Within this area and on the irrigated alluvial fan, at a distance of about 12 km from the mountains, the city was built on a site (the Bālā Ḥeṣār of today) which was probably coextensive with a slight rise in the plain and perhaps adjacent to an old arm of the river.


BALḴ, city and province in northern Afghanistan.

i. Geography.

ii. From the Arab conquest to the Mongols.

iii. From the Mongols to modern times.

iv. Modern town.

v. Modern province.

vi. Monuments.

(For the ancient history of Balḵ, see BACTRIA.)

i. Geography

The city of Bactra, later Balḵ, owed its importance to its position at the crossing of major routes: the west-east route along the foot of the Khorasan and Hindu Kush mountains from Iran to Central Asia and China, and the route by left bank tributaries of the Oxus and passes through the mountains of central Afghanistan to northwestern India. The river of Balḵ (Balḵāb) gives easy access by the valley of its tributary the Dara-ye Ṣūf and the Qarā Kotal pass to the Bāmīān basin and thence to Kabul. This route has the advantage of being the westernmost of the roads over the Hindu Kush and thus the shortest for travelers from the west, as well as one of the easiest. Its existence must have been the main reason why a great city arose in the area where the Balḵāb debouches into the plain.

Within this area and on the irrigated alluvial fan, at a distance of about 12 km from the mountains, the city was built on a site (the Bālā Ḥeṣār of today) which was probably coextensive with a slight rise in the plain and perhaps adjacent to an old arm of the river. This is only a supposition, because adequate archeological exploration has not yet been carried out. In any case, the site subsequently grew higher through the gradual accumulation of the debris left by successive human occupants.



A. Foucher, La vieille route de l’Inde de Bactres à Taxila, MDAFA 1, 2 vols., Paris, 1942-47.

(X. de Planhol)

ii. History from the Arab Conquest to the Mongols

Information on the process of the Arab conquest of Balḵ is somewhat vague. According to Balāḏorī (Fotūḥ, p. 408), Aḥnaf b. Qays raided Balḵ and Ṭoḵārestān in ʿAbd-Allāh b. ʿĀmer b. Korayzδs governorship of Khorasan during the caliphate of ʿOṯmān (32/653), but further attempts at controlling the city were not possible until Moʿāwīa had restored a measure of peace and stability to the troubled Arab empire. In 42/662-63 ʿAbd-Allāh b. ʿĀmer nominated Qays b. Hayṯam over Khorasan, who in turn sent ʿAbd-al-Raḥmān b. Samora into Khorasan and Sīstān, conquering Balḵ and, allegedly, Kabul. But the people of Balḵ renounced their peace agreement with the Arabs, and in 51/671, Rabīʿ b. Zīād had to reappear at Balḵ; it is clear that no firm or enduring Arab control over the city was ever established in the early Omayyad period. It was, however, during these raids under ʿOṯmān and Moʿāwīa that the great Buddhist shrine of Nowbahār, situated in the rabaż (suburb) of the city according to the classical Arabic geographers, was despoiled and destroyed, although it long remained a sacred site; the northern Hephthalite prince Ṭarḵān Nīzak (q.v.) went to pray there and to derive blessing when he rebelled in Gūzgān and lower Ṭoḵārestān against the Arab governor Qotayba b. Moslem Bāhelī (q.v.) in 90/709 (Ṭabarī, I, p. 1205), necessitating Qotayba’s dispatching 12,000 men to Balḵ.

From its strenuous opposition to the Arabs on various occasions, and the latter’s vengeful reprisals, Balḵ is described as being largely ruinous in the mid-Omayyad period, so that the Arabs built for themselves a new military encampment two farsaḵs away, called Barūqān, where what was normally a comparatively small Arab garrison (at least in comparison with that of Marv) was installed, until in 107/725, after an outbreak of feuding amongst the Arab troops at Barūqān (represented in such sources as Ṭabarī, perhaps misleadingly, as a tribal clash of Qays and Yaman), the governor Asad b. ʿAbd-Allāh Qasrī (q.v.) restored Balḵ on its former site, employing as his agent for this Barmak, the somewhat shadowy father of the early ʿAbbasid minister Ḵāled Barmakī (Ṭabarī, II, pp. 1490-91); Barūqān now drops out of mention. A few years later, Asad temporarily transferred the capital of Khorasan from Marv to Balḵ, giving the latter city an access of prosperity.

The last Omayyad governor in Khorasan, Naṣr b. Sayyār Kenānī (q.v.), built Balḵ up into a significant military base. In 116/734, according to Ṭabarī, II, pp. 1566-67, he had there an army of 10,000 men, composed of the Arab tribesmen of Khorasan and also probably of Syrian forces, which he used against the rebel Ḥāreṯ b. Sorayj. During the ʿAbbasid daʿwa in Khorasan led by Abū Moslem, Balḵ was strongly defended for Naṣr and the Omayyads by Zīād b. ʿAbd-Allāh Qošayrī. Abū Moslem sent against him and against other loyal government forces of Ṭoḵārestān, including the local Iranian princes, his lieutenant Abū Dāwūd Ḵāled b. Ebrāhīm Bakrī. Possession of the city oscillated between the Omayyad defenders and Abū Moslem’s commanders Abū Dāwūd and ʿOṯmān b. Kermānī, until it was secured for the revolutionaries at the third attempt (130/747-48). See for this early period of the consolidation of Arab control and of islamization, Markwart, Ērānšahr, index s.v.; J. Wellhausen, The Arab Kingdom and Its Fall, Eng. tr., Calcutta, 1927, index s.v.; P. Schwarz, “Bemerkungen zu den arabischen Nachrichten über Balkh,” in Oriental Studies in Honour of Cursetji Erachji Pavry, London, 1933, pp. 434-43; M. A. Shaban, The ʿAbbāsid Revolution, Cambridge, 1970, index s.v.

