Section 7 – The Kingdom of Juli (the ‘Eastern Division’ of the Kushan Empire)

7.1. Juli 車離 [Chü-li]. The character 車 is correctly transcribed as chē, but is pronounced jū in modern Mandarin, I have transcribed the name of this kingdom as Juli. Ju (or che) means a ‘chariot’ or a ‘vehicle.’
          The Hou Hanshu gives Dongli 東離 [Tung-li] for the name of this kingdom. Ju車, in the Weilue therefore, may well be an error for the easily confused character, dong 東 = ‘east’.
          The character li 離 can mean: ‘distant’ or ‘division.’ It seems unlikely that Dongli was intended as a transcription of some foreign name. It translates literally as the ‘Eastern Section’ or ‘Eastern Division;’ from: dong = ‘east’ + li = ‘division.’ Dongli 東離 [Tung-li], therefore = the ‘Eastern Division,’ or ‘The Far East’ (of the Kushan Empire).
          I have rendered Juli (= Dongli) as the ‘Eastern Division’ (of the Kushan Empire), as it clearly refers (in both texts) to the Kushans’ newly conquered territories in eastern India.
          In the excellent article by F. W. Thomas: “Sandanes, Nahapāna, Caṣṭana and Kaniṣka : Tung-li P’an-ch’i and Chinese Turkestan,” New Indian Antiquary VII. 1944, pp. 81-100, Thomas also concludes (ibid. pp. 90-92) that the name Dongli was probably not a transcription of a local name but, rather, should be translated as ‘Eastern Division.’ He notes that the second Chinese character in the name, li:

“ancient lyie < lyia (KARLGREN no. 533 [Analytical Dictionary – in his later Grammata Serica, no. 23f – *lia] meaning ‘oriole’, ‘leave,’ ‘quit,’ ‘separated,’ ‘pass through,’ etc., is frequently used in rendering Sanskrit expressions denoting ‘separation,’ ‘lack,’ etc., especially compounds with vⁱ- (including vibhāga, ‘division’), it seems possible that Tung-li is not a transaction, but a translation, meaning ‘Eastern Division,’ in Sanskrit prācya (or pūrva)- vibhāga or prāg-deśa, an expression which by reason of its intelligibility would be specially likely to be rendered by a translation. [Thomas notes here that: “In later times the Chinese uses the expression ‘Tung T‘ien-chu’, ‘Eastern India.’]. Now prāg (or pūrva) -deśa is a regular term for the eastern half of Hindustan, and its popular use, so as to cover the whole country from Magadha in the east to the borders of the Panjāb, appears from the fact that Alexander’s Indian campaign, if continued further east, would have brought him into collision with the Prasioi, the Prācya people, sc. the Magadha empire.” Thomas (1944), p. 91.

Conversely, some writers believe Juli 車離 is an attempt to transcribe Kosala, and that Dongli is a mistake for Juli 東離. For example:

“Again the name Kosala may be compared with Ch’ê-li, the main name which the Wei-liao gives for the country. Now, the first character of the latter [sic] ch’ê 車 has two pronunciations, tsai and kio, and using the latter we find that the whole name, pronounced Kiao-lei, could well be a contraction of Kosala. If this is so it would suggest that the Hou-han-shu name Tung-li is also a corruption.” Shiratori (1956b), p. 40.

However, I find the argument that ju 車 (K. 74a *ki̯o / ki̯wo; EMC kɨə̆) could be used to transcribe the Kos of Kosala unconvincing.
          Eitel (1888), p. 68, in discussing “Prācya or the eastern country,” places it to the east of Madhyadeśa (in which Śāketa and the kingdom of Kosala were located). Its western borders did change and, according to some ancient authors, stretched at times almost as far west as Prayāga, which is south of Śāketa. Other ancient authorities place it further east, in the catchment area of the Brahmaputra River.

7.2. This phrase, 車離國一名禮惟特, 一名沛隸王 Juliguo yi ming li wei te, yi ming pei li wang, Chavannes translates as: “Le royaume de Kiu-li est aussi appelé Li-wei-t’o 禮惟特, ou encore P’ei-li-wang 沛隸王;...”. Chavannes (1905): 551.
          The reconstructed transcription of Liwete [Li-wei-t’o] is:

li 禮 K. 597d: *liər / liei; EMC lεj

wei 惟 K. 575n *di̯wer / i̯wi; GR 12144: sgi̯wər / i̯wi; EMC jwi

te 特 K. 961hⁱ * d’ək / d’ək; EMC dək

          “According to the account in the Life it was from Kanauj that Yuan-chuang [Xuanzang] went 600 li south-east to Ayudha [阿踰陀 – Ayutuo. W-G: A-yü-t’o]. The capital of this country, which was about a mile [1.6 km] to the south of the river, has been identified with the Ayodhyā of other writers, the old capital of Oudh. On account of difficulties of direction and distance Cunningham proposes a different site for Yuan-chuang’s Ayudha. But it seems better to adhere to Ayodhyā, and to regard Yuan-chuang’s Ganges here as a mistake for a large affluent of the great river. The city was on the south bank of the river, and about 120 miles [193 km] east-south-east from Kanauj. Its name is found written in full A-yü-t’ê-ye (阿喩駃也)[Ayujueye. W-G: A-yü-chüeh-yeh], Ayudhya (Ayodhyā), and the city is said to have been the seat of government of a line of kings more or less mythical. We know also that to the Hindus Ayodhyā was the old capital of Rāma and the Solar race. It is possible that an old or dialectic form of the name was Ayuddha, and the Chinese translation of the Sanskrit word, which suits either form, means invincible or irresistible. Moreover we find that Yuan-chuang makes his Ayudha the temporary residence of Asanga and Vasubandhu, and other authorities represent Ayodhyā as a place of sojourn for these two illustrious brothers. Then the Ayudha of Yuan-chuang is apparently the Sha-ki or Saket, that is Ayodhyā, of Fa-hsien ; this was ten yojanas south-east from the Holi village which was three yojanas south from Kanouj. Alberuni makes Ayodhyā to have been about 150 miles [241 km] south-east from Kanauj, being 25 farsakhs down the Ganges from Bāri, which was 20 farsakhs east from Kanauj. It is the Sākētā or Oudh of the Brihat-sanhitā which merely places it in the “Middle country”.” Watters (1904-05), I, pp. 354-355.

7.3. Peiliwang 沛隸王 [P’ei-li-wang]. This term or phrase is not clear. It could indicate that an alternative name for the kingdom is Peiliwang, as implied by the Chinese sentence construction; or it could mean that the king’s name is Peili (= Pala or Bala)?
          Pāla in Sanskrit is not necessarily a name, it can also mean, ‘governor,’ ‘guard,’ ‘protector,’ ‘king,’ ‘prince,’ – which raises new possibilities. See Monier-Williams, p. 627.

Pei 沛 – K. 501f *p’wâd / p’uâi; EMC p^haj^h

li 隸 – K. 1241m – liei; EMC lεj^h

wang 王 – K. 739a *gi̯wang / ji̯wang; EMC wuaŋ^h

7.4. Tianzhu 天竺 [T’ien-chü] = North(west)ern India. See note 5.15.

7.5. Shaji or Shaqi 沙奇 [Sha-chi or Sha-chih] = Sāketa.

Sha 沙 – K. 16a *sa / ṣa; EMC: ʂaɨ / ʂɛː

qi, ji 奇 – K. 1s *g’ia / g’jie̯; EMC: giə̆ / gi, and kiə / ki

zhi 祗 – K. 590p *t̑̑i̯ǝr / tśi; EMC: tɕi

da 大 – K. 317a *d’âd / d’âi, and t’ âd / tâi, and *d’ âr / d’ â; EMC: da’ / dài, and – da^h / daj^h

The Chinese form of this name Shaqi 沙奇 is identical in both the accounts of the Hou Hanshu and the Weilue. Fa Xian’s account of c. 405 CE, gives Shazhi 沙祗 [Sha-chih] for the city and, possibly, Shazhida 沙祗大 [Sha-chih-ta] for the name of the kingdom.
          This latter reading would make a fuller and more accurate transcription for Sāketa but, unfortunately, it is impossible to tell whether the last four characters in the sentence, 沙祗大國, should be read: “the kingdom of Shazhida” or “the great kingdom of Shazhi.” The Rabatak Inscription writes: [ζ]αγηδ̣ο = [z]agēḍo for Sāketa – see Sims-Williams (1998), p. 81.

The Hou Hanshu says:

“The town of Shaqi (Sāketa) is the capital of the kingdom of Dongli東離 (‘the Eastern Division’). It is more than 3,000 li (1,248 km) to the southeast of Tianzhu (Northwestern India). It is a large kingdom.
          The climate and products of the country are similar to those of Tianzhu (Northwestern India). There are several tens of towns of the first rank whose leaders give themselves the title of king. The Da Yuezhi (Kushans) attacked and enslaved it.”

By modern maps it is about 1,250 km from Sāketa to Taxila, the ancient capital of Gandhāra, which lends credibility to the accounts of both the Hou Hanshu and the Weilue.
          From this it follows that Dongli was composed of a number of small kingdoms with Sāketa as the administrative centre. It would seem fairly safe to assume that this “Eastern Division” included the cities such as Kausambi, Pataliputra and, probably, Champa, that we now know from the Rabatak inscription, were conquered by Kanishka by the first year of his era (127 CE).
          Sarnath (near modern Varanasi) should be included in this list, as two inscriptions from year 3 of the same era have been found there. See: Sims-Williams and Joe Cribb (1995/96), p. 83; Sims-Williams (1997), p. 2; Falk (2001), pp. 121-136; Kumar (1973), pp. 42, 245.
          The Sūtrālakāra or Kalpanāmaktikā, ascribed to Aśvaghoṣa, but probably by Kumāralāta, and translated by Kumārajīva in the early fifth century CE, mentions a ‘Dong Tianzhu,’ or ‘Eastern India,’ which must surely refer to the same division or region as Dongli. See: Zürcher (1968), p. 385. Also see: Chavannes (1905), p. 551, n. 1; Chavannes (1907), p. 194, n. 5; and Mukherjee (1968), p. 35. Faxian, in his account of c. 405, says:

“. . . they came to the great kingdom of Shâ-che.” Beal (1885), p. 54.

This passage could equally well have been translated as beginning: “. . . they came to the kingdom of Shazhida. . . . ” It is not clear whether the character 大 da is part of the name Shazhida (which makes for an excellent and more complete transcription of Sāketa), or whether it relates to the following character, 國 guo, making it read: “they came to the great kingdom of Shâ-che,” as Legge has interpreted it in his translation above. Either interpretation is acceptable.
          In either case, it almost certain they refer to the same town. Both names share the same first character, and 祗 – zhi gives a reconstructed pronunciation of K. 590p *t̑i̯ər / tśi; EMC tɕi – which is close enough to the reconstructed qi in the Weilue’s Shaqi (= K. 1s *g’ia / g’jie̯; EMC: giə̆ / gi). The Chinese name Shaqi 沙奇 [Sha-ch’i] has been identified with Śāketa by Thomas (1944), p. 90:

“Returning to Tung-li (Dongli), we may note with some considerations in favour of an identification with the central region of northern India, madhya-deśa, the ‘mid India’ of Chinese writers. It was a great country, extending over ‘several thousand li’ from north to south and from east to west ; it had dozens of great cities, each with a king ; nevertheless it was a unity having a capital city. This cannot fail to recall to mind the fact that from the time of the Nandas and Mauryas the great central part of Hindustan had continued to constitute an imperial state, which in the period of Aśvaghos̲a and Kaniṣka had two capital cities, namely Śāketa/Ayodhyā and Pāṭaliputra. As regards Śāketa, LÉVI has noted (pp. 90-1) that sometimes the Chinese transcriptions of its name, Sha-chi resemble the Chinese form, Sha-ch’i, of the name of the capital of Tung-li ; but, since one of LÉVI’s Sha-chi forms should in fact be Sha-chi’s [sic – should read Sha-ch’i] (KARLGREN, no. 879), there is rather identity than similarity in the two cases ; and, if it is urged that the Chinese ch’ should represent an Indian g rather than a k, that is no difficulty since the Indian name would naturally have been heard in the Prākrit form Sha(sa) geda, which is the one reproduced in Ptolemy’s Σαγδα [Sagda].”

There has often been confusion between Ayodhyā and Sāketa, which some writers consider the same, and, others, as separate cities. Ayodhyā is considered by Hindus to be the birthplace of the god Rama and is, therefore, a particularly holy place. A dispute between Muslims and Hindus in recent years (in which Hindus insisted that the famous Babari mosque in modern Ayodhyā was built on the site of the original temple marking the birthplace of Rama) finally led to a violent confrontation which resulted in the destruction of the mosque by orthodox Hindus.
          The destruction of the mosque apparently uncovered a large (5 ft by 2.5 ft) slab of sandstone with an inscription in Sanskrit proving the existence of a Hindu temple (a Vaishnava temple – Vishnu being regarded as an incarnation of Rama) “during the closing years of the eleventh century AD and its reconstruction/renovation sometime about the mid-twelfth century AD.” Shastri, Ajay Mitra. (Date unknown), p. 3.
          It also makes clear that, at this time, the “temple city of Ayodhyā [was] situated in the Saketamandala (district, line 17), showing that Ayodhyā and Saketa were closely connected, Saketa being the district of which Ayodhya was a part.” Lal (Date unknown), p. 4.

The phrase: qi wang zhi Shaqi zheng 其王治沙奇城, in the Weilue, translates as ‘the seat of the king (of this country – i.e. Juli 車離 or Dongli 東離) is the city of Shaqi.’ Shaqi 沙奇 is undoubtedly a transcription of Sāketa, one of the two capital cities of Kośala (the other being Sāvatthī or Śrāvastī).
          The name 沙奇 Shaqi in the Weilue is very close to the Shazhi(da) 沙祗(大) [Shâ-chih(ta)] that Fa Xian [c. 400 CE] uses for the country of Sāketa (Beal rendered it Sha-che). Shiratori (1956b), p. 40, says:

“Fa-hsien’s 法顯 Fo-kuo-chi 佛國記 puts Shê-wêi 舍衞, the then capital of Kosala, at a distance of eight yojana 由延 from a city called Sha-chih-to 沙祗多. This name, evidently transcribed from the Sanscrit name Sāketa, may be confidently connected with Sha-ch’i, the name given in the Wei-liao to the capital of the country.”

I believe that in the accounts of Faxian and Xuanxang we have proof that Sāketa and Ayodyha were either identical or very close to one another. Faxian, speaking of Shazhi (Sāketa), says:

“As you go out of the city by the southern gate, on the east of the road (is the place) where Buddha, after he had chewed his willow branch, stuck it in the ground, when it forthwith grew up seven cubits, (at which height it remained) neither increasing or diminishing. . . . Here also is the place where the four Buddhas walked and sat, and at which a tope was built that is still existing.” Legge (1886), p. 54.

Xuan Zang [Hsüan-tsang], speaking of Ayoutuo(ye) [A-yü-t’o-(yeh)] or Ayodhyā says:

“To the north of the city 40 li (but only 4 or 5 li according to Hui-li - see the next quote), by the side of the river Ganges, is a large sanghârâma in which is a stûpa about 200 feet high [46.2 metres or 152 English feet], which was built by Aśôka-râja. . . . By the side is a stûpa to commemorate the place where are traces of the four past Buddhas, who sat and walked here.” Beal (1884), p. 225.

The account in the biography by Xuanzang’s companion Huili says:

“North-west of the city [Ayodhyā] four or five li, and by the side of the river Ganges, is a great Sanghârâma, in which is a Stûpa about 200 feet high. This was built by Aśôka râja on the spot where Buddha in old days delivered the Law for three months. By the side of this Stûpa is a spot where the four Buddhas of the past age walked for exercise.” Beal, (1884), p. 85.

It is clear from the above accounts that Shazhida and Ayodhyā are either identical, or so close to each other as to be considered twin cities. Even if we take the maximum distance given of 40 li between them, this is considerably less than 20 km (no matter which of the various suggested measurements we take for the Tang li). If, on the other hand, we take Huili’s account as more accurate, it is a matter of less than 2½ kilometres. 

Sāketa is included among the six great cities of early Buddhism in the early Buddhist scriptures of Sri Lanka:

“When this had been said, the venerable Ānanda, spoke to the Bhagavā, saying: ‘Let it not be, Lord, that the Bhagavā should pass away in this mean place, this uncivilised township in the midst of the jungle, a mere outpost of the province! There are great cities, Lord, such as Champā, Rājagaha, Sāvatthi, Sāketa, Kosambi and Banares - let the Bhagavā have his final passing away in one of those! For in those cities dwell many wealthy nobles and brahmanas and householders who are devotees of the Tathāgata, and they will render due honour to the remains of the Tathāgata.” From the Mahā-Parinibbāna Sutta translated in: Vajira (1961), p. 68.

The texts of the Hou Hanshu and the Weilue specifically state that Śāketa was under the control of the Da Yuezhi or Kushans. It is known from the Rabatak Inscription that Śāketa was one of the conquests of Kanishka in the first year of Kanishka’s era, see Sims-Williams & Cribb (1995/6), especially pp. 78 and 83; Sims-Williams (1998), p. 83.
          There is now seemingly convincing new evidence that the Kanishka era began in March circa 127 CE – see Falk (2001), especially p. 130. There are also inscriptions dated in year 2 of this era at both Kosam (ancient Kauśāmbī – to the southwest of Sāketa) and at Sarnath, well to the east of Sāketa. See: Kumar (1973), pp. 244-245.
          Finally, if Kanishka’s era started as late as 127 CE, the information on Śāketa and Dongli in the Hou Hanshu and the Weilue must necessarily have been collected some time after Ban Yong’s report to the Emperor c. 125 CE. Mac Dowall (2002), pp. 163-164, states that:

“Fussman reminds us that the Rabatak inscription by itself gives no assurance that Kanishka’s grandfather bore the name of Vima Taktu. The restoration was apparently suggested not by anything visible on the stone itself but by comparison with (a) the Brahmi inscription from the devakula at Mat near Mathura in India and (b) the trilingual inscription in Bactrian, Kharoshthi and an unknown language from Dasht-e Navur near Ghazni in Afghanistan.” Also: “Fussman pointed out that the inscription does not derive from Kanishka himself- the first person is never employed: and it does not date from year one of Kanishka–this is the hypothetical restoration of Sims-Williams whereas the stone is completely effaced at this point. But whatever his name was there was a prince between Kujula and Vima Kadphises.” See also Appendix N.

Sims-Williams, however, has never dated the Rabatak Inscription to year one of Kanishka’s era, although, as he said in a recent email: “The Rabatak Inscription is largely concerned with events of the first year of Kanishka.”
         I am very pleased and grateful to Professor Sims-Williams for his kind permission to reproduce here the revised second edition of the translation of this remarkable inscription as presented in his article, “Further Notes on the Bactrian Inscription of Rabatak, with an Appendix on the names of Kujula Kadphises and Vima Taktu in Chinese.” See Sims-Williams (1998), pp. 81-90. It is quite clear from this translation that the inscription must have been written some time after the events it mentions. The revised translation of the inscription says:

“. . . . of the great salvation, Kanishka the Kushan, the righteous, the just, the autocrat worthy of divine worship, who has obtained the kingship from Nana and from the gods, who has inaugurated the year one as the gods pleased. And he *issued a Greek *edict (and) then he put it into Aryan. In the year one it has been proclaimed unto India, unto the *whole of the realm of the *kṣatriyas, that (as for) them – both *Wasp, and Sāketa, and Kauśāmbī, and Pāṭaliputra, as far as Śri-Campā – whatever rulers and other powers (they might have), he had submitted (them) to (his) will, and he had submitted all India to (his) will. Then King Kanishka gave orders to Shafar the karalrang *at this . . . to make the sanctuary which is called B . . . ab, in the plain of the (royal) house, for these gods, whose *service here the . . .  *glorious Umma leads, (namely:) the above-mentioned Nana and the above-mentioned Umma, Aurmuzd, the Gracious one, Sroshard, Narasa, (and) Mihr. And he gave orders to make images of the same, (namely) of these gods who are written herein, and he gave orders to make (them) for these kings: for King Kujula Kadphises (his) great grandfather, and for King Vima Taktu (his) grandfather, and for King Vima Kadphises (his) father, and for himself, King Kanishka. Then, as the king of kings, the scion of the race of the gods . . . had given orders to do, Shafar the karalrang made this sanctuary. [Then . . . ] the karalrang, and Shafar the karalrang, and Nukunzuk the ashtwalg [performed] the (king’s) command. (As for) *these gods who are written here––may they [keep] the king of kings, Kanishka the Kushan, for ever healthy, fortunate, (and) victorious, and [may] the son of the gods *rule all India from the year one to the year *one thousand. . . . the sanctuary was founded in the year one; then in the *third year [it was] completed . . . according to the king’s command, also many *rites were endowed, also many *attendants were endowed, also many . . . . . . the king gave an *endowment to the gods, and for these . . . *which [were given] to the gods . . .”

