasis has long been confidently established – the evidence from the ancient histories is overwhelming. See,
for example, Pulleyblank (1963), pp. 91, 100, 213. Other accounts include:
“Khotan (Khotana in Kharos̱t̲hī script, Hvatäna in Brāhmī and Hvamna or Hvam in the later Khotanese
texts) was known throughout its 1,200 years as a kingdom (Hvatäna-kshīra).” Zhang (1996), p. 284.
          “As the main source of jade, the Khotan envoys carried pounds (kīna) of the stone for presentation to
the heads of states. They used the word īra- ‘stone’ for their stone par excellence, the ‘jade’, and the Bud. Skt.
word śilā. Hence they could call their river the Śailodā and in Khotan-Saka ranījai ttāji ‘the river of precious
stone (ratna-)’. The īra- in the adjective form īrīnaa- was also used to render the mountain name Vajraka.
The rivers in Khotan were in Turkish times called the yörüng qaš öküš and qara qaš öküš ‘the white and black
jade rivers’.” Bailey (1985), p. 14; also p. 58.

“The major oasis of the southern Tarim basin, Khotan, was favourably set amidst the Yurung-kāsh and the
Qara-qāsh, the only two rivers to carry the melt waters of the Qurum (Kunlun) Mountains northwards to
join the Tarim rather than, like so many others, dissipating into a sea of sand. The fertile loess soils of the
Khotan oasis ensured that its agricultural foundation would support a major settlement, and when Aurel
Stein and other explorers visited it at the turn of the [20th] century, they observed that the region was
underpopulated given its agricultural potential (the population at that time was estimated roughly at c.
200,000, approximately ten times larger than that given in the Hanshu). But despite all its fertility, like all
the other oases its agricultural potential depended entirely on irrigation.
          Khotan was also the centre of silk production in the Tarim Basin and Stein suggested that it might
have been the actual Serindia of the ancient geographers (rather than China) whence the West learned of
the product itself. The legend tells that at the time when the Chinese prohibited the export of silk worms,
mulberry trees and the knowledge of the manufacture of silk, a wily king of Khotan requested the hand of a
Chinese princess in marriage. Before she departed to her new land, her husband made it clear that if she
expected to be kept in silks, she had better procure what was necessary for their production, so she
secreted silkworms’ eggs and seeds of the mulberry tree in her headdress and carried them to Khotan.”
Mallory and Mair (2000), p. 77.
For a broader discussion of the early introduction of silk cultivation to Khotan, see: TWR Appendix A: The
Introduction of Silk Cultivation to Khotan in the 1st Century CE.
“The paper mulberry tree (Broussonetia papyrifera) is also found primarily in the region of Khotan
and its bark was pulped into the earliest paper in the region [c. 105 CE], another gift of Chinese technology
to the West. Both cotton and wool production have been major products of Khotan since antiquity, while
Khotan was also a major supplier of jade to China (the ‘Jade Road’ between Khotan and China is
considerably older than the Silk Road).
          Khotan also occupied a remarkably strategic position. To its south, the forbidding Qurum and
Qaraqurum ranges were absolutely desolate and Stein could count but a mere 400 people scattered across a
territory of 9,000 sq. miles [23,310 sq. km]. To its east one could follow the Silk Road, but beyond Niya
(Minfeng) the oases were so few and far between that it would have been difficult to facilitate any major
approach to Khotan other than one that had been highly organized, such as might be found in Chinese
military operations. To the north lay the full expanse of the Täklimakan Desert. Only the west provided a
relatively easy route through which populations might have entered this region in deep antiquity. Khotan
itself, despite its prestige, was surprisingly small. The sole historical source (Beishi) to provide a dimension
of the town reckons its circuit at 8 or 9 li [3.4 to 3.7 km] and this is roughly confirmed by Aurel Stein’s own
excavations at Yōtkan which discovered that the circumference of the town was merely about 2.5 to 3.2 km
(1.5 to 2 miles).” Mallory and Mair (2000), pp. 77-78.
The invention of paper is attributed by the Hou Hanshu (chap. CVIII, p. 2b) to 蔡倫 Cai Lun in 105 CE (although
there is a heated debate presently underway as to whether paper was, in fact, first invented some time earlier):  
“In the first Huanxing year (105 CE) he [Cai Lun] offered his invention to the Emperor who praised his
skilfulness. From this moment there was nobody who did not adopt the use (of his paper), and this is why
throughout the Empire gave the name of 紙 zhi (to paper) from the honourable 蔡 Cai.” Translated from
Pelliot (1905), p. 6.
          “Old Khotan experienced an economic golden age in those days [c. 4th century], thanks to silk
production and exports to the west. The mulberry trees in the plantations not only furnished the leaves to
feed the silkworms, but from the bark valuable paper was manufactured, an export article also in great
demand. This kind of paper manufacture was widespread in Khotan even at the beginning of the 20th
century, as recorded in the diary of the British Consul-General in Kashgar, C. P. Skrine. . . .
          Khotan’s wealth was not based on silk and paper alone, but also on the production of felt and woollen
carpets and on the export of precious jade. Khotan was indeed predestined to be the centre of jade
commerce, for jade was found along the upper course of both rivers that surround the city. In the west
runs the Yurung Kash which means “white stone” in Uighur and where white jade is found, while in the
east flows the Kara Kash – the “black stone” – which yields green jade.
          Apparently, a lively jade exchange with central China had taken place from the 3rd BC onwards. . . . ”
Baumer (2000), p. 59.
 “I have left to the last the mention of the gold produce of Khotan ;  for if we except the little gold washed
from the sand of the Yurung-kāsh, the precious metal with which the name of Khotan is often associated is
found in its natural state only at places situated at considerable distances from the oasis. The gold mines of
Surgak, Kapa, near the headwaters of the Cherchen river and on the high plateaus of the Arka-Tāgh
towards Tibet, may well have been worked in ancient times. But no mention is made of them in the old
Chinese notices of Khotan ;  and it is doubtful whether, with the exception of the first-named place (on the
upper course of the Niya river), any of these localities ever fell within the political boundaries of the
Khotan kingdom. That the gold extracted from them must have helped to increase the commercial
importance of Khotan, as the nearest emporium for its disposal, may, on the other hand, be considered
certain.” Stein (1907) Vol. I, p. 136.
“Landsell [H. Landsell, Chinese Central Asia. London] (1893) noted, among other occurrences, the presence of
gold at twenty two places in Khotan.” Lahiri (1992), p. 79.
Note: Unfortunately, since late in the 19th century the identical name, Yutian 于寘, has been used for the
subprefecture centred in Keriya, which has, naturally, caused considerable confusion. See note 3.1 and Stein (1907),
pp. 166-172; CICA p. 96, n. 147.
5.12. Refer to Appendix K.
         