Little is heard of Balḵ during the early ʿAbbasid period, but it was a base for Hārūn al-Rašīd’s commander ʿAlī b. ʿĪsā b. Māhān in the operations against the rebel Rāfeʿ b. Layṯ b. Naṣr b. Sayyār, and the fact that Balḵ suffered from a violent earthquake in 203/818-19 is mentioned. Soon afterwards, it came within the vast governorship of the East held by the Taherid family from the ʿAbbasid caliphs. But with the seat of the Taherids’ power at Nīšāpūr, 500 miles to the west, Balḵ seems to have been left, according to the general pattern of Taherid overlordship in the east, to local princes. These were from the Abu Dawudid or Banijurid family, most probably of Iranian stock. Dāwūd b. ʿAbbās b. Hāšem b. Banījūr was governor in Balḵ from 233/847-48 onwards, in succession to his father, and was the builder of the village and castle of Nowšād or Nowšār near Balḵ. He was still there when the Saffarid Yaʿqūb b. Layṯ destroyed Nowšād and temporarily captured Balḵ before going on to Kabul (in 256/870 according to Gardīzī, ed. Nazim, p. 11, in 257/871 according to Ebn al-Aṯīr, ed. Beirut, VII, p. 247). Dāwūd fled to the Samanids in Samarqand, returning to Balḵ and retaking it soon afterwards and dying there in 259/873. His kinsman (nephew?) Abū Dāwūd Moḥammad b. Aḥmad ruled in Balḵ from 260/874, and was involved in the complex power struggle between rival condottieri for control of Khorasan after the Taherids’ loss of Nīšāpūr to the Saffarids in 259/873. Abū Dāwūd was immediately besieged in Balḵ by 5,000 troops under Abū Ḥafṣ Yaʿmar b. Šarkab, and then soon afterwards was again attacked by Abū Ḥafṣ’s brother Abū Ṭalḥa Manṣūr after the latter had been expelled from Nīšāpūr (see Ebn al-Aṯīr, VII, pp. 296, 300, giving the data for the second attack as 265/878-79 or 266/879-80). This Abū Dāwūd also controlled Andarāb and Panjhīr in Badaḵšān, where he minted coins from the local silver, and was still ruling in Balḵ in 285/898 or 286/899, when the Saffarid ʿAmr b. Layṯ summoned him and the other local potentates of northern Khorasan and Transoxania to obedience. ʿAmr’s plans of extending his control to these regions were of course speedily dashed by his defeat near Balḵ, after fortifying that city with a moat and rampart, at the hands of the Samanid Esmāʿīl b. Aḥmad (q.v.) (287/900). See for these events, Gardīzī, ed. Nazim, pp. 11-19; Naršaḵī, Tārīḵ-eBoḵārā, tr. Frye, pp. 87ff.; Markwart, Ērānšahr, pp. 301-02; Barthold, Turkestan3, pp. 77-78, 224-25; C. E. Bosworth “Banīdjūrids,” in EI2, Suppl.

The late 3rd/9th- and 4th/10th-century geographers expatiate with enthusiasm on the amenities and the flourishing state of Balḵ at that time, calling it Omm al-belād “the greatest of the cities of Khorasan” from the populousness of the region (Yaʿqūbī, Boldān, p. 287; tr. Wiet, p. 100) and Balḵ al-bahīya “Splendid Balḵ” (cf. Moqaddasī, p. 302); it was equal in size to Marv and Herat, and according to Moqaddasī again, rivaled Bukhara in size. It stood on a river, the Balḵāb (or as Ebn Ḥawqal, ed. Kramers, p. 448, names it, the Dah-ās “[turning] ten mills”), which came down from the Hindu Kush but which did not, in Islamic times, actually reach the Oxus, petering out in the sands. The Balḵāb divided at the city into twelve branches to irrigate the surrounding countryside; among the products of this agricultural area are mentioned citrons, oranges, water-lilies, and grapes, in sufficient quantities for export, whilst the nearby open steppes were used for rearing an excellent strain of Bactrian camels. Outside these domains, however, lay salt marshes and deserts. The ruins of Nowbahār were apparently still impressive, and the author of the Ḥodūd al-ʿālam (372/982) mentions wall-paintings and other wonders there; by his time, construction of the original building was attributed to the Sasanian emperors. Balḵ had the usual tripartite plan of an inner citadel (qohandez), an inner city (madīna or šahrestān), and an outer city or suburb (rabaż or bīrūn). There were mud brick walls (mud brick being also the normal material for the houses of Balḵ) around both the madīna and the rabaż, with a ditch beyond the outer wall; in earlier times, there had been a wall twelve farsaḵs long, with twelve gates, enclosing both the city and adjacent villages, as a protection from nomads and other marauders, but by the 3rd/9th century this no longer existed. In the next century, the rabaż seems to have had seven gates and the madīna four, the latter a number characteristic of a number of other Persian cities. The seven rabaż gates included the Bāb Hendovān, attesting the presence nearby of a colony of Indian traders, and the Bāb al-Yahūd, showing the existence of a Jewish community also (both these groups were still of significance in Balḵ at the end of the nineteenth century, despite the complete eclipse of Balḵ as a trading center; see C. E. Yate, Northern Afghanistan or Letters from the Afghan Boundary Commission, Edinburgh and London, 1888, p. 256). The Ḥodūd al-ʿālam, indeed, describes Balḵ as the emporium (bārkaḏa) of India. The markets were mainly situated in the madīna, where stood the main Friday mosque; according to Yaʿqūbī, there were forty-seven mosques with menbars in the moderate-sized towns of the Balḵ region. See for the information of the Arab geographers, Le Strange, Lands, pp. 420-22, to which should be added the Persian Ḥodūd al-ʿālam, tr. Minorsky, p. 108; Barthold, “Istoriko-geograficheskiĭ obzor Irana,” in his Sochineniya VII, Moscow, 1971, pp. 41-44, 47-49, tr. S. Soucek, Historical-Geographical Survey of Iran, Princeton, 1983, pp. 25-26; Barthold, Turkestan3, pp. 76-79.