I should also mention that Professor Sims-Williams is currently working on a third edition of this inscription “based on examination of the inscription itself rather than just photos.” It is to be expected that his examination of the actual inscription may lead to some changes and/or additions, giving us a better understanding of the inscription.

7.6. This conquest of Sāketa apparently occurred just prior to, or during, 127 CE, as outlined in note 7.5.

7.7. “The people, the men and women are all eighteen chi tall”. This would indicate an unbelievable height of 4.158 metres or 13 feet 9 inches. This is clearly a mistake. The character for 10 appears to have been inserted in front of the account in the parallel text of the Hou Hanshu which says “The men and women there are eight chi tall.” This latter gives the far more credible height of 1.85 metres, or just over 6 English feet. See also: Chavannes (1905), p. 551, n. 2.

7.8. This statement shows that Ju- (or Dong-)li was, therefore, fully subservient, and considered part of the Kushan Empire, not just a tributary state. This is probably also implicit in the name ‘Dongli’ which may be read as the “Far East (of the Kushan Empire).”

Section 8 – The Kingdom of Panyue 盤越 [P’an-yüeh] = The kingdom of Pandya at the southern tip of India.

8.1. The kingdom of Panyue 盤越 [P’an-yüeh] = Pandya – is also called Hanyue wang 漢越王.

“Pandian kingdom. – This was Pāndya, the southernmost, and traditionally the earliest, of the three Tamil states. Roughly it coincided with the modern districts of Tinnevelly and Madurā; at the time of the Periplus it extended beyond [sic] the Ghāts and included Travancore. The capital, originally at Korkai ( the Colchi of § 59, which see ) had been removed to Madurā ( 9^o 55’ N., 78^o 7’ E. ).

Here too, as in the Chēra kingdom, the name is used for the country and as a dynastic title, not as the name of any king.” Schoff (1912), p. 211. Pelliot discussed the name ‘Hanyue wang’, and the name 磐起 ‘Panqi’, which is used in the Hou Hanshu for this kingdom:

“...but there can be no doubt that the form Panqi is due to a copyist’s error, and it is probably quite late, as the name Panyue was used again in the 7th century when the Tang reorganised the western countries using the names from previous histories in a whimsical manner (Cf. Chavannes, Documents sur les Tou-kiue occidentaux, p. 68). And, on the other hand, Panyue, with the same orthography, is also found in the Liang shu (ch. 54, folio 7 a, of the edition of the Tushujicheng Library).
          The alternation of pan for han in the Weilue is interesting. . . . the alternation of pan and han in transcriptions is established through other examples. . . . ” Translated from Pelliot’s review of Chavannes’ translation of the Weilue. Pelliot (1906), p. 371, n. 2.

“Towards the end of the first millennium BC south India moved from pre-history into history, and literary records reflecting contemporary events are available. Ashoka in his inscriptions refers to the peoples of south India as the Cholas, Cheras, Pandyas and Satiyaputras – the crucible of the culture of Tamilakam – called thus from the predominant language of the Dravidian group at the time, Tamil. The use of the suffix putra in some of these names would suggest a system of clans and chiefs. The first three chiefdoms became almost generic to societies based on clans and lineages in the area and acquired the status of kingdoms in a later period. The Cholas and the Pandyas were located in the eastern area, with a Chola concentration in the lower Kaveri. Korkai and Alagankulam are recently excavated sites, thought to have been exchange centres in Pandyan territory. The first is linked to pearl fisheries and the second developed as a port. Karur on the banks of the Kaveri was an important inland centre, as was Kodumanal, with excavated evidence of working semi-precious stones. Gradually, over time, the Cheras were associated with the western coast. The Satiyaputras, with a more limited history, have been identified through being mentioned in a local inscription in Tamil-nadu.” Thapar (2002), p. 229.

“Muchiri or Muziris, located perhaps in the vicinity of Kodangallur/Cranganore (near Kochi) [recently shown to be at Pattanam, north of Paravoor, on an ancient branch of the Periyar River – see the news items in note 12.12 (5)], was linked to the trade in pepper, spices and beryl. A recently discovered Greek papyrus of the second century AD, documents a contract involving an Alexandrian merchant importer and a financier that concerns cargoes, especially of pepper and spices from Muziris, which provides evidence of the large volume of this trade. References to the rich pepper trade with Malabar continue for centuries, up to the time of the Portuguese. Location of coin hoards suggest a link from Muziris via the Palghat Gap – tapping the beryl mines – and along the Kaveri Valley to the east coast.” Thapar (2002), p. 241.

“The earliest Roman coin [found in the Tamil kingdoms of the southern part of the peninsula] is that of Emperor Augustus 31 b.c. to 14 a.d It is towards the end of the reign of Augustus, contact between the Tamils and the Roman empire, is recorded. Even then trade and contact remained along the West coast, keeping the land and its vicinity and through middlemen. No Roman coin seems to have reached the Tamil land, before the Christian era. In all probability they reached the Tamil Kingdoms in the reign of Tiberius, in whose time there was a great increase in trade. The coins of Tiberius in gold and silver found in the South are quite numerous⁽¹⁸⁾. It is only after the establishment of direct trade between the Roman world and the Tamil Kingdoms, as a result of the discovery of the monsoon by Hippalus, Roman artefacts are found in increasing number in the Tamil soil. Though Romans themselves never seemed to have reached the Tamil Kingdom it is the Greeco [sic – should read “Graeco”] Romans from Egypt who represented them. Hence for the identification of the imported pottery found in Roman context in Tamil Nadu, one has to look to Egypt. The identification of the red slipped ware as the African red ware thus confirms that they were brought by the Egyptian Greeco [sic] Romans.” Nagaswamy (1995), pp. 76-77.

“Roman coins have been found in large numbers in and near Coimbatore. They point: (1) active overland trade between the Malabar coast and the eastern coast at Arikamedu via Coimbatore (Palghat) gap thus avoiding the sea voyage around cape Comorin and (2) to a natural concentration of the trade of the three Tamil Kingdoms in Coimbatore District” Nagaswami (1995), p. 79.

“It should be noted that in the extreme south of the Indian peninsula, the Pandyas played a dominant role in sea trade and even had control over the Sri Lankan trade as well as the pearl fishery. The location of a Roman trading center on the Pandyan coast [Alagankulam -at the mouth of the Vaigai river, southeast of Madurai] is quite understandable from the geopolitical point of view. The mention in Tamil texts of the Pandyan king’s special liking for yavana wine also becomes significant in this context. . . .
          The extreme south of Tamil Nadu was the ancient Pandya-nadu comprising the districts of Madurai, Tirunelveli, and Ramanathapuram. The capital was inland at Madurai and the important ports were Korkai, Tondi and Kumari. Excavations at Alagankulam provide important archaeological evidence for trade between the Mediterranean and Pandyan country. We may also note that some “hoards” of Roman coins have been recovered from different parts of the Pandyan country such as Madurai, Kaliyamputhur, and Karivalamavandanallur. The last mentioned is close to the newly discovered site of Alagankulam.” Raman (1991), pp. 129, 130.

“The classical geographers and the Tamil literature of the ‘Śaṅgam’ age have familiarized historians with the outlines and some of the details of Indian trade with the West in and after the first century A.D. At its prime, the trade was extensive. It included as Indian exports pepper, pearls, gem-stones, muslin, tortoise-shell, ivory and silk; and as imports from the West coral, lead, copper, tin, glass, vases, lamps, wine and, at first, coined money.
          By the latter part of the first century A.D. the literary evidence makes it clear that this trade was organized on lines not unlike those of the European ‘factories’ established in India from the sixteenth century onwards. The Periplus of the Erythraean Sea (c. A.D. 60-100) speaks of ὲμπόρια νόμιμα, the unqualified ὲμπόρια of Ptolemy (c. A.D. 150), which may fairly be described as treaty-ports. That is to say, permanent lodges of Western traders were settled in them under formal agreement with the appropriate Indian ruler, and were visited at the proper seasons by convoys of deep-sea merchantmen. And just as the agents of the Portuguese, Dutch, French or British establishments in the days of the Companies travelled widely in the interior to circumvent the middlemen, so we may with probability recognize in some of the Roman coin-hoards far from the sea (fig. 48) the penetration of earlier Western traffickers with similar interest.
          On the scale indicated, this organised interchange implies, a knowledge of the periodicity of the monsoons. Without that knowledge, which, according to Pliny, could restrict the crossing of the Indian Ocean to forty days the laborious coastal voyage to India or the still more precarious overland routes must have prevented the development of regular and direct trade with southern or eastern India. The establishment of permanent agencies in those parts must therefore have post-dated the discovery of the so-called ‘Hippalus’ or north-western monsoon ; a discovery of unknown date but appreciably earlier than Pliny and the Periplus, and possibly, though not certainly, earlier than c. A.D. 21, when Strabo speaks of 120 ships sailing for India from Myos Hormos on the Red Sea. On the other hand, there is no hint that the monsoon was familiar to Mediterranean merchants before the time of the Roman principate. The unification of the western world under Augustus (23 A.D. – A.D. 14) and the recorded reception by him of at least two Indian delegations (c. 25 B.C. and 21 B.C.) provide an obvious context, if not for the actual discovery, at least for its diffusion beyond the corporations of Arab sailors and other agents who had previously monopolized the Indian trade thitherto.
          Further than this the literary evidence fails to carry us. Warmington conjectures from it that the direct route to the Malabar coast, i.e. the full use of the monsoon, was introduced ‘soon after 41‘. Archaeology now indicates a rather earlier date. Some considerable time prior to the abandonment of the manufacture of Italian red-glazed pottery (Arretine and related fabrics) – an event which is unlikely to be later than A.D. 50 – a Roman emporium is now known to have been firmly established far up the eastern coast of India, near Pondicherry. We may infer that at least as early as c. A.D. 30, and possibly before the death of Augustus, regular monsoon-trade had been established between the Mediterranean and western India, with a coastwise or overland extension to the Coromandel coast. Whether at this early date the route was carried to its logical conclusion in Malaya and China is at present unknown.” Wheeler, Ghosh and Deva (1946), pp. 18-19.

“Arabian and Indian traders had doubtless been crossing the Indian Ocean for years, and it is hardly conceivable that they had not learned to use the periodic winds, the chief meteorological phenomenon of this sea. Indeed, some believe that voyages westward were made as early as the seventh century B.C., that they had become common by the time of Nabonidas of Chaldea (556-539 B.C.), when Indian and Chinese ships reached Babylonia, and that this trade continued to flourish without interruption down to Ptolemaic times [“..., see Kennedy in J.R.A.S., 1898, 241-87”]. At any rate, Hippalus’ discovery – whether made in the age of Augustus or, as some think, in that of Claudius (perhaps c. A.D. 45) – at once affected eastbound trade from Rome and Greece, and in this respect he may justly be said to have been quite as great a discoverer as if he had really been

the first that ever burst

Into that silent sea.

Merchants now spoke of a wind called the hippalus, but no geographer mentioned him as a person, or knew that a mariner of that name had discovered a new route to India.
          The chief depot for pearls was Ceylon. For this reason and because the route up the Nerbudda to Barygaza was hard to navigate and the Gulf of Cutch a little to the north was full of shoals, the successors of Hippalus soon learned to cut across the sea from Syagrus, and later even from Guardafui, to Melizigara on the site of the present fishing village of Jaigarh or to Rajapur farther south, between Bombay and Goa. Still later mariners cut across to the coast of Damarike in the land of the Tamil even farther south, the usual route being to Muziris or Nelcynda below modern Calicut. As we have already noted, the voyage to Muziris from Berenice below Myos Hormos took seventy days [Pliny, 6, 103-4]. The voyage to India from Puteoli, the southern port of Italy, took 112 days, which means the round trip took nearly two thirds of a year. Lucian, the second-century satirist, says that within two Olympiads three return trips of 16 months each could be made from the Pillars of Heracles to India with time off for exploring [Hermotimus, 4].” Hyde (1947), pp. 206-207. See note 12.12(6) for the recent discovery of ancient Muziris, just south of the mouth of the Periyar river mouth in Kerala State, southwestern India.
          Before we turn to the Pondicherry site, the literary picture of these trading-cities may be completed in outline by reference to one or two familiar passages in the Tamil literature. This literature is very insecurely dated, but the relevant passages, which are numerous, may be ascribed on general grounds to the early centuries A.D. . . . And Tamil rajas employed bodyguards of western mercenaries, ‘the valiant-eyed Yavannas whose bodies were strong and of terrible aspect’ and who, equipped with ‘murderous swords’, were ‘excellent guardians of the gates of the fort-walls’. In this capacity they are said to have been employed at Madurā [the inland capital of the Pandyan kingdom]. Yavana craftsmen were also sought after in southern India especially for the manufacture of siege engines. In one way and another, the Yavana in partibus [i.e. in inhospitable or hostile surroundings] enjoyed a considerable prestige whether as trader or as settler.” Wheeler, Ghosh and Deva (1946), pp. 19-20.

“Alagankulam⁽¹⁾, a village about 20 km from Ramanathapuram, near Ramesvaram in Ramanathapuram District of Tamil Nadu, is situated on the banks of the river Vaigai. It is virtually at the mouth of the river about two kilometres inland from the sea. Originally the sea was skirting the village during ancient times but now has receded far. Now a tiny village Arrankarai is situated on the coast that overlooks the Sri Lankan coast. It is claimed that the opposite side could be reached in about 25 minutes by country boats from this village.

The river Vaigai passes through the city of Madurai, the ancient capital of the Pandyas. It is dry for nearly half of the year and empties itself in a big tank near Ramanathapuram town, and beyond that, the river bed has now dried up totally, only traces of the bed passing through Alagankulam and joining the sea are visible, to show that it ever reached the sea some centuries ago.

However, there are copious references in the Sangam Tamil literature to this river since it passed through the Pandyan capital Madurai. The Pandyas were called the rulers of the Vaigai Vaigai Koman, just as the Cholas were called the rulers of the Kaveri. The Sangam work Maduraik kanchi⁽²⁾, calls it a perennial river (line 356). The commentator Naccinarkkiniyar also confirms it.”

1. Nagaswamy, R., Alagankulam on [sic – should read “an”] Indo-Roman Trading Port.

2. Maduraikkanchi in ‘Patthu Pattu’, Dr. U.V. Swaminatha Iyer, (ed) Madras, line 356.

Nagaswamy (1995), p. 70 and nn. 1-2.

”With the discovery of Alagankulum, indicating profuse Roman contacts in the later period, the centre of attraction now shifts to the Pandyan Kingdom. situated on the river Vaigai, the river of the Pandyas, it is tempting to identify the site with Saliyur mentioned in the accounts of the classical geographers. “The far-famed Saliyur (Selur, Salur, Delur in Ptolemy) opposite the north end of Ceylon, was a similar mart overcrowded with ships which had crossed the dangerous ocean and from which costly wares were landed”⁽²⁴⁾

24. Warmington, op. cit., p. 62.

Nagaswamy (1995), pp. 78-79.

There are also a number of references in early Tamil literature to the “Yavanas” (Romans or Roman subjects) playing an integral part Tamil Nadu life – not only in trade, but as bodyguards and craftsmen. Some of the references specifically mention the Pandyan Kingdom. For details see: Nagaswamy (1995), pp. 96-102.

8.2. Yibu 益部 [I-pu] translates as ‘Yi Circuit’, also known as 益州 ‘Yi Province.’ The sentence in Section 8 of the Weilue: [盤越] 與益部相近, should be read as “[Panyue] is in contact with Yi Circuit.”
          This phrase has regularly been translated by earlier scholars as “it [meaning Panyue] was near [or nearest] Yibu.” The Chinese text uses the word 與 yu here. This certainly can be used to translate the English word ‘near,’ or ‘to be close to,’ but it can also represent, ‘to associate with’, ‘to make contact with’, ‘to frequent’, ‘an allied country.’ I feel certain that, in this context, it should be read ‘in contact with’ or ‘in communication with.’ See note 15.5 for a detailed discussion of a similarly constructed sentence using this same word 與 yu.
          Chavannes (1905), p. 552 n. 1, suggests that 益部 Yibu (‘Yi Circuit’) should be read as, 益郡 Yiqun: “the commandery of Yi, during the Han period, which had its centre to the east of the secondary prefecture of 普寧 Puning (Prefecture and Province of Yunnan).” However, I see no need for this. The region (basically the region of present-day southern Yunnan) was commonly known as Yi Circuit during the Han Dynasty.

          “By and large, during Later Han, it appears that the terms circuit and province were essentially interchangeable. Strictly, however, a circuit was supervised by an Inspector, while the term province implied that the same region was controlled by a Shepherd (mu, also rendered as Governor), an official of senior rank with effective executive powers. During the reign of Emperor Ch’eng, in 8 B.C., the office of Inspector of a circuit had been changed to Shepherd of a province. The title and functions were restored to their former situation in 5 B.C., under the government of Emperor Ai, but shepherds were again established in 1 B.C., and the office continued in force until the end of Former Han.” Holmgren (1980), p. 55.

8.3. Chavannes (1905), p. 552, translated this passage to read: “Jia Si 賈似, who was a man of the country of Shu 蜀 (Western Sichuan), went there.”
          I believe, however, that the sentence: Shurengushu zhiyan 蜀人賈似至焉 [Shu-jen-ku-shu chih-yen], should be read “Traders from Shu (Western Sichuan) travel this far.” I believe the third character 賈, is intended here in its form of gu – a ‘trader’ or ‘merchant’ rather than in its other form, jia – which is frequently used as a personal name.
          If traders were coming to India from Western Sichuan (i.e. from well inland, near the border of Assam and Tibet), they presumably travelled there by one of the several overland routes. For detailed discussions of these routes see Pelliot (1904), pp. 131-413.

“Shu 蜀 (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.

It is quite likely that things were much the same to the end of the 19^th century as there were overland trade routes to India at this time from western Sichuan [Szechwan]. There were certainly routes that a loaded porter or ‘coolie’ could negotiate as recounted by Mesny (1896), pp. 122-123:

“A coolie will carry 200 lbs of brick tea destined for Tibet and the West through the province going at a rate of fifteen or twenty miles a day over terrible mountain passes, with which the reports of various travellers, such as the late lamented Mr. E. C. Baber, have made us acquainted.”

Mesny (ibid., p. 141), adds:

“The military jurisdiction of this Viceroy [of Sichuan] extends across Tibet to the very confines of Assam, Nepaul, India, Kashmere and Kashgaria. Chinese garrisons having been established in those remote regions for the past century and a half; the troops being periodically relieved by detachments from the Ssŭ-ch’uan forces. The soldiers in those garrisons are allowed to have Tibetan wives.”

          “Finally, if one adds the less commonly used, extremely dangerous overland route through the Chinese province of Sichuan southwards to Burma and East India (an early predecessor of the notorious World War II Burma Road), we have three large arteries for east-west trade, of which the Silk road is the most famous.” Baumer (2000), p. 11.

Section 9 – The Central Route. (Refer to Appendix A for details).

9.1. The kingdom of Weili (or Yuli) 尉梨 [Wei-li or Yü-li] = Korla.

Wei or yu 尉 – K. 525a,b: *i̯wəd / jwe̯i; EMC ?uj^h, or as yu: ?ut

li 梨 – K. 519h,i: *li̯ər / lji; EMC li.   

Stein (1921), Vol III, pp. 1230 ff., and (1928), Vol. II, pp. 724, 777 ff, locates Weili at Kara-kum (actually marked Weili, some 40 km south of Korla, on modern maps), and Weixu at Korla itself.
          The Hanshu places Weili 100 li (42 km) south of Yanqi, or Karashahr, and Weixu the same distance away, but no direction is given. Kara-kum is twice as far from Karashahr as the Hanshu indicates Weili and Weixu are.
          By modern road, it is about 47 km southwest from Yanqi (Karashahr) to Korla. The small difference of 5 km from the measurement of the Hanshu between Yanqi and Weili can be explained by changes in the route, or the town centres since the Han period.
          Korla has long been the largest centre in the region after Karashahr itself, having abundant water and extensive farmlands, and control of the main routes to the south and west of Karashahr. Weili is given a population of 9,600 compared to only 4,900 for Weixu in the Hanshu, which also mentions that it adjoins Shanshan and Qiemo (Charchan) to the south.
          The Hanshu states it is 300 li (125 km) west from Weili (Korla) and 350 li (146 km) east of Kucha to the seat of the Protector General at Wulei. Measured either west from Korla or east from Kucha this brings one to the oasis of Yangisar [or Yanghi-hissār].
          Because Stein was not aware of the true length of the Han li (he thought it to be about 1/5 of a mile, or 322 metres, instead of the true length of about 416 metres), he was unable to properly choose between the three oases of Bugur, Yanghi-hissār, and Chādir, as the site of modern Wulei. See Stein (1928). Vol. II, p. 794. He favoured the Bugur oasis as the probable site because it is the largest oasis between Korla and Kucha (ibid. 796). However, he also gave Yanghi-hissār serious consideration and noted its strategic importance:

          “The importance of Yanghi-hissār is increased by the fact that a route leads from it across the high range northward of the Yulduz plateau at the head of the Kara-shahr valley. It was stated to be the first practicable route east of Kuchā to the plateau, and to be much used by Mongols taking supplies from the oasis to their grazing grounds. The pass crossing the watershed was said to retain snow all through the year ; but these hardy Mongol customers, I was told, find it practicable even during the winter months.” (Ibid. 791).