5.13. Daxia 大夏 [Ta Hsia] = Bactria – derived from Old Persian Bākhtri-, an Iranian but non-Persian form of the
name. Frye (1963), p. 69. The Avesta gives the form Bāxδi (or ‘Bachdi’). Negmatov (1994), p. 442. For other possible
derivations of this name see Bailey (1985), p. 130.
          There can be no doubt that Daxia referred to the ancient region of Bactria. It was taken over by the Da Yuezhi
and other nomad hordes in the late second century BCE. The previous rulers were of Greek descent and heritage
and had been there since Alexander’s conquest c. 328 BCE. It had become independent of the Seleucids about the
middle of the third century CE but had retained its largely Greek ruling class and was heavily influenced by
Hellenistic culture.
          Bactria is best described as a region (rather than a state) consisting of the fertile plains on either side of the
Amu Darya or Oxus River, and was known to the Persians as the Jayhun. It is usually thought to have included most
of northern Afghanistan, including Badakhshān in the east, and what is now southern Tajikistan and Uzbekistan, at
least as far west as the region of Termez.
“It should be emphasized that Bactria never resembled Parthia in being a unified state. Bactria is above all a
historico-geographical term, rather than a political one. During these nearly five hundred years various
states were formed in this area – the Graeco-Bactrian state, the empire of the Kushans (which continued to
exist for a while after the fall of the Parthian state, and the various principates of the Great Yüeh-chih.”
Rtveladze (1995), p. 181. 
Bactria’s major city, under both the Persians and Greeks (and probably the Kushans), was Zariaspa or Bactra
(modern Balkh). It was situated south of the Oxus, 84 km southwest of Termez, and about 15 km northwest of
modern Mazar-e Sharif. It is a very ancient city, still known throughout the region as the ‘Mother of Cities.’
          It is not clear whether the Greeks managed to retain control of the city or whether, as some claim, it was
taken from them by the Parthians:
“The root of the name Aspionus [an eastern district of Bactria taken by the Parthians probably between 160
and 150 BCE] is clearly the word asp (horse), which was used to form many toponyms in Central Asia. In
Bactria in particular, it was one of the main components of the name of the town Bactra-Zariaspa (golden
horse), which is mentioned by Strabo and Pliny. In view of the linguistic similarities, it is a reasonable
hypothesis that the satrapy of Aspionus was connected with the region of Bactra-Zariaspa. If this is true,
during the reign of Mithradates I the Parthians wrested from the Graeco-Bactrian kingdom of Eucratides
the western territories of Bactria, including Bactra.” Rtveladze (1995), p. 185.
Mark Passehl commented (personal communication July 7, 2003) on the two quotes from Rtveladze above, and I
believe his criticisms are worthy of serious attention:
“I don’t understand Rtveladze’s distinctions between Parthia and Baktria. . . .
          Both were former Persian satrapies which became the “home territories” of successful conquests
states/dynasties (Parthian Empire of the Arsakids, Bactrian dominion of the Diodotids, Euthydemids, etc.)
          Next page the comments about the Arsakid seizure of Baktra seem quite wrong. The Arsakids
probably took the two satrapies right near the end of Eukratides’ reign when he was campaigning in India
(ca. 146 BC), but the archaeology (Rapin’s article) seems to say that even when the great nomad invasions
came in the 140s-130s BC Baktra held out longest as a Greek-dynasty outpost. So either at their weakest
they retook it from the Parthians (unlikely!) or never lost it when they lost the two westernmost
provinces.”
The middle reaches of the mighty Oxus River, which presently forms the northern boundary of Afghanistan, is
frequently up to a mile [1.6 km] wide. It, however, has a number of important fordable points, each of which,
naturally, became the site of fortifications to control traffic across the river. The system used to ford the river
during the Kushan period probably hadn’t changed very much by the time the 1911 version of the Encyclopædia
Britannica was compiled:
          “The principle on which the Oxus ferries are worked is peculiar to these regions. Large flat-
bottomed boats are towed across the river by small horses attached to an outrigger projecting beyond the
gunwale by means of a surcingle or bellyband. They are thus partially supported in the water whilst they
swim. The horses are guided from the boat, and a twenty- or thirty-foot [6-9 metres] barge with a heavy
 load of men and goods will be towed across the river at Kilif [approximately 100 km west of Termez]
(where, as already stated, the width of the river is between 500 and 600 yards [457–549 metres]  – only)
with ease by two of these animals. The Kilif ferry is on the direct high-road between Samarkand and Akcha.
It is perhaps the best-used ferry on the Oxus.” EB – under ‘Oxus’.
By the time of Fitzroy Maclean’s visit to the region in 1938, horses had given way to steamboats:
          “The frontier post is situated at Patta Hissar [near Termez]. Along the river, stretch for a mile [1.6
km] or so in a narrow strip the barracks of the frontier troops, the officers’ bungalows and piles of
merchandise awaiting transhipment; then, as far as the eye can reach, a jungle of reeds ten or eleven feet
high [3-3.6 metres], reputed to harbour tigers as well as a great deal of smaller game. The Oxus must at this
point be almost a mile wide, a vast muddy river full of mud flats and sandbanks, flowing between low mud
banks. I have seen more exciting rivers, but its name and the knowledge that very few Europeans except
Soviet frontier guards have ever seen it at this or at any other point of its course, made up for its rather
drab appearance. In the distance there were some blue mountains. . . .
          The crossing took half an hour or more, the sandbanks making navigation rather complicated. From
the upper floor of the two-storeyed cottage which combined the functions of bridge, engine-room and
sleeping-quarters for the crew, I commanded an extensive view of the river and of the jungle on both
shores, with, on the Soviet side, watch towers at intervals and a patrol of frontier troops setting out to look
for Diversionists. On the Afghan side there was, as far as I could see, nothing except jungle.” Maclean
(1949), pp. 129, 131.
           “After a few miles’ riding we emerged from the reeds of the jungle into the desert. It was very much
like any other desert in Central Asia, with its dunes of drifting sand and shrivelled tamarisk bushes.
Marmots with their short forelegs, long hind-legs and bushy tails whistled petulantly and scuttled in and
out of their holes. From time to time we came upon the bleached skeletons of horses and camels. Then,
after some miles of crawling up sand dunes and slithering down the other side, we came out on to a flat,
completely barren plain with absolutely nothing in sight. Underfoot was hard white clay. There was no
road, but something approaching a track had been worn by the caravans making their way down to the
Oxus. . . .
          [After a stop at a small mud fort] An hour or two later we sighted a small earth-coloured hump on the
horizon. The drab, khaki-coloured desert was absolutely flat and it was a very long time before we came
near enough to see that it was the immense dome of a ruined mosque, apparently of very great age. From
now onwards the plain was scattered with ruins, sometimes a few crumbling stones, at others, whole cities
with mosques and watch towers and city walls stretching for miles. Away to the west lay what is left of
Balkh, the ancient Bactria, the Mother of Cities. . . . .  There were no signs of vegetation near any of these
ruins and any water supply there may have been must have dried up or been diverted.
          Towards sunset we came to the cultivated fields and plantations of the oasis of Seyagird [about 25 km
north of Mazār-e Sharif], the first we had seen since the Oxus. Here a large military fort, with crenellated
mud wall, towers over a cluster of houses and gardens surrounded by high walls and a small mosque, all
built of the mud bricks used throughout Turkestan. In a large open space before the fort the camels of a
number of caravans were resting, before setting out once more.” Maclean (1949), pp. 134, 135.
Bactria, with its major trade emporium of Balkh or Baktra, was a key centre on the extensive trade routes
developed to transport lapis lazuli, spinel rubies and, quite possibly, emeralds – from the mines in the mountains –
see Giuliani et al (2000), pp. 631-633; Giuliani et al (2000b), pp. 58-65; Schwarz and Giuliani (2001), pp. 17-23;
Bowersox (1985), and refer to Appendix K.
          Lapis lazuli from Badhakshan was being traded to Mesopotamia, and Egypt from the second half of the fourth
millennium and to the Indus River cultures by the third millennium BCE. Sarianidi (1971), pp. 12-15. These routes
were later to form the basis of the networks we now call the ‘Silk Routes.’
“Daxia (Bactria) is described as lying more than 2,000 li [838 km] southwest of Ferghana, south of the Gui
(Amu Darya). Like the people of Ferghana, its occupants were a settled people living in walled towns. They
lacked powerful chiefs and rather were divided into small individual towns with their own leaders. Their
armies are described as insignificant and cowardly, a clear come-down from their reputation when they
faced Alexander, but they excelled in commerce with enormous markets, especially in their capital
Lanshicheng (Bactra). They numbered about a million people. While in Bactria, Zhang saw trade goods
from Sichuan and asked how they had come there. He learned that they were obtained from a land called
 Shendu (i.e. Sind, the Punjab), which lay in the region of a great river (the Indus) and was occupied by a
people who employed elephants in warfare.” Mallory and Mair (2000), p. 59.
“Archaeological evidence reveals intensive exploitation of new agricultural land and the expansion of
agricultural oases at the beginning of the Christian era in the river valleys and ancient agricultural oasis
areas of Central Asia, especially in the southern regions, even though the best and most suitable croplands
were by that time already under cultivation. It has also been established that, with the opening up of new
regions and the extension of crop-farming to the northern provinces of Central Asia on the lower reaches
of the Zerafshan, on the middle reaches of the Syr Darya and in the Tashkent oasis, large numbers of
nomadic livestock-breeders switched to a settled way of life and new centres of urban civilization were
formed. As a result of the extensive development of irrigation networks, practically all the main provinces
of Central Asia were brought under cultivation during this period and the establishment of the major crop-
growing oases was completed. The extent to which northern Bactria was populated and brought under
cultivation at this time can be judged from the 117 archaeological monuments of the Kushan period
recorded in recent years in the territory of the Surkhan Darya province. A major channel, the Zang canal,
leading from the Surkhan river, was constructed. In the zone irrigated by it a new oasis, the Angor, was
established around the town of Zar-tepe. The founding of Dalverzin-tepe as a major urban centre also dates
back to this period. The Surkhan Darya and Sherabad Darya valleys, with their flourishing agricultural
oases, fortified towns and extensive grazing lands, were able to provide a strong base for unifying the
domains of the Yüeh-chih on the right [northern] bank of the Amu Darya. When they were unified by the
ruler of the Kuei-shuang [Kajula Kadphises], who subjugated the four other Yüeh-chih principalities, the
nucleus of the Kushan Empire was formed.” Mukhamedjanov (1996), pp. 265-266. 
Strabo (c. 23 CE), XI. xi. 1, also describes the exceptional fertility of ancient Bactria and proves that its reputation
had spread as far as the Mediterranean world:
“As for Bactria, a part of it lies alongside Aria towards the north, though most of it lies above Aria and to
the east of it. And much of it produces everything except oil. The Greeks who caused Bactria to revolt grew
so powerful on account of the fertility of the country that they became masters, not only of Ariana, but also
of India, as Apollodorus of Artemita says: and more tribes were subdued by them than by Alexander. . . . ”
“The coexistence of Hellenistic traditions might have continued after the Yuezhi-Kushan entered into
Daxia. One Tang Dynasty scholar, who also annotated Sima Qian’s History, quoted from the a now-lost text
[the Yiwuzhi by the 3rd century scholar, Wan Zhen] as saying:
“The Great Yuezhi is located about seven thousand li north of India. Their land is at a high altitude;
the climate is dry; the region is remote. The king of the state calls himself “son of heaven.” There
are so many riding horses in that country that the number often reaches several hundred
thousand. City layouts and palaces are quite similar to those of Daqin (the Roman empire). The
skin of the people there is reddish white. People are skilful at horse archery. Local products,
rarities, treasures, clothing, and upholstery are very good, and even India cannot compare with
it.”36
It is difficult to verify the sources of this record about the Kushan, since the quoted book appears
to be lost.37 The descriptions, however, accord very well with the horse-riding Kushan who ruled a
formerly Hellenistic country. The climate and location sound like Bactria; the kings of the Kushan did
indeed call themselves devaputra, meaning “son of heaven” or “son of god.” They owned numerous good
horses and cultivated nomadic skills and cultures. Yet they ruled a country with a population of Greeks and
other immigrants from the Mediterranean, so that the architecture of the country combined Greco-Roman
style with local materials and flavor. At least it looked similar to the Roman style in Chinese eyes, and the
people looked fairer than Indians and some other Central Asian populations.”
36. Sima Qian, Shiji, 123/3164.
37. The book entitled Nanzhouzhi, literally “the history of the southern states,” authored by Wan Zhen
[3rd century CE – see Leslie and Gardiner (1996), p. 333], was available to Zhang Shoujie, the Tang
scholar who annotated the History by Sima Qian, as it was listed in the bibliographies of the Tang History
with the title of Nanzhou Yiwuzhi, meaning “history of exotic things in the south states.” However, it did
not appear in the bibliographies of later official histories. Liu (2001), pp. 278-279.
 Liu (2001), p. 278.
The Da Yuezhi overran Bactria and settled there in the late second century BCE. This gave them control of the main,
and increasingly busy, overland trade routes between China, India and the West. This not only quickly made them
rich and powerful, but their exposure to Persian, Hellenic and Indian cultures helped them become a more
sophisticated and effective force. It is thought that before they entered Bactria they were not literate. By the time
they invaded northern India in the first century CE they had become capable administrators, traders and scholars.
The ‘Iron Gates’ – the northern border of Daxia
About 107 km northwest of Termez (13 km west of Derbent) the main trade route passed through a formidable, very
narrow and easily defended gorge known in antiquity as the “Iron Gates” which has traditionally marked the
boundary between Sogdiana and Tokharistan and almost certainly marked the frontier between Kangju and Kushan
territory during the period of the Kushan Empire.
“When he spoke of borders, Euthydemus [Graeco-Bactrian monarch, late 3rd century CE] probably meant a
dense ridge of mountains consisting of a spur of the southern part of the Hissar chain together with the
adjacent Baysuntau and Kughitang Mountains. In this area, near the settlement of Darband, a monumental
defensive wall of the Kushan period (1st–2nd centuries A.D.) has been discovered. This wall (Fig. 1) was
probably built to block the main entry route into Bactria and also the gateway which in early medieval,
especially Chinese texts, is known as the «Iron gates». Further investigation of the wall and of the adjoining
fortifications has brought to light fragments of pottery of the Graeco-Bactrian period, a fact which may
indicate that the wall was already in use in the preceding period, i.e. in the early Graeco-Bactrian period. It
is possible that after Euthydemus’s political successes and the consolidation of his power, he and the later
kings of Graeco-Bactria managed to defend this part of the border against the onslaught of the nomads.
The most valuable part of the border, the one about which the Graeco-Bactrian kings were worried, was in
my opinion the north-western side of the country, the area along the middle reaches of the Amu-darya
(the area of modern Gaurdak, Mukry, Kerki and Chardjow), where entry into the centre of Bactria was
facilitated by the ford over the river at Kerki and not impeded by impassable mountains. At any rate it is
precisely this region that Strabo means when he tells of the Parthians annexing «part of Bactria, driving
back the Scythians, and even earlier Eucratides and his successors» (STRAB. XI, 9, 2). In the same passage,
listing the principal towns of Bactria, Strabo mentions Eucratidea (Dilberdjin) [identified in Rapin (2001),
pp. 217-218 however, as Ay Khanum]. «After seizing this region the Greeks divided it into satrapies; of
these, the Parthians took the satrapies of Aspionus and Turiva from Eucratides» (X, 9, 2). When he speaks of
Sogdiana, «which is situated above Bactria», the ancient author is referring to the region known to modern
scholars as southern Sogdiana (the western and south-western parts of the modern region of Kashka-
darya).” Abdullaev (1995). In: ITLOTG, pp, 151-152.
When Xuanzang passed through here in 630 CE he described it as having iron or ironclad gates with numerous small
bells suspended on it. Later writers make no mention of actual gates. Clavijo, the Spanish ambassador to the court
of Timurlane passed through the Iron Gates in August 1405 CE. He said the ravine looked:
“as if it had been artificially cut, and the hills rise to a great height on either side, and the pass is smooth,
and very deep. In the centre of the pass there is a village, and the mountain rises to a great height behind.
This pass is called the Gates of Iron, and in all the mountain range there is no other pass, so that it guards
the land of Samarkand in the direction of India. These Gates of Iron produce a large revenue to Timūr, for
all the merchants who came from India pass this way.” Quoted in Verma (1978), p. 39, [from G. Le Strange,
The Lands of the Eastern Caliphate. Cambridge, 1930, pp. 441-2.]
          “The texts in this place [in the Tarikh-i-Rashidi], have Darband-i-Ahanin, or “Iron Gate,” but in all other
places Kulugha, the name by which this pass was usually known. It is often mentioned by ancient travellers,
but has very rarely been visited in modern times, at any rate up to within the last few years, or before the
Russians became possessors of Samarkand and Khokand. The gates, in ancient times, were a reality, for the
Chinese traveller Hiuen Tsang, who passed the Darband in 630 A.D., describes the defile as “closed by
folding gates clamped with iron.” (See Sir H. Yule in Woods Oxus, 1872, p. lxix.) At the time of Chingiz Khan,
when Chinese travellers frequently went backward and forwards between China and the conqueror’s camp
in various parts of Central Asia, the pass of the Iron Gates is frequently mentioned under the name of Tie-
men-Kuan (literally, Iron Gate barrier) ; and one of them, the Taoist monk Cháng Chun, describes his
passage through the defile in 1222, with carts and an escort of a hundred Mongol and Muhammadan
 soldiers : “We crossed the mountains in a south-east direction and found them very high. Masses of rock
were lying scattered about. The escort themselves pulled the carts and took two days to pass to the other
side of the mountains.” (Chinese Mediæval Travellers to the West, by Dr. E. Bretschneider 1875, pp. 41, 42) The
gates themselves seem, thus, to have disappeared by the thirteenth century, and they had certainly done
so at the beginning of the fifteenth, when Ruy Gonzalez Clavijo visited the spot, in the course of his
embassy (1403-5) from Henry III, of Spain to Amir Timur. He wrote : “These mountains of the Gates of Iron
are without woods, and in former times they say that there were great gates covered with iron placed
across the pass, so that no one could pass without an order.” (See Embassy of Ruy G. Clavijo to Court of Timur,
by C. R. Markham, Hakluyt Series, 1859, p. 122.) From the time of Don Ruy down to 1875, when the Russian
Hisar Expedition passed the Darband, no European appears to have seen (or, at any rate, to have described)
the defile. Mr. N. A. Mayef, who accompanied the Russian Expedition, described the spot thus : “The famous
ravine of the Iron Gate winds through a high mountain chain, about twelve versts to the west of Derbent. It
is a narrow cleft, 5 to 35 paces wide and about two versts long. It is known now as Buz-ghala Khána (i.e., the
House of Goats). Its eastern termination is 3540 feet above the sea ; its western termination 3740 feet. A
torrent, Buz-ghala Khána bulák, flows through it . . .” (Geogr. Magazine, Dec., 1876, p. 328).” Elias (1885), p.
20, n. 3.
After passing through the Gates of Iron one could either head north to Samarkand through Kesh (modern
Shahrisabz), or northwest towards Bukhara. From Bukhara one route led southwest via Merv into Parthia – the
other avoided Parthian territory by heading northwest along the Syr Darya (or Jaxartes) to the Aral Sea and then
continued around to the north of the Caspian before reaching Tanais, the port on the Sea of Azov which gave
maritime access via the Black Sea to the Mediterranean.
The derivation of the name ‘Daxia.’
The derivation and significance of the Chinese name for Bactria, Daxia 大夏 [Ta Hsia], is still being contested. I
tend to favour a connection with the name ‘Tajik’ (Tibetan ‘Tzag-zig’), as proposed by Charles Allen:
          “The Persian-speaking peoples of the upper Oxus or Amu Daria are known as the Tajiks, a name
preserved in the former Soviet republic and now the state of Tajikistan, which borders the northern shores
of the Oxus. The name ‘Tajik’ carries a special resonance for followers of Bon because, as Tzag-zig, it is
linked inextricably with Olmo-lungring, the homeland of their religion.
          The Tajik country is set amongst the almost impenetrable mountain barriers of the Hindu Kush and
the Karakoram but softened by broad, fertile valleys which give access from the west. These valleys form
the main migratory and trade routes of the region, a frontier between the settled agricultural peoples of
the south and the nomads of the inner Asian steppes. For those seeking to break through the mountains
and plunder the fertile Indian plains they provided a natural gateway, which is why they have so often
been shaken by the passage of invaders. Between about 500 BCE and 500 CE this Tajik country – known to
the Chinese as Ta-hsia and to the Greeks as Bactria and Sogdiana – was ruled over in turn by Achaemenid
Persians, Mauryans, Alexander’s Greeks, Parthians, Scythians, Sasanid Persians and Huna – as well as a
nomadic people known to the Chinese as the Yuezhi, who came to call themselves the Kushans.” Allen
(1999), p. 184.   
Following are accounts of some of the other main theories:
“Haloun (1926), pp. 136, 201-202, has made it clear that the term Ta Hsia originally referred to a mythical or
fabulous people, vaguely located in the North (but eventually shifted to the West and even to the South).
He states that it was Chang Ch’ien personally who identified the Bactrians with the Ta Hsia, the
westernmost people he knew, but that he did not use the words ta and hsia to reproduce their actual name.
Haloun rightly stresses this last point, viz. that the pronunciation of this old-established, mythological
term need not have been anything like an approximation of the name of the actual country. Henri Maspero
completely endorses Haloun’s views in his review of the latter’s work in JA 1927, pp. 144-152.” CICA: 145, n.
387.