This commercial and economic prosperity was reflected in Balḵ’s role in nurturing ulema (ʿolamāʾ ) and other scholars, whom Samʿānī, Ansāb, ed. Hyderabad, II, pp. 303-35, describes as innumerable. In fact, these included such figures as the early Sufi Abū Esḥāq Ebrāhīm b. Adham (d. 161/778), who stemmed from Balḵ before he went westwards to Syria (cf. Ebn al-Aṯīr, VI, p. 56), the geographer and astronomer Abū Zayd Aḥmad Balḵī (d. 322/934), and the Muʿtazilite philosopher Abu’l-Qāsem ʿAbd-Allāh Balḵī (d. 319/931). Scholars like these, and especially traditionists, theologians, and religious lawyers, were surveyed and classified in the local histories and ṭabaqāt books on the notable men of Balḵ, one of which, a Ketāb fażāʾel Balḵ, was apparently written by Abū Zayd Balḵī himself (see Bibliography).

Thus under the Samanids, Balḵ was especially flourishing, although the warfare of rival military factions in the last decades of the emirate affected it on certain occasions. The Ḥājeb Fāʾeq Ḵāṣṣa was governor there during the ascendancy of the Sīmjūrīs in the 370s/980s, and in 381/991 he was besieged in Balḵ by Abu’l-Ḥasan Ṭāher b. Fażl, of the Muhtajid family of Čaḡānīān; the latter was, however, killed, and Fāʾeq was confirmed in the governorship of Balḵ and Termeḏ in 382/992 by the Qarakhanid invader of Transoxania, Boḡra Khan Hārūn. When Maḥmūd of Ḡazna and the Qarakhanids partitioned the Samanid empire between themselves, the lands north of the Oxus fell to the former, although the Qarakhanids for long coveted also northern Khorasan. Hence in 396/1006 the Ilig Khan Naṣr sent his general Čaḡritigin or Jaʿfartigin into Ṭoḵārestān. The population of Balḵ resisted fiercely, and the city was plundered before Čaḡritigin was forced to retreat to Termeḏ on Maḥmūd’s return from India, the Ilig’s ambitions here being finally quelled by Maḥmūd’s overwhelming victory at Katar, 12 miles from Balḵ, in 398/1008. It was during Čaḡritigin’s occupations of Balḵ that the Bāzār-e ʿĀšeqān or “Lovers’ market” built there by the sultan himself was destroyed; Maḥmūd later censured the people for resisting the enemy and so causing the loss of his lucrative property. We have other information about Ghaznavid constructions in the city, including mention of a fine garden laid out by Maḥmūd, whose upkeep was a burden on the local people until the sultan grudgingly transferred the onus to the local Jewish community. We also learn that the raʾīs or civic head of Balḵ, Abū Esḥāq Moḥammad b. Ḥosayn, supplied money to Maḥmūd for his campaigns when the flow of taxation revenues from Khorasan dried up after the exactions of the vizier Esfarāʾenī; doubtless these subventions were made by the Balḵ merchant community as a whole. See on this period, Barthold, Turkestan3, pp. 253-54, 259, 272, 276, 280, 288-89, 291; M. Nāẓim, The Life and Times of Sulṭān Maḥmūd of Ghazna, Cambridge, 1931, pp. 31, 39, 42-43, 48-50, 154, 166; Bosworth, Ghaznavids, index s.v. Balkh.

Although threatened by the incursions of the Saljuqs during the latter years of Masʿūd of Ḡazna’s reign, Balḵ did not, like Nīšāpūr and Marv, fall immediately into the Turkmen’s hands, even after Masʿūd’s disastrous defeat at Dandānqān in 431/1040. There seems nevertheless to have been a disaffected element in the city’s population who probably wished to reach an accommodation with the Saljuqs, for Masʿūd’s vizier reported the presence of large numbers of “corrupt persons, evil-wishers and malevolently-inclined people” there, and at one point it was in fact briefly occupied and plundered by the Turkmen. But Balḵ was a key point in the Ghaznavid defense system for northern Afghanistan, protecting the capital Ḡazna itself, and resistance there was organized against Čaḡri Beg Dāwūd by the local ṣāḥeb-e barīd Abu’l-Ḥasan Aḥmad ʿAnbarī, called Amīrak Bayhaqī.

Despite his efforts, Balḵ seems to have passed definitely to the Saljuqs early in Mawdūd of Ḡazna’s reign, for in 435/1043-44 Čaḡri Beg’s son Alp Arslān, based on Balḵ, fended off a Ghaznavid attempt to reconquer northern Afghanistan. Alp Arslān was now formally invested with the governorship of all northeastern Khorasan, including Balḵ and Ṭoḵārestān, as far as the Oxus headwaters, the day-to-day running of administration here falling to Čaḡri Beg’s vizier Abū ʿAlī Šāḏān; and on his accession in 451/1059 the sultan Ebrāhīm b. Masʿūd of Ḡazna made a peace treaty with Čaḡri Beg at last recognizing Saljuq control of these regions. During Alp Arslān’s reign, the governor here was the sultan’s son Ayāz, who was momentarily ejected from Balḵ in 456/1072 by the Qarakhanids when his father died and was soon afterwards succeeded by the new Saljuq sultan’s other brother Tekiš (466/1073-74). The allocation of this northeastern corner of the Saljuq empire to princes of the ruling family not infrequently led ambitious princes into rebellion against the sultan in distant western Iran. Thus in 490/1097 Berk-Yaruq (Barkīāroq) had to spend seven months at Balḵ suppressing the outbreak of a Saljuq claimant, Moḥammad b. Solaymān b. Čaḡri Beg, called Amīr-e Amīrān, whose father had at one time been governor of Balḵ and who had received military help from the Ghaznavids.

