Hudud al-'Alam, The Regions of the World
V. Minorsky





- __I_ (Early Islamic scholars)

- __II_

- __III_ (Abū Zayd Balkhī)

- __IV_ (Jayhānī, Gardīzī)

- __V_

- __VI_ (Description of the Muslim world)

- __VII_ (The non-Muslim provinces)


(P 3)

THE present edition is intended to discharge an obligation under which Russian science has long lain; namely, that of rendering available to the specialists the important work dealing with Muslim geography which was discovered in 1882 in Bukhārā [1] through the initiative of the Russian orientalist, A. G. Toumansky (d. 1920).




The activity of the early Islamic scholars, who wrote almost exclusively in Arabic, is known to us not only by their original works that have reached us, by references to the books that have disappeared, and by quotations from them, but also through bibliographical surveys, of which the necessity was felt even then. Only five years [2] after the date of the treatise preserved in the Toumansky MS., al-Nadīm composed his Fihrist, from this work and from later bibliographical compilations [3] European scholars have culled most of their information as to what works, known to be important and not yet to be found in European libraries, must still be sought for. Such quests, even if successful in bringing to light desired volumes, have sometimes brought disillusion as well, even in the cases when the book was linked with a great name.


The Persian Abul-Fal Gulpāyagānī, who had the luck to discover the present precious MS., was searching in Bukhārā, on behalf of