“Further to the west the Chinese name for Ferghana, “Dawan,” and that for Bactria, “Daxia,” were also
variations of Tuhara.15 Bactria, a name given by the Greeks to northern Afghanistan and Uzbekistan, was
known as the “land of the Tuharans” as late as the seventh century C.E., according to the Chinese pilgrim
Xuanzang.16
 15. Yu Taishan. A Study of Saka History, p. 72.

16. Ji Xianlin, Da Tang Xiyuji Jiaozhu (An Edited Edition of the Travelogue of the Western Region by
Xuanzang of the Great Tang Dynasty) Beijing: Zhonghua Shuju, 1985, p. 100.
Liu (2001), p. 268.

Taishan Yu also has some interesting comments to make on “Daxia” and its history:
          “In the “Xirongzhuan” 西戎傅 of the Weilue 魏略 it is recorded: “the states of Jibin, Daxia, Gaofu
and Tianzhu are all subject to the Da Yuezhi.” “Da Yuezhi” here also refers to the Guishuang Kingdom. If
the Guishuang Kingdom was established by the Daxia, the record of the Weilue would be tantamount to
saying that the Da Yuezhi were both the conqueror and the conquered.
          In my opinion, “Da Yuezhi” here actually refers to the Guishuang Kingdom. However, “Daxia” here
must refer to Tukhārestān. Therefore, the statement that the state of Daxia was subject to the Da Yuezhi
only shows that Tukhārestān (the land of the former state of Daxia) was a part of the Guishuang kingdom,
namely the Xihou of Guishuang, was established by the Daxia, but it was not equal to the state of Daxia, and
that the territory of the Guishuang Kingdom far exceeded the boundary of the former state of Daxia.” Yu
(1998), pp. 31-32.
          “Daxia” was a transcription of “Tochari”, but there were some differences between “Daxia” as
described in the Shiji, ch. 123 and the Hanshu, ch. 96 and “Daxia”, in the pre-Qin books. The latter was
referring to the Tochari. The former had in fact included the Asii, the Gasiani and the Sacarauli. As far as
the Tochari, those who had migrated west to the valleys of the Ili and Chu and then to Tukharestan should
be different from those who remained in the Hexi region, due to being affected by different surrounding
tribes. More accurately, there must have been some differences in language, custom and physical
characteristics between them.
          Also, there must have been differences between the Tochari who moved south into the Pamir region
from the valleys of the Rivers Ili and Chu and then spread east to the Tarim Basin, and those who entered
Tukharestan from the northern bank of the Syr Darya.
          For the same reason, though “Yuezhi” “Guishuang”, and “Jushi” and “Qiuci” all were transcriptions of
“Gasiani”, there must have been some differences between those who migrated west in late [sic] of the 7 th
century B.C. and those who migrated west in c. 177/176 B.C. The former had divided into two groups later.
One of them entered Tukharestan, and the other entered the Tarim Basin. There must have been some
differences between the two groups. The circumstances of the Asii and the Sacarauli may be explained at
the same time.” Yu (1998), p. 35.
“Markwart (1901), p. 206, suggests that the Tochari must have been identical with the Daxia. The Hellenic
Kingdom of Bactria was destroyed by the Daxia, and the latter was destroyed by the Yuezhi. I think his
theory is correct. . . . ” Yu (1998), pp. 38-39, n. 18.
“. . . .  Tarn suggests that “Asii”, whose adjectival form was “Asiani”, may have been identical with
“Kushān”. I disagree. Yu (1998), p. 40, n. 30.
5.14. The kingdom of Gaofu 高附 [Kao-fu] = Kabul. Gaofu is almost universally taken to represent Kabul or, rather,
Kabulistan. The Greek form of the name is usually given as Kophen, although Strabo writes it Kophes – the Vedic
form was Kubha. See Lèvi (1895), pp. 372-373. This identification as the phonetic resemblance plus the geographic
indications make, I believe, a convincing case. Pulleyblank (1963), p. 223, reconstructs the name as *kauĥ-bōh =
Kabul, Κάβουρα [Kaboura]. See also, Chavannes (1907), p. 192; CICA, p. 122, n. 296.
          Bailey (1985), p. 10, gives: “Kābul, Καβουρα, Zor. Pahl. K’pwl k’wl *kāpul, kāvulastān, N Pers. Kābul. For more
details, see ibid., p. 119.
          Kabul naturally related more to the west and south than to the valley containing Kapisha/Begram which joins
the lower Kabul River valley on the way to Jalalabad and Peshawar. The easily-defended and extremely narrow
gorge which the Kabul River runs through to the west of the city was always difficult to traverse and often
completely flooded.
5.15. The kingdom of Tianzhu 天竺 [T’ien-chu] = Northern India). These names for India: Tianzhu 天竺 [T’ien-
chü] – a transcription of the Iranian, ‘Hinduka, – and Juandu 身毒 [Chüan-tu] – a transcription of Sanskrit,
‘Sindhu,’ seem to be merely different forms of the same name and are practically interchangeable.
Tian 天 K.361a *t’ien / t’ien; EMC: thɛn
zhu 竺 K.1019f *tô / tuok; EMC: truwk
Juan 身 K.386a *śi̯ĕn / śi̯ĕn; EMC: ɕin
du 毒 K.1016a *d’ôk / d’uok; EMC: dəwk
The name Juandu is sometimes (less correctly) rendered Shendu. They are both ultimately derived (perhaps via
Iranian Hinduka) from Sanskrit Síndhu – a river or stream – especially the Indus. A form of this name was used in
very early Indian literature to refer to the country around the lower Indus – known today as ‘the Sind’ or ‘the
Sindh.’
          Juandu seems to be used here in a more general sense than Tianzhu, which is specifically stated to be beside a
‘great river,’ i.e. the Indus. However, Tianzhu is also frequently used in later times in a much broader sense –
sometimes referring to the whole of northern India, including the Ganges valley and sometimes, even, to the whole
sub-continent. For a detailed discussion of the derivation of these names see Bailey (1985), pp. 22-24. Also see:
Pulleyblank (1963), pp. 108, 117.
          Mukherjee (1988), pp. 297-303) argues quite convincingly that the name Juandu referred only to the region of
the lower Indus at the time of Zhang Qian’s report included in the Shiji (completed c. 100 BCE). Later, as Chinese
knowledge of the subcontinent expanded, Juandu, and by association, Tianzhu, came to include lands further and
further east until, by the time of Kang Dai’s mission to Funan (c. 245-250 CE), Tianzhu referred to the whole of
northern India and even included the kingdom of Danmei (Tāmralipti) at the mouth of the Ganges on the Bay of
Bengal. By the time of Xuan Zang’s visit in the 7th century it included “roughly the whole of the subcontinent.” See
also: CICA, p. 97, note 154 and the discussions in Leslie and Gardiner (1996), pp. 257-258.
          “We next come to T’ien-chu (天竺) and T’ien-tu (篤) said to represent only one name pronounced
something like Tendu or Tintok. We are told by one Chinese writer that the name Tien-chu was first
applied to India in the Han Ho-Ti period (A.D. 89 to 106) but the authority for the statement is not given.
Another account makes Mêng K’an (about A.D. 230) the first to identify T’ien-chu with Shên-tu, but this
likewise is unsupported by authority. We are also told that the chu (竺) of T’ien-chu is a short way of
writing tu (篤), a statement which is open to very serious doubt. This word tu occurs in the ancient
classical literature, and native students declare that it represents an earlier chu. This is specially noted with
reference to the occurrence of tu in a wellknown passage of the “Lun-Yü”. Then as to the first part of the
name there seems to have been an old and perhaps dialectical pronunciation of the character as Hien or
Hin. This pronunciation is found at present in the dialect of Shao-wu foo in the Province of Fuhkeen in
which 天竺 is read Hien-tu.
          But what was the sound originally represented by the character now read Chu in the compound T’ien-
chu? It seems that no satisfactory and decisive answer can be given at present to this question. We find
that in the Han period the character represented several sounds which cannot be said to be very like each
other. The upper part chu meaning bamboo is not significant here, we are told, but only phonetic; and the
lower part is significant, and refers the word to the category earth. The character might then be read
something like du. But this account of the syllable may be doubted, as we learn also that the character was
read like tek, an old and still current pronunciation of the word for bamboo. Then this same character was
also read as chah, tuh, kat, and kc or gou. . . .
          T’ien-tu, on the other hand was the name of a place in the Eastern Sea mentioned in the “Shan-hai-
ching” along with Chao-hsien or Korea. This place was afterwards identified wrongly with the T’ien-chu of
writers on India and Buddhism. But we find mention also of another T’ien-tu (written in the same way), a
small country to the west of China, which has been supposed by some to be the Shên-tu of Chang Ch’ien.
          Whatever the name T’ien-chu may have signified originally, however, it came to be given by the
Chinese in their literature to the great extent of territory between the Bay of Bengal and the Arabian Sea,
and reaching from the Kapis country in the north to Ceylon in the south. Thus used it supplanted the old
Shên-tu, and all other names for India among the Chinese; and it continued to be the general literary
designation for that country down to the T’ang period when the new name Yin-tu was brought into fashion.
We even find the term T’ien-chu used with a wider application, and it is employed as a synonym for
 “Buddhist countries”, for example, in a title given to the “Fo-kuo-chi” of Fa-hsien.” Watters (1904-05), I, pp.
135-137.
5.16. The kingdom of the Da Yuezhi 大月氏 [Ta Yüeh-chih] = the Kushans. There is a translation of this whole
passage, plus others on the Kushans, in the useful and interesting article by Zürcher (1968), pp. 346-390. See also:
Enoki (1968), pp. 1-13.
          Theories abound regarding the possible connections of this name, Yuezhi. Translated literally, it means
something like “The Moon People.” This explanation seems to lead us nowhere, although among the many gods
represented on the coins of Kanishka and his successors was Mao, the Iranian moon goddess, partner of Miiro /
Mihr, the sun god. The moon goddess is also found represented by the name of the Greek goddess, Salene or Selene.
          Of more interest, perhaps, are the theories connecting the Chinese name (Da) Yuezhi with one or the other
tribes or peoples mentioned by Classical and Indian writers as invading first the Bactrian region and, later, India
itself.
          The first theory, developed by W. B. Henning in his 1965 paper, “The first Indo-Europeans in history,” is
discussed at some length in Mallory and Mair (2000), pp. 281-282. They explore Henning’s suggestion that the
ancient pronunciation of ‘Yuezhi” could be approximately reconstructed as *Gu(t)-t’i and related it to the ‘Guti’
people who began harassing the western borders of Babylon from c. 2100 BCE.
          According to Assar (2003), the Parthian king Mithradates II mounted a major campaign into the “Gutian
country” circa 120 BCE and there is a reference to actions by Parthia involving the Guti as late as circa 77 BCE.
          Apparently, Henning believed that Guti in the ‘Kuchean-Agnean’ or ‘KA’ language “would have been rendered
Kuči, and hence be equivalent to Kuchean. As for the toχri mentioned in the Uighur colophon, Henning believed one
need look no further than the name of the Tukriš who had been neighbours of the Guti in western Persia and hence
had given their name both to the toχri of the northern Tarim and the Tocharians of Bactria.”
          Unfortunately, for this theory, Mallory and Mair find his supposed support on the basis of similar ceramics
unconvincing but, “Of greater detriment to such a theory is that Henning accepted a reconstructed Chinese
pronunciation of Yuezhi as *Gu(t)-t’i, when, in fact, it is commonly reconstructed now as *ngi̯ wāt-ti̯ĕg – which
makes it a far less transparent correspondence.”
A far more convincing argument is made in the detailed essay on the name ‘Gara’ in Bailey (1985), pp. 110-141 and
his ‘Epilogus’ on p. 142. I will have to summarize the development of his position by quoting brief excerpts from his
text:
“In Khotan-Saka script this name is written gara, inflected as an -a- stem, plural gara, gen. plural –garām̲,
loc. plural garvā, garrvā, and allative (‘towards’) garvās̲t̲ä. . . .
          Below, reasons are given for equating the Khotan-Saka gara- with the γαρα of Greek Θογαρα, and
Tibetan -gar in to-gar. The -a- is always an essential part of the name, and was emphasized by the long -ā- in
Bud. Skt. tukhāra, N Persian tuxāristān and Khotan-Saka ttahvāra. . . .  
          The development of g, γ, χ (stop, fricative, unvoiced fricative) is important. Tibetan had -gar, -ggar in
tho-gar and thog-gar, but also bh̲o-gar for Bukhāra, and could put -d-k, -dk- in place of -g-.
          The replacement of voiced γ by unvoiced χ is fairly common in various languages. . . .
          The forms of the name Gara involve many complex differences. In the ninth- to tenth-century
Khotan-Saka texts, when the Turks of various tribes are reported in the Chinese cities of S̲acū and Kamcū
(Θροανα and Θογaρα) in good orthography, the people of Gara are cited: KT 2.113.102 mājā gara ‘our Gara
(allies)’. . . .
          The Chinese records report a people whom they named with the syllable 月 (with added suffix 支
or, with the same pronunciation according to an old gloss, 氏) one of whose centres was the very Čaʼn-ie,
the centre also of (θο-)γαρα, Tibetan hgar and, as proposed above, of the Khotan-Saka gara- in the region of
Kamcū (Θογαρα). . . .
          This 月, if it can in any way be found to indicate such a syllable as this gara-, will easily express the
same ethnic name in the very place of its base. This can in fact be shown. . . .
          For 月 the Tibetans spelt hgvyar, hgyar, hgvar in which the laryngeal h- could also indicate a nasal
sound, as in Ga-hǰag for Kančaka-, the name of Kāšγar. . . .
          The importance of the unaccepted transcriptions of 月 by G. Haloun (sgu), K. Enoki (sguǰa), and Ed.
Pulleyblank’s (i̭at-), lies in their recognition that the name began with 月 and that consequently the 大
t’ai, ta ‘great’ placed before was an adjective epithet. When later two divisions of these people were known,
 besides the 大 ta ‘great’ they employed also 小 ‘small, little’ for the group remaining beside them in
the Nan-şan and in the Köke-nagur (Kokonor) region.
          The Chinese quoted this name adding to 月 ( = γar) a syllable 支 K 1212 t̲s̲ï < tśie̯ (from t’a), G 864 a t̑i̯ě,
and 氏 K 879 s̲ï < tśie̯, G 864 a t̑i̯ĕg. The syllable is then a foreign -čik, -jik to be read -čik, with either -i- or -ī-.
          To an Iranist the -čik is the commonest of suffixes to form ethnic names. Three forms are known. . . .
          The base tau-: tu- ‘to increase in size, strength or number’ is very widely attested in Indo-
European. . . .
          For the present problem of the gara- it is important to recognise Iranian tu- ‘great’. . . .
          In the θο- of θογara (second century A.D.) and το- of τόχαροι of 300 years earlier (second century B.C.)
is transmitted and Iranian tu- ‘great’ (from earlier tuυi-, as in Old Ind. tuυi-). Note that Old Iranian did not
have the graphic means to distinguish ŭ from ǒ, so that foreigners recorded Iranian u as either u or ǒ. With
u and o distinct, Greek τόχαροι, Armenian toxara-stan, touuxrstan, touxari-k’ (ou = u), Bud. Skt. tukhara-, Old
Ind. tokşāra-, Kuči-Skt tokharika, Arabic script tuxāristān. . . . ” Bailey (1985), pp. 110-115 + 118-119, 123. For
those with a special interest in the issue, I recommend a thorough study of his original essay.
 