          “Wei-li 尉犂, GSR 55b and 519g : *iwcd/*jwei - licr/liei or li̯cr/lji. The 19^th and 20^th century Han shu commentators locate it around Bugur. Chavannes (1905), p. 552, note 5, and (1906), p. 234, note 2 locates Wei-li – and Wei-hsü . . . [see note 9.2] – in the vicinity of Lake Bagrash or Bostang nor. He criticizes Wylie for following the Hsi-yü t’ung-wen chih 西域同文志 of 1766, and consequently locating Wei-li at Kalgan-aman, close to and NE of Korla. Huang Wen-pi (1958), pp. 6-7, suggests that the extensive ruins NE of Korla and South of Ssu-shih li ch’eng 四十里城 (marked on his “additional map 5” at c. 86° 28’ E and 41° 55’ N) might still be the capital of Wei-li, containing a Han settlement. Shimazaki (1969), p. 44, still places Wei-li at Kalgan-aman.” CICA: 177, n. 585.

          “The prosperous town of Korla lies on the Baghrach Kol, a large lake, through which the Kaidu river pursues its course. The water of the lake is of fabulous transparency, and enlivened by endless numbers of large fish, most of them belonging apparently to the barbel family. There are, however, shad as well – ugly creatures as long as a man and with enormous mouths.
          Herr Bartus, as an old sailor, could not resist throwing his line in here. . . . He had flung into the water a pound of meat on a gigantic hook and strong line, and an antediluvian monster had swallowed the bait. With great effort he dragged it out of the water, to the intense delight of the entire population, who were watching the visitors’ doings. It weighed about fifty pounds, had a smooth skin – brown spotted with white – and was something like our eel-pout. In spite of my warning – for some of the fish here are dangerous eating – Herr Bartus persisted in having some of it for dinner and found it excellent. . . .
          There are only two districts in the whole country where fish are often eaten, viz. round about Maralbashi, where the River Tarim brings down enormous quantities, which are enjoyed by the Dolans living there ; and, secondly, in the neighbourhood of Lake Lop-nor, where the whole population, apparently differing in many respects from the other Turks, live chiefly on fish, either fresh or dried. It is remarkable that both the Dolans and the dwellers round Lop-nor are looked upon as people of another race by the Turks.
          The lake at Korla is the playground, too, of innumerable flocks of water-birds, and is the breeding-place of swans, whose plumage is in much demand by the Chinese as an edging for valuable robes. Geese and ducks of different kinds frequent the shores and surface of water in great quantities, and we always saw numbers flying in their hook-shaped flocks across the sky. Herons of every kind are also to be found there, but we could never inspect them closely as they always took to timid flight at the approach of men on horseback.” von Le Coq (1928), pp. 109-110.

Put together, I think this evidence confirms the identification of Weili as Korla, and Yangisar [Yanghi-hissār] as the seat of the Protector General during the Former Han.

9.2. The kingdom of Weixu 危須 [Wei-hsü] = Hoxud = Chokkur.

          Stein (1921), Vol III, pp. 1230 ff.; and (1928), Vol. II, pp. 777 ff) places Weixu 危須 at Korla. However, the Hanshu states that Weixu is 200 li (83 km) farther from the seat of the Protector General at Wulei 烏壘 than Weili (500 li instead of 300), and places it 100 li (42 km) from Yanqi (Karashahr). One must, therefore, assume it is further east than Karashahr, on the route to Turfan.
          This identification is supported by the Shuijingju [Shui-ching chu] which indicates that the Yulduz river used to have a northern branch, flowing into the northwest of Lake Bostang, to the west of Weixu. This old course of the river is now indicated by the network of irrigation channels, to the west of the present town of Hoxud, that service this region, the water being used up before it can flow into the lake.

“Wei-hsü 危須, GSR 29a and 133a : ngwia/ngwie̯ - si̯u/si̯u. The Shui-ching chu 2.30ff., says that the Tun-hung river’s ... eastern tributary flows southeast and then divides into two [although the present-day Yulduz River (= Tun-hung) apparently does not bifurcate]; coming from Yen-ch’i (i.e. Karashahr), it is led West of Wei-hsü and then flows southeast to end in the Tun-hung Marsh. . . . The latter is identified with Bostang Lake or Bagrash Kul and the former with the Hai-tu or Yulduz. Hsü Sung locates Wei-hsü to the Southeast of Bostang Lake; Chavannes (1905), p. 552, note 6, seems to accept this localization, criticizing Wylie for following the Hsi-yü t’ung-wen chih (see note 585 above) and placing Wei-hsü at Chagan-tungi, Northeast of Karashahr.” CICA, p. 177, n. 587.

These descriptions and the distance, (travelling east from the seat of the Protector General at Wulei) 100 li (42 km) past Yanqi (Karashahr), makes it very probable that Weixu was located somewhere closer to the present town of Hoxud.

9.3. The kingdom of Shanwang 山王國 [in the western Kuruk mountains]. This placename could just as well be translated as “the kingdom of the King of the Mountains.”
          Chavannes (1905), p. 552, n. 7, points out that this kingdom is undoubtedly the same as the kingdom of Shan 山國 in the Hanshu and the kingdom of Moshan 墨山國 [‘(Black) Ink Mountain’] in the Shuijing. He says that it must have been located between Lake ‘Bagrach’ (‘Bagrax’, ‘Bostang’, or ‘Bosten Hu’) and Lop Nor and that Grenard’s proposal to locate it at Kyzyl sanghyr, 130 km southeast of Korla is “very plausible.”
          The 6^th century Shuijing places Weili, which I identified as Korla, 240 li (100 km) to the west of Moshan [ ‘(Black) Ink Mountain’]. See Stein (1928) Vol. II, p. 724.
          The Hanshu places the kingdom of Shan only 160 li (67 km) southeast of Yanqi (Karashahr) so it must be located near the extreme western end of the Kuruk-tāgh, although its exact position remains to be determined.
          Stein (1921) Vol. I, p. 334, says it “can only roughly be located in the Western Kuruk-tāgh,” although he does consider the possibility that it might have been located at Singer (= Kyzyl sanghyr); but this is much further than 67 km to the southeast of Karashahr. See also: ibid., p. 420; CICA, pp. 85, n. 85, and 182, n. 615.

9.4. The kingdom of Yanqi 焉耆 [Yen-ch’i] = Karashahr.
          Yanqi 焉耆 [Yen-ch’i] has long been confidently identified with the region of modern Karashahr. The Buddhist Sanskrit name was Agni-. For detailed discussions of the derivation of the name and its likely associations, see Bailey (1985), pp. 1-2; 137-138, and Pulleyblank (1963), pp. 99; 123. During the time of the Moghuls it was called Chálish (Jálish) – see, for example, Elias (1895), pp. 99, 100, 102, 122 of the Introduction.

          “The whole of this district round Kara-shahr and Korla is, from a geographical and political point of view, both interesting and important ; for whilst all other parts of Chinese Turkestan can only be reached either by climbing high and difficult passes – the lowest of which has the same elevation as Mont Blanc – or traversing extensive and dangerous waterless deserts of sand-hills, here we find the one and only convenient approach to the land through the valleys of several rivers in the neighbourhood of Ili, where plentiful water abounds in the mountain streams on all sides, and where a rich vegetation makes life possible for wandering tribes. Such Kalmuck tribes still come from the north-west to Tal. They are Torgut nomads who pitch their yurts round about Kara-shahr and live a hard life with their herds. . . .
          Just as these Mongols wander about here at the present day, so the nomadic tribes of an earlier period must have used this district as their entrance and exit gate. The Tochari (Yue-chi), on their way from China, undoubtedly at that time passed through this gate to get into the Ili valley. . . . ” von Le Coq (1928), pp. 145-146.

The Hanshu mentions that Yanqi (Karashahr) “adjoins Wu-sun on the north.” CICA, p. 177, n. 588, p. 178. This was of particular concern to the Chinese as Stein makes plain in the following passage:

“These observations on the present conditions of Kara-shahr will make it clear that, while the territory has been favoured by nature in various ways, its geographical position must at all times have exposed it to a very serious drawback. I mean its close vicinity to, and its easy access from, mountain tracts which, as far back as history takes us, have always had a particular attraction for nomads. It is unnecessary here to explain in detail how the famous grazing uplands of Yulduz have been cherished haunts for all the great nomad nations, from the Wu-sun and Huns downwards, which held sway along the T’ien-shan, that natural spina, as it were, in the cycle of Central-Asian migrations. Situated as Kara-shahr is at the very mouth of the big valley leading down from Yulduz, it must have been like a gate specially inviting those who had their favourite summer camps on those grassy plateaus and necessarily looked to the oases on the south as their richest grounds for raids and exactions. Whenever Chinese power was firmly established from Turfan to Kashgar or beyond, the gate might be kept safely closed. The same is likely to have been possible during periods while internal feuds or conflict with nomad aggressors weakened the tribes in the north. But the danger must always have been close at hand, and from time to time Kara-shahr was bound to suffer from its onset. The oases further west would then be exposed, too, to plunder and heavy exactions of tribute. But the additional risk of prolonged occupation would be reserved for Kara-shahr, which alone could offer grazing grounds adequate for the maintenance of large nomad hosts.” Stein (1921), p. 1180.

“Some distance before the town of Yanqi, soda-whitened marshes, tall grasses and grazing cattle indicate the proximity of the vast Baghrash Lake. Though today Yanqi is only the country seat of the Yanqi Hui Autonomous County, where one of the main industries is the making of reed screens for fencing and roofing, historically it was the very important oasis of Kara-shahr (Black Town), which in AD 11 revolted against Han domination by murdering the Chinese protector-general. The revolt was ruthlessly stamped out by the Han-Dynasty general Ban Chao, who sacked the town, decapitating 5,000 inhabitants and carrying away 15,000 captives and 300,000 head of livestock.” Bonavia (1988), p. 147.

“On MARCH 26 we passed along the northern outskirts of the Korla Oasis, and turned northeast following the courses of the Konche darya. The trail led over a plain of piedmont gravel, gradually rising toward the mountains. . . .
The road entered a narrow river gorge. Not far from the entrance stood a Chinese outpost and we had to produce our permits to travel to Karashahr. For some reason the Torgut region was carefully watched by the Chinese.
We camped on the banks of the river not far from a grove of trees. The air was wonderfully cool and refreshing, and we rested from the dusty atmosphere of the desert. In the early morning we broke camp and started for Karashahr, still some twenty miles distant. Not far from our camp, the Konche darya turned its course eastward and our route crossed a low hill and then descended into the basin of the Lake Baghrash. The road lay over a sandy tract covered by tamarisk shrubs. . . .
We passed by the village of Shorchuck, in the vicinity of which are situated important ruins, explored by German and British expeditions. After five hours’ ride, we reached the Karashahr darya and had to spend a considerable time waiting for rafts to come from the other bank of the river.” Roerich (1931), p. 102.

“Yen-ch’i 焉耆, GSR 200a and 5521 : *ian/*iän or gian/jiän - g’i̯ɛr/g’ji, traditionally identified with Karashahr. Huang Wen-pi (1958), p. 7, suggests that “the old walled town of Ha-la-mu-teng” 哈拉木登, a few li South of the modern settlement of that name and North of the Haidu River might have been the administrative centre of Yen-ch’i; the site is located on Huang’s map nr. 2 at c. 86^o 5’ E and 42° 16’ N. – For different ancient misspellings of this name see Chavannes (1905), p. 564, note 2. Wang Ching-ju (1944), p. 91, believes that in Han-times Yen-ch’i was pronounced *ārgi, leading to a later *arśi; it is to be noted that the Αoρσoι mentioned by Strabo are usually identified with the people of Yen-ts’ai. . . . ” CICA: 177, n. 588.

9.5. The kingdom of Gumo 姑墨 [near modern Aksu].         

Gu 姑 – K. 49g: *ko/kuo; EMC: kɔ

mo 墨 – K.904c:   *mək/mək; EMC: mək

Although many writers assume that the largest centre of the ancient kingdom of Gumo is identical with modern Aksu, I suggest that, while it may have included the site of Aksu, its seat of power was considerably to the south.
          The main town of the kingdom of Gumo in the Hanshu is called Nancheng 南城 ‘Southern City,’ which, by its very name, suggests it was located near the southern end of the region watered by the Aksu River. CICA, p. 162 and n. 498. The exact site is yet to be located, but Nancheng was likely in the strategic region of modern Wuxuntamu or Aral, approx 40^o 26’ N, 80^o 51’ E, more than 100 km southeast of modern Aksu, near the junctions of the Khotan and Aksu rivers with the Kashgar River.
          From here, one had many choices of routes across the desert. One could follow the Kashgar river west to Kashgar, or the branch tributary, the Yarkand River to Yarkand. Alternatively, one could follow the Khotan River due south to Khotan, or the main Tarim River east to Loulan and Lop Nor.
          Grenard was the first to say that the earlier identification of Wensu with ancient Aksu was wrong, and showed that Aksu was represented in the Han period by Gumo. Chavannes (1905), p. 553, n. 1, has outlined strong evidence in favour of the latter identification which is backed up by Stein’s observations (1921) Vol. III, pp. 1297-1298.
          The Hanshu says that Gumo was 670 li (279 km) west of Kucha (the modern road is 262 km) and 1,021 li (425 km) from the seat of the Protector General stationed at Wulei 烏壘 = Yengisar (see CWR note 20.10). It adjoined the Wusun to the north, and was 15 days by horseback south to Khotan. CICA, p. 162, and n. 497.
          Stein (1921), p. 1299, notes that it was a “seven-miles [11.3 km] ride from the ‘New’ to the ‘Old Town’ of Ak-su. . . . ” Given that we still don’t know the exact positions of the Han settlements at either Kucha or Aksu, the distance of 273 km along the road from modern Kucha to ‘Old Aksu’ is remarkably close to the distance of 279 km given in the Hanshu.

“The area is frequently visited by light dust storms generating an eery, creeping ‘fog’ around the base of the sand mounds and the occasional ruins of the Han-Dynasty (206 BC–AD 220) beacon towers.

Neolithic artefacts from 5000 BC have been discovered in the Aksu area. By the first century BC news had reached the Chinese imperial court of the Kingdom of Baluka, one of the 36 kingdoms of the Western Regions. The kingdom, aided by the Xiongnu, held out against the Chinese army under General Ban Chao for a time, only to have him march upon the capital city in AD 78 and execute 700 inhabitants.” Bonavia (1988), p. 160. The Hanshu – see CICA, p. 162 – says that Gumo produced copper, iron and orpiment.

“The 262 kilometre (162 mile) journey from Kucha to Aksu [by bus] takes between five and six hours. The area is frequently visited by light dust storms generating an eery, creeping ‘fog’ around the base of the sand mounds and the occasional ruins of the Han-Dynasty (206 BC–AD 220) beacon towers.” Bonavia (1988), p. 160.

9.6. The kingdom of Wensu 溫宿 [Wen-su] = modern Uqturpan or Wushi 烏什.

wen 溫 – K. 426c *·wən / ·uən; EMC ?wən

su 宿 – K. 1029a *si̯ôk / si̯uk; EMC suwk

The Hanshu (CICA, p. 163) places Wensu at an impossible 2,380 li (990 km) from the seat of the Protector General at Wulei (Yengisar), which was only 350 li (146 km) east of Kucha.
         However, it states that it was 270 li (112 km) west of Aksu (Kumo), 610 li (254 km) south of 赤谷 Chigu, the capital of the Wusun near lake Issik-kol, and 300 li (125 km) east of 尉頭 Weitou.
          These indications place it firmly near modern Uqturpan (= Uch-Turfan) or Wushi 烏什, in the valley of the Toxkan (or Toshigan) River, further up the Aksu River from the Kingdom of Aksu or Gumo.
          It must be emphasized that several earlier writers confused Uqturpan with modern Wensu, which is written with the same characters as ancient Wensu. Modern Wensu is less than 20 kilometres northwest of Aksu whereas, ancient Wensu was about 90 km to the west, on the site of modern Uqturpan or Wushi 烏什. See: Chavannes (1905), p. 553, n. 1; 1906: 224, n. 3); Stein (1921) Vol. III, pp. 1299-1301; Pelliot (1959), p. 492; CICA, p. 162, n. 502.
          Ancient Wensu controlled access up the Toshigan valley and also the approach to the Bedel Pass, the main route north to Issik-kul. Stein (1921) Vol. III, pp. 1300-1301, remarks:

“The usual dust haze of the spring was hiding the view of the great snowy range of the T’ien-shan northward. It was thus impossible to obtain even a distant glimpse of the Bedel Pass, by which Hsüan Tsang had once gained the Issik-kul region and thence Sogdiana. But even without that imposing background Uch-Turfān presented itself to me as the most picturesque and pleasant of any district headquarters I had visited in Chinese Turkestān. The view of the fertile green valley, set off vividly by the chain of barren grey hills which encircle the town from the south, was particularly striking from the height of the Chinese citadel. . . . This crowns the top of a precipitous rocky spur, which adjoins the west wall of the town and projects beyond it like a huge natural ravelin, rising with its westernmost cliffs to a height of some 250 feet. The citadel and the flanking defences joining it to the town walls are recent, having been built in the place of fortifications destroyed when Uch-Turfān was besieged and taken during the Muhammadan rebellion. But this natural stronghold is bound to have been utilized since early times.
          . . . . the ‘kingdom’ and town are referred to [as Wen-su] in the Former Han Annals, the Hou Han shu, and the Wei lio. The former Han Annals ascribe to it a population of 2,200 families, which seems proportionate, and indicate its position quite correctly with reference to the Wu-sun capital which lay 610 li to the north, to Ku-mo, or Ak-su, and to Wei-t’ou 尉頭, 300 li westwards. . . . ”

9.7. The kingdom of Weitou 尉頭 [Wei-t’ou] = modern Karaqi, to the west of Akqi.

          Weitou is located in the Han shu (CICA: 163), as being 300 li (125 km) west of Wensu or Uqturpan [Uch-Turfan]. The only place it could possibly be, assuming the distance of 125 km west of Uqturpan is correct, is the small community called Karaqi or Ha-la-ch’i.
          The Times Comprehensive Atlas of the World, Tenth Edition, 1999, places Karaqi at 40^o 44’ N; 77^o 59’ E, while the U.S. Defence Mapping Agency Aerospace Center ONC, 1:1,000,000 map (Sheet F6, Edition 5, revised 1981) places Ha-la-ch’i at 40^o 50’ N; 77^o 55’ E.
          Karaqi controls a key fork in the routes of the region, the main summer route from Aksu to Kashgar heads southwest from here, while another route leads almost due west towards Ferghana. Its importance undoubtedly derives from its strategic position.
          Most previous writers have mostly located it at Safyr Bai or Akqi, but both of these places are far too close to Wensu to explain the statement in the Hanshu that it was 300 li (125 km) to the west of that place.

“[Wei-t’ou] is mentioned by the Later Han Annals as on the road leading from Su-lê or Kāshgar north-eastwards to Wên-su, Ku-mo, and Kuchā, and may be located with much probability about Akche on the upper Tushkan-daryā (map No. 14. A. 2). This place is passed by the much-frequented summer route between Ak-su and Kāshgar, and the adjoining area shows now a good deal of Kirghiz cultivation.” Stein (1921), Vol. III, p. 1301.

“The distance of about 70 miles [113 km*] between Uch-Turfān and Akche agrees well with the 300 li indicated by the Former Han Annals between Wên-su and Wei-t’ou. Safar-bai, which M. Grenard, Mission Dutreuil de Rhins, ii. p. 61, has suggested for the name of the latter place, is too near; see Map No. 14. c. 2. It is worth noting that Akche is the first place with agricultural resources which the traveller by this route reaches after leaving the Kashgār district.” Ibid., p. 1301, n. 26. *Note that with the most careful measurements on Stein’s own maps, I cannot get any distance further than 63 miles or 101 km between Uch-Turfān and Akche.

“Wei-t’ou 尉頭, GSR 525b and 118e : *•i̭wəd / *•ὶjwḙi - d’u/d’ə̭u; [Note: Karlgren’s reconstructions here are incorrectly given in CICA] Wang Hsien-ch’ien identifies it with Wu-shih County, at approx. 41° N, 79° E. Chavannes (1905), p. 554, note 1, refers to Grenard, Mission scientifique en Haute Asie II, p. 61, who localizes Wei-t’ou at Safyr bay, Southwest of Uch Turfan.” CICA, p. 142, n. 374.