During the first half of the 6th/12th century, Balḵ came within the extensive sultanate of the east held by Sanjar. The city remained flourishing, not least intellectually; a Neẓāmīya madrasa had been built there, either by the great vizier Neẓām-al-Molk himself or with his encouragement, and in the later part of the century, the poet Anwarī (d. 585/1189-90?) spent his last decades there. Towards the end of Sanjar’s reign, however, Saljuq power in Khorasan was challenged by external rivals such as the Ḵᵛārazmšāhs and the Ghurids, and by the internal malcontent element of the Oghuz nomads who pastured their flocks in the upper Oxus region and who chafed under the heavy hand of Saljuq taxation and officialdom, including that of Sanjar’s governor in Balḵ, ʿEmād-al-Dīn Qamāč. In 547/1152 the Ghurid ʿAlāʾ-al-Dīn Ḥosayn occupied Balḵ for a while with Oghuz help. In the next year the Oghuz offered conciliatory terms to Qamāč, which he shortsightedly rejected; he attacked them outside Balḵ, but was routed by them and had to flee to Sanjar’s capital at Marv, leaving Balḵ to be plundered by the Oghuz, with considerable destruction of public buildings. The Oghuz now installed themselves at Balḵ, offering their obedience to Sanjar’s nephew, the Qarakhanid Maḥmūd Khan, and held the city for several years. Later, suzerainty over it passed to the Qarā Ḵetāy of Transoxania, until in 594/1198 the Ghurid Bahāʾ-al-Dīn Sām b. Moḥammad of Bāmīān occupied it when its Turkish governor, a vassal of the Qarā Ḵetāy, had died, and incorporated it briefly into the Ghurid empire. Yet within a decade, Balḵ and Termeḏ passed to the Ghurids’ rival, the Ḵᵛārazmšāh ʿAlāʾ-al-Dīn Moḥammad, who seized it in 602/1205-06 and appointed as governor there a Turkish commander, Čaḡri or Jaʿfar.

In summer of 617/1220 the Mongols first appeared at Balḵ. It seems that the city surrendered peacefully to the incomers, but in spring 618/1221 Jengiz Khan himself arrived there, and Balḵ was subjected to a frightful sacking, conceivably after a revolt of the populace against the Mongol garrison. Whether Balḵ did indeed have a population of 200,000 before the Mongol massacres, which last involved a large part of the populace, is unconfirmed, but certainly the agricultural and commercial activities on the eve of the invasion described by Yāqūt (Moʿjam al-boldān I, p. 713), when Balḵ supplied produce to Khorasan and Ḵᵛārazm, was dealt a severe blow, from which the city did not recover till Timurid times. See, for the Saljuq period and after, the standard sources for Saljuq and Mongol history (Bondārī, Rāvandī, Ebn al-Aṯīr, Jovaynī, etc.); of secondary literature are Barthold, Turkestan; Bosworth and Boyle, in Camb. Hist. Iran, V; and Bosworth, Later Ghaznavids.



This is substantially given in the article. It should be noted that Balḵ, like other cities of Khorasan, seems to have had a lively genre of local histories and works on the excellencies and merits of the city, many of these being biographical in approach. Virtually all of these are apparently lost, but material from several of them was used by the Šayḵ-al-Eslām Abū Bakr ʿAbd-Allāh b. ʿOmar Balḵī for his Ketāb fażāʾel Balḵ (610/1214), of which a Persian translation by ʿAbd-Allāh b. Moḥammad Ḥosaynī was made at Balḵ in 676/1278 (ed. ʿAbd-al-Ḥayy Ḥabībī, Fażāʾel-e Balḵ, Tehran, 1350 Š./1971; cf. Storey, I, pp. 1296-97).

(C. E. Bosworth)


iii. From the Mongols to Modern Times

The medieval and modern history of Balḵ, which has been filled with breaks and recoveries, offers a prime opportunity for a new approach to the study of the post-Mongol period in arid Central Asia. The political history and ethnic evolution of the Balḵ oasis have essentially shared with Mā Warāʾ al-Nahr (Transoxania) frontier and population movements that can be traced until the middle of the nineteenth century. The final integration of Balḵ into the Afghan domain was then hastened by the Anglo-Russian accord of 1873, which established the Amu Darya as the boundary between the zones of influence of the two empires.

Balḵ belonged to the Mongol Empire after its surrender to Jengiz Khan in 617/1220 and, with Bactria, formed the southern part of what became the khanate of Chaghatay. The destruction resulting from the Mongol conquests was very severe at Balḵ, and the city remained in ruins for more than a century (Ebn Baṭṭūṭa, p. 299); for some time, however, hypotheses about the long-term consequences of this destruction have been debatable, for Balḵ did recover some prosperity in the course of the eighth/fourteenth century. Subsequently it was a valued appanage in the territorial system of the different Jengizid ruling houses until the twelfth/eighteenth century. Thus a long period of conflicts began, on the background of the disputes over the succession and revolving around real or nominal control of these appanages. In this way the Mongol princes of the khanate of Chaghatay vied with one another, whether directly or indirectly through the intermediary of local dynasts, like the Kart rulers (maleks) of Herat, who were involved on several occasions.

The territorial changes brought about by the formation of Tīmūr’s (Tamerlane’s) empire initiated long periods of stability, which, however, began with the devastation caused by the Balḵ campaign in 771/1369. The city was included successively in Tīmūr’s, Šāhroḵ’s, and Oloḡ Beg’s possessions, then, after more than twenty years of internal struggle, belonged to Sultan Ḥosayn Bāyqarā, who ruled southern Turkestan between 872/1468 and 911/1506 and established his brother Bāyqarā at Balḵ. The former died in combat against the Uzbek, who were ultimately victorious, after the short reigns of two of his sons, and established themselves permanently as far as the Hindu Kush. The period of Tīmūr and his descendants, the Timurids, was recognized from the beginning as favorable to the development of urban civilization (Clavijo, pp. 141-48).