Following from this, it is likely the ‘Kara’ mentioned of some of Kujula’s coins denotes that he belongs to the Gara
people = the Yuezhi.
          There is also a possibility worth considering that the name Yuezhi is related to that of the Άσιοι or Asiani
mentioned in Classical sources along with the Tochari as one of the tribes who invaded ancient Bactria:
          “Pelliot cited this example apropos of the famous and controversial name Yüeh-chih 月氏 M.
ŋi̯wαt-ci̯e, pointing out that the initial ŋ- was unlikely to have represented a foreign g-, as has been
generally assumed, before the mid-T’ang period. Pelliot did not himself make any proposal as to the true
equivalent of the name but his argument greatly strengthens the case for one of the many proposals that
have been made, namely that of the ’Ιάτιοι found on the north side of the upper Yaxartes in Ptolemy. The
initial of the second syllable would have been still unpalatalized *t- ath the beginning of the Han dynasty
when the Yüeh-chih first appear. The labial element in the Chinese transcription remains unexplained. The
true initial may have been the yw- found in some Tocharian words (= I.P. ´?) which could not have been
exactly represented in any other way in Greek. The question as to whether the ’Ιάτιοι are the same as the
Άσιοι or Asiani, as has often been stated, must be left aside for the moment. The equation seems highly
probable on historical grounds.” Pulleyblank (1963), pp. 93-94. For a discussion of the occasional later
replacement of the second character in Yuezhi (zhi: rad. 83) by other characters (zhi: rad. 65; zhi: radical 75-
4), see ibid. pp. 106-107.  
Other proposals and quotes of interest on this subject follow:
“As we have just mentioned, the people who emerge as Tocharians in Western sources are often
equated with a branch of the Yuezhi of Chinese sources who were driven first from the Gansu borderlands
by the Xiongnu, then further west by the Wusun, arriving at the Oxus, and going on to conquer Bactria and
establish the Kushan empire. Narain argues that once one accepts the equation Tocharian = Yuezhi, then
one is forced to follow both the Chinese historical sources (which for him would propel the Yuezhi back to
at least the 7th century BC) and the geographical reference of their first cited historical location (Gansu) to
the conclusion that they had lived there ‘from times immemorial’. Narain infers that they had been there
at least since the Qijia culture of c. 2000 BC and probably even earlier in the Yangshao culture of the
Neolithic. This would render the Tocharians as virtually native to Gansu (and earlier than the putative
spread of the Neolithic to Xinjiang) and Narain goes so far as to argue that the Indo-Europeans themselves
originally dispersed from this area westwards. Seldom has a tail so small wagged a dog so large.” Mallory
and Mair (2000), p. 281.

“By the third century B.C.E., when the Xiongnu became a real threat to the border of the Chinese empire,
the Yuezhi were better known as suppliers of horses.” Liu (2001), p. 272.
“In the Sanguozhi 三國志, ch. 3, it is recorded that on the date of Guimao 癸卯 of the 12th month, in the
third year of Taihe 太和 (i.e., A.D. 229), “The king of the Da Yuezhi, Bodiao 波調 (Vāsudeva), sent his
envoy to present tribute and His Majesty granted him a title of “King of the Da Yuezhi Intimate with Wei
 魏.” If the Guishuang Kingdom was established by the Daxia, it would not have accepted this title.
            In my opinion, the so-called Da Yuezhi actually [by this time] included the Asii, the Tochari, the
Gasiani and other tribes. The Xihou of Guishuang may have been the Gasiani, because “Guishuang” can be a
transcription of “Gasiani”. As mentioned above, the Gasiani and the Yuezhi had the same origin, thus
“Guishuang” and “Yuezhi” were objectively different transcriptions of one and the same name. Therefore,
there was no difference between “the king of the Da Yuezhi” and “the king of the Great Guishuang”. Why
should Podiao not have gone ahead to accept?” Yu (1998), p. 31.
“The Yuezhi resided on the border of agricultural China even earlier that the Xiongnu. While the Xiongnu
were famous in history because of their conflicts with Chinese empires, the Yuezhi were better known to
the Chinese for their role in long-distance trade. Ancient economist Guan Zhong (645 B.C.E.) referred to the
Yuezhi, or Niuzhi, as a people who supplied jade to the Chinese. It is well known that ancient Chinese rulers
had a strong attachment to jade. All of the jade items excavated from the tomb of Fuhao of the Shang
dynasty [a royal consort of the early 12th century BCE], more than 750 pieces, were from Khotan in modern
Xinjiang. As early as the mid-first millennium B.C.E. the Yuezhi engaged in the jade trade, of which the
major consumers were rulers of agricultural China.” Liu (2001), p. 265.

Section 6 – The Kingdom of Linni (Lumbini)