9.8. Qiuci 龜玆 [Ch’iu-tz’u] is universally accepted as referring to the Kucha oasis. See note 4.16.

9.9. Zhenzhong 楨中 [Chen-chung] = Arach (near Maralbashi). Not much is known of this place during Han times – there is no mention of it in the Hanshu, and no major entry in the Hou Hanshu. It appears to have been close to Kashgar.
          The first brief reference to this “kingdom” seems to be in the biography of Ban Chao, under the name of Sunzhong 損中 [Sun-chung]. Zhong 忠 [Chung], a king of Kashgar who had been unseated and replaced by Ban Chao with the help of Kangju and the Kushans, returned and established himself here in 86 or 87 CE. He tried to make an alliance with Kucha, but Ban Chao tricked him to a meeting, and had him beheaded. Chavannes (1906), pp. 230-231.
          The next reference to Zhenzhong 禎中 [Chen-chung] (note the slightly different initial character) dates to 169 CE in the Cefuyuangui [Ts’e fu yüan kuei] (chap. 973, p. 8b):

“The Prefect of Liangzhou, Meng Tuo, ordered the Congshi Ren She to take five hundred soldiers from Dunhuang (Shazhou) and, with the Wuji Marshall (Wu-chi Szu-ma) Cao Kuan and the Chief Administrator of the Western Regions (Xiyu Zhangshi) Zheng Yan who was at the head of more than thirty thousand men from Yanqi (Karashahr), Qiuci (Kucha) and the Anterior and Posterior tribes of Jushi (Turfan and Jimasa) to go to attack Sule (Kashgar). They attacked the town of Zhenzhong for more than ten days without being able to subdue it. Then they withdrew.” Translated from Chavannes (1905), p. 554, n. 2.

Zhenzhong 禎中 (or Sunzhong 損中) was obviously a very well-fortified site to withstand such a substantial siege. We have scant information except that, as it was attacked by troops from Karashahr, Kucha, Turfan and Jimasa, who were en route to conquer Kashgar, it was probably some distance to the east of Kashgar, guarding the main route via Aqsu. That it was mainly a defensive position, rather than a thriving oasis town, is probably the reason it is not included in Han itineraries.
          The most likely place is Arach (39^o 58’ N; 78^o 40’ E) on the “Old road to Tumshuk,” where Stein found remains of a ruined fort and watchtowers controlling the only viable pass though the Achal-tāgh range. This would have controlled the main routes from the northeast to both Kashgar and Yarkand:

          “Having thus traced the ancient road from Ak-su as far as the Lāl-tāgh site, there still remains for us to consider the line of its probable continuation to the south-west. Such a line had necessarily to cross the chain of hills represented by the Bēl-tāgh and its southern extension, the Achal-tāgh or Ōkur-mazār-tāgh ; and the map shows that only two passages were available for it. One is represented by the gap in the Bēl-tāgh above described ; the other leads through the defile that separates the southern end of the Bēl-tāgh from the northern offshoot of the Achal-tāgh formed by the low rocky spur of Arach. I had visited this defile on my first rapid reconnaissance from Marāl-bāshi in May, 1908. I ascertained on that occasion that the high road from Tumshuk to Marāl-bāshi had in quite modern times, until after the Chinese reconquest of the Tārīm basin in 1877, passed through it, the present line past the Ōkur-mazār being then impracticable owing to great undrained marshes in the area now occupied by the lands of Chār-bāgh. I had been able to trace there the remains of old watch-towers and of other fortifications meant to guard the defile, and their presence left no doubt that an important road must have passed there in ancient times.
          I had not time then to examine the desert ground to the east of the defile, and this was an additional reason for now returning to Marāl-bāshi by this passage, appropriately known as Achal, ‘the opening’. I found no reason to regret the decision. We passed numerous lines of dead Toghraks with shallow dry channels between them for the first two miles from the Lāl-tāgh site, clearly showing by their direction that running water must have reached this part from the south, i.e. from the area west of Tumshuk still liable to inundations from the Kāshgar-dārya. Then, after crossing a belt of tamarisk-cones, we emerged upon a bare clay steppe undergoing wind-erosion, and here for more than a mile came across frequent patches of ground covered with ancient potsherds and other ‘Tati’ remains. In the midst of them we crossed the unmistakable embankment of an old canal about 12 feet across at its top and owing to erosion of the surrounding land raised some 5 feet above the present ground level.” Stein (1928), pp. 77-78 and map MARAL-BĀSHI, Serial No. 8. See also: Stein (1921), III, pp. 1311-1312, and map 15.

          “At the small town of Sanchakou, 214 kilometres (133 miles) southwest of Aksu, is a turnoff for Bachu, called Maralbashi [also known as Tumschuk] in the records of the 19^th- and 20^th-century European explorers. Sir Aurel Stein traced the walls of a fort and the structures of an extensive city, both long abandoned. A direct desert route along the Yarkand River linked it with Yarkand (Shache), a journey accomplished by Stein in five days.” Bonavia (1988), pp. 162-163.

          “In view of the old remains traced at and near the gap of Achal, it is, I think, safe to assume that the ancient route passed on from Lāl-tāgh to this defile, as did the modern route via Tumshuk until some forty years ago. But beyond it, topographical facts, combined with such archaeological indications as I have discussed above in connexion with my journey from Kāshgar, point to a bifurcation. On the one hand it is on general grounds highly probable that there was a south-western continuation of the ancient route to the present Marāl-bāshi ;  for though the present town is of avowedly modern origin, and probably the surrounding oasis also, yet a look at the map shows that the branch of the road which forms the easiest and most direct connexion between Ak-su and Yarkand must always have passed the site in question. Whether there existed any large settlement during ancient times corresponding to the present Marāl-bāshi is doubtful, seeing that the detailed accounts of the ‘Western kingdoms’ contained in the Han and T’ang Annals make no mention of a special territory or tract at this place.” Stein (1928), p. 80.

“The distance from Faizabad [the last settlement of the Kashgar Oasis] to Maral-bashi was covered in four stages. The road ran through a sandy country covered by jungle, said to abound in game. In the past this must have been a huge forest area, but nowadays the jungle is retreating to the north and the sands are rapidly encroaching on the forest. The road was made difficult by numerous dry tree stumps and roots completely buried under the sand. . . .

On March 3, we entered the large and important Oasis of Maral-bashi, connected by two important routes with Kashgar and Yarkend. The oasis itself is not large but it is well irrigated and has an abundant supply of subsoil water and several lakes. The Yarkend darya approaches it from the south. To the northwest and north rise low, rocky ridges – southern offshoots of the outward T’ien Shan. . . .

Our next stop was Ak-tumshuk or simply Tumshuk. On leaving the cultivation zone of Maral-bashi and Char-bagh Bazaar, a small hamlet some fifteen miles northeast of Maral-bashi, we emerged on a vast sandy plain covered by shrubs. The stage was a very long one, about thirty miles, and we reached the village of Tumshuk at about eleven o’clock at night. Not far from the village stands a Chinese inscription dating back to the time of the reconquest of the New Dominion in 1877. North and southeast of Tumshuk lie important ruins. The ancient sites were excavated by the eminent French sinologist, Professor Paul Pelliot, during his expedition in 1906-8.
From Tumshuk it took us four long stages to reach Aksu. . . .” Roerich (1931), pp. 95-96.

9.10. The kingdom of Suoju 莎車 [So-chü] = Yarkand.

Suo 莎 – K. 16f: *swâ / suâ; EMC: swa.

OR Sha – K. 16f: *sa / ṣa; EMC: ʂaɨ / ʂɛː

ju 車 – K. 74a: + *ḳi̯o / ki̯wo; EMC: kɨǝ̆

          OR che – *ẗ’i̯å / ẗś’ia; EMC: tɕ^hia

There has never been any question that the Suoju (or Suoche) 莎車 [So-chü or So-ch’e] of the Han period is to be identified with modern Yarkand or Suoche (also written Shache). Pelliot (1963), pp. 876-885, discusses the identification of Yarkand in detail and, as he notes in an earlier work:

“All scholars, both Chinese and European, agree that the kingdom of So-chü of Han times must be identified with Yarkand.” Pelliot (1959), p. 879.

            “For this city Ptolemaios’s text has four variants σοÃτα, σοÃγα, σότα, σάγα. Here the fourth is nearest to the Saka word. The modern, half-Turk name is Yār-kand. The Turkish yăr is ‘cleft’ and ‘rock’ as in yar tuzï ‘rock salt’. The Turkish has kept the meaning of the original name and added the word kand ‘city’. Here yar is from yarmaq ‘to split’.” Bailey (1958), p. 73.

“Yarkend impressed us as being a much more active and thriving place than Khotan. It is the chief entrpôt of Indian and Afghan trade and its roofed bazaars and chaikhaneh or restaurants were thronged by different nationalities. Yarkend is the largest oasis in the country. It is well watered by irrigation channels and has an extensive cultivation of rice, which is exported to Khotan and Kashgar. Besides rice, the oasis produces corn, wheat, barley, cotton, flax, hemp, sesamum, and tobacco. According to recent official estimates, the population of the oasis is over 200,000. The Mohammedan town or Kone-shahr is surrounded by a brick wall with several towers. The Han-ch’eng or Chinese section lies to the northwest of the native quarter. The Afghan Government maintains a consul at Yarkend who, although not recognized by the Chinese as such, supervises the Afghan colony and its interests in the oasis. The most costly products of Afghan export are opium and Badakhshani horses. The trade is clandestine and the offenders are usually punished. For the past few years the export of Badakhshani horses has been prohibited by the Amir, and Afghan traders, travelling to Chinese Turkestan or elsewhere, have to sign a document that they will return to the country with their horses. Notwithstanding these strict measures, Badakhshani horses are constantly seen on the Yarkend market, the horse trade being carried on through Kirghiz traders.
Most of the Yarkand trade is conducted with India and the Indian colony is the largest foreign colony in the oasis. . . .” Roerich (1931), pp. 86-87.

Also see: Stein (1907), p. 88; CICA, p. 139, n. 361; Bailey (1985), p. 73; Bailey (1958), p. 133; Daffinà (1982), pp. 325-326.

9.11. The kingdom of Jieshi 竭石 [Chieh-shih]. Possibly one of the smaller towns in the Kashgar oasis. Stein (1907), p. 15, believed it may have represented an early attempt to render the name Kāshkār in the Kashgar oasis, and in Chitral (also known as Kāshkār), where we find an identical, or very similar, name:

          “The open and fertile part of the main valley, containing the large villages which bear collectively the name of Chitrāl and form the political centre of the Kāshkār or Chitrāl State, answers remarkably well to the description given in the T’ang Annals of the mild climate and rich produce of Chieh (Chieh-shih). And in view of the topographical arguments already adduced for its identification, we need not hesitate to suggest also that it was the local name Kāshkār, or an earlier form of it, which the Chinese endeavoured to reproduce by Chieh-shih or Chieh-shuai. The application of the term Kāshkār to the territory of Chitrāl is well attested from Muhammadan sources, and its use is still current throughout those regions. Chieh-shih, as an attempt to represent Kāshkār by Chinese sounds, would have a parallel in the name Ch’ia-sha which Hsüan-tsang gives to the present city and oasis of Kāshgar, in Chinese Turkestān.” Stein (1907), p. 15.

“Chavannes (1906), p. 554, note 3, believed that the place name Chieh-shih 竭石, GSR 313r and 795a : g’i̯at/g’i̯ät - d̑‘ǎi̯k/z’i̯äk in the Wei-lüeh should be compared to the T’ang names 佉沙 ki̯wo-ṣa and 迦帥, Karlgren (1923), no. 342 and GSR 559a : ka-ṣi, both indicating Kashgar; the Wei-lüeh term would then provide the necessary authority to identify Kashgar with Ptolemy’s κασια χώρα. Pulleyblank (1963), p. 219. . . reconstructing an “Old Chinese” pronunciation *sa̲ĥ-gleats, believes this to be an early form of transcribing Soghd, to be identified with Kesh.” CICA: 130, n. 320.

9.12. The kingdom of 渠沙 Qusha [Ch’ü-sha]. Literally ‘Big Sands’ or ‘Spacious Sands’.

“Qusha 渠沙 appears identical to Qusuo 渠莎 [also pronounced Qusha] that the Pei shi (chap. XCVII, p. 3 b) mentions as occupying the ancient town of Suoju莎車 (Yarkand). Qusha or Qusuo was therefore not a principality distinct from Yarkand.” Translated from Chavannes (1905), p. 554, n. 4.

“But a much earlier source [than the Weishu or the Peishi], the Wei-lio, which was written in the second quarter of the 3^rd cent., mentions in succession “the kingdom of So-chü (Yārkänd), the kingdom of 碣石 Chieh-shih (*G’i̯ät-źi̯äk; perhaps read 竭石Chieh-shih, *G’i̯ät-źi̯äk, and cf. CHAVANNES, Doc. sur les Tou-kiue, 69), the kingdom of 渠沙 Ch’ü-sha (*G’i̯^wo-a), the kingdom of 西夜 Hsi-yeh, ...” (cf. CHAVANNES, in TP, 1905, 554). The Hou-Han shu (118, 4 a) has a notice on “the kingdom of Hsi-yeh, also called 漂沙 P’iao-sha” (*Pi̯äu-a; CHAVANNES, misled by a misprint in the T’u-shu-chi-ch’êng Shanghai edition, gives 虜沙 Lu-sha in TP, 1907, 174; but all the ancient editions have P’iao-sha, adopted in DE GROOT Chin. Urkunden, II, 79). The Hou-Han shu owes much to the Wei lio, and I hold it for certain that its P’iao-sha is merely a copyist’s error for the Ch’ü-sha of the Wei lio. It is no less clear that the Ch’ü-sha of the Wei lio and the Ch’ü-so of the Pei shih also represent one and the same name. In all likelihood, Ch’ü-sha, indirectly supported by P’iao-sha of the Hou-Han shu, is the correct form, and sha was altered to so in the Pei shih under the influence of the following So-chü. But even the identification of Ch’ü-sha = Ch’ü-so with So-chü = Yarkand is unreliable, since both the Wei lio and the Hou-Han shu (this with the corrupt form P’iao-sha) agree in mentioning Ch’ü-sha quite apart from So-chü (Yarkand). The wrong identifications of ancient names in this chapter of the Pei shih are numerous.” Pelliot (1963), p. 880.

My own suggestion is that Qusha referred to a settlement on the main route between Yarkand and Khotan, possibly in the region of modern Kizil Bazaar, two stages towards Khotan from Yarkand, and a notoriously sandy region. Aurel Stein (1912), pp. 138-139 reports finds of ancient Uighur manuscripts and the remains of ancient dwellings and a part of a leather slipper in this region, although he was unable to provide chronological evidence for the finds.

“Next day we were in Kizil Bazaar – a half-ruined village with a vast desert plain called Karakum or “Black Sands.” The next stage to Yangi-hissar lay across a desert country intersected by hamlets with small patches of cultivation.” Roerich (1931), PP. 87-88.

9.13. The kingdom of Xiye 西夜 [Hsi-yeh] = Karghalik.
          Chavannes (1903), p. 397, n. 4, (1905), p. 554, n. 5, and (1907), p. n. 3, identified this kingdom with modern Yularik, south of Yarkand. However, the directions given in the Hanshu and the population figures given in the Hou Hanshu make it almost certain that it refers to modern Karghalik, as Aurel Stein first pointed out:

“. . . we are necessarily led to identify Hsi-yeh as Karghalik ; for only on the assumption that this great oasis is meant can we account for the striking difference in population which the notice of the Later Han Annals indicates by stating the number of households as 2,500 at Hsi-yeh and only 350 at Tzu-ho. The proportion is about the same as a modern census would be likely to reveal between the oasis of Karghalik proper and the Beg-ship comprising Kök-yar, Yül-arik and Ushak-bashi. The identification of Hsi-yeh with Karghalik is in striking agreement with the statement in the Ch’ien Han shu that Hsi-yeh joined P’i-shan on the east and So-ch’ê on the north ; for Guma and Yarkand are the neighbours on these sides exactly as here represented.” Stein (1921a), pp. 86-87.

My only qualification to Stein’s analysis is that he locates Zihe [Tzu-ho] too close to Kharghalik, partly due to his under-estimation of the Han li (at about 322 metres instead of the true 415.8 metres).
          But, even using his figures, the text would place Zihe 322 km from Kashgar – a figure impossible to reconcile with him situating Zihe not far south of Karghalik, around Kokyar (which is only about 50 km south of Karghalik), and the tiny oasis of Yularik, about 10 km further east. Kokyar lies on the winter route (zamistani) between Yarkand and the Karakoram Pass. See: von Le Coq (1928), p. 153.
          The text states that Zihe was in a gorge, some 1,000 li, (about 416 km from Kashgar) – which makes it almost exactly at modern Shahidulla (or Xaidulla = Pinyin: Saitula). Shahidulla is a strategically important centre on the way to the Sanju and Karakoram Passes on the main route south to Ladakh. This shortcut to India via Ladakh was in regular use until the Chinese closed the borders soon after the Communist victory in 1949. 
          The main obstacle on the route was the high and dangerous Karakoram Pass (5,575 m or 18,291 ft). Caravans would often rest and graze their animals in the fertile valley near Shahidulla until conditions were favourable to cross the less rigorous Sanju Pass, also called the “Suget Pass’ (5,364 m or 17,598 ft), and then the notorious Karakoram.
          Shahidulla controlled the route to the north of the Sanju Pass from where one could head northwest to Pishan and Yarkand or northeast towards Khotan. Although the Suget is a difficult pass, it is possible to take laden yaks across it:

“On the 5th of November, after passing through the gorge of the Karakash River, at the foot of the walls of the fortlet of Shaidullah, built by the Kashmiraians and long since abandoned by them, we came at Toghrusu, into the midst of the gay tumult of a Kirghiz wedding. . . .
          On the 7th, we left the Karakash Daria, the valley of which is impracticable in the downward direction, and began to ascend the gorge of one of its affluents which runs down from the Sanju Pass. Imagine an exceedingly narrow gorge, whimsically tortuous, deeply confined within tall peaked rocks, bare and strangely hewn and slashed and the whole gorge obstructed by flint rubbish. On reaching the end of this gorge, we found ourselves as though at the bottom of a well. With the assistance of some Kirghiz oxen, we scaled one of the walls of the well and thus reached the summit of the Sanju Pass, which is at a height of 16,800 feet. From there, according as one turns to the north or the south, the view offers a striking contrast. In the south is a monstrous chaos of gigantic snow mountains and dazzling glaciers, which the rays of the sun sometimes cause to look like great blue lakes slumbering amid a polar whiteness ; in the north, a few brown hills, beyond which stretches something like a vast ocean wrapped in a shroud of grey mist : this is the Kashgarian plain and its atmosphere laden with dust.
          The ascent of the pass was not easy, but the descent was worse. The slope is so steep that, in a league of horizontal projection, one descends 1,880 metres and, for a distance of 800 metres, the slope, at 45 degrees, is covered with a thick layer of ice. The yaks are really wonderful animals which, descending a mountain like this, carry over two hundred pounds on their backs without stumbling. Our horses, although carrying no burden, did three-fourths of the road in some other way than on their feet : one of them slipped so badly that it was hurled to the bottom of the valley and broke its spine. . . .
          From the foot of the pass, one follows a deep, grassy valley, here and there meeting the round tents of Kirghiz herdsmen. Little by little, the mountains grow lower, the valley wider, the grass disappears, the sand shows itself and one sees, between two dusky hills, the trees of the oasis of Sanju. Here there are some thousand houses, scattered on every side, and a considerable amount of ground under cultivation ; and it is easy for the traveller to procure all that he wants provided that his wants are modest.” Grenard (1904), pp. 28-30.

Notes on Sanju Pass adapted from Merzliakova (2003):

“The Pass leads to Shahidula. The uplift starts in the valley of Chibra.” “The uplift is not steep, but the terrain is rocky without any vegetation, covered with snow in some places. It is practicable for laden animals. The way down the Pass is steep into the deep and narrow valley. . . . The ancient Lekh [sic] trade route lead over the Pass toward Kargalyk between India and China through Karakorum, Chuchu-Dawan, Sasser. . . .”

Shahidulla also controlled access to a lesser-known route which headed west, eventually passing by a settlement called Mazar and then through the Shimshal Valley and over the Shimshal Pass into the upper Hunza Valley or into Wakhan. This provided an alternative to the route through Tashkurghan to northern Hunza or Wakhan, and would have been shorter and more convenient for travellers coming from the east or south.

Notes on the Shimshal Pass adapted from Merzliakova (2003):

“Height: 4735 m [or 15,535 ft]. [From another source]: 4420 m [or 14,501 ft]. “The route is not very difficult, it is practicable for ponies, but goes along the precipice in 800 m over the river. . . . The Pass is not under permanent snow. In winter the route to the Pass goes along the river Khunza, but in summer there is no road at all, because it is flooded. According to Younghusband the Pass is used in summer. [At present]:"There is an overgrazed high pasture around the Pass. The women of Shimshal graze here about 1000 yaks and 3000 sheeps and goats in July and August. . . . . There is a collection of shepherds huts in one mile from summit. There are also two small lakes.”