The subsequent Uzbek period lasted three centuries, the longest in the post-Mongol history of Balḵ. The establishment of the Uzbeks was reflected in major construction activity at Balḵ (Mukhtarov, pp. 17-97), which became the third or fourth most important city of their empire. The written reports on Shaibanid and Janid Balḵ are quite numerous, and many contemporary authors came from this center of power or lived there (Akhmedov, pp. 3-14; Mukhtarov, pp. 8-16). The position of Balḵ in relation to Bukhara improved in the eleventh/seventeenth century: It became the second most important city in the Bukharan domain and the capital of the heirs to the Janid throne. This important position, however, attracted invaders and led to redefinition of international frontiers in the region.

From the west the Safavids installed themselves in Khorasan; the Uzbeks recaptured Balḵ from them in 922/1516. From the southeast came the Mughals; their occupation of Balḵ, from 1051/1641 to 1057/1647, under the command from 1056/1646 of Awrangzēb, who then became emperor, represented a last attempt to restore the old domain of Bābor. The episode of Nāder Shah a hundred years later was equally transitory. On the other hand, the birth of Dorrānī Afghanistan turned the Amu Darya into a frontier, where first atalïks, then Mangit amirs of Bukhara struggled with the Sadōzay and Moḥammadzay rulers of Afghanistan for a century. In 1164/1751 Aḥmad Shah incorporated Balḵ into a political entity unconnected with Mā Warāʾ al-Nahr for the first time since the Mongol conquest. In 1257/1841 the Afghans permanently recaptured the city from the Bukharans, who had reestablished themselves there in 1241/1826 (Ivanov, pp. 107ff.). The suzerainty of the latter did not come to an end, however, until Bukhara itself lost its sovereignty in 1285/1868. Balḵ, which had shrunk to a large village during the twelfth/eighteenth century, finally lost its status as an administrative center in 1282/1866, in favor of Mazār-e Šarīf. Reduced to 500 households by the beginning of the twentieth century, the population of Balḵ has since increased but is still only one tenth that of its neighbor.

The conditions of recent decline at Balḵ show that standard explanations of the frequent periods of crisis in the history of the Central Asian oases must often be revised. At Balḵ, both the population and the number of canals have diminished since the twelfth/eighteenth century, the latter dropping from eighteen to eleven. These facts, along with the importance of nomads around Balḵ and the supposed drying up of the Balḵāb, could all be taken as evidence of the evolution of a typical post-Mongol Central Asian city. “It is only within the last 750 years that Balkh has fallen on evil days” (Toynbee, p. 95). The decline of Balḵ in favor of Mazār-e Šarīf must be viewed aside from the question of the so-called tomb of ʿAlī, within the framework of solidarities resulting from the irrigation networks: The two cities form part of the same oasis and depend on the same supply line through the canals from the Balḵāb. It thus seems more significant for the history of the development of the oasis to emphasize the migration of urban population from there to Mazār-e Šarīf, via Taḵta Pol, rather than contrast the modern village with the large ancient city. In fact, with about 30,000 inhabitants in 1295/1878 and 100,000 today, Mazār-e Šarīf demonstrates the capacity of the irrigation system in the oasis, where present population density is between 30 and 100 inhabitants per square kilometer (Tübinger Atlas, A VIII 3), to continue to support the largest city in Afghan Turkestan, as it has done in the past.

The cultural character of the Balḵ oasis today reflects the ethnic and political shifts in its post-Mongol history. The Turkish populations, especially the Uzbeks but also the Turkmen, predominate over the Tajiks. There are also colonies of Pashtun, though fewer than in the Maymana and Tāšqorḡān oases; one Jewish community; and some Arabic-speaking villages (Tübinger Atlas, A VIII 16). The linguistic picture is differentiated, including an important component of the Fārsī of Balḵ, but it corroborates the profound Uzbekization of the region (Tübinger Atlas, A VIII 11).

See also balḵāb.



An initial attempt to make use of the Arabic geographers to follow the continuous course of the history of Balḵ was that of V. V. Barthold, Istoriko-geograficheskiĭ obzor Irana I: Baktriya, Balkh i Tokharistan, Sochineniya 7, Moscow, 1971, pp. 39-59.

In this article Barthold throws doubt on the assertion that in antiquity the Balḵāb flowed into the Oxus. For a history of Balḵ on the eve of the Mongol conquest see Abū Bakr Wāʿeẓ Balḵī, Fażāʾel-e Balḵ, Pers. tr. ʿAbd-Allāh Moḥammad Ḥosaynī-Balḵī, ed. ʿA. Ḥabībī, Tehran, 1351 Š./1972.

The situation of Balḵ after the Mongol conquest is described by Ebn Baṭṭūṭa (Paris) II, pp. 299.

The ruling dynasties of the khanate of Chaghatay have been reconstructed from the Chinese and Islamic lists by L. Hambis, “Le chapitre VII du Yuan Che,” T’oung Pao 38, supplement, 1945, pp. 57-64.

A report on the prosperity of Timurid Balḵ is furnished by Ruy Gonzalez de Clavijo, Embajada a Tamorlan, ed. by F. Lopez Estrada, Madrid, 1943, pp. 141-48.