Regarding the kingdom of Linni (Lumbini), 1 the Buddhist books say:
“The king of this country fathered Futu (the Buddha). 2 The Buddha was the heir apparent. His
father was called Xietouye (Suddhodana). His mother was called Moye (Maya).
The Buddha wore yellow clothes. His hair was silky black. The hair on his chest was black;
his complexion a coppery-red.3
Initially Moye (Maya) dreamed of a white elephant and became pregnant. When the Buddha
was born, he emerged from the left side of his mother. 4 At his birth, he had a topknot (the
ushnisha) of hair.5 As soon as he touched ground, he was able to take seven steps 6.”
This kingdom is in the centre of the towns of Tianzhu (Northern India). Also, there was
another holy man named Shalü (Sāriputra) 7 in Tianzhu (Northern India).
Previously, in the first Yuanshou year (2 BCE), during the reign of Emperor Ai of the Han
dynasty, the National University Student, 8 Jing Lu, received verbal instructions from Yicun, the
envoy of the king of the Da Yuezhi (Kushans),9 on the Buddhist sūtras which say this man
(the Buddha) is the one who is reincarnated.10
The Buddhists mention linpusai (upâsaka – a male lay disciple),11 sangmen (śramaṅa –
monks, ascetics),12 bowen,13 shuwen (śrāvaka – ‘a hearer’, a follower of the Hīnayāna), 14
baishuwen (‘white’ or ‘pure’ or ‘elder’ śrāvaka’),15 biqiu (bhiksu – an ordained monk),16
chenmen (‘Guardian of the Gate’),17 which are all terms for disciples.
The Buddha’s [teachings] are related to, but different than, the scriptures of Lao Zi of the
Middle Kingdom. Indeed, it is believed (by the Taoists) that Lao Zi left the passes and,
heading west, crossed the Western Regions to Tianzhu (Northern India), where he taught the
Hu (Westerners).18
There are, altogether, twenty-nine titles for disciples of the Buddha, which I am not able to
give in detail, so I have summarised them as above. 19
 Section 6 – The Kingdom of Linni (Lumbini)
6.1. Linni (Lumbini). Lin-ni 臨兒.
臨 – K669e *bli̭əm / pi̭əm; EMC lim
兒 – K873a *ńi̭ĕg / ńźie̯; ni/ er; EMC ŋgj
The character 兒 – ni, is also pronounced er, and this is how Chavannes transcribed it in his translation.
Chavannes (1905), p. 539, n. 2, and others have noted, it is clearly meant to be a transcription of the name of
Lumbini, the garden in which the Buddha was born and, therefore, should be read ni not er.
The famous Chinese Buddhist monk, Faxian, in c. 405 CE , visited Lumbinī which he transcribed as 論民 –
Lunmin. 
論 – K470b *li̯wən / li̯uĕn; EMC lwənh
民 – K457a  *mi̯ən / mi̯ĕn; EMC mjin.
          “Lumbinī is Rumminideī in the Nepalese Terai, 2 miles to the north of Bhagavanpur and about a mile to the
north of Paderia” Law (1932), p. 45.
          “It was not till the nineties of the last century [i.e. the 19th century] that Lumbinī came to light once
again. In the forest the wood-cutters were plying their axes as usual, felling timber-tress. Through an
opening in the trees, something strange and man-made showed itself, – a yellowish pillar of sandstone cleft
down to the middle by a stroke of lightening and the top of it shattered and largely embedded in the
accumulated debris. Below the crack made by the lightening, the ruined pillar showed some strange
unintelligible inscriptions.
          The existence of the inscribed pillar had been known to the foresters of the Terai for some years
before it attracted in 1894 the attention of an official archaeologist Dr. Führer, – just 2,175 years after
Emperor Asoka had set it up. On 1st December, 1895, it was identified as Asoka’s monolith, and the ‘river of
oil’, Hsüan-tsang had heard of centuries before, it trickled down still within sight. Hill-men still called it by
the same ancient name, Tilaur, the ‘river of Til (oilseed)’. Buried in a thicket and perched on a mound was
also a small brick-built shrine to a goddess unknown to Hindu or Buddhist mythology. The shrine had been
kept up by local hill-men since forgotten antiquity.
          The inscription below the crack in the pillar was deciphered and edited by Dr. Bühler in 1898. It is in
five lines containing three sentences only : 
(a) when king Devānampiya Piyadasi had been anointed twenty years, he came himself and
worshipped (this spot), because the Buddha Sākymuni was born here.
(b) (He) both caused to be made a stone bearing a horse (?) and caused a stone pillar to be set up (in
order to show) that the Blessed One (Bhagavaṁ) was born here.
(c) (He made the village of Lumminī [in later readings – Lumbinī] free of taxes, and paying (only)
an eighth share (of the produce).
          The language of the inscription was an eastern dialect, possibly the court language of Pāṭaliputra in
Asoka’s time, of which the principal peculiarity was the tendency to convert ‘r’ into ‘l’. Thus in the
inscription the ‘Rājina’ (by the king) is modified as ‘Lājina’, and this phonological peculiarity was
particularly helpful in equiparating the name, Rummin, by which the hill-men called the find-place of the
pillar, with the Lumbinī of the legends.
          The ‘goddess of Rummin’ also was identified through the discovery near the shrine of a much defaced
relief in stone showing the Buddha’s nativity as given in the legends, – Queen Māyā holding the branch of a
tree and the divine child, just delivered, standing by her side. The nativity panel was of yellowish
sandstone like the Asokan pillar, though its age and original emplacement are unknown, the likelihood
being that it belonged to the original shrine.” Dutt (1955), pp. 21-22. Another (later) translation of the
inscription on the pillar is given in Thapur (1961), p. 261:
Rummindei Pillar Inscription
“The beloved of the Gods, the king Piyadassi, when he had been consecrated twenty years, came in person
and reverenced the place where Buddha Śakyamuni was born. He caused a stone enclosure to be made and
a stone pillar to be erected. As the Lord was born here in the village of Lumbinī, he exempted it from tax,
and fixed its contribution [i.e. of grain] at one-eighth.”
6.2. Futu 浮屠 [Fu-t’u] = the Buddha. An early and common transcription of the Buddha’s name into Chinese. See,
for example, GR Vol. II. p. 709, No. 3615.
6.3. The body of the Buddha was endowed with a yellow colour. His hair was blue like blue silk. His breast was blue,
his body-hair red like copper.
“The “golden colour” (suvarṇa-varṇa) of the Buddha is one of his 32 characteristics (lakṣana).” Zürcher
(1972), p. 383, n. 168.
Regarding the blue breast and hair of the Buddha; blue represented ‘refined’ and ‘true’ – see Zürcher (1972), pp. 178
and 384, n. 181. It was commonly used in illustrations of Buddha (and other holy personages):
“In early Buddhist texts liu-li also denoted blue or green precious stone, primarily lapis lazuli. The Buddhist
literature which was translated into Chinese in the fourth and early fifth centuries describes the
enlightened Buddha showing his hair, as beautiful as liu-li, to his father; devotees who did not hurt others
would be born with hair the colour of liu-li. From the earliest times up to now Buddhist artists always paint
the hair of the Buddha, Bodhisattvas and even other heavenly beings, sky-blue. . . .  At least in the Buddhist
context of the early centuries AD, vaidūrya in Sanskrit and liu-li in Chinese meant lapis lazuli. Probably
because the liu-li sold to the Chinese included not only lapis lazuli and similar precious stones but also blue
or green glass, the Chinese gradually, certainly by the fifth century, found out that certain kinds of liu-li
could be made by melting different kinds of stones together (WS: CII, 2275).” Liu (1988), p. 59.
6.4. At first I assumed that the statement that the Buddha was born from the left side of his mother was a simple
mistake as he is usually always represented as having emerged from the right side of his mother, if not from her
womb itself. This apparent mistake was noticed by Chavannes (1905), p. 545, n. 2, and p. 540, line 5 of the note.
          However, there may be some basis for it in Chinese literature. There is a Taoist tradition dating back at least as
far as the 4th century CE that Laozi [Lao-tzu] was born from his mother’s left arm-pit. See the discussion in Zürcher
(1972), p. 433, n. 70, where he explains: “The change from left to right is understandable: in general, left is the
direction which corresponds with the male principle (yang) (cf. M. Granet, Pensée chinoise, p, 369).”
6.5. The “topknot” or ushnisha is commonly represented on Buddha images and signifies the additional wisdom and
knowledge a Buddha possesses.
“The representations of the Buddha must always have either the chignon or the protuberance on the skull
which is presumably the seat of the manas, or divine mind (soul) of the Buddha.” Getty (1928), p. 18.
For more details on the Buddha’s ushnisha see: Banerjea (1931); and Chandra (1934).
6.6. “As soon as he touched ground, he was able to take seven steps.” This is a condensed version of the popular
accounts of the birth of the Buddha as recounted in many early Indian Buddhist texts such as the Mahāvastu (c. end
of 2nd century BCE), Nidānakathā and Lalitavistara, and the celebrated Buddhacarita or The Acts of the Buddha by
Ashvaghosha, who wrote in the first or early second century CE.
           “There lived once upon a time a king of the Shakyas, a scion of the solar race, whose name was
Shuddhodana. He was pure in conduct, and beloved of the Shakyas like the autumn moon. He had a wife,
splendid, beautiful, and steadfast, who was called the Great Maya, from her resemblance to Maya the
Goddess. These two tasted of love’s delights, and one day she conceived the fruit of her womb, but without
any defilement, in the same way in which knowledge joined to trance bears fruit. Just before her
conception she had a dream. A white king elephant seemed to enter her body, but without causing her any
pain. So Maya, queen of that god-like king, bore in her womb the glory of her dynasty. But she remained
free from the fatigues, depressions, and fancies which usually accompany pregnancies. Pure herself, she
longed to withdraw into the pure forest, in the loneliness of which she could practise trance. She set her
heart on going to Lumbini, a delightful grove, with trees of every kind, like the grove of Citraratha in
 Indra’s Paradise. She asked the king to accompany her, and so they left the city, and went to that glorious
grove.
          When the queen noticed that the time of her delivery was approaching, she went to a couch
overspread with an awning, thousands of waiting-women looking on with joy in their hearts. The
propitious constellation of Pushya shone brightly when a son was born to the queen, for the weal of the
world. He came out of his mother’s side, without causing her pain of injury. His birth was as miraculous as
that of Aurva, Prithu, Mandhatri, and Kakshivat, heroes of old who were born respectively from the thigh,
from the hand, the head or the armpit. So he issued from the womb as befits a Buddha. He did not enter the
world in the usual manner, and he appeared like one descended from the sky. And since he had for many
aeons been engaged in the practice of meditation, he now was born in full awareness, and not thoughtless
and bewildered as other people are. When born, he was so lustrous and steadfast that it appeared as if the
young sun had come down to earth. And yet, when people gazed at his dazzling brilliance, he held their
eyes like the moon. His limbs shone with the radiant hue of precious gold, and lit up the space all around.
Instantly he walked seven steps, firmly and with long strides. In that he was like the constellation of the
Seven Seers. With the bearing of a lion he surveyed the four quarters, and spoke these words full of
meaning for the future: ‘For enlightenment I was born, for the good of all that lives. This is the last time
that I have been born into this world of becoming’.” From: Conze (1959), pp. 35-36. For another translation
of the same text see: Johnston (1936), Part II, pp. 1-4.
6.7. Shalu 沙律 [Sha-lü] = Sāriputra.
“Mr. SYLVAIN LÉVI (Journal As. Jan.-Feb. 1897, p. 16, n., and May-June 1900, p. 461–462), has shown that the
characters Shalu could be a vernacular transcription of the name of Sāriputra.” Translated from Chavannes
(1905), p. 546, n. 2.
“According to Ch’en Tzu-liang 陳子良, as quoted in Fa-lin’s 法琳 Pien-cheng-lun 辨正論 (Treatise on
the Discernment of the Right), Sha-lü was aged and white-haired, constantly instructing men to construct
Buddha (stūpas) 沙律年老髮白, 常敎人浮圖. Sha-lü appears to be Śāriputra, one of the Buddha’s most
renowned disciples. While the Lao-tzu hua-hu-ching 老子化經 (the Scripture of Lao-tzu’s Conversion of
the Barbarians) is said to be the work of Taoist practitioner Wang Fu 王符, composed during the reign of
Emperor Ch’eng of the Eastern Chin 東晉成帝, its origins can be detected as early as in Hsiang K’ai’s 襄楷
memorial; in sum it is a work of considerable age.” Wada (1978), p. 34, n. 8.
Note the phonetically almost identical to the first part of the Tibetan name, Sharu, in the following quote:
“S’ÂRIPUTRA or S’arisuta or S’aradvatiputra (Pali. Sariputta. Singh. Seriyut. Burm. Thariputra. Tib. Sharu by
or Saradwatu by or Nid rghial). . . .  One of the principal disciples of S’âkyamuni, whose “right hand
attendant” he was; born at Nalandagrama, the son of Tichya (v. Upatichya) and S’ârika, he became famous
for his wisdom and learning, composed 2 works on the Abhidharma, died before his master, but is to re-
appear as Buddha Padmaprabha in Viradja during the Maharatna pratimandita kalpa.” Eitel (1888), pp. 148-
149.
The Singhalese name, Seriyut, is almost identical to the hypothesized Prakrit form, *Sariyut, proposed by Zürcher:
“S. Lévi (in J. As. 1897, p. 16 and 1900, p. 461-462) has demonstrated that this sha-lü 沙律 (Arch. *sa.bli̯wət >
Anc. *sa.bli̯wĕt) must be a very archaic rendering of the name Sāriputra or of a corresponding Prākrit form
*Sariyut.” Zürcher (1972), p. 428, n. 23.
Apparently, the Yellow Turban rebels who led a rebellion from 184 to c. 204 CE under the leadership of a Daoist
faith healer named Zhang Jue [Chang Chüeh], claimed that Shalu or Sāriputra, the Buddha’s elder disciple, was
actually Laozi. For a detailed discussion of this story and Chavannes’ rather convoluted attempts to explain it, see
Zürcher (1972), pp. 391-392 (5); 428, n. 25. 
6.8. boshidizu [po-shih ti-tzu].
“HAN: lit., disciple of the Erudites (po-shih), first appointed in 124 B.C.: National University Student, a
promising man admitted to the National University (t’ai-hsüeh) at the dynastic capital on the basis of a
recommendation by a territorial administrator; pursued studies of classical texts for one year; if successful
in examinations given then, became a qualified member of the official class and might join the pool of
 expectant appointees to office called Court Gentlemen (lang) at the capital or might seek an appointment
on the staff of a District Magistrate (hsien-ling) or a higher territorial administrator. Comparable to chien-
sheng of the late imperial dynasties. Commonly abbreviated to ti-tzu. . . . ” Hucker (1985), p. 390, No. 4753.
6.9. Yicun 伊存 [Yi-ts’un], the envoy of the king of the Da Yuezhi (Kushans).
“If we accept the [Weilue’s] text as it stands, this doubtless means that Ching Lu obtained this instruction in
China, most probably at the capital, from a Yüeh-chih who had come to China as an envoy.” Zürcher (1972),
p. 24.
This interesting passage has appeared in a number of later Chinese texts in a bewildering variety of forms which
are frequently contradictory. This has caused much discussion among scholars; see, for example, the discussions in
Chavannes (1905), p. 547, n. 1, and Zürcher, (1972), pp. 24-26.
          I personally feel that there is nothing improbable in having a Da Yuezhi envoy instruct a Chinese scholar in
Buddhist teachings in 2 BCE. The Da Yuezhi by then had been in control of the trade routes through Bactria for well
over a century and would have been in close contact with Indian thought and philosophies throughout this period.
          As the account in the Weilue predates the others by a couple of centuries, and contains nothing that would
seem implausible, it deserves serious consideration. However, Zürcher, (1972), p. 25, says:
“. . . but if this tradition after more than two centuries of silence turns up in some seven versions which are
partly unintelligible and in which neither the name of the Chinese scholar nor the function of the Yüeh-
chih nor the place of action appears to be fixed, we are no longer allowed to use it as reliable material for
historical research.”
Zürcher (1972), pp. 24-25 translates the term koushou 口授 [k’ou-shou] in this sentence as ‘oral instruction’ or
‘orally to instruct,’ but on page 31 he gives a fuller picture of its meaning:
“The master either had a manuscript of the original text at his disposal or he recited it from memory. If he
had enough knowledge of Chinese (which was seldom the case) he gave an oral translation (k’ou-shou 口授),
otherwise the preliminary translation was made, “transmitted”, by a bilingual intermediary (ch’uan-i 傳
譯). Chinese assistants – monks as well as laymen – noted down the translation (pi-shou 筆受), after which
the text was subjected to a final revision. . . . ”
6.10. “… Buddhist sūtras which say this man (the Buddha) is the one who is ‘reincarnated’ – 復立 fuli.” The
meaning of the term fuli here is unclear. Literally, it means something like ‘the reappeared.’
          Zürcher (1972), p. 429, translates it as “the reinstated.” Rather, it could be an early attempt to represent the
Buddhist concept of ‘reincarnation’ in Chinese and that it refers here to the Buddha. The word 復 fu was used in
descriptions of the “calling back the soul” ritual and, although it was not used in the combination fuli, it adds
weight to my hypothesis that fuli was probably an early attempt to represent the foreign Buddhist concept of
reincarnation in Chinese.
“The Northern equivalent to the ritual of “Calling back the soul” would be the fu “return-ritual” mentioned
repeatedly in the three classics on rites. In the Liyun chapter, the ritual is connected with the beginning of
ritual practices (compare Ruan Yuan, Shisanjing zhushu, (Zhonghua shuju 2. vol version), p.1415. Here only
po is mentioned, zhi qi “knowing/aware? Qi” (de Groot “sentient afflatus”) seems to be used instead of hun.
In the following text hun is mentioned too, but the connection between the ritual and hun and po are not
clear. A more detailed description of the performance of fu is contained in sangdaji, SSJZS, p.1572.
Unfortunately no mention of hun and po here. Even more details in Yili, SSJZS, p.1128. The locus classicus
for hun and po is the discussion between Zhao Yingzi and Zi Chan regarding the return of the soul of Bo
You contained in the Zuozhuan, Zhao 7, SSJZS, p.2050.
          Unfortunately no mention of the ritual here. This is where the story almost ends. In the preparation
of my master thesis on the “Zhaohun” I came across more references to hun and po in WS texts, but to my
knowledge the fu ritual is only mentioned in Zhouli, Yili and Liji. Yet the details mentioned suggest a clear
awareness of the ritual practice on the side of the authors.
          Hope this information is of some help.” Michael Schimmelpfennig (University of Erlangen-
Nuremberg). Posted on wsw@yahoogroups.com on April 24, 2003.
“I am glad my comment was of some use to you. Let me add one remark. In the Shuowen jiezi under the
 entry of yun grass or Rue (Matthews’, 7749) Xu Shen adds a quotation from Liu An saying that Rue grass
can bring the dead back (fu) to life. When I came across the remark, I surmised that it could indicate that
the Han lacked the concept of unconsciousness which is sort of supported by the fact that the Chinese
language lacks genuine expressions for the loss of consciousness. But here, Don Harper may know more. At
the same time fu is the central expression in and designation of the “Northern” Calling back the Soul ritual,
if such a procedure was ever practiced.” (A reply from Michael Schimmelpfennig on 21 August, 2003 kindly
giving me permission to use the information).
“The following custom exists in China: When someone has just died, a fine new garment, for example, is
shown to the soul from the housetop and it is implored to return to the body. This ritual is abundantly
attested in the classical texts80 and has continued to our day;81 it even supplied Sung Yü with the subject of
a long poem entitled, precisely, “Calling Back the Soul.” Sickness, too, often involved the flight of the soul,
and then the sorcerer pursues it in ecstasy, captures it, and replaces it in the patient’s body. 83
80 Cf. S. Couvreur, tr. Li Ki; ou, Mémoires sur les bienséances et les cérémonies ( 2nd edn. ), I, 85, 181, 199 ff.; II,
11, 125, 204, etc.; J. J. M. de Groot, The Religious System of China, I, 245 ff. On Chinese conceptions of the
life after death, cf. E. Erkes: “Die alt-chinesischen Jenseitsvorstellungen”; “The God of Death in Ancient
China.”
81 Cf., for example, Theo Körner, “Das Zurückrufen der Seele in Kuei-chou.”
82 Erkes, Das “Zurückrufen der Seele” (Chao-Hun) des Sung Yüh. Cf. also Maspero, Les Religions chinoises, pp.
50 ff.
83 This type of cure is still practiced today; cf. Groot, VI, 1284, 1319, etc. The sorcerer has the power to
call back and replace even the soul of a dead animal; cf. ibid., p. 1214 (the resurrection of a horse). The
Thai sorcerer sends some of his souls to search for the strayed soul of the patient, and he does not fail to
warn his souls to take the right road when they come back to this world. Cf. Maspero, p. 218.
Eliade (1964), pp. 447-448 and nn.
Chavannes (1905), p. 547 and n. 2 translated the term as “the reappeared” and believed it referred to the Taoist
story of Lao Tzu travelling to the West and being reincarnated in India as the Buddha. This hypothesis seems
unnecessary and unconvincing in light of the above discussion.
          Wada (1978), p. 33, gives fudu 復豆 as an alternative to fuli 復立 and seems to treat it as another attempt to
transcribe Buddha into Chinese. First, of all, I am unable to determine where he found 復豆, or on what grounds he
proposes to replace fuli 復立, with 復豆.
          Additionally, the Buddha’s name is clearly transcribed as Futu, a regular transcription of the name, at the
beginning of this same passage and I find it hard to believe that the author would have used a completely different
transcription, which is otherwise unknown, in the same chapter for the Buddha.
6.11. linpusai 臨蒲塞 [lin-p’u-sai] = Sanskrit upasâka – a male lay disciple. “UPÂSAKA (Singh. Upasika. Tib. Dge
snem. Ming. [sic] Ubaschi)… lit. male devotees. Lay-members of the Buddhist church who, without entering upon
monastic life, vow to keep the principal commandments. If females, they are called Upâsikâ (Sing. Upasikawa. Tib.
Dge snen ma. Mang. [sic] Ubaschanza).” Eitel (1885), p. 187.
          See also: Chavannes (1905), p. 550 and n. 1; Zürcher (1972), p. 27. The first character, lin 臨, in linpusai is
probably, as Sprecht first pointed out, and Chavannes explains in the note, a mistake for yipusai 伊蒲塞 [i-p’u-sai],
or youpose 優婆塞 [yu p’o sai], the initial characters of which are all quite similar and are attempts to phonetically
transcribe Sanskrit upâsaka – a faithful lay Buddhist).
           The form yipusai is first found in Hou Hanshu zhuan 72, referring to the gift of food in 65 CE toyipusai 伊蒲塞
[i-p’u-sai] and sangmen 桑門 [sang-men] (phonetic transcription of Sanskrit śramaṇa or monk) by Ying, the king of
(the dependent kingdom of) Chou.
“Let the cloth be returned, therewith to supplement the feasts of the i-p’u-sai 伊蒲塞 and sang-men 桑門.
I-p’u-sai is the same as yu-p’o-sai 優婆塞 (upāsaka), translated into Chinese as chin-chu 近住 (dwelling
close by). It means that, undertaking ascetic behaviour, he is allowed to approach the dwellings of the
Saṃgha. Sang-men [桑門] is the same as sha-men 沙門 (śramaṇa). The decree was distributed and
 displayed throughout the realm.” Quoted from Wada (1978), pp. 32-33.
As Chavannes notes, this text proves that the Buddhist church in China was well enough established at this early
date to have both monks and lay disciples.
6.12. sangmen 桑門 (śramaṇa – monks, ascetics).
“(Pali. Saman. Burm. Phungee. Tib. Dges by ong). . . .  Ascetics of all denominations, the Sarmanai or
Samanaioi or Germanai of the Greeks. (2.) Buddhist monks and priests “who have left their families and
quitted the passions.” Eitel (1888), p. 157. See also the previous note, 6.11.
6.13. bowen 伯聞 [po-wen].
“This term and the two following are very obscure. SYLVAIN LÉVI (J. A., May-June 1900, p. 463) has
proposed to see them as translations of the word “śrāvakas” (the hearer) – The Dongdiangives the reading:
伯開 [bokai], 疏間 [shujian], 白間 [baijian]; the Taipinghuanyuji writes: 伯聞 [bowen], 疏閒 [shuxian],
白閒 [baixian]. Meanwhile, although the Shanghai edition (1888) of the Sanguozhi, presents for the last of
these three terms, the reading 白疏聞 [baishuwen], I find in the Baorentang, the reading 白疏間
[baishujian].” Translated from Chavannes (1905), p. 550, n. 2.
6.14. shuwen 疏聞 [shu-wen] = śrāvaka – literally, ‘a hearer’, a follower of the Hīnayāna. It seems probable that
shu(su)-wen 疏聞, like the better-known form, sheng-wen 聲聞, was meant to represent Sanskrit śrāvaka, lit. ‘voice-
hearer’, later used for followers of the Hīnayāna. From DEABT:
“聲聞 [py] shēngwèn [wg] sheng-wen [ko] sŏngmun [ja] ショウモン shōmon ||| śrāvaka. ‘voice-hearer’;
originally, a disciple of the Buddha (who heard his voice); later, a follower of Hīnayāna. . . . ” DEALT. See
also GR, Vol. V, No. 9701, p. 268
“(Pali. Savako. Sing. Srawaka. Tib. Nan thos. Mong. Scharwak). . . .  (1.) All personal disciples of Śàkyamuni,
the foremost of whom are called Mahâśrâvakas. (2.) The elementary degree of saintship, the first of the
Triyâna, the śrāvaka (superficial yet in practice and understanding) being compared with a hare crossing Sañsara by
swimming on the surface.” Eitel (1888), p. 157.
6.15. baishuwen 白疏閒 [pai-shu-wen] – literally: ‘White (or ‘pure’) śrāvaka’.
6.16. biqiu 比丘 [pi-ch’iu] – a regular transcription of the Sanskrit bhiksu – a mendicant monk. (Pâli. Bhikkhu.
Singh. Bhikchu. Tib. Dgeslong. Mong. Gelong). Eitel (1888), p. 31.
6.17. chenmen 晨門 [ch’en-men]. According to GR No. 637, chen-men was an ancient term meaning, ‘guardian of the
door.’ It is equivalent to the Sanskrit Dvārapāla – ‘Keeper of the Gate.’
          “It appears from the legends that a functionary whose designation was ‘Guardian of the Gate’ acted
at both Nālandā and at Vikramaśilā. The gate-keeper of Nālandā, evidently a learned monk of high status, is
designated as ‘men-che’ [門者 = ‘doorkeeper’ – Pinyin: menzhe] and of Vikramaśilā as ‘go-srun’ in the
Tibetan. The Chinese and the Tibetan expressions are synonymous. Nālandā had one gate, while
Vikramaśilā had six, each ‘kept’ by a Go-srun, equivalent to Sanskrit Dvāra-pāla (Keeper of the Gate).
            The function at Nālandā is reported in the Chinese records to have been to judge the qualifications of
persons intending to join one of its ‘schools of discussion’. Whether the Vikramaśilā ‘gate-keeper’ exercised
the same function is not known, but those named in the Tibetan legends as incumbents of the office of Go-
srun or Dvāra-pāla were all scholars of high eminence and celebrity, holding the office on royal
commission.” Dutt (1962), pp. 360-361.
          “We have referred already to another important officer – the ‘Door-keeper’ who used to hold the
‘screening examination’ of candidates seeking admission to one of the ‘schools’ of Nālandā.” Ibid., p, 338.
See also ibid., p. 351.
6.18. Fotu suosaiyu zhongguo Laozijing xiangchru 佛屠所載與中國老子經相出入 . For an alternate translation of
this rather difficult paragraph see Zürcher (1972), p. 291 (4) and for a discussion of its possible meanings and Daoist
influences, see p. 428, n. 19.
          This passage has been the subject of lengthy discussion by a number of eminent scholars including Sylvain
Lévi, Paul Pelliot, and E. Zürcher. All these discussions, however, have been complicated by the faulty rendering of
the term churu 出入 [ch’u-ju] as “analogies” rather than “discrepancies” or “differences.” (See: GR, p. 115, No. 2512;
ABC, p. 137). This has led to much fruitless speculation.
          The text now becomes much easier to comprehend – it is clearly a statement intended to discredit the Daoist
legend that Laozi [Lao-tzu] went to the West and converted the Barbarians and was later reincarnated as the
Buddha. This finds confirmation in the quotation from the Xiyu zhuan of the Weilue by Chen Ziliang [Ch’en Tzu-
liang] in his commentary to Falin’s [Fa-lin’s] Bianzhenglun [Pien cheng lun] of 626 CE. This apparently refers to the
same original as our text – the Xiyu zhuan of the Weilue quoted by Pei Songzhi [P’ei Sung-chih] in his commentary
to the Sanguozhi [San kuo chih]. See especially Zürcher (1972), p. 428, nn. 19 and 25.
6.19. One of my correspondents kindly sent me the following information on this passage:
“In the text Anavatapta-gatha, twenty-nine disciples reported their previous life stories to the Buddha in
the Anavatapta Lake. The text is in the Taisho vol. 4, text no. 199, page 190-202. Since you have mentioned
that Yu Huan, the author of the Wei lue, composed the text between 239-265 A.D., he is not likely to have
referred to the version of Anavatapta-gatha I just mentioned, because it was translated by Dharmaraksa in
303 A.D. However, there is another text closely related to AG, translated by Kang Mengxiang, in the Hou
Han period (25-220 A.D.). In other words, if Yu Huan did refer to the text AG in his mention of 29 disciples
of the Buddha, there must be some materials existing preceding the received text in Taisho, on which Yu
Huan based his comments.”
This would seem very likely. Zürcher (1972), pp. 36, and 333, n. 99, makes the point that the Indian Tanguo [T’an-
kuo], who is said to have come from Kapalivastu and worked together with another Indian, Zhu Dali [Chu Ta-li = ?
Mahābala] and the Sogdian Kang Mengxiang [K’ang Meng-hsiang] translated: “the earliest extant Chinese accounts
of the Buddha : the Chung pen-ch’i ching . . . (T 196) and the Hsiu-hsing pen-ch’i ching . . . (T 184).”
          “The poverty of our information on Buddhism during the Han period is certainly due to the terrible
troubles which ravaged the whole Chinese Empire at the end of the 2nd and the beginning of the 3rd
century of our era. Certain paragraphs of Muzi prove that at this period a great many Buddhist works had
already been translated into Chinese. But the same writer shows all the provinces in the hands of
governors who had, in fact, become independent sovereigns who did not allow passage through their
territory by subjects of the neighbouring provinces. Long and bloody wars were needed to reduce all these
principalities. Still unity was unable to be recovered, and three sovereign princes had to be allowed to
secure for each one a relative calm in a portion of the Empire. As a result of these constant quarrels, in
which every town was taken and pillaged many times, innumerable documents perished. The same held
when, at the beginning of the 4th century, the Tartar invasions took the dominance of the Huang He basin
away from the Chinese for about three hundred years. We have some short works on this subject. M.
Chavannes has spoken very recently (Mémoires historiques, bk. V, p. 465) of the two works known to have
been lost during the troubles which obliged the Qin to leave Luoyang in Henan to establish their capital at
Nanjin in 417. Buddhist literature refers to the destruction it underwent then. We know that one of the two
monks who arrived at Luoyang in 67 AD, Zhu Falin, besides the Sūtra in 42 Sections, translated four, and
perhaps five, other works of Indian Buddhism. Mr. Nanjio (Catalogue, Appendix II, No. 2) states that all those
works were lost by the 8th century, but it can be specified more exactly, for the author of the Gaosengzhuan,
Bojiao, says in his biography of Zhu Falin that, since the troubles which marked the change of capital these
works were lost, and did not reach the “left of the River.” In other words, from the beginning of the 5 th
century, the revolutions had destroyed almost all the work of the first translator whose memory has been
preserved by Chinese Buddhism.” Translated from Pelliot (1906), p. 395, n. 6.   