In the 19th and early 20th century this route was closed as it became the base for a band of notorious robbers who preyed on the caravans travelling between India and the Tarim Basin via the Karakoram Pass. See the chapter on “The Raiders’ Stronghold,” in Younghusband (1924), pp. 127-141.
          Another piece of supporting evidence that Zihe was in the region of modern Shahidulla – the Hanshu remarks that “the soil of Tzu-ho produces jade-stone.” CICA, p. 101.
          In 1868, Robert Shaw, a British trader based in India, set out on a journey from Ladakh to Yarkand. On his way, not far from Shahidulla, on the upper courses of the Karakash river, he passed a group of stone huts:

“We found out afterwards that this valley had formerly been frequented by the Chinese, who obtained jade from hence. This industry is now extinct, as the Mussulmans of Toorkistân have no taste for ornaments of this stone. A line in the Chinese ‘Thousand Character Classic,’ enumerating various productions, says, “Jade comes from the Kuen-lun Mountains” (which are those in question). I am indebted to Mr. Aston, of the India Museum, for this quotation.” Shaw (1871), p. 98, n. 1.

“To the west of Khotan was the small kingdom of Cu-gon-pan or Karghalik,³ which was important as the usual starting-point for intercourse with the western Himalaya, India, and the Pamir countries, via Sarikkol or Tashkurghan, and later by the Karakoram passes. The high mountain valleys south of the main watershed which led to these passes were, no doubt, dependent upon Cu-gon-pan rather than Khotan, so long as it existed as a separate entity ; but it is to be suspected that this kingdom was in general overshadowed by Khotan, though clearly it would derive support from its relations to its northern neighbours Yarkand and Kashgar.”

³ [Chavannes, Documents], pp. 124-5; Stein, Ancient Khotan, pp. 91-3.

Thomas (1935-1963) Part I, p. 150.

9.14. The kingdom of Yinai 依耐 [I-nai]. I have been unable to locate this tiny community with any confidence. Undoubtedly it was close to Tashkurgan as Chavannes (1905), p. 555, n. 1, notes.

“ ..., and it [Yinai] is distant by 10150 li [4,597 km] from Ch’ang-an. There are 125 households, 670 individuals with 350 persons able to bear arms. To the north-east it is a distance of 2730 li [1,136 km] to the seat of the protector general. It is a distance of 540 li [225 km] to So-chü [Yarkand] and 540 li [225 km] to Wu-lei; to the north it is a distance of 650 li [270 km] to Shu-lo [Kashgar]. To the south it adjoins Tzu-ho and the way of life of the two places is similar. There are few cereals and the state hopes to obtain [the produce of] cultivated fields from Shu-lo [Kashgar] or So-chü [Yarkand].” CICA, p. 102.

“I-nai 依耐, GSR 550f and 982h : *i̯ər/*jc̯i nəg/nậi. – I-nai is supposed to have been situated in the same area as P’u-li, see note 180.” CICA, p. 102, n. 182.

“The exact location of P’u-li and I-nai, the other small territories which Chu-chü-po or Karghalik had absorbed in T’ang times, cannot be determined at present. But the mention made in the Chien Han shu of their position north of Tzŭ-ho, and of their dependence on So-ch’ê or Yarkand for agricultural produce, suggests that they may represent the isolated hill settlements found in those little accessible valleys like Asghan-sal, Öch-b’ldir, Tong, which are drained by the middle course of the Zarafshān, or Yārkand, River. . . . ” Stein (1921), Vol. 1, p. 87.

9.15. The kingdom of 滿犁 Manli [Man-li] = Tashkurghan. Chavannes made the point that this place is identical with the 蒲犂 Puli of the Hanshu:

“In place of Manli 滿犂 (the second character being written 犁 in the edition of the Sanguozhi said to be by Bao Rentang), the Qian Hanshu (chap. XCVI, a, p. 4 b) gives the reading 蒲犂 Puli; in place of 億若 Yiruo, the Hou Hanshu (chap. CXVIII, p. 4 a) gives the reading 得若 Deruo. – The three kingdoms of Yinai, Puli and Deruo should be in the region of Tashkurgan; cf. BEFEO, III, p. 397, n. 4.” Translated from Chavannes (1905), p. 555, n. 1.

“The Tashkurghan kingdom of Sarikkol adjoined Cu-gon-pan [Karghalik] on its own eastern side, while more northwards it was connected by direct routes with Kashgar : its rulers were more in relation with the latter. On the west the Sarikkol kingdom extended up to the Wakhan Pamir, where its neighbour may be designated by the general term Tokharestan.⁴”

⁴ [Chavannes, Documents], p. 155.

Thomas (1935-1963) Part I, p. 150 and n. 4.

“P’u-li 蒲犂, GSR 102n and 519g : b’wo / b’uo - liər / liei. – Wang Hsien-ch’ien refers to the Shui-ching chu 2.12a which says that the Southern River, after having flowed eastward through P’u-li, descends northward to the country of P’i-shan; he then quotes the Shui-ching t’u-shuo ... which says : “The waters that flow separately between Yengihishar and Yarkand from here on flow eastward to unite with the Yarkand at Yarkand.” CICA, p. 101, n. 180.

“The seat of the king’s government is at the valley of P’u-li, and it is distant by 9550 li [3,973 km] from Ch’ang-an. There are 650 households, 5000 individuals with 200 persons able to bear arms. To the northeast it is a distance of 540 li [225 km] to So-chü [Yarkand] and to the north of 550 li [229 km] to Shu-lo [Kashgar]. To the south it adjoins Hsi-yeh and Tzu-ho, and to the west it is a distance of 540 li [225 km] to Wu-lei. [There are the following officials] : a noble and a commandant. The state hopes to obtain [the produce of] cultivated fields from So-chü [Yarkand]. Its race and way of life are similar to those of Tzu-ho.” CICA, pp. 101-102.

The identification of Manli/Puli with modern Tashkurghan appears certain. The directions in the Hanshu provide accurate directions and distances from both Suoju or Yarkand – 540 li or 225 km to the east – and Shule or Kashgar – 550 li or 229 km north. These distances are very accurate when checked on modern maps. There is no other settlement on these routes or in this region that it could be. 
          The dependence on food supplies from Yarkand (referred to in the Hanshu) and the size of the population (about 5,000) have not changed since the time of the Former Han, thus providing additional confirmation of its identification. See CICA, pp. 101-102. There is only one piece of evidence that does not fit. In the Hanshu’s description of Suoju [Yarkand], it is said that Puli is 740 li (308 km) to the southwest (CICA, p. 140). This must be a simple copyist’s error for the 540 li listed in the section on Puli (CICA. p. 101 – see above).
          Heading south from Tashkurgan the traveller in ancient times had two main choices. The first route (and the only one practicable for laden pack animals, led over the Neza Tash Pass and southwest through the Ak Tash (‘White Stone’) Valley along the Aksu or Oksu (‘White Water’) River – which here flows to the northeast – until one entered the Wakhan Valley proper. The second route, through Hunza and Gilgit was not passable for beasts of burden.

“Our first day’s journey [from Tashkurgan] was to the foot of the Neza Tash pass, sixteen miles [26 km] in a south-westerly direction up the Shindan stream, which flows through the defile of the same name and falls into the Sirikol river. The defile at several places is extremely narrow, and shut in by precipitous rocks and bold steep hills which rise high above it. The fallen stones and stream boulders make the road particularly bad for many miles. Willows and thorn bushes grow plentifully at the head of the defile, and the hills there lose their bold character, and become rounded and sloping. Our camp was in snow, but large patches of grass free from it were found in the vicinity sufficient for our horses, which ate it greedily, preferring it greatly to the chopped straw we carried for mixing with their grain. This grass was similar to what we found in many parts of the Pamirs, and in the Aktash valley, rich and sweet to the smell, resembling English meadow hay, and relished immensely by our animals. Judging from what we saw of it in the end of winter, it is easy to believe in its fattening properties in summer, as related by Marco Polo and other travellers, and also told us by the Wakhis. Neza Tash, meaning spear-stone, is named from a spear-like pointed rock near the place.
          On the second day we crossed the Neza Tash pass (14,920 feet)[4,548 m], leading over a high range running about north-west, and encamped at the mouth of the ravine leading down from it to the Aktash valley, travelling a distance of seventeen miles [27 km] in a general westerly direction. Snow fell in the night time, and our journey for this and the following three days, covering a total distance of seventy-eight miles [126 km], was made mainly through snow. We found plenty of grass in scattered patches and brushwood fuel at this day’s camping place. We were here joined by a party of Sirikolis with yaks and ponies carrying supplies sent by Hussan Shah to accompany us to Wakhan.” Gordon (1876), pp. 123-124.

Notes on the Neza Tash Pass adapted from Merzalikova (2003):

“Height: 4328 m [14,199 ft]. The Pass leads down Karasu stream to valley Aksu in “Sarez Pamir.” “On the west side it is very easy to ascent. Descent more difficult steep and stony.” From there one travelled past the Little Pamir Lake roughly 80 km to Langar:

“…twenty-five miles [40 km] from the lake. A deserted village and traces of cultivation were observed here, and numerous yaks and cattle were seen grazing on the opposite side of the valley. A stream of considerable size also joins at Langar, flowing from the south-east, and a road goes by it to Kunjut, over the Kura pass.” Gordon (1876), p. 129.

When one reached Sarhad (literally, ‘frontier’), approximately another 40 km west from Langar, one could cross south over the relatively easy Baroghil Pass (3,798 m; 12,460 ft) towards Mastuj which leads to either the Chitral Valley. From here there was relatively easy access to both the region around modern Jalalabad in Afghanistan, with somewhat more difficulty, through Swat to the region of ancient Gandhāra near Peshawar.
          If, on the other hand, one headed due west from Sarhad through the Wakhan Valley, one travelled through Badakshān to Bactra (Balkh) and beyond.
          The quickest route into northern India, usually open all year, was extremely dangerous and only suited for travellers on foot. From Tashkurgan one travelled just over 70 km south to the junction of the Minteke River. Heading west up this valley one reached the Kilik and Mintaka Passes which both led into upper Hunza from which one could travel over the infamous “hanging passages” to Gilgit and on, either to Kashmir, or to the Gandhāran plains.

9.16. The kingdom of Yire (or Yiruo) 億若 [I-je or -jo] = Mazar, also known as Tágh Nák and Tokanak.

This is undoubtedly the same small kingdom, Dere (or Deruo) 德若 [Te-je or –jo] that is mentioned in the Hou Hanshu (note the very similar first characters) – see Chavannes (1905), p. 555, n. 1.

億 yi; K. 957e: *i̯ək; EMC ?ik

德 de; K. 919k: *tək; EMC tək

若 re or ruo.; K. 777a: *ńi̯ak; EMC: ɲia’, also ɲɨak (commonly used to transcribe the Sanskrit sounds of: j, ya, and jña.)

The reconstructed Chinese pronunciations above show a remarkably close resemblance to the name Tágh Nák used twelve centuries later in the Tarikh-i-Rashidi, which was completed in 1547 CE.

“The Kingdom of Dere controls more than 100 households, 670 individuals, and 350 men able to bear arms. On the east, it is 3,530 li (1,468 km) to the residence of the Chief Scribe [at Lukchun]. It is 12,150 li (5,052 km) from Luoyang. It borders on Zihe (Shahidulla), and their customs are the same.” Hou Hanshu Chap. 118 (TWR, Section 7).

For some years I had suspected that the site of the tiny ‘kingdom’ of Dere mentioned in the text of the Hou Hanshu after Zihe (which I had confidently identified as Shahidulla – see note 7.13) was in the region of the hamlet marked Mazar [Ma-cha] on modern maps (approx. 77^o 0’E, 36^o 26’N – about 90 km west of Shahidulla).
          The reasons for my identification of Dere with Mazar includes the fact that Mazar, like Shahidulla, was of strategic importance and that they were not far apart and were connected by a broad, fertile river valley and, as said in the Hou Hanshu, shared a border. Settled communities near Shahidullah were, few.
          Mazar was important strategically because it controlled the junction of the route via the Shimshal Valley to Hunza; the one which led to Wakhan and, thus, to Badakshan; as well as an another important route which led directly north to Kokyar and Karghalik which then branched either northwest to Yarkand and Kashgar, or southeast to Pishan and Khotan.

“We had intended to travel by the Kokyar route to Yarkend and then to Khotan. Nazar Bai strongly advocated the alternate route across the Sanju Pass and we decided to follow his advice. The Sanju route was shorter by six days and there were fewer streams to cross.” Roerich (1931), p. 44.

The name ‘Mazar’ refers to a shrine of a Muslim saint and had no connection with the ancient name of Dere (or Yire) of the ancient Chinese texts.
          Elias’ 1895 translation of the Tarikh-i-Rashidi, contains a note that, in 1535, the author of the work, Mirza Muhammad Haidar, had been travelling north from Tibet with a few remaining men, but decided it would be too dangerous for him to head to Yarkand and that he would head for Badakshan instead, after being told (ibid., pp. 464-465):

“. . . that from a place called Tágh Nák,¹ there was a bye-path leading to the Pamirs of Badakhshán.” It appears that this Tágh Nák is identical with the modern “Mazar.”

“¹ Mirza Haidar’s spelling of this name is probably the right one. It appears on our latest maps as Tokanak, and is a spot on the Yárkand river just below Kulan-uldi, where the track to Kugiar and Yárkand leaves the valley of that river. Mirza Haidar’s party (it will be seen by the map) branched of [sic] from the direct route to Yárkand at Ak-tágh [just south of the difficult Suget Pass leading to Shahidullah], then followed down the Yárkand river past Kulan-uldi, Tágh-nák, etc., first into the district known as Ráskám and eventually on to the Pamir of Taghdumbásh. The route is an exceedingly difficult one, on account of the river crossings, and is seldom or never followed by traders or travellers.”

Elias (1895), p. 464, n. 1.

Suffice to say that, after a difficult passage, the travellers finally reached Wákhán, “which is the frontier (sar-hadd) of Badakhshán” (ibid., p. 467), and then into Badakshan proper were he and his followers spent the winter safely before heading on to Kabul and Hindustan in the spring.
          Even more than the geographic indications, what is of great interest here is the fact that the name Tágh Nák is almost perfectly rendered by the phonetic reconstructions of the Chinese name used in the Hou Hanshu some 1,400 years previously: *tək – *ńi̯ak (and see above for other closely related possibilities).
          Taken together I believe there is no doubt that the Yire/Dire of the Weilue and Hou Hanshu represents the tiny, but strategically-placed, hamlet marked on our modern maps as Mazar.

9.17. The kingdom of Yuling 榆令 [Yü-ling].

yu (GR 13100), literally ‘elm’ (Ulmus pumila L.); EMC: jǔa; K. 125g: *di̯u

ling (GR 7153) EMC: liajŋ, also liajŋ^h; K. 823a: *li̯ěng

I am unable to find any other reference to this place and have to leave it unidentified, as did Chavannes (1905), p. 555, n. 2.

9.18. The kingdom of Juandu 捐毒 [Chüan-tu], literally, ‘Tax Control’ – near modern Irkeshtam). See CICA, p. 139 and n. 358, and also the discussion in Appendix A: (b): “The Central Route”.

“All this leads one to search for the Scythian Emod and the Homodotes of that region on the upper Qizil Su-Vakhsh. Chinese sources contain additional information that makes it possible to locate the Emod and the Homodotes more precisely. Here, it seems relevant to propose a correlation between the Scythian Emod and the Juandu (ancient form: i̯wan-d’uok), the nomadic confederation of “ancient Saka tribes.” The basic source of information about Juandu is the Qian Hanshu. As for the location of this confederation, it is generally agreed that it occupied territories west and south-west of Kashgar, evidently in the basin of the Kashgarian Qizil Su and its tributaries. On the west, its lands extended up to the crest of the Congling, probably up to the Taunmurun pass, which is situated somewhat west of Irkeshtam at the headwaters of the Qizil Su-Vakhsh. Thus, the Homodotes might correspond to the Juandu; and Mount Emod, as the place where the Oxus rises, might refer to the Congling, by which in this instance the eastern areas of the Alai and Trans-Alai ranges are meant.” P’iankov (1994), pp. 40-41.

9.19. Xiuxiu 休脩 [Hsiu-hsiu ], literally, ‘Good Rest Stop,’ must have been near modern Karakavak. See CICA, pp. 138-139 and n. 355; Shiratori (1957), p. 27.
Xiu 休 (GR No. 4562) means ‘to rest,’ ‘to stop for a few moments;’ 脩 hsiu (GR No. 4579) carries among its meanings: ‘beautiful,’ ‘good,’ ‘excellent,’ ‘long,’ ‘high,’ ‘big;’ 循 xun (GR No. 4770) can mean ‘to walk’ ‘to console’, ‘comfort, or ‘good’ . So, I think I am justified to translate the name in both cases as ‘Good Rest Stop’. See also CICA, pp. 138, 139 and nn. 355-358.
          Stein (1928), Vol. II, pp. 849-851 makes a very strong case for placing Juandu in the region of Irkeshtam, about 200 km west of Kashgar, on the modern border between China and Kyrgyzstan, and Xiuxiu not too far to the west (260 li or 108 km), on the Alai Plateau. Stein places Xiuxiu/Xiuxun near modern Chat, but this is too far from Juandu, being about 155 km southwest of Irkeshtam.
          Instead, measuring it on a modern map, Xixun/Xiuxiu corresponds closely to the small modern settlement of Karakavak (Turkic for: ‘Black Poplar’ – Populus nigra L.), about half way along the fertile pasturelands of the Alai Valley at approximately 39^o 39’ N; 72^o 42’ E.
          The Alai Valley through which the Kizil Su (‘Red River’) runs, is the favoured summer pasture grounds of the local Kirghiz. This fits well with the description of Xiuxiu in the Hanshu as a very small settlement of only 1,030 pastoral nomads and adds that “in company with their stock animals they go after water and pasture.” It would have been an ideal place for caravans to exchange goods and rest and refresh their animals after the long haul from Bactra or the Tarim Basin. Interestingly, they, and the inhabitants of Juandu are both said to be originally of the “Sai race.” For detailed discussions see Yu (1998), pp. 86-90.
          Irkeshtam is near a major fork in the route from Kashgar to the west. One branch headed over the Terek Pass to Ferghana; the other led down the Alai Valley past Karakavak, Daraut-kurghān and Chat (where Stein locates Xiuxiu/Xiuxun), along the valley of the Surkhab (or Kizil-su) and on to Termez, where there was a famous crossing of the Oxus River (or Amu Darya) which led to ancient Bactra (modern Balkh).

          “Chinese and medieval Islamic sources are helpful in tracing the further history of the Komedes. Chinese sources reflecting the situation during the first century B.C.– third century A.D. locate the Xiuxun [= Xiuxiu – see CICA, p. 138, n. 355], a nomadic confederation of ancient Saka tribes similar to the Juandu, in the areas where the Classical sources place the Komedes. They lived west of the Congling, i.e., of the Taunmurun pass–and they were centred in the valley known as the Migration of the Birds (the Alai valley), whose ancient name by then was only a toponymic relic of the Amyrgians who had formerly inhabited it. There are also indications that the Xiuxun territory extended about as far south as the lands of the Komedes. It is evident that one and the same Saka tribe was called Xiuxun by the Chinese and Komedes by the Greeks.” P’iankov (1994), p. 41. Note: For the “valley where the birds fly” or valley of the “Migration of the Birds.” For more details see CICA, p. 138, n. 356.

I do not agree with some of the identifications made by P’iankov below [indicated by my notes in square brackets] but, in other respects, his geographical discussions are well worthy of consideration:

          “The reports from the Classical authors can be refined and supplemented by Chinese sources, in particular the Qian Han Shu. Jibin [Kaspiria] beyond the Hanging Passage undoubtedly corresponds to Kashmir [Refer to Appendix K. which shows that Jibin at this period refers to Kapisha–Gandhāra – not Kashmir]. However, there is no exact information on the site of Jibin, and it is possible that the source sometimes refers to a more extensive territory than modern Kashmir. Northeast of Jibin and subject to it lay the territory of Nandou [the Chitral valley]. To the east lay Wucha [= Wuhao – see TWR note 8.2 where it is located near modern Ghujak Bai (Aijie Keboyi), at the junction of the Mintaka and Tashkurgan Rivers, about 70 km south of Tashkurgan]; the road from Wucha to Jibin must have run through Nandou. In the north Nandou bordered on Xiuxun [or Xiuxiu]; in the west, on Wulei and the Great Yuezhi; and in the south on Chuo Qiang (i.e. the Tibetans of Baltistan). . . .
          It is more difficult to interpret the report that Xiuxun [=Xiuxiu] was Nandou’s [Chitral Valley’s] neighbour to the north. Nevertheless, this appears to be correct. The Xiuxun/Komedes, who had settled far up the Panj River, might have come as far south as Shughnan and Ishkashim and penetrated into the domains of the Great Yuezhi. In this area, they would have bordered on Nandou. In addition, the road to the center of Xiuxun in the Alai valley actually did lead directly north from that point.” P’iankov (1994), p. 43. [This identification of the Chuo Qiang as “the Tibetans of Baltistan” must be challenged. Tibet, as a nation did not exist at that time. One can only assume that some group with affiliations or similarities to other Qiang tribes had been reported to the south of Nandou – perhaps in the upper Swat Valley. See the discussion of the Chuo or Er Qiang (literally, the ‘Unruly’ or ‘Unsubdued’ Qiang) in note 3.3.]