A history of the Timurid period is the Maṭlaʿ-e saʿdayn, by ʿAbd-al-Razzāq Samarqandī. Since the first great Uzbek chronicles were published by A. A. Semenov, more and more works of commentary and editions of Shaibanid and Janid texts have been issued. Particularly noteworthy is Baḥr al-asrār fī manāqeb al-aḵyār, a work by Maḥmūd b. Amīr Walī, prepared on the orders of the Janid governor of the town, Nāder-Moḥammad, translated by Riazul Islam, Karachi, 1980; and the publication of part of the eighteenth-century Tārīḵ-eraḥīmī, of which only two of the many manuscripts, mss. D. 710 and C. 1683, contain the list of the eighteen medieval and modern irrigation canals; cf. M. A. Salakhetdinova, “K istoricheskoĭ toponomike Balkhskoĭ oblasti,” Palestinskiĭ sbornik 21/84, 1970, pp. 222-28; the most recent bibliographies of the published and unpublished Timurid, Uzbek, and Afghan sources on Balḵ can be found in B. A. Akhmedov, Istoriya Balkha, Tashkent, 1982, and A. Mukhtarov, Pozdnesrednevekovyĭ Balkh, Dushanbe, 1980.

The former work also represents the most thorough study on the Uzbek khanate of Balḵ and the latter provides the most complete description of the evolving topography of the city and the transition from the Timurid to the Uzbek period; it also gives a list of the eighteen nahr and the jūy connected with each, cf. pp. 99-109.

For the entire Uzbek period in central Asia, see I. P. Ivanov, Ocherki po istorii Sredneĭ Azii, Moscow, 1958.

For the historical ethnography and Uzbekization of the area, see B. K. Karmysheva, Ocherki etnicheskoĭ istorii yuzhnykh rayonov Tadzhikistana i Uzbekistana, Moscow, 1976.

For geography, see J. Humlum, La géographie de l’Afghanistan, Copenhagen, 1959; and Tübinger Atlas des Vorderen Orients, Section 9, Series A, Wiesbaden, 1984.

See also A. Toynbee, Between Oxus and Jamna, London, 1961.

(V. Fourniau)


iv. Modern Town

The crisis in Balḵ’s urban evolution came in the mid-13/19th century. Great damage was done to the town and the surrounding area in the troubled times following its destruction by the amir of Bukhara in 1840 and its recapture by the Afghans of Dōst Moḥammad in 1850, which gave rise to an exodus of many of its Uzbek inhabitants. A further cause of decline was lack of maintenance of the irrigation canals. One of the results was that Balḵ became a very unhealthy place, and it is therefore not surprising that the Afghans, when again in control after 1850, preferred to base their governorate of Turkestan at Taḵta Pol near Mazār-e Šarīf, later at Mazār-e Šarīf itself. The exact date of the move from Balḵ is uncertain (in 1866, according to Barthold, EI1 III, p. 430; during Moḥammad-Afżal Khan’s governorship of Afghan Turkestan, according to Peacocke, in Gazetteer of Afghanistan IV, p. 110; after 1878, according to Grodekoff, Ride from Samarcand to Harat, London, 1880, p. 80, quoted by Centlivres, p. 124). It may be that at first only a temporary move was intended. In any case the transference was complete and final when the British frontier delimitation commission passed through in 1886. Balḵ’s population was then reckoned to be some 600 families of Tājīks, Uzbeks, and Arabs, of whom 100 were old local families, together with 40 Jewish families and a community of 20 Hindu families originally from Shikarpur in Sind. All lived in the southeastern quarter of the old town inside the wall. The bāzār then had 60 shops. In addition to the permanent inhabitants, there was a floating population of about 1000 Pashtun families in the town and outside the wall. Another source, however, speaks of only 200 Tājīk families (Gazetteer of Afghanistan IV, p. 112; cf. C. E. Yate, Northern Afghanistan, Edinburgh and London, 1888, pp. 255ff.). The decline continued in the following decades. The sketches of Balḵ in the first world war by Niedermayer (p. 48) and in 1924 by Foucher (I, p. 59) depict a mean village of hovels situated to the south of the citadel with a still existing Jewish quarter to the west.

A new phase set in when work on the construction of a new town began in 1934. It was laid out geometrically in concentric circles around a central square with eight radial arteries. The initial plan was overambitious, providing for 1,270 houses together with a large bāzār of some 400 shops and 32 sarāys (K. Ziemke, Als deutscher Gesandter in Afghanistan, Berlin, 1939, p. 229). Actual achievement fell far short; in 1973 (according to Grötzbach, p. 105) only 430 houses had been built and demand for them was weak, the attraction of Mazār-e Šarīf still being dominant throughout the region. According to the preliminary returns of the 1979 census, Balḵ then had only 7,242 inhabitants (communication from D. Balland). Even so, its economic role was by no means negligible. It became an important market for agricultural produce (cotton, melons, almonds, karakul pelts). Buyers from Mazār-e Šarīf came on the market days (Monday and Thursdays) to take advantage of the lower prices, and two-way business with Mazār-e Šarīf grew after the start of a regular bus service. Four cotton firms, two of which had ginneries, were located in Balḵ and its outskirts. After the opening of the asphalted Mazār-e Šarīf-Šeberḡān highway with a 2 km branch to Balḵ in 1970, Balḵ began to attract tourists. From 1972 onward it had the benefit of electricity generated by gas from fields in the region. It also possessed a good primary school and a small hospital. Though only 20 km from Mazār-e Šarīf, Balḵ ranked as a small independent center.



L. W. Adamec, ed., Gazetteer of Afghanistan IV, 1979, pp. 98-112.

O. von Niedermayer and E. Diez, Afghanistan, Leipzig, 1924.

A. Foucher, La vieille route de l’Inde de Bactres à Taxila, MDAFA 1, 2 vols., Paris, 1942-47.

M. Le Berre and D. Schlumberger, “Observations sur les remparts de Bactres,” in B. Dagens et al., Monuments préislamiques d’Afghanistan, MDAFA 19, Paris, 1964, pp. 61-105.

P. Centlivres, “Structure et évolution des bazars du Nord Afghan,” in E. Grötzbach, ed., Aktuelle Probleme der Regionalentwicklung und Stadtgeographie Afghanistans, Afghanische Studien 14, Meisenheim am Glan, 1976.