Section 7 – The Kingdom of Juli (or, rather, Dongli)

The kingdom of Juli (should read ‘Dongli’ = ‘Eastern Division’ of the Kushan Empire 1) is also
called Liweite (Ayodhyā),2 and Peiliwang.3
It is more than 3,000 li (1,247 km) to the southeast of Tianzhu (Northern India). 4 This country
is low, humid, and very hot.
The king rules from the town of Shaji (or Shaqi = Sakēta). 5 There are several tens of other
towns.
The people are cowardly and weak. The Yuezhi (Kushans) and Tianzhu (Northwestern India)
attacked and conquered them.6
This territory is several thousand li from east to west, and north to south. The men and
women of this nation are all eighteen chi tall [mistake for 8 chi = 1.85 metres, or just over 6
feet, as in the Hou Hanshu].7 They ride elephants and camels into battle. Currently they
provide military service and taxes to the Yuezhi (Kushans). 8
 
Section 8 – The Kingdom of Panyue (Pandya)
The kingdom of Panyue (Pandya) is also called Hanyuewang. 1 It is several thousand li to the
southeast of Tianzhu (Northern India), and is in contact with Yi Circuit. 2 The inhabitants are
small; they are the same height as the Chinese. Traders from Shu (Western Sichuan) travel
this far.3
The Southern Route, after attaining its most westernmost point, turns southeast until it
reaches its end.
 
Section 9 – The Central Route
The Central Route goes west to:
• the kingdom of Weili (near modern Korla), 1 the kingdom of Weixu (Hoxud =
Chokkur),2 the kingdom of Shanwang (in the western Kuruk mountains), 3 which are all
dependencies of Yanqi (Karashahr).4
• the kingdom of Gumo (south of modern Aksu), 5 the kingdom of Wensu (Uqturpan),6
the kingdom of Weitou (Karaqi),7 which are all dependencies of Qiuci (Kucha). 8
• the kingdom of Zhenzhong (Arach?), 9 the kingdom of Suoju (Yarkand),10 the kingdom
of Jieshi,11 the kingdom of Qusha,12 the kingdom of Xiye (Khargalik),13 the kingdom of
Yinai (Tashkurghan),14 the kingdom of Manli (modern Karasul),15 the kingdom of Yire
(Mazar – also known as Tágh Nák and Tokanak),16 the kingdom of Yuling,17 the
kingdom of Juandu (‘Tax Control’ – near Irkeshtam), 18 the kingdom of Xiuxiu (‘Excellent
Rest Stop’ – near Karakavak),19 and the kingdom of Qin,20 which are all dependencies
of Shule (Kashgar).21
Leaving there (Kashgar), and going west, you reach Dayuan (Ferghana), 22 Anxi (Parthia),23
Tiaozhi (Characene and Susiana),24 and Wuyi (Arachosia and Drangiana – capital,
Kandahar).25 Wuyi is also called Paizhi.26 These four kingdoms succeed each other to the
west. These are kingdoms that existed previously and have not been modified.
 
Section 10 – Previous Misconceptions
In earlier times, it was mistakenly thought that Tiaozhi (Characene and Susiana) was west of
Da Qin (Roman territory). Now it is known to be to the east. 1
In earlier times it was also mistakenly thought to be more powerful than Anxi (Parthia), but it
has been changed into a dependency said to mark the western frontier of Anxi (Parthia).
In earlier times it was also mistakenly thought that the Ruo Shui (‘Weak River’) was west of
Tiaozhi (Characene and Susiana). Now it is (thought to be) west of Da Qin (Roman territory). 2
In earlier times, it was also mistakenly thought that if you left Tiaozhi (Characene and
Susiana), and travelled more than two hundred days to the west, you reached the place
where the sun goes down. Now, (it is thought that) you travel west from Da Qin (Roman
territory) to reach the place where the sun sets. 3
 
Section 11 – Da Qin (Roman territory/Rome)
The kingdom of Da Qin (Rome)1 is also called Lijian.2 It is west of Anxi (Parthia) and Tiaozhi
(Characene and Susiana), and west of the Great Sea. 3
From the city of Angu (Gerrha)4, on the frontier of Anxi (Parthia), you take a boat and cut
directly across to Haixi (‘West of the Sea’ = Egypt).5 With favourable winds it takes two
months; if the winds are slow, perhaps a year; if there is no wind, perhaps three years. 6
The country (that you reach) is west of the sea (haixi), which is why it is called Haixi (literally:
‘West of the Sea’ = Egypt). There is a river (the Nile) flowing out of the west of this country,
and then there is another great sea (the Mediterranean). The city of (Wu) Chisan
(Alexandria)7 is in Haixi (Egypt).
From below this country you go north to reach the city of Wudan (Tanis?). 8 You (then) head
southwest and cross a river (the Sebannitus branch of the Nile?) by boat, which takes a day.
You head southwest again, and again cross a river (the Canopis branch of the Nile?) by boat,
which takes another day.9 There are, in all, three major cities [that you come to]. 10
Now, if you leave the city of Angu (Gerrha) by the overland route, you go north to Haibei
(‘North of the Sea’ – the lands between Babylonia and Jordan), then west to Haixi (Egypt), 11
then turn south to go through the city of Wuchisan (Alexandria). After crossing a river, which
takes a day by boat, you circle around the coast (to the region of Apollonia, the port of
Cyrene). (From there, i.e. the region of Apollonia) six days is generally enough to cross the
(second) great sea (the Mediterranean) to reach that country (Da Qin = Rome). 12
This country (the Roman Empire) has more than four hundred smaller cities and towns. It
extends several thousand li in all directions.13 The king has his capital (that is, the city of
Rome) close to the mouth of a river (the Tiber).14 The outer walls of the city are made of
stone.
This region has pine trees, cypress, sophora, catalpa, bamboo, reeds, poplars, willows,
parasol trees, and all sorts of plants.15 The people cultivate the five grains [traditionally: rice,
glutinous and non-glutinous millet, wheat and beans], and they raise horses, mules, donkeys,
camels and silkworms.16 (They have) a tradition of amazing conjuring. They can produce fire
from their mouths, bind and then free themselves, and juggle twelve balls with extraordinary
skill.17
The ruler of this country is not permanent. When disasters result from unusual phenomena,
they unceremoniously replace him, installing a virtuous man as king, and release the old king,
who does not dare show resentment.18
The common people are tall and virtuous like the Chinese, but wear hu (‘Western’) clothes.
They say they originally came from China, but left it. 19
They have always wanted to communicate with China but, Anxi (Parthia), jealous of their
profits, would not allow them to pass (through to China). 20
The common people can write in hu (‘Western’) script.21 They have multi-storeyed public
buildings and private; (they fly) flags, beat drums, (and travel in) small carriages with white
roofs, and have a postal service with relay sheds and postal stations, like in the Middle
Kingdom (China).
From Anxi (Parthia) you go around Haibei (‘North of the Sea’ – the lands between Babylonia
and Jordan) to reach this country.22
The people (of these countries) are connected to each other. Every 10 li (4.2 km) there is a
ting (relay shed or changing place), and every 30 li (12.5 km) there is a zhi (postal station).23
There are no bandits or thieves, but there are fierce tigers and lions that kill those travelling
on the route. If you are not in a group, you cannot get through. 24
This country (Rome) has installed dozens of minor kings. The king’s administrative capital
(Rome) is more than 100 li (42 km) around.25 There is an official Department of Archives.
The king has five palaces at 10 li (4.2 km) intervals. He goes out at daybreak to one of the
palaces and deals with matters until sunset and then spends the night there. The next day he
goes to another palace and, in five days makes a complete tour. They have appointed thirty-
six leaders who discuss events frequently.26 If one leader does not show up, there is no
discussion. When the king goes out for a walk, he always orders a man to follow him holding
a leather bag. Anyone who has something to say throws his or her petition into the bag. When
he returns to the palace, he examines them and determines which are reasonable. 27
They use glass to make the pillars and table utensils in the palaces. 28 They manufacture
bows and arrows.
They divide the various branch principalities of their territory into small countries such as that
of the king of Zesan (Azania?),29 the king of Lüfen (Leucos Limen),30 the king of Qielan (Wadi
Sirhan),31 the king of Xiandu (Leukê Komê),32 the king of Sifu (Petra),33 (and that of) the king
of Yuluo (Karak).34 There are so many other small kingdoms it is impossible to give details on
each one.
 
Section 12 – Products of Da Qin (Roman territory)
This country produces fine linen.1 They make gold and silver coins. One gold coin is equal to
ten silver coins.2
They have fine brocaded cloth that is said to be made from the down of ‘water-sheep’. It is
called Haixi (‘Egyptian’) cloth. This country produces the six domestic animals, which are all
said to come from the water.3
It is said that they not only use sheep’s wool, but also bark from trees, or the silk from wild
cocoons,4 to make brocade, mats, pile rugs, woven cloth and curtains, all of them of good
quality, and with brighter colours than those made in the countries of Haidong (“East of the
Sea”).5
Furthermore, they regularly make a profit by obtaining Chinese silk, unravelling it, and making
fine hu (‘Western’) silk damasks.6 That is why this country trades with Anxi (Parthia) across
the middle of the sea. The seawater is bitter and unable to be drunk, which is why it is rare for
those who try to make contact to reach China.
The mountains (of this country) produce nine-coloured jewels (fluorite) of inferior quality. They
change colour on different occasions from blue-green to red, yellow, white, black, green,
purple, fiery red, and dark blue.7 Nowadays nine-coloured stones of the same type are found
in the Yiwu Shan (a mountain range east of Hami). 8
In the third Yangjia year (CE 134), the king of Shule (Kashgar), Chen Pan [who had been
made a hostage at the court of the Kushan emperor, for some period between 114 and 120,
and was later placed on the throne of Kashgar by the Kushans], 9 offered a blue (or green)
gem and a golden girdle from Haixi (Egypt).10
Moreover, the Xiyu Jiutu (‘Ancient Sketch of the Western Regions’) now says that both Jibin
(Kapisha-Gandhāra) and Tiaozhi (Characene and Susiana) produce precious stones
approaching the quality of jade.11
 