9.20. The kingdom of Qin 琴 [Ch’in]. One of the meanings given for this name which is, perhaps, significant, is ‘tumulus’ or ‘tomb’ – see GR 2023, 3. Like Chavannes (1905), p. 555, n. 5, however, I am unable to find this name mentioned elsewhere.

9.21. Shule 疏勒 [Shu-lo] = Kashgar – or Qäshqär – Mallory and Mair (2000), p. 69. There can be no doubt that Shule = Kashgar. See, for example, Stein (1907), pp. 47-57; CICA, p. 141, n. 373, and the detailed discussion of the various names for the town in Bailey (1958), pp. 50-54; also note 1.58. 

Shu 疏– K. 90b: *ṣi̭o / ṣi̭wo; EMC: ʂɨə̆ or ʂɨə̆^h
le 勒 – K. 928f: *lək / lək; EMC: lək

“The pilgrim Xuanzang tells us that its name in Sanskrit was Śrīkrīrāti which means something like ‘Fortunate Hospitality’ ; the local name was transcribed in Chinese as Shule..., which provides fairly dramatic evidence for what happens when a Chinese tongue tries to articulate Indo-European clusters of sound. From the perspective of the Chinese traveller, Shule was a main emporium en route to Ferghana and Bactria; it is highly likely that General Li Guangli led his forces through it in his quest for the ‘heavenly horses’ of Ferghana. During the Han period its population was initially recorded as about a quarter of the size of Kucha, i.e. 1,510 households, 18,647 people of whom 2,000 could bear arms (but the town was booming by the 2nd century AD when the number of families was about 21,000 and it was fielding ten times the earlier number of soldiers). We are informed that there were markets with stalls in the town. It was an important garrison town in the Western Han dynasty (206 BC–AD 9), but early in the 1st century AD it fell to Khotan only to be retaken by the Chinese under General Ban Chao. Thus, the trade route west was secure at the time that Marinus of Tyre was gathering information about the Silk Road through the agents of Maës the Macedonian in the early 2nd century AD.” Mallory and Mair (2000), pp. 69-70.

From Kashgar the main route headed west to Irkeshtam (= Juandu – see note 9.18), on the present border between Xinjiang and Kygyzstan. From this point there were two main choices for major caravans, (although smaller parties could cross via Tashkurgan to the Pamirs and then descend through Wakhan into Badakhshan):
          One route headed over the Terek Pass entering the rich Ferghana Valley near modern Osh, and then through the valley towards Samarkand. The alternative was to head down the Alai Valley towards modern Dushanbe, and then on to Termez and Balkh.
          The reason the route from Kashgar through the Ferghana Valley was so popular was that the Terek Pass was open all year and there was an abundance of fodder and fuel en route:

          “But Osh [the first major town in the Ferghana valley for caravans approaching from the east] to-day, as of yore, is a centre for more than health seekers. It is, and has been for centuries, the caravan town whence camels and donkeys and horses set forth for Kashgar and China, and south to Tibet and Hindustan. Even the coming of the railway to Andizhan in the valley below has not changed the caravan centre from Osh to the railway. For the mountains about Osh are plentifully covered with grass and camping costs nothing. But most of a trip’s profits might well be squandered for the cost of hay in Andizhan.
          The presence of grass near Osh, and on the trail from Osh through the passes, is more important than the casual visitor guesses. For centuries gone this simple fact has drained far-away Europe of gold and heaped up gold in India, so that Indian princes became world-famous for their treasuries of precious metal, though India has no gold mines of her own. The reason lies in the grass of Osh and the high breathless pass of the Karakorum.
          Since time unknown a three-cornered trade has been carried on between Chinese Turkestan, India and the land now known as Russia, which leads to Europe. The exports of Chinese Turkestan are heavy–sheepskins, furs, raw cotton. The road from Kashgar to India is long and high and most of it grassless. To go and return takes for a caravan an entire season, and much of the transport must be wasted carrying fodder. But the road to Russia by Osh has grass all the way, and the hardy animals travel without cost to their owners. To Osh and back to Kashgar the trip may be made two or three times in a long summer season.
          So the heavy export from Kashgar goes regularly to Russia, locking these two countries in a natural economic alliance. But Kashgar desires the silky and light-weight cotton weaves of India ;  her luxury imports come from the south. So Chinese Turkestan carries the heavy wares to Russia, and takes her pay in gold, which she carries in turn to India. Even the grassless pass of the Karakorum is insufficient bar against valuable luxury goods.” Strong (1930), pp. 51-52.

“The abundance of grazing [in the Alai valley] was bound to be appreciated by caravans, particularly those coming from the arid valleys on the Kāshgar side. Quite as important is the fact that places permanently occupied, and hence capable of offering shelter and some local supplies, could be found on either side up to an elevation of about 9,000 feet ;  for some cultivation exists not only at Irkesh-tam, but also above it at a point known as Nōraning-sōwa on the route to Taun-murum. Thus the distance on the Alai route over which habitations were not to be found is reduced to less than 70 miles or three easy marches. The route remains open for laden animals, including camels, during eight or nine months of the year. Even in the months of December to February, when it is reported to be closed by deep snow, it would probably be made practicable in the same way as the route from Irkesh-tam across the Terek pass (12,700 feet above sea-level), provided there were sufficient traffic to tread a track through the snow and keep it clear.” Stein (1928), Vol. II, p. 848.

Stein (1928), Vol. II, pp. 849-851 makes a very strong case for placing the Juandu of the Hou Hanshu (see note 9.18) in the region of Irkeshtam, about 200 km west of Kashgar, on the modern border between China and Kyrgyzstan, and Xiuxiu not too far to the west (260 li or 108 km), on the Alai Plateau. Stein places Xiuxiu/Xiuxun near modern Chat, but this is too far from Juandu, being about 155 km southwest of Irkeshtam.
          Measuring it on a modern map, we find that Xixun/Xiuxiu coincides with the small modern settlement of Karakavak (Turkic for: ‘Black Poplar’ – Populus nigra L.), about half way along the fertile pasturelands of the Alai Valley at approximately 39^o 39’ N; 72^o 42’ E.
          The Alai Valley through which the Kizil Su (‘Red River’) runs, is the favoured summer pasture grounds of the local Kirghiz. This fits well with the description of it in the Hanshu as a very small settlement of only 1,030 pastoral nomads and adds that “in company with their stock animals they go after water and pasture.” A perfect place for caravans to exchange goods and rest and refresh their animals after the long haul from Bactra or the Tarim Basin. Interestingly, they, and the inhabitants of Juandu are both said to be originally of the “Sai race.” For detailed discussions see Yu (1998), pp. 86-90.
          Irkeshtam is near a major fork in the route from Kashgar to the west. One branch headed over the Terek Pass to Ferghana; the other led down the Alai Valley past Karakavak, Daraut-kurghān and Chat (where Stein locates Xiuxiu/Xiuxun), along the valley of the Surkhab (or Kizil-su) and on to Termez, where there was a famous crossing of the Oxus River (or Amu Darya) which led to ancient Bactra (modern Balkh).

Notes on the Terek Pass adapted from Merzliakova (2003):

The Terek Pass: “Height = 3,871 m [12,700 ft]. The Pass leads from Irkeshtam to the valley of Kush-Aba. . . . The main trade route linking Kashgar and Fergana went over the Pass. It was used in winter. An alternative summer route went through Alai valley and Passes: Taldyk [11,200 ft or 3,414 m], Archan [or Archat: 11,600 ft or 3,536 m] and Shart [14,000 ft or 4,389 m]. This road was the shortest one free of any natural obstacles. x_coord = “73.666664”, y_coord = “39.950001”.
          “The road from Irkeshtam to the Terek Pass was passable only by laden animals. No carriage could go there. . . . “Permanent snow.”

The 1911 Encyclopædia Britannica gives its height as 12,205 ft or 3,720 m and records that it is open all year round. Other sources on the internet give its height as 3,730 m or 12,238 ft.
          I am deeply indebted to Professor Merzliakova, who very kindly sent me an excellent map illustrating the old route from Irkeshtam to Osh over the Terek Pass, with the main points marked in English, and she had measured the route accurately to 156.5 km.
          The Hanshu (CICA, p. 139) states that it is 1,030 li (428 km) from Juandu northwest to Dayuan. If we take the distance of 156.5 km from Irkeshtam to Osh from the total, we get another 271.5 km. If we measure this distance westwards from Osh through the Ferghana Valley, we come to Khujand – or ‘Northern Wuyi’ – see note 24.1 for details on this key town which, apparently, had become an independent kingdom by the time of the Weilue.
          The Hanshu says that it was 1,610 li (670 km) west to the Da Yuezhi from Xiuxun. Measured on modern maps this is the distance from Balkh (Bactra) via Termez and Shahr-i Nau (40 km west of the modern Dushanbe), to Karakavak. This provides convincing evidence of the identifications of the “capital” of the Da Yuezhi as being Bactra/Balkh, as well as ancient Xiuxun being in the region of Karakavak, and Irkeshtam representing ancient Juandu.
          The construction of the major fortified Kushan town of Shahr-i Nau in the 1^st or 2^nd century CE, is shows the increasing importance of the Alai valley to Termez route at this time:

“The huge town of Shahr-i Nau (40 km west of the modern Dushanbe), which came into being in Kushan times under Vima Kadphises or Kanishka I and was surrounded by a strong defensive wall that was 7 km long and more than 8 m high, with towers every 25 m, was abandoned at the turn of the fifth century like many other settlements in the Hissar valley.” Zeimal (1966), p. 126.

The Hanshu also gives a distance of 690 li (287 km) from Dayuan southwest to the Da Yuezhi. Unfortunately, as Yu (1998), p. 59, points out, this is far too short a distance, even if the Da Yuezhi were presumed to still have their main settlement on the north bank of the Oxus.
          I must assume there was a scribal error here – the most likely one being caused by a copyist dropping out the character qian 千 for a thousand. If this was the case, the original text would have instead read 1,690 li (703 km). This is very close indeed to the 716 km I measure on modern maps from Balkh (Baktra) via Kara Tepe (just to the west of Termez), the Iron Gates, Guzar and Samarkand, to modern Kokand.

          “But during the centuries before and after the beginning of the Christian era, when Baktra was a chief emporium for the great silk trade passing from China to Persia and the Mediterranean, all geographical factors combined to direct this trade to the route which leads from Kāshgar to the Alai valley and thence down the Kizil-su or Surkh-āb towards the Oxus. Nature has favoured the use of this route, since it crosses the watershed between the Tārīm basin and the Oxus where it is lowest. Moreover, it has, in Kara-tēgin, a continuation singularly free from those physical difficulties which preclude the valleys draining the Pāmīrs farther south from serving as arteries of trade. According to the information received at Daraut-kurghān and subsequently on my way through Kara-tēgin, the route leading mainly along or near the right bank of the Kizil-su is practicable for laden camels and horses at all seasons right through as far as Āb-i-garm. From there routes equally easy lead through the Hissār hills to the Oxus north of Balkh.” Stein (1928), Vol. p. 848.

For details on the ‘Iron Gates,’ which apparently formed the northwestern frontier of Yuezhi territory, see note 5.13 under the subheading: The ‘Iron Gates’ – the northern border of Daxia.

9.22. The kingdom of Dayuan 大宛 [Ta-yüan] = Ferghana is generally accepted as being centred in the Ferghana valley. See, for example: Negmatov (1994), pp. 454-455; CICA, p. 131, n. 325; Tarn (1984), pp. 474-477, Appendix 10, “Ta-Yuan.”
          Dayuan appears as a quite powerful state in both the Shiji and the Hanshu. However, by the time of the Hou Hanshu it appears to have weakened with Xian, the king of Yarkand, conquering it just after the middle of the 1^st century CE, and then it gradually fades from view, although according to the Hou Hanshu it sent tribute and offerings to the Chinese court in 130 CE along with Kashgar and Yarkand: “In the fifth year [130 CE], Chen Pan sent his son to serve the Emperor and, along with envoys from Dayuan (Ferghana) and Suoju (Yarkand), brought tribute and offerings.” See TWR Section 21.
          I suspect that Dayuan came more under the influence of the Kangju, and the caravans from China tended to head down the Alai valley and through Kushan territory to Termez and Balkh, rather than via Ferghana and Kangju. Therefore, it became of less importance to the Chinese, and received only scant attention in the later chronicles. By the time of the Weilue, the old capital, Khojend had become a separate kingdom – see note 24.1.
          Dayuan 大宛 is sometimes written Dawan – the latter character 宛 can be read either way, though the former is preferable, and more commonly employed – see GR No. 10210 (on p. 688). Pulleyblank (1963), p. 90, writes that Dayuan was, “…the first western country which Chang Ch’ien [Zhang Qian] visited (Shih-chi 123) = Greek Τόχαροι, Τάχαροι, Latin Tochari, Sanskrit Tukhara, Tuṣara, etc., based on an original which Henning reconstructed as *Taxwār (1938). . . . ” Also see Ibid. p. 224.
          The capital of Dayuan is named Guishan 貴山 [Kuei-shan] in the Hanshu (CICA, p. 131 and n. 326). Taishan Yu (1998), pp. 69 and 92, n. 22, makes a strong case for identifying it with Khojend / Kujand / Northern Wuyi – see note 24.1.

          “On the location of the town of Guishan, the seat of the royal government of Dayuan, there are five theories. They are: a) Kokand, b) Ura-tübe, c) Akhsikath, d) Kāsān and e) Khojend. Up to now, the first three have already been discarded. But which of the last two is correct has not been determined. I believe that Khojend is better than Kāsān.” Yu (1998), p. 69.

Yu refers the reader to Kuwabara (1934-2, 3, 4) for critiques of the first three theories he mentions (see above). He then outlines his reasons for choosing Khojend. (However, note that by a slip of the pen he has, in his second point, mistakenly quoted from Shiji, ch. 123: “Dayuan is situated more than 2,000 li southwest of Dayuan. . . . ”. This should, of course, read: “Daxia is situated more than 2,000 li southwest of Dayuan. . . . ”).
          Khojend (ancient Alexandria Escharte or ‘Alexandria the Furthest’, during Soviet times, Leninabad) is very strategically placed and not only guards the entrance to the fertile Ferghana Valley, but controls the main trade route from the east which branches here either southwest towards Samarkand, or north towards Tashkent. Yu (1988), p. 69, no. 3, correctly emphasizes that:

“3. In the Hanshu, Ch. 96A, it is recorded that “[to the northwest of the state of Xiuxun] is a distance of 920 li to the state of Dayuan;” and that “[to the northwest the state of Juandu] is a distance of 1,030 li from “Dayuan.’ The “920 li” and “1,030 li” were equal to the distances from Khojend to the Alai Plateau and the upper reaches of the Kizil River respectively.”

“In the mid-second century B.C. the Yüeh-chih tribes passed southwards through Ferghana and Usrushana, and subsequently conquered Bactria. It seems likely that the far-flung, wealthy and densely populated state of Ta-yüan arose about the same time. Much detailed information about this state is given by the Chinese chronicler Szü-ma Ch’ien, who passed through Ta-yüan in the latter half of the second century B.C. The name Ta-yüan was used until the second century A.D., when it was replaced by Pu-han and Pa-han-na (fifth century A.D.) – the Chinese transcriptions of the name ‘Ferghana’. The identification of Ta-yüan with Ferghana is firmly established in historical literature.
          According to the Chinese sources, the country had many large and small towns and settlements, numbering over seventy. The population was 300,000 and the inhabitants had deep-set eyes and thick beards; they were skilled merchants and held women in high esteem. The country’s army numbered 60,000 fighting men armed with bows and spears, skilled in shooting from horseback. It was a land of highly developed agriculture; both wheat and rice were grown; there were large vineyards, wine was made and stored for dozens of years, and much mu-su (lucerne) was sown. Particularly famous were the Ferghana horses, highly prized in neighbouring lands and especially in China. They were said to ‘sweat blood’ and were considered ‘heavenly’. Emperor Wu-ti was particularly keen to have these blood-sweating horses. At one time they were worshipped in China and poets wrote odes to them.
          Ta-yüan also included Khojand and Usrushana. To the north and west it bordered on K’ang and to the south on the Yüeh-chih or Kushan possessions. Its capital was the city of Ershi, identifiable either with the ancient site of Markhamat in Andizhan District or with Khojand or Ura-Tyube. Its rulers also had a residence in the city of Yu-chen, possibly present-day Uzgen.” Negmatov (1994), pp. 454-455. Note that Pulleyblank (1963), p. 120 identifies Ershi with “Nesef, Naskhšab, present Karchi in Sogdiana. Historical grounds for this identification of the capital of Ta-yüan which the Chinese besieged and captured in 101 B.C. will be given elsewhere (Shih chi 123).”

“The powerful state of Dawan in modern Ferghana was similar to the Yuezhi in custom and style, according to the description in the History of the Han Dynasty.²⁷ Dawan was famous for its grape wine and for its horses. Grape wine might be one of the legacies of Hellenistic influence or Hellenization of the region before the Tuharan speakers took over. The name Dawan, as mentioned above, was a variation of Tuharan. The horses of Dawan were so famous that Wudi sent two major military expeditions to defeat the king and obtain horses. As for the Yuezhi, who lived further west now, their major trading item with the Han was probably no longer horses. They now controlled the resources not only of Central Asia, but also those on the fertile agricultural land of Bactria, they were not poor nomads in tatters, but rich, proud horse-riding people skilful at trade.”

27. Ban Gu, Hanshu (History of the Han Dynasty) (Beijing: Zhonghua Shuju, 1964), 96a/3894.

Liu (2001), pp. 274-275.

“1. In the Shiji, ch. 123, it is recorded that “Wusun is situated some 2,000 li northeast from Dayuan.” The “2,000 li” [832 km] was roughly equal to the distance from Khojend to the town of Chigu, the seat of the royal government of Wusun.” Yu (1998), p. 69.

As described in Appendix A, under the subheading, “(d) The “New Route of the North”– the route from Jeti-Öghüz [which I identify as Chigu, the capital of the Wusun] around the north of the Lake (still, today, the most practicable route) to the western end of Issyk-kol and then past Lake Ozero Sonkel and through the Ferghana Valley to Khojend was about 820-830 km between Jeti-Öghüz and Khojend by this route. So, it seems fair to assume that this is route mentioned in the Shiji, ch. 123, in the quote above.
          The capital of the main branch of the Wusun, referred to as the ‘seat of the Greater Kunmi’ in the Hanshu, was named Chigu 赤谷; literally, ‘Red Valley’ [chi = ‘red’ + gu = ‘valley’. Ch’ih-ku]. I believe it is now possible to locate Chigu with a fair degree of precision and certainty.
          First, there is the name of the town, Chigu which means, ‘Red Valley.’ There is, in fact, a very dramatic and famous red-coloured mountain and valley not far west of the present town of Karaköl:

“About 25 km west of Karakol, at the mouth of the Jeti-öghüz canyon is an extraordinary formation of red sandstone cliffs that has become a kind of tourism trademark for Lake Issyk-Kul.
          A village of the same name is just off the main around-the-lake road. Beyond it the earth erupts in red patches, and soon there appears a great splintered hill called Razbitoye Serdtse or Broken Heart. (Legend says two suitors spilled their blood in a fight for a beautiful woman; both died, and this rock is her broken heart.)
           Beyond this on the west side of the road is the massive wall of Jeti-Öghüz. The name means Seven Bulls, and of course there is a story here too – of seven calves growing big and strong in the valley’s rich pastures. Erosion has meant that the bulls have multiplied. They are best viewed from a ridge to the east above the road. From that same ridge you can look east into Ushchelie Drakanov, the Valley of Dragons.
           Below the wall of the Seven Bulls is one of Issyk-Kul’s surviving spas, the ageing Jeti-Öghüz Sanitorium, built in 1932 with a complex of several plain hotels, a hot pool, a restaurant and some woodland walks. . . .  
           From here you can walk up the park-like lower canyon of the Jeti-Öghüz river to popular summer picnic spots. Some five km up, the valley opens out almost flat at Dolina Svetov, the Valley of flowers. . . . There are also said to be pre-Islamic petroglyphs up here, similar to those at Cholpon-Ata.” King, et al. (1996), p. 392.