E. Grötzbach, Städte und Basare in Afghanistan: Eine stadtgeographische Untersuchung, Beihefte zum Tübinger Atlas des Vorderen Orients, series B, no. 16, Wiesbaden, 1979.

A. Mukhtarov, Pozdnesrednevekovyĭ Balkh (Materialy k istoriciheskoĭ topografii goroda v XVI-XIII vv.), Dushanbe, 1980.

B. A. Akhmedov, Istoriya Balkha (XVI-pervaya polovina XVIII v.), Tashkent, 1982.

(X. de Planhol)


v. Modern Province

Balḵ is a province (welāyat) of northern Afghanistan which covers 11,833 km2. In 1363 Š./1984 it was divided into seven districts (woloswālī) and four subdistricts (ʿalāqadārī). The main town and provincial capital is Mazār-e Šarīf (q.v.) and three more localities within the province have urban status (Balḵ, Dawlatābād and Šōlgara or Boynaqara).

Balḵ province was created in 1343 Š./1964 out of the former and much larger province of Mazār-e Šarīf.

See Table 16 and Table 17 for compilation of main available data about population and land use in the province, districts, and subdistricts.

(D. Balland)


vi. Monuments of Balḵ

The successive city-walls. The mud ramparts of Balḵ which still survive, superimposed one upon the other, at an impressive length and height, more than 20 m at the citadel (Bālā-Ḥeṣār) and on the southern side, are the most substantial remains of the ancient periods of the “Mother of Cities.” Archeological examination of these ramparts has provided the key to the successive stages of the topographical development of the town (see Le Berre and Schlumberger). The initial limit is represented by the Bālā-Ḥeṣār (“Balḵ I”); its circular plan is probably inherited from the Achaemenian period, while its present Timurid circuit-wall largely reuses the massive Greek rampart which, in 208-06 b.c., withstood the attack of the Seleucid Antiochus III. From the Greek period also dates a gigantic wall built against the nomadic incursions along the northern edge of the oasis, where its remains have been traced for a length of 60 km (Kruglikova; Pugachenkova, 1976); it is mentioned as still in use by Yaʿqūbī (fl. 276/889), and it sheltered other important towns, mainly Delbarjīn (Greek-Kushan period) and Zādīān-Dawlatābād (Saljuq period) (see Figure 15).

The development of a southern suburb of Balḵ along the caravan-road to India led to a first extension of the walled city (“Balḵ IA”: late Greek or Kushan period). At some time between the Kushans and the Islamic conquest it was further enlarged to the east (“Balḵ II”). These walls with square towers remained in use until Balḵ was thoroughly destroyed in 617/1220 by the Mongols of Jengiz Khan.

In 765/1363 the Bālā-Ḥeṣār was reoccupied by Amir Ḥosayn, after which Tīmūr and his successors completely refortified the whole city while slightly moving it to the west, probably because the eastern part had become marshy after the destruction of the irrigation system. This last rampart (Balḵ III), made of heterogeneous materials extracted from the ruins left by the Mongols, had semi-circular towers, and was adorned at its southern side by the monumental Bābā-Kōh gate (or Nowbahār gate; now destroyed) and by the Borj-e ʿAyyārān, an eight-arched belvedere (Foucher, p. 164, pl. VI; Mukhtarov, pp. 21-42).

The Buddhist remains. Apart from the ramparts, the only monuments which have survived from pre-Islamic Balḵ are Buddhist stūpas, which owed their preservation to the massivity of their mud-brick masonry. Four, all standing along the roads on the outskirts of the city, were identified by A. Foucher in 1924-25; the Top-e Rostam, in the south, was the only one he excavated. Although greatly ruined and stripped of all its decoration, it can be reconstructed as the most monumental stūpa witnessed north of the Hindu Kush (dimensions: square platform 54 x 54 m, cylindrical dome 47 m in diameter, total height probably ca. 60 m). Its location and size correspond to those of the “New Monastery” described in the 7th century by the Chinese pilgrim Hsüan-tsang (Th. Watton, On Yuan Chwang’s Travels in India I, London, 1904, pp. 108-09); otherwise known as the Nowbahār (from Sanskrit nava-vihāra), it is renowned in the Islamic sources because the Buddhist ancestors of the Barmakids had been its administrators. But the neighboring Taḵt-e Rostam, a steep mound sometimes considered the remains of the convent itself, looks rather like the mud platform of an early medieval manor (kōšk). Two other stūpas, the Čarḵ-e Falak and the Āsīā-ye Kohnak, have mosques grafted onto their remains, a clear indication of the continuity of cult-places (Foucher, pp. 83-98, 168-69, pls. XIX-XXII; Mélikian-Chirvani, 1974).

Pre-Mongol Islamic monuments. A building of outstanding interest is an ʿAbbasid suburban mosque known locally as Noh Gombad or Ḥājī Pīād, which was discovered in 1966, at a short distance south of the Top-e Rostam. Built of baked bricks, it consists of four round pillars standing in the center of a square (20 x 20 m) formed by three curtain walls and an open façade which is articulated on two more pillars; the pillars were linked to each other and to coupled columns attached to the walls by perpendicular arcades, the inner space being thereby divided into nine equal squares, each of which originally supported a dome. A deeply carved stucco ornamentation still covers the capitals, imposts, and bases of the pillars, as well as the spandrels and soffits of the arches; the motifs include grape leaves, freely moving vine-scrolls, fir-cones, palmettes, rosettes, set in interlacing straps and thickly packed so as to fill up the panels almost entirely. Neither the architectural composition nor the decoration have their direct origin in the Central-Asian tradition (which, for example, ignored the open arcatures); they rather represent the direct transposition of a model which took shape in the heart of the ʿAbbasid empire and from there spread both east and west (where the clearest examples now surviving are to be found, especially some religious monuments of Tulunid Egypt). The stucco ornament has its closest parallels in the styles A and B of Samarra, which indicates the first half of the 9th century as the most probable date of construction (Pugachenkova, 1968; Golombek; differently Mélikian-Chirvani, 1969).