Product List12
Note: The translator has added the numbering in brackets for the convenience of the reader
in checking the notes on the various items. For information on any of the items mentioned in
the list, please click on the blue superscript No. 12 after “Product List” above, and then scroll
down the page of notes until you come to the number you are looking for. For instance, if you
want to check the notes on tin, scroll down until you reach note number 12.12 (6).
Da Qin (the Roman Empire) has plenty of:
(1) gold
(2) silver
(3) copper
(4) iron
(5) lead
(6) tin
(7) ‘divine tortoises’ – tortoises used for divination
(8) white horses with red manes
(9) fighting cocks
(10) rhinoceroses
(11) sea turtle shell
(12) black bears
(13) ‘red hornless (or immature) dragons’ (which produced the famous “dragons’ blood” resin)
(14) ‘poison-avoiding rats’ = mongooses
(15) large cowries
(16) mother-of-pearl
(17) carnelian
(18) ‘southern gold’
(19) kingfisher feathers
(20) ivory
(21) coloured veined jade
(22) ‘bright moon’ pearls
(23) luminescent ‘pearls’ or pearl-like jewels (probably large diamonds)
(24) genuine white pearls
(25) yellow amber
(26) (red) coral
(27) ten varieties of glass: red, white, black, green, yellow, blue-green, dark blue, light blue,
fiery red, purple
(28) a magnificent jade
(29) white carnelian?
(30) rock crystal or transparent glass
(31) various semi-precious gems
(32) realgar
(33) orpiment
(34) nephrite
(35) multicoloured jade or gemstone
(36) ten sorts of wool rugs: yellow, white, black, green, purple, fiery red, deep red, dark blue,
golden yellow, light blue and back to yellow
(37) finely patterned multicoloured wool carpets
(38) nine colours of multicoloured lower quality wool carpets (kilims rather than knotted
carpets?)
(39) gold threaded embroidery
(40) polychrome (warp twill) fine silk or chiffon
(41) woven gold cloth
(42) purple chi cloth
(43) falu cloth
(44) purple chiqu cloth
(45) asbestos cloth
(46) fine silk gauze cloth
(47) shot silk, ‘clinging cloth’ or ‘cloth with swirling patterns’?
(48) dudai cloth
(49) cotton-wool cloth?
(50) multicoloured tao cloth
(51) crimson curtains woven with gold
(52) multicoloured ‘spiral curtains’?
(53) yiwei
(54) myrrh
(55) storax
(56) diti
(57) rosemary
(58) probably dhūṇa – an incense made from the resin of the Indian Sal tree.
(59) bai fuzi – lit. ‘white aconite’ – but it is not clear what plant this refers to here. See notes.
(60) frankincense
(61) turmeric, saffron or tulips
(62) rue oil
(63) Oriental lovage – Lysimachia foenum-graecum Hance
Altogether (they have) twelve types of aromatic plants. 13
 
Section 13 – The Sea Route to Da Qin (Roman territory)
As well as the overland route from Da Qin (Roman territory) through Haibei (‘North of the Sea’
– the lands between Egypt and Parthia), one can also follow the sea south along the seven
commanderies of Jiaozhi (stretching down the north Vietnamese coast), 1 which are in contact
with foreign countries.2 Nearby (or ‘North’)3 is a waterway (the Red River)4 which leads to
Yongchang5 in Yizhou (a commandery in present-day southern Yunnan). 6 That’s why rare
items come from Yongchang.
In early times only the maritime routes (to Da Qin) were discussed because they didn’t know
there were overland routes.7
 
Section 14 – Roman Dependencies
Now, (the Roman Empire) can be summed up as follows: the number of people and families
cannot be given in detail. It is the biggest country west of the Bai Congling (‘White Pamir
Mountains’).1 They have installed numerous minor kings so only the bigger dependencies are
noted here:
 
Section 15 – The Kingdom of Zesan (Azania)
The king of Zesan (Azania)1 is subject to Da Qin (Rome). His seat of government is in the
middle of the sea.2 To the north you reach Lüfen (Leukê Komê).3 It can take half a year to
cross the water, but with fast winds it takes a month.4
(Zesan) is in close communication with Angu city (Gerrha) in Anxi (Parthia). 5 You can (also)
travel (from Zesan) southwest to the capital of Da Qin (Rome), but the number of li is not
known.6
 
Section 16 – The Kingdom of Lüfen = Leukê Komê or modern Al Wajh
The king of Lüfen (Leukê Komê)1 is subject to Da Qin (Rome). It is 2,000 li (832 km) from his
residence to (the nearest) major city (= Daphnae) of Da Qin (the Roman Empire). 2
From the city of Lüfen (Leukê Komê) going west to Da Qin (alongside the Butic Canal), you
cross over the sea by an ‘elevated bridge’ 230 li (96 km) long;3 then you take the sea route
southwest, travelling around the sea (coast), and then head west (to reach Da Qin). 4
 
Section 17 – The Kingdom of Qielan (Wadi Sirhan)
The king of Qielan (Wadi Sirhan)1 is subject to Da Qin (Rome). From the kingdom of Sitao
(Istakhr, Stakhr)2 you go south, cross a river (the Rūd-i Kor), then head west 3,000 li (1,247
km) to go to Qielan (Wadi Sirhan).3 The route leaves south of the river (the Rūd-i Kor), only
then do you head west.4
From Qielan (Wadi Sirhan) you again travel west 600 li (250 km) to the kingdom of Sifu
(Petra).5 The Southern Route joins (this east-west route) at Sifu (Petra). Also, (a route) goes
southwest to the kingdom of Xiandu (‘Aynūnah). 6
Due south from Qielan (Wadi Sirhan) and Sifu (Petra) is Jishi (‘Rock Piles’). 7 To the south of
Jishi (‘Rock Piles’) there is a big sea (the Red Sea) which produces coral and pearls. 8
North of Qielan (Wadi Sirhan), Sifu (Petra), Sibin (Susa) 9 and Aman (Ariana)10 there is a
mountain range (the Taurus mountains)11 running east to west.
East of both Da Qin (Roman territories) and Haixi (= Egypt) 11 there is a mountain range (the
Jibāl ash Sharāh Range or Mount Seir)12 running north to south.
 
Section 18 – The Kingdom of Xiandu (‘Aynūnah = Leukos Limên)
The king of Xiandu (‘Aynūnah = Leukos Limên) 1 is subject to Da Qin (Rome). From his
residence it is 600 li (250 km) northeast to Sifu (Petra).2
 
Section 19 – The Kingdom of Sifu (Petra)
The king of Sifu (Petra)1 is subject to Da Qin (Rome). From his residence northeast to Yuluo
(Karak),2 you go 340 li (141 km), and cross over a sea (mistake for ‘river’ = the Wadi al-
Ḥesa).3
 
Section 20 – The Kingdom of Yuluo (Karak)
Yuluo (Karak)1 is subject to of Da Qin (Rome). The seat of government is northeast of Sifu
(Petra) across a river (the Wadi al-Ḥesa).2 From Yuluo (Karak) you go northeast, and again
cross over a river (River Arnon).3
 
Section 21 – The Kingdom of Siluo (Sura)
Northeast of Siluo (Sura)1 you again cross over a river (the Euphrates). The kingdom of Siluo
(Sura) is subject to Anxi (Parthia), and it borders on Da Qin (Roman territory).
 
Section 22 – The Far West
West of Da Qin (Rome) is sea water. West of the sea water are rivers. West of the rivers there
are big mountains running south to north. 1 West (of this) is the Chi Shui (‘Red River’ =
Kāshgar-daryā?).2 West of the Chi Shui (‘Red River’ = Kāshgar-daryā?) are the Baiyu shan
(‘White Jade Mountains’).3
In the Baiyu Shan (‘White Jade Mountains’) lives Xi Wangmu (‘Spirit-Mother of the West’). 4
West of Xi Wangmu are the long Liusha (‘Shifting Sands’). 5
To the west of the Liusha (‘Shifting Sands’) is the kingdom of Daxia (Bactria), the kingdom of
Jiansha (‘Stable Sands’),6 the kingdom of Shuyao (Sogdiana),<7 and the kingdom of the
Yuezhi (Kushans).
West of these four kingdoms is the Hei Shui (‘Black River’), 8 which is as far west as I have
heard of.
 
Section 23 – The New Route of the North
The New Route of the North1 goes west reaching the kingdom of Eastern Jumi (near modern
Dashito),2 the kingdom of Western Jumi (near modern Mulei), 2 the kingdom of Danhuan,4 the
kingdom of Bilu,5 the kingdom of Pulu,6 and the kingdom of Wutan,7 which are all
dependencies of the king of the Further Jushi Section (near Jimasa).
The king has his capital in the city of Yulai.8 The Wei (dynasty) conferred the title of
‘Probationary Wei Palace Attendant’9 on Yiduoza, the king, with the honorific name of ‘Great
Defender of the Wei’.10 He received the ‘Seal of King (appointed by the) Wei’. 11
The (New Northern) Route then turns northwest to reach Wusun (Issyk-kol and
Semirechiye),12 and Kangju (Tashkent plus the Chu, Talas, and middle Jaxartes basins). 13
These kingdoms existed previously and have neither grown nor shrunk. 14
 
Section 24 – The Kingdom of Northern Wuyi (Khujand – Alexandria Escharte)
Northern Wuyi (modern Khujand)1 is a distinct kingdom in the northern part of Kangju.
 
Section 25 – The Kingdom of Liu (Turkestan? Kzyl-Orda?), Yan (north of Yancai), and Yancai
(= the Alans between the Black and Caspian Seas).
Then there is the kingdom of Liu (between Kangju and Yancai?), 1 the kingdom of Yan (to the
north of Yancai),2 and the kingdom of Yancai (between the Black and Caspian Seas), 3 which
is also called Alan.4 They all have the same way of life as those of Kangju.
To the west, they border Da Qin (Roman territory), to the southeast they border Kangju
(Tashkent plus the Chu, Talas, and middle Jaxartes basins).
These kingdoms have large numbers of their famous sables. 5 They raise cattle and move
about in search of water and fodder. They are close to a big marsh (to the northeast and north
of the Aral Sea).6 Previously they were vassals of Kangju (Tashkent plus the Chu, Talas, and
middle Jaxartes basins). Now they are no longer vassals. 7
 
Section 26 – The Kingdom of Hude
The kingdom of Hude is north of the Congling (the Pamirs), northwest of Wusun (Issyk-kol
and Semirechiye), northeast of Kangju (Tashkent plus the Chu, Talas, and middle Jaxartes
basins). They have more than 10,000 men able to bear arms. They follow their cattle. They
produce excellent horses, and have sables.
 
Section 27 – The Kingdom of Jiankun (Kirghiz)
The kingdom of Jiankun (Kirghiz)1 is northwest of Kangju (Tashkent plus the Chu, Talas, and
middle Jaxartes basins). They have more than 30,000 men able to bear arms. They follow
their cattle and have lots of sables and excellent horses.
 
Section 28 – The Kingdom of Dingling (Around Lake Baikal and on the Irtish River)
The kingdom of Dingling is north of Kangju (Tashkent plus the Chu, Talas, and middle
Jaxartes basins).1 They have 60,000 men able to bear arms. They produce famous sable
pelts,2 as well as white and blue Arctic fox pelts.3
Of these three states Jiangun (the Kirghiz – in the region of Omsk?), is the central one. It is
7,000 li (2,911 km) from the court of the Chanyu (Shah)4 of the Xiongnu (which was close to
modern Ulan Bator) on the Anxi River (Juul Gol?).
It is 5,000 li (2,080 km)5 south to the six kingdoms of Jushi;6 3,000 li (1,247 km) southwest to
go to the frontier of Kangju (Tashkent plus the Chu, Talas, and middle Jaxartes basins); 8,000
li (3,326 km) west to go to the capital of the king of Kangju.
It was thought that perhaps these Dingling were the Dingling to the north of the Xiongnu
(around Lake Baikal), but the Northern Dingling are west of Wusun (and north of Kangju), and
it seems they are of another race. Furthermore, north of the Xiongnu (near modern Ulan
Bator) is the kingdom of Hunyu, the kingdom of Qushi, the kingdom of Dingling, the kingdom
of the Gekun, and the kingdom of the Xinli. 7
It is known that if you go south from Bei Hai (‘Northern Sea’ = Lake Baikal) you find the
Dingling again. They are not the same as the Dingling to the west of the Wusun (Issyk-kol and
Semirechiye).8
The Wusun elders say that north of the Dingling is the kingdom of Majing (‘Horses Shanks’).
These men make sounds like startled wild geese. From above the knee, they have the body
and hands of a man, but below the knees, they grow hair, and have horses’ legs and hooves.
They don’t ride horses as they can run faster than horses. They are brave, strong, and daring
fighters.9
 
Section 29 – The Kingdom of Duanren (‘Short Men’)
The kingdom of Duanren (‘Short Men’)1 is northwest of Kangju (Tashkent plus the Chu, Talas,
and middle Jaxartes basins). The men and women are all three chi tall [0.693 metres or 2.27
feet]. They are very numerous.
It is a long way from Yancai (at the mouth of the Syr Darya near the Aral Sea) and the other
kingdoms. The elders of Kangju (Tashkent plus the Chu, Talas, and middle Jaxartes basins)
say that merchants frequently cross this kingdom.2 It is possibly more than 10,000 li (4,158
km) from Kangju (Tashkent plus the Chu, Talas, and middle Jaxartes basins).
 
Section 30 – Yu Huan’s Comments
Yu Huan (the author) observes: It is commonly believed that a fish living in a little stream does
not know the size of the Qiang (Yangtze River) and the sea. 1 The mayfly,2 for that matter,
does not know of the changing of the four seasons. Why is this so? Because one lives in a
small place, and the other’s life is short.
I am, at the moment, intensively examining Da Qin (the Roman Empire) and all the other
foreign kingdoms. Still, it seems to me that I am neglecting to (fully) instruct the uninformed.
Moreover, as to the speculations of Zou Yan,3 or the hypotheses of the Dayitai xuan, “The
Great Mystery of the Noble Yi (-jing)”,4 alas, I am limited to travelling by foot, and living in the
puddle left in the hoof print of an ox.5 Besides, I don’t have the longevity of Peng Zu.6
It has not been my fate to see things first hand, travelling with the rapid winds, or enlisting
swift horses to view distant vistas. Alas, I have to strain to see the three heavenly bodies [the
sun, moon, and stars] but, oh, how my thoughts fly to the eight foreign regions! 7