As there are no other noteworthy red-coloured rock or cliff formations around Issyk-köl, it seems probable that Jeti-öghüz is identical to the ‘Red Valley’ and ‘Red Mountain’ of the early Chinese accounts. This is confirmed by the distances contained in the Hanshu between Chigu and the town of Wensu, to the south of the mountains.
Secondly, the Hanshu (CICA: 162) gives the distance from Wensu to Chigu as 610 li (254 km). Wensu, was located in the valley of the Dashigan He (also known as the Taushkan Darya), and is usually identified with the region of modern Wushi (Uch Turfan or Urqtur pan), about 85 km west of Aksu (see note 9.6).
          My measurements range from 230 to 270 km from the Jeti-Öghüz Sanatorium, over the relatively low (4,284 m or 14,055 ft) Bedel Pass to the town of Wushi, depending which of several possible routes are chosen through the mountains. (Checked on Russian Military 1:100,000 topographic maps, 1970 and 1973, and the U.S. Defence Mapping Agency Aerospace Center map ONC, Sheet F6, 1:1,000,000, revised Feb. 1981). This confirms the likelihood that Chigu was located somewhere in the upper Jeti-Öghüz Valley.

9.23. The kingdom of Anxi 安息 [An-hsi] = Parthia. GSR 146a and 925a : *·ân/*·ân – si̯ək/si̯ək “M. *·an-si̯ək = *Arśak (Parthia).”

“Kingsmill (Journ. R. Asiatic Soc., vol. XIV, 1882, p. 81) appears to have been the first to see the transcription of the term Arsak in the Chinese words An-xi as designating the Parthian sovereigns called Arascides. Hirth has confirmed and rendered this identification scientific by showing that the consonant n can correspond to an r, and that, on the other hand, the word xi was previously pronounced sak; thus Anxi is the strict equivalent of Ar-sak (cf. Hirth, Syrisch-chinesiche Beziehungen, p. 438, n. 2.” Translated from Chavannes (1907), p. 177, note 1. See also: CICA, p. 115, n. 267; Pulleyblank (1963), pp. 77, 221; Yu (1998), p. 173.

          “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.

          “Wu Di, around 115 or 105 BC, sent another delegation to the Parthians with the task of establishing direct contacts. The Chinese ambassadors were received by King Mithridates II, who was to establish diplomatic relations with Rome in 92 BC. These circumstances lead one to suppose that direct Sino-Parthian trade contacts originated with the journey of this delegation and that they laid the foundations for trade between east and west that later on was to grow so extensively.” Baumer (2000), p. 36.

“The Kingdom of Parthia emerged as a result of the socio-economic crisis affecting the Seleucid state in the mid-third century B.C. In the course of that crisis the governors of the extreme eastern satrapies – Diodotus in Bactria and Andragoras in Parthia – seceded from the Seleucid kingdom. While in Bactria an independent Graeco-Bactrian kingdom came into being, the situation in Parthia was much troubled by incursions of nomads belonging to the confederation of the Parni who had occupied land along the edges of the agricultural oases from the Caspian Sea to the River Tedzhen. The Parni, with Arsaces at their head, invaded Parthia. In the ensuing struggle Andragoras was killed and control of the country passed to the nomadic aristocracy of the Parni headed by Arsaces. The Parni soon seized Hyrcania (a region on the southern and south-eastern Caspian seaboard), and this boosted the economic and military potential of the infant state.
          . . . . The long and variable struggle between Rome and Parthia over this [control of Armenia] ended with an agreement in A.D. 63 that the brother of the Parthian king Vologases should be proclaimed King of Armenia and crowned in Rome by the Roman emperor Nero. This agreement was extremely important since it led to a long period of peace on the frontier between Rome and Parthia interrupted by only minor disputes.
          Peace was next broken in A.D. 114 when the Roman emperor Trajan began his carefully prepared campaign against Parthia. The Roman army marched to the Persian Gulf, and the Roman fleet sailed down the Tigris. The success of the Romans owed much to the bitter conflicts within Parthian society between rival claimants to the Arsacid throne, and to the revolts that had broken out in Elymais and Persia [i.e. Persis]. But at the height of the Roman success the situation radically changed. In all the Parthian territories conquered by the Romans, insurrections broke out, triggered off by the introduction of the Roman system of provincial administration, which strictly controlled towns, taxes and requisitions, and by the discontent of the petty rulers who had recognized Rome’s authority and had subsequently been stripped of the remnants of their independence. The rival representatives of the Arsacid house united against the invader and in A.D. 117 the Romans were compelled to abandon all their conquests in Parthia. Although the Roman frontier was peaceful again, Parthia was still not secure and faced severe complications on its northern and eastern borders. It appears that Hyrcania finally achieved independence; the separatist trends of other regions of the state became more marked; and Parthia’s northern provinces suffered incursions from the Alani. The emergence and growth of the powerful Kushan Empire created a permanent danger in the East. Exhausted by internecine wars and constant difficulties with Rome, Parthia sought to reduce tension in the East to a minimum. The stumbling block in relations between Rome and Parthia, however, remained Armenia, where in the time of Vologases III there was a bitter clash in A.D. 161-63. The northern flank of the Roman defence collapsed and Parthian troops invaded Syria. Rome, alarmed that there might be a general uprising against its rule in the East, mustered its strength to stabilize the situation and then to launch a counter-offensive. The peace treaty concluded at the end of the war was harsh for the Parthians, since the whole of Mesopotamia as far as the River Khabur was ceded to Rome. Even harsher for Parthia were the consequences of the war which broke out in A.D. 195. The Romans found that Vologases IV (A.D. 191-207), who seemed to have invaded eastern Iran, had at the same time to oppose the large-scale revolts that had broken out in Media and Persia. The Roman military expedition dealt a heavy blow to Parthia: the richest parts of the country were devastated and some 100,000 inhabitants were taken to Syria and sold into slavery. The last war between Rome and Parthia began in A.D. 216. The conflict between Vologases V and Artabanus V, the two pretenders to the Parthian throne, made the conditions ripe for Roman intervention. The Romans, under their emperor Caracalla, invaded Parthian territory and laid waste much of Mesopotamia and part of Media. In the summer of A.D. 217, Artabanus V, who had mustered sizeable forces, started to wage a resolute campaign against the Romans. Caracalla fell at the hands of conspirators and Macrinus became emperor. After a decisive battle at Nisibis the Romans had to sue for peace. However, this was the Parthians’ last success. The ruler of Persia [i.e. Persis], Ardashir, united with a number of other local rulers to raise a revolt against the Arsacids. In 223 he defeated and killed Vologases V. A few years later Artabanus V was defeated and killed at the battle of Homizdagan, and the entire territory of the Arsacids soon passed into the hands of the new dynasty of the Sasanians.” Koshelenko and Pilipko (1994), pp. 131, 135-136.

“The origin of the Parthians is traced to the Parni or Aparni, a Scytho-Sakian nomad tribe identified as part of the Daha. Their main occupation had been pillaging along the by-paths leading to the Siberian gold fields and the caravan routes connecting Syria with the oasis settlements of Sogdiana and Media. Following these major revolts [of Bactria and Parthia against the Seleucids] a spate of minor rebellions ensued, and in the general turmoil many Greek and Macedonian settlers returned to western Iran or to their native homelands. Only Bactria was strong enough to retain its Greek traditions and Graeco-Macedonian settlers and develop eventually a new Graeco-Bactrian tradition of its own. Although Seleucid power declined, hellenization continued. Even in regions under Parthian control, Greek replaced Aramaic as the language of the élite and the court in all of the urban centres.
          By 250 BC the Parthians under the Arsacids had developed into a powerful force which gradually gained control of the northern two-thirds of Iran. Eventually even Bactria lost its nominal independence and the centre of control shifted from Persis in the south-west to Parthia in the north-east – a shift that profoundly influenced the development of Persian civilization.” Bowles (1977), p. 107.

“About 18 km. north-west of Ashkabad two towns, Old Nisa and New Nisa, were excavated. It is possible that Nisa was really the first capital, or at least the home town of Arsaces I [r. circa 238 BC]. The size and splendour of the excavated halls attest to the wealth of the Parthian rulers. It is not known whether the original Parthian name of Old Nisa was Mithradatkirt, or whether this was a renaming of the old site by Mithradates I. . . . ” Frye (1966), p. 210.

9.24. The kingdom of Tiaozhi 條支 [T’iao-chih] = Characene and Susiana. Almost all recent writers agree that this territory of Tiaozhi – the furthest point reached by the Chinese envoy Gan Ying in 97 CE, which bordered on the ‘Western Sea’ and was a dependency of the Parthians at the time – must refer to the region near the head of the Persian Gulf. It was first mentioned in the Shiji and again in the Hanshu where it presumably referred to the Seleucid territories in the lower Tigris-Euphrates region.
          I tend to agree, on the whole with Chavannes’ notes on the identification of this kingdom, although I would extend it to include Susa and the surrounding region, known as Susiana :

“Tiaozhi appears to me to correspond to the Arab kingdom of Characene which was founded between 130 and 127 BCE in Mesene, at the mouths of the Tigris. Mesene is called Dest Misau in a fragment of Ibn Qutaybah [828-829], and Amru, quoted by [Joseph] Assemani [1687-1768], simply calls Desht the country of Desht Misan; this name of “Desht”, is the Persian word desht which signifies “plain” [or “desert”]. Perhaps it is this word which is hidden in the Chinese transcription of Tiaozhi 條支. The Characenes were subject to the Parthians during the reign of Trajan (98-117 CE), for we see this emperor waging war against the Parthians and the Arabs at the same time. The Chinese historian tells us in fact several lines later on that Tiaozhi (Desht Misan) was subject to Parthia.” Translated from Chavannes (1907), p. 176, n. 3.

There have been a number of suggestions for the derivation of the name Tiaozhi but the question remains unresolved. Some of the more plausible suggestions by recent writers are:

“For SC and HS, T’iao-chih refers to the Seleucid Empire. By the later Han period, with HHS and HHC, we can hardly accept this identification, for the Seleucid Empire had long ceased to exist. Though occasional references seem to be echoes of earlier information, we must look for a more contemporary country.
          We are inclined to follow the view of Chavannes and Shiratori in particular that T’iao-chih must be Characene (or Mesene), with capital Charax, in the delta of the Tigris and Euphrates. This requires the concomitant identification of the Western Sea (sometimes the Great Sea) which it overlooks as the Persian Gulf leading to the Indian Ocean.” Leslie and Gardiner (1996), p. 260.

Pulleyblank (1999), pp. 73-74 makes a case for the phonetic derivation of Tiaozhi from Seleukia:

          “Tiáo- in Tiáozhī 條支 EMC dεw tɕia̭, with yōu 攸 EMC juw as phonetic, shows the same pattern [as in the note on Wuyishanli in note 9.25 by Pulleyblank]. Moreover, the Middle Chinese palatal initial of the second syllable can be shown to go back to an earlier *k. Compare the Amoy colloquial pronunciation [ki]. The name was reconstructed as *δeuɦ kēɦ in Pulleyblank (1962: 101). At the time, unfortunately, I did not recognize the obvious similarity to (Se)leukia but accepted the identification of Herrmann and Fujita with Taoke, a place near Bushire on the Persian Gulf. The connection with Seleukia was, however, later made by Wu Chi-yu (1977). Wu took my 1962 form, *δeuɦ kēɦ at face value and made the connection with (Se)leukia through a change of [l] to [δ] in Middle Iranian. If *δ is replaced by *l- in the reconstruction, this roundabout route is unnecessary.
          It should be further noted that the xiéshēng series of tiáo 條 includes words with initial s- like xiū 修 EMC suw and xiăo 篠 EMC sεw’. The rules governing initial clusters in Old Chinese in such cases are still uncertain but these xiéshēng connections clearly support the possibility of reconstructing tiáo 條 with initial *sl- rather than simply *l-. In view of the perfect geographical as well as phonetic fit, we can feel very confident that, as Wu argued, Tiáozhī stood for Seleukia and there is no need to look for a connection with Antioch, whether the more famous city of that name in Syria or, as claimed by Shiratori, a more shadowy Antiochia in Persis at the head of the Persian gulf.” Pulleyblank (1999), p. 74.                                                

David Graf (1996), p. 203, presents an intriguing alternative:

“It seems far more likely that T’iao-chih is simply an attempt to transcribe the word “Tigris” (Assyrian-Babylonian Idiglat; Old Persian Tigra). Support for this view can be found in the rendering of the Ganges river valley as Huang-chih in CHS [Hanshu] (ch. 188/32ab), suggesting that the character chih in the name T’iao-chih was pronounced ga in the Han period. T’iao-chih can then be considered as the Chinese transcription for the Persian form of the name for the Tigris. Just as the Chinese name for the Ganges designated the kingdom on the Indian seacoast, so T’iao-chih represents the kingdom on the Tigris near the coasts of the Persian Gulf. In fact, in the later Chinese account of Persia by Ma Tuan-lin (Po-ssū ch. 339/6), the region south of Su-li on the banks of the Ta-ho-shui (i.e. Seleucia on the Tigris) is equated with the territory of ancient T’iao-chih. All of this territory may have earlier been under the administration of Charax Spasinou, the central city of the Lower Tigris.”

It is generally agreed that Tiaozhi must have included the lands at the head of the Persian Gulf, and that its main city and port was Charax Spasinou, the ‘capital’ of the semi-autonomous territory of Characene.
          On the other hand, I believe Sōma (1978), pp. 1-26, is probably correct when he argues that the big city, described as 40 li (or over 16 km) around in our text, could not possibly be Charax Spasinou, as we know from classical sources that it was much smaller than this. It does, however, admirably fit with what we know of Susa – the second largest city in the region (after Seleucia/Ctesiphon). Susa used Charax Spasinou as its port. We know Susa retained its importance throughout the Roman period and retained a considerable degree of autonomy from the Parthian capital at Ctesiphon, though the details are anything but clear.

“The region of Susiana is distinguished from Elymais by Strabo XVI.1.8, 17, 18 and Pliny, NH VI.135-136. For the absorption of Susiana and its capital by the kingdom of Elymais, see U. KAHRSTEDT, Artabanos III, 40-47 and G. L. RIDER, Suse, 426-430, who dates the end of Parthian Power in Susa to c. A.D. 45 and places a mint of Elymais in the city by c. A.D. 75. Possibly at this time Susa became the capital of Elymais.” Raschke (1976), p. 817, n. 721.

Pliny the Elder seems to indicate that Charax Spasinou was (at least at his period) considered to be under Arab control at this time:

“A particularly inaccessible part of it [the coast at the head of the Persian Gulf] is called Characene, from Charax, a town of Arabia that marks the frontier of these kingdoms [Elymais and Farsistan]. . . . ” Pliny NH (b), p. 136. (VI. Xxxi).

“After Petra the country as far as Charax was inhabited by the Omani, with the once famous towns of Abaesamis and Soractia, founded by Samiramis ; but now it is a desert. Then there is a town on the bank of the Pasitigris named Forat, subject to the king of the Characeni; this is resorted to by people from Petra, who made the journey from there to Charax, a distance of 12 miles [17.6 km] by water, using the tide. But those travelling by water from the kingdom of Parthia come to the village of Teredon below the confluence of the Euphrates and the Tigris ; the left bank of the river is occupied by the Chaldeans and the right bank by the Scenitae.” Pliny NH (b), pp. 145-146. (VI. Xxxii).

Although I have only tentatively identified Tiaozhi as Characene and Susiana, it clearly referred to the region about the mouth of the Tigris River, at the head of the Persian Gulf. See also: CICA, p. 113, n. 253.
          Other indications that the identification is correct can be found in the Hanshu which points out that Tiaozhi “is warm and damp, and the fields are sown with rice ; there are bird’s eggs as large as [water] jars.” CICA p. 113. It is well known that rice was cultivated in this region during ancient times. See for example the short but excellent: “A note on rice cultivation in Mesopotamia and Susiana” by Potts (1991-2).
          Also, ostriches were still living in the wild at this time throughout the dryer parts of the Middle East, including Parthia:

“The Jewish historian Flavius Josephus recorded that during a day’s hunt King Herod the Great killed 40 different kinds of animals, from lions to wild boars, from gazelles to ostriches.          
          Ostriches were common then and their range immense: from today’s Morocco to Egypt, from southernmost Africa through the Middle East to Mesopotamia (modern-day Iraq), ancient Persia (Iran) and Arabia. They were avidly hunted. Their plumes were in great demand for the majestic fans of the pharaohs and as panache for the headdresses of nobles and knights. Ostrich eggshells, set in gold or silver, became the precious goblets of the rich. The Egyptians, noting the perfect balance and symmetry of the ostrich feather, revered it as a symbol of truth and justice. . .
          The Middle Eastern, or Syrian, ostrich – smallest of the six ostrich races – was hunted mercilessly with cars and guns, a thrilling “sport” that quickly eradicated the great birds. The last ostrich of the Middle East drowned in a flash flood in southern Jordan in 1966.” Bruemmer (1997).

9.25. The kingdom of Wuyi 烏弋 [Wu-i] = Kandahar (and Arachosia, of which it was the capital).
         The full form of this name, Wuyishanli 烏弋(山離) [Wu-i-shan-li], is found in the Hou Hanshu but it is also found in the same shortened form in the Weilue – see CWR note 8.5. It is probably a transliteration of Alexandria or Alexandropolis and, according to most authorities, stands here for Kandahar, the chief city of the province of Arachosia. See the discussions in Daffinà (1982), p 319; Chavannes (1905): 555, n. 6; Pelliot (1959), p. 29; Pulleyblank (1963), pp. 116, 128; and CICA, p. 112, n. 250.

          “A good example of a Han transcription that shows both Old Chinese *r > Middle Chinese *l and Modern Chinese l-, corresponding to foreign –r-, and Old Chinese *l > Middle Chinese *j- corresponding to foreign –l- is the name Wūyìshānlí 烏弋山離 EMC ?ɔ jik ʂəɨn lia̯, long accepted as equivalent to the name Alexandria (not the great metropolis in Egypt but one of the other cities by this name founded by Alexander in present Afghanistan). Greek –r- is correctly represented by Middle and Modern Chinese l-, while Greek l- is represented by the initial consonant of the second syllable. The graph yi 弋 (a Type B syllable) is phonetic in dài 代 EMC dəj^h (a Type A syllable) showing the typical pattern of the Han dynasty *l-.” Pulleyblank (1999), pp. 73-74.

Alexandria-in-Arachosia or Alexandropolis, from which the modern name of Kandahar was most probably derived, was probably founded in the spring of 329 BCE. 

          “Arachosia, Persian HARAUVATISH or HARAHVATISH, in ancient times a province of the Achaemenid, Seleucid, and Parthian empires. It occupied southern Afghanistan and was bounded on the south by Gedrosia (Baluchistan). The capital city, Alexandria-of-the Arachosians, was founded by Alexander the Great and is usually identified with Qandahār. Arachosia was famous for its ivory and elephants.” NEB, I, p. 471.

A number of scholars have suggested that Wuyi could refer to both Arachosia and Drangiana (modern Seistan), to the west. I agree with them as the descriptions in both the Hanshu and the Hou Hanshu mention only one state (Wuyishanli) between Jibin to the northeast and Tiaozhi and Lijian to the west.
          Arachosia and Drangiana were closely associated from early times, forming two adjoining provinces of the Achaemenid empire, and later, in the time of Aśoka and the Seleucid king, Antiochus II, “it is clear that the two kingdoms were then contiguous with a frontier west of Kandahar. . . . ” Sherwin-White and Kuhrt, p. 80. Michael Witzil’s excellent article, “The Home of the Aryans” provides further confirmation of the combination of these two satrapies later on:

“The extent of Arachosia has shifted over time (cf. also GNOLI 1980: 36), see the distinction in the O.P. inscriptions and Greek sources which distinguish between Arachosia and Drangiana. . . . Strabo [writing circa 23 CE] 11.560, on the contrary, has both Drangiana and Arachosia within one satrapy; the Avestan *Drangiana/Sistan and Arachosia indeed share the same x^v dialect, as is clear from the very name, Harax^vaitī, and not the usual Avestan –huu. . . . ” Witzel (2000), p. 26, n. 59. See also, Yu (1998), pp. 168-169.

No discussion of Wuyi(shanli) here identified as Kandahar would, however, be complete without giving space to some of the dissenting opinions and other suggestions:

“Alexandropolis near Sacastene can only be Kandahar and this settles the question of its foundation, for a place called Alexandropolis cannot have been a city founded by Alexander (p. 7), and in fact there is no record that he founded Kandahar; Alexandropolis at best was a military colony which (possibly quite correctly) attributed its settlement to him. I need not consider conflicting opinions about the name Kandahar, whether it be derived from Alexander (Iskandr), from Gandhāra (for which there seems no historical reason), or (most probably) from the Parthian Gondophares (Gundofarr). . . . ” Tarn (1984), p. 471 (which see for more details).

          “CANDAHAR, n.p. Kandahār. The application of this name is now exclusively to (a) the well-known city of Western Afghanistan which is the object of so much political interest. But by the Ar. geographers of the 9^th to 11^th centuries the name is applied to (b) the country about Peshāwar, as the equivalent of the ancient Indian Gandhāra, and the Gandharitis of Strabo. Some think it was transferred to (a) in consequence of a migration of the people of Gandhāra carrying with them the begging-pot of the Buddha, believed by Sir H. Rawlinson to be identical with a large sacred vessel of stone preserved in a mosque of Candahar. Others think that Candahar may represent Alexandropolis in Arachosia. We find a third representation of the name (c) in Ibn Batuta, as well as in earlier and later writers, to a former port on the east shore of the Gulf of Cambay, Ghandhar in Broach District.” Yule and Burnell (1886), p. 154.

Yu (1998), p. 169, proposes that Wuyishanli probably referred to Alexandria Prophthasia (Farāh), south of Herat. However, I feel that his arguments could better be applied to Kandahar.
          Kandahar has for long been the larger and more prosperous region, and famous as one of the few places where elephants could be bred and raised successfully in captivity (meaning this was the coveted production area of the ancient equivalent of army tanks), I am inclined to accept, along with most other authorities, that Wuyishanli here referred primarily to Arachosia, a kingdom centred on Kandahar and probably including Drangiana.

“Arachosia (Old Persian harahuvati, corresponding to Sanskrit sarasvati ‘rich in rivers’) was the well-named land of present southern Afghanistan, the valley of the Upper Helmand (Avestan Ha’tumant ‘rich in dams’) and the tributaries where the Thamani (Herodotus III.93, 117) lived.., the people of Arachosia must have been settled agriculturalists from an early time in this fertile land comparable to Bactria in the north. Similar to Bactria in the north, Arachosia was the centre of Achaemenid rule over neighbouring tribes to the south and east and Darius was fortunate to have a loyal satrap who, after a number of battles with the rebels sent against him from the west, was able to consolidate the rule of the new king.
          The lower course of the Helmand river and the Hamun lake was occupied by the Zrangi (Old Persian Z(a)ra(n)ka, with local zB for Old Persian d-), which name has been explained as ‘sea land’ by many scholars, unsuccessfully, I believe. The name survived into Islamic times as Zarang, the capital of the country. The Hamun lake area played an important role in Zoroastrian tradition and as the homeland of the hero Rustam. By geography and history it has been connected with Arachosia and the upper Helmand rather than with Fars province or the west. The invasion of Saka tribes in the second and first centuries before our era undoubtedly changed the population for their name was applied to the land which has held to this day, Seistan. In pre-Achaemenid times as today it is a land where the steppe and sown are intermingled and nomads are on all sides of the lake which is large in winter while almost vanishing in the late summer.” Frye (1963), pp. 71-72.

Pavel Lurje of the Department of Ancient Near East, St. Petersburg Branch of the Institute of Oriental Studies, Russian Academy of Sciences, kindly wrote on 18 May 2002, informing me that the form Ha’tumant in the above quotation is incorrect: “Avestic Haetumant-, (Greek Etymandres), not Ha’tumant. The Kandahar inscription of Ashoka seems to be not in Aramaic language, but some local Iranian or Indian language in Aramaic script. . . . ”
          The rich agricultural lands along the Helmund River centred on Kandahar have, since ancient times, formed an essential link and staging post on the shortest and easiest route between southern Persia and India. From northern Parthia one travelled to Herat and then on to Kandahar. From southern Parthia the route travelled through Persepolis to Kandahar. The Persian name for the city was Kapisakaiš but it was renamed Alexandria after Alexander’s visit in 329 BCE and the present name, Kandahar, is ultimately derived from Alexandria.
          Kandahar and the associated district of Arachosia were of great importance because of its fertility in an otherwise barren region and its strategic position in controlling the main southern route between India and Persia. It also provided the rare combination of hot and steamy conditions and abundant fodder needed for the breeding and raising of herds of war elephants along the Helmand River to the west of Kandahar.

“Suitable sites for elephant-parks are rare in both Syria and Afghanistan. And the Seleucids’ war-elephants, like the Ghaznavids,’ were sinews of war in the literal sense. In discussing Ashoka’s inscription at Qandahar, I have recalled that the first Seleucus ceded all his provinces west of Qandahar and south of the Hindu Kush to Chandragupta Maurya in exchange for 500 of the Indian emperor’s elephants; and the price in terms of ceded territory turned out not to be excessive from Seleucus’s point of view. Those 500 elephants were trumps. They won him his victory over his rival Antigonus ‘One-Eye.’ In fact, they won him his empire. No wonder that he and his successors took trouble to provide their elephants with congenial accommodation.” Toynbee (1961), p. 72.  

Here is the account of the Hanshu on the Kingdom of Wuyishanli:

“The seat of the king’s government is at . . . [the text seems to be defective here] and it is distant by 12,200 li [5,073 km] from Ch’ang-an. It is not subject to the protector general . . . [There are many] households, individuals and persons able to bear arms, and it is a large state. To the north-east it is a distance of 60 days’ journey to the seat of the protector general. It adjoins Chi-pin in the east, P’u-t’ao in the north, and Li-kan and T’iao-chih in the west ; after travelling for some one hundred days one then reaches T’iao-chih. . . . [a passage on Tiaozhi (T’iao-chih) is unaccountably inserted here in the Chinese text – in the middle of the section on Wuyi]
          The land of Wu-i is very hot ; it is covered in vegetation and flat. For matters such as grass, trees, stock-animals, the five field crops, fruit, vegetables, food and drink, buildings, market-stalls, coinage, weapons, gold and pearls, [conditions] are identical with those of Chi-pin, but there are antelope, lion and rhinoceros.
          The way of life is such that a serious view is taken of arbitrary murder. The obverse of the coins shows only a human head with a rider on horseback on the reverse. Their staves are embellished with gold and silver.
          [The state] is cut off and remote and Han envoys reach it only rarely. Proceeding by the Southern Route from the Yü-men and the Yang barriers, and travelling south through Shan-shan one reaches Wu-i-shan-li, which is the extreme point of the Southern Route; and turning north and then proceeding eastward one arrives at An-hsi.” CICA, pp. 112-115.

This description of the region of Kandahar (Arachosia) seems valid. Certainly it is hot and fertile. The mention of pearls is presumably a reference to an active trade in pearls, though it is not clear whether they originated in the Persian Gulf or in Indian waters. (Pearls are also mentioned in the account on Jibin – which was even further inland).
          The reference to the “serious view taken of arbitrary murder” is perhaps reflective of the influence of Buddhism. The early arrival of Buddhist ideals in the region is confirmed by a bilingual inscription (in both Greek and Aramaic) by Ashoka (died c. 238 BCE) which advocates a vegetarian diet and the avoidance of, or at least restraint in, the hunting and killing of animals.
          I have been unable to find references to coins from Kandahar with the portrait of the king on one side and a horseman on the other, but the coinage of the area is poorly known. Certainly silver coins that would fit this description were issued by Eucratides I (c. 170-145 BCE) – probably from Taxila, and by the late 1st century CE Kushan monarch, who according to Cribb (Sims-Williams and Cribb (1995/96), pp. 97-99) now known as Wima Tak[to] or Soter Megas [but see Mac Dowall (2002), p. 163], apparently issued from Balkh and/or Kapisha. So, it is quite possible that similar coins were circulating in, or were issued from, Kandahar during the time of the informant for the Hanshu.
          The reference to lions and rhinoceroses in the Hanshu is accurate. Lions were found in southeastern Iran until recent times. Rhinoceros, though now extinct in the region, were still being hunted in the Afridi hills northeast of Kandahar in the sixteenth century:

“c. 1555. – “We came to the city of Purshawar, and having thus fortunately passed the Kotal we reached the town of Joshāya. On the Kotal we saw rhinoceroses, the size of a small elephant.” – Sidi ‘Ali, in J. As. Ser. 1. tom. ix. 201.” Yule and Burnell (1886), p. 700.

“1519. – “After sending on the army towards the river (the Indus), I myself set off for Sawâti, which they likewise call Karnak-Khaneh (kark-khâna, ‘the rhinoceros-haunt’), to hunt the rhinoceros. We started many rhinoceroses, but as the country abounds in brushwood, we could not get at them. A she rhinoceros, that had whelps, came out, and fled along the plain; many arrows were shot at her, but... she gained cover. We set fire to the brushwood, but the rhinoceros was not to be found. We got sight of another, that, having been scorched in the fire, was lamed and unable to run. We killed it, and everyone cut off a bit as a trophy of the chase.” – Baber, 253.” Quoted from Baber in: Yule and Burnell (1886), p. 762. Also see Chandra (1977), p. 9.

The so-called “antelope” is discussed at some length in CICA, p. 114, n. 262. Unfortunately, the authors seem to have made a mistake with the Chinese and rendered it tiaoba 挑拔 [t’iao-pa] – first character GSR 1145o, instead of taoba 桃拔 [t’ao-pa] – first character GSR 1145u. In any case, the important thing is that they point out that it was said to be another name for the fuba, a specimen of which was sent as a present to Emperor Zhang in 87 CE by the king of Parthia.
          The fuba 符拔 [fu-pa] was, according to Von Gutschmid to be identified with the βούβαλς antelope [the Bubal antelope], and this identification was noted by Chavannes (1906), p. 232, n. 1. The Bubal antelope (Alcephalus boselaphus)is a purely African species and is most unlikely to have been sent to China by Parthia and even less likely to have been sent by the Kushans in 88 CE, as the biography of Ban Chao notes. It also sports conspicuous horns in both sexes, contradicting the information given in the text.
          The text of the Hou Hanshu says the fuba was similar to a lin 鳞, but without a horn. Unfortunately, except for the fact that the fuba does not have a horn, this doesn’t help us very much. The lin is usually referred to as a female Chinese unicorn (commonly drawn with a scaly body). The Digital Dictionary of East Asian Literary Terms describes it as “an auspicious, mythical East Asian horse-like animal,” but Williams (1909), p. 527, notes that this character “seems to have also been intended for a large elk.” GR Vol. II, p. 716, No. 3631 says (translated from the French): “(Myth.) A fabulous animal represented either in the form of a stag endowed with a long tail and with one or two horns, or in the form of a mythical horse.”
          As this animal was sent to China, it was definitely a real, and not an imaginary animal. And, as the text specifies that it did not have any horns, it most probably was the common and very graceful Persian or Goitered gazelle, Gazella subgutterosa, found from Asia Minor to Mongolia. The female has only rudimentary or no horns, the only member of this family showing this feature. It is called the Goitered gazelle because the larynx of the male swells in the breeding season.

The city of Kandahar itself is located on the Tarnak River, a tributary of the Helmand and has, since early times, been a major centre of trade. The ancient city, which was sacked by Nadir Shah in 1738, was located about 6 km west of the present city and had been the capital of the whole region since Achaemenid times. Its pomegranates, melons and grapes are still deservedly famous throughout Afghanistan. There is also an abandoned gold mine about 3 km north of the city that may have had some importance in antiquity.
          Apparently, the Kushans did not conquer this region and it remained under a series of semi-autonomous or autonomous Parthian or “Indo-Parthian” rulers until the arrival of the Sasanians about 240 CE. See: Simonetta (1978), p. 186.

          “Alexander moved on swiftly to Kandahar [from the west] to Kandahar, where he founded [or, rather, renamed!] another town, Alexandria in Arachosia (the part of Baluchistan which lies behind the Quetta Hills). Again, this has been a strategic site throughout Afghan history, and Kandahar has been occupied from then till now. In the old citadel, a temple to the deified Alexander has been discovered, along with an inscription in Greek and Aramaic by the Indian emperor Ashoka who lived a few decades after Alexander. (This is a place where the Indian and West Asian culture zones have always overlapped.) In the bazaar in the old town, the hakims (traditional doctors) claim descent from the doctors who went with Alexander – descendants of the physicians Philip and Critobulos. They still practise the Yunnani (Greek) herbal medicine which can be found right across Pakistan and North India.” Wood (1997), p. 136.

From Kandahar there were two main routes to the Indian subcontinent. One route ran northeast upstream along the Tarnak River past modern Ghazni to Kabul and, from there, through either the Khyber Pass (1,067 m or 3,500 ft) into the Gandhāran plains of northern Pakistan, or one of the several other passes – said to be more frequented in ancient times than the Khyber – such as the Nawa (Nawar or Spinasuka Pass) through Swat to the region of the ancient capital of Puṣkalāvatī or Chārsaddā some 18 miles (29 km) northeast of Peshawar. See Verma (1978), pp. 52-56 and nn.
          The other ran southeast, via the relatively easy Khojak (2,707 m or 8,881 ft) and Bolan (1,798 m or 5,899 ft) passes, to the lower Indus River, in what is now southern Pakistan. All three passes are open all year.

“1552 – Those who go from Persia, from the kingdom of Horaçam (Khoraṣan), from Bohára, and all the Western Regions, travel to the city which the natives corruptly call Candar, instead of Scandar, the name by which the Persians call Alexander. . . . ” – Barros, IV, vi. 1.” Yule and Burnell (1886), p. 154.

“A Persian army was reported to be massing for an attack on Herat in western Afghanistan [in 1836]. Encouraged, partly financed and probably officered by Russians, it looked as if the hour of reckoning might already be nigh. Herat could not be expected to hold out for long and, as Burnes knew only too well from his strategic studies, once Herat fell the easiest of approaches to India, that via Kandahar, would lie right open.” Keay (1977), pp. 142-143.

“Zaranj-Kandahar-Ghazni Route: The other important route from Zaranj [in the Helmand basin in Seistan] was the famous Kandahar route which still plays an important role in the political and economic system of modern Afghanistan and Pakistan. The main stages on this route were Bust Banjaway and Ghazni. This route went through the Garmsil region, that is, Zanbuk, Sanizan and Haruri on the left bank of the river Khwash. Between Haruri and Bust it crossed a desert. Between Bust and Banjaway of Rukhaj it crossed two tributaries of the Helmand namely the Arghandab and the Tarnak and reached Ghazni which was situated to the northeast of Banjaway.” Verma (1978), pp. 46-47.

9.26. Wuyi is also called Paizhi 排持 [P’ai-chih]. This may have been transcribed from a foreign name, but the reconstructed ancient pronunciations do not resemble any name known from other sources and neither of the characters in it are commonly used to transcribe foreign sounds. According to K. 579x + 961p, 排持 should have been pronounced something like: *b’εr-di̯əg. See CICA, p. 112, n. 250. However Yu (1998), p. 168, suggests:

“In the Houhanshu, ch. 88, it is recorded: “[The state of Wuyishanli, which] covers several thousand li 里, has changed its name into Paite 排特.” Similarly, the “Xirongzhuan” of the Weilue records that “Wuyi’s other name is Paichi” (持 is noted mistakenly as 持 in the original text). “Pai-te” [buəi-dək] can be read as a short transcription of “Prophthasia”.

I am not sure from which edition of the Hou Hanshu Yu got the form of the name: 排特 Paite, although it is known in other sources:

“This information [i.e. 排持 as an alternate name for Wuyi(shanli)] is found again in the Hou Hanshu (chap. CXVIII, p. 4b). The edition of the Sanguozhi, said to be of the Northern Song gives the reading Paite 排特 (critical notes of Qianlong).” Translated from Chavannes (1905), p. 555, n. 7.

The Shanghai edition of 1888 which Chavannes used; the Shanghai: Zhonghua shuju, Min guo 25 [1936] edition, and the Shanghai Zhonghua Shuju 1965 edition (reprint 1973), all have the same form: 排持 Paichi, which is also found in the Weilue. This form does not suggest the reconstruction of the name “Prophthasia.”
          Another suggestion has been that 排持 Paichi might have been a “slip of the pen” for 塞持 Saichi (literally: “Governed by the Sai/Saka”). I consider this unlikely, as the characters for Pai 排 and Sai 塞 not only look so different, but the reconstructed ancient pronunciations are quite dissimilar, and were unlikely to have been confused.
          The name could be a literal translation. The character pai 排 (GR 8449) can have the meaning of ‘shield’ or ‘platoon(s), and chi 持 (GR 1872) means ‘to take in hand,’ ‘to govern,’ ‘maintain,’ or ‘guard with firmness.’ Therefore, an explanation of the name Paichi, is that it could have represented something like, “Guarded or governed by platoons,” or “Military Post,” being similar to the English military term, “Cantonment”.
          Finally, see Chavannes (1907), p. 176, n. 1, for a very speculative association with the feichi bu 緋持布 [fei-ch’ih pu], or ‘scarlet chi cloth,’ mentioned as a product of Da Qin in the Weilue.

Section 10 – Previous Misconceptions

Translator: The reader should be aware that there is a real possibility of some missing text in this section which might explain some of the geographical absurdities it contains. As Chinese geographical knowledge kept expanding, legends traditionally located in the west were continually moved further to the west when proof of them was not found. It is also possible that one or more bamboo strips were mislaid (bamboo strips were still in use for centuries after the traditional invention of paper in 105 CE). If paper was used by Yu Huan one or more pages could have been lost.

10.1. “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.”

I believe this statement originates from a deduction made on the basis of the various accounts of the position of the so-called “Weak Water” and “Queen Mother (or ‘Spirit Mother’) of the West.” It can be seen from the texts that, as the Chinese knowledge of the true geography of Parthia and other western countries advanced, these legendary places were thought to be further west. A similar situation occurs in regard to the place where it was thought that “the sun sets.” See also note 10.3.

10.2. “In earlier times it was also mistakenly thought that the Ruo Shui 弱水 [Jo-shui. Literally, ‘Weak River’], was west of Tiaozhi (Characene and Susiana). Now it is (thought to be) west of Da Qin (Roman territory).”

“The “Weak Water”, Jo-shui, and the “Queen Mother of the West”, Hsi Wang mu, are situated in the extreme West of the world by a number of ancient Chinese texts. Both were specifically Chinese concepts and so they cannot have been known to “the elders of An-hsi”. Their presence in this text is presumably to be explained by a Chinese having asked an elderly man from An-hsi about these - to us mythological - geographical features and the elder replying that they might be there, this evasive answer being turned again into a more positive statement, which the author of this section of HS 96 considered worthwhile to insert into his account.
To the Chinese, the Weak Water continued to be a reality ; the Wei lüeh, apud Chavannes (1905), p. 556, corrects the Han shu, saying that the Weak Water was not to the west of T’iao-chih..., but to the West of Ta Ch’in, i.e. the Roman Orient. Chavannes adds that in T’ang times the Weak Water was identified with the river Yasin, referring to his Documents sur les T’ou-kiue occidentaux (Petersburg, 1903), pp. 153 and 313.” CICA: 114, n. 260.

It is of interest to note that there was, apparently, another Ruo Shui [Jo-shui], to the north of China:

“Jo River 弱水. (Kansu) TSFYCY 45.14a-b. North of the desert above Shensi.” Rogers (1968), p. 320. See also ibid. p. 140 – this is where the king of Dai [Tai], Sheyijian [She-i-chien] led when he was defeated in 376 CE.

10.3. The Hou Hanshu says:

“It is said that to the west of this kingdom is Ruoshui (the ‘Weak River’) and Liusha (the ‘Shifting Sands’) which are close to the place where Xiwangmu (‘Spirit-Mother of the West’) lives, and which go almost as far as the place where the sun sets.”

 The Hanshu says:

“Leaving Tiaozhi (Characene and Susiana), if you head west for more than two hundred days, you approach the place where the sun sets.” This does not agree with the books of today. [The reason is that] the Han envoys under the first [Han] dynasty all returned after reaching Wuyi (Arachosia and Drangiana),²⁰ and none of them went as far as Tiaozhi (Characene and Susiana).

Chavannes points out that that the Hou Hanshu misquotes the Hanshu here:

“The actual wording of the Qian Hanshu (chap. XCVI, a, p. 6a) is a little different: “On leaving Tiaozhi, when on goes by water about a hundred days, one reaches the spot where the sun goes down”.” Translated from Chavannes (1907), p. 185, n. 4.

Another translation of this same passage (from CICA, p. 115) throws extra light on this quote, making it clear that this was very much a second-hand story passed on by the elders of Parthia:

“It is said : “The elders of An-hsi have learnt by hearsay that in T’iao-chih there is the Weak Water and the Queen Mother of the West ; but they have all the same never seen them. If you travel by water westward from T’iao-chih for some hundred days you draw near the place where the sun sets.”

The Weishu, “written previous to A.D. 572, and embracing the period A.D. 386-556, ch. 102; Hsi-yü-chuan. With one exception, this account is identical with one contained in the Pei-shih, a revised history of this same dynasty”, says:

“Although in that country [i.e. Da Qin] sun and moon, and the constellations, are quite the same as in China, former historians say that going a hundred li west of T’iao-chih you come to the place where the sun sets ; this is far from being true.”

See also TWR, Section 12 and note 12.19.