The only other monument which can be ascribed to the pre-Mongol Islamic period is the plain, single-chambered, domed mausoleum known as Bābā Rōšnay (at the southwest of the Bālā-Ḥeṣār; first half of the 11th century; Pugachenkova, 1978, pp. 31-32).

Timurid and Ashtarkhanid monuments. 3 kms to the east of the outer wall stands the mausoleum locally known as Mīr-e Rūzadār, surrounded by ornamented brick burial enclosures. The mausoleum preserves an elaborate interior decoration (angular interlacing ribbed design on the dome and niches, enhanced by painting); but the outer dome and exterior facing are lacking, which has led to the supposition that the monument remained unfinished because of the political troubles of the 1440s (Pugachenkova, 1978, pp. 33-35; Mukhtarov, pp. 75-83). Its architectural composition expresses the Timurid taste for the octagonal tomb-chamber, with external vaulted niches hollowed in the facets and angles, and projection entrance-room. The same composition is repeated, with variations, at the later mausoleums of Ḵᵛāja Bajgāhī (eastern edge of the town; 17th cent.) and Ḵᵛāja Akāša; it is also to be found, in a more sophisticated form, at the funerary mosque of Ḵᵛāja Abū Naṣr Pārsā, perhaps the most famous monument of Balḵ. It was erected in 867/1462-63, shortly after the death of the theologian, who is buried in the platform which lies in front. The usual entrance-room is replaced here by a tall pēštāq flanked by two minarets, each of which is preceded by slender corkscrew pillars. The whole of the façade and the fluted outer dome were veneered in kāšī whose predominant tint is a cold silvery blue; their manufacture was of the best quality, but due to an inadequate mode of fixation large surfaces have collapsed. The interior, lighted by sixteen lattice openings at the basis of the drum, is richly ornamented by a well-preserved angular interlace of stucco, completed by painted floral motives (Pugachenkova, 1970). Together with the contemporary mosque at Anau (Turkmenistan), this monument represents one of the finest examples of late Timurid memorial architecture.

Balḵ had a late flourish under the Ashtarkhanid dynasty, when it formed the apanage of the heirs to the throne of Bukhara (1007-1164/1599-1751). From this time dates the madrasa built by the Sayyed Sobḥānqolī Khan in the last years of the 11th/17th century; only the tiled entrance ayvān remains, facing the mosque of Abū Naṣr Pārsā in the garden which is now the center of the town. The ruins of the governor’s palace, including a small mosque, which were excavated by Foucher in the Arg of the Bālā-Ḥeṣār, cannot be precisely dated but obviously belong to the late Islamic period also (Foucher, pp. 98-112, 165-66, pls. XI-XVIII).



The successive phases of the archeological exploration of Balḵ are described in: O. von Niedermayer and E. Diez, Afghanistan, Leipzig, 1924, pp. 204-05 (brief description of the main Islamic monuments); A. Foucher, La vieille route de l’Inde de Bactres à Taxila, MDAFA 1, vol. 1, Paris, 1942, pp. 55-121, 163-70, pls. V-XXVI (general survey; excavations at the Top-e Rostam and at the citadel); M. Le Berre and D. Schlumberger, “Observations sur les remparts de Bactres,” in Monuments préislamiques d’Afghanistan, MDAFA 19, Paris, 1964, pp. 61-105 pl. XXXII-XLV, figs, 10-19 (study of the successive city-walls; supersedes R. S. Young, “The South Wall of Balkh-Bactra,” American Journal of Archaeology 59, 1955, pp. 267-76; completed by J. Cl. Gardin, Céramiques de Bactres, MDAFA 15, Paris, 1957).

The wall of the oasis is studied by I. T. Kruglikova, Dil’berdzhin [I], Moscow, 1974, pp. 9-15, and by G. A. Pugachenkova, “K poznaniyu antichnoĭ i rannesrednevekovoĭ arkhitektury Severnogo Afganistana,” in Drevnyaya Baktriya I, ed. I. T. Kruglikova, Moscow, 1976, pp. 137-41.

On the Nowbahār/Top-e Rostam see also A. S. Mélikian-Chirvani, “L’évocation littéraire du bouddhisme dans l’Iran musulman,” Le monde iranien et l’Islam 2, Geneva and Paris, 1974, pp. 10-23 (discusses the Islamic sources); K. Fischer, Indische Baukunst islamischer Zeit, Baden-Baden, 1976, p. 131.

The Islamic monuments have been seriously studied for 20 years only. On the ʿAbbasid mosque: G. A. Pugachenkova (Pougatchenkova), “Les monuments peu connus de l’architecture médiévale de l’Afghanistan,” Afghanistan 21/1, 1968, pp. 17-27; A. S. Mélikian-Chirvani, “La plus ancienne mosquée de Balkh,” Arts Asiatiques 20, 1969, pp. 3-19; L. Golombek, “Abbasid Mosque at Balkh,” Oriental Art 25, 1969, pp. 173-89.

On the mosque of Abū Naṣr Pārsā: G. A. Pugachenkova, “A l’étude des monuments timourides d’Afghanistan,” Afghanistan 23/3, 1970, pp. 33-37; idem, Zodchestvo Tsentral’noĭ Azii: XV vek, Tashkent, 1976, pp. 30 and 61.

On other monuments recently discovered: Idem, “Little Known Monuments of the Balkh Area,” Art and Archaeology Research Papers, London, June, 1978, pp. 31-40.

A. Mukhtarov, Pozdnesrednevekovyĭ Balkh, Dushanbe, 1980.

(F. Grenet)

(X. de Planhol, C. E. Bosworth, V. Fourniau, D. Balland, F. Grenet)

Originally Published: December 15, 1988

Last Updated: December 15, 1988

This article is available in print.
Vol. III, Fasc. 6, pp. 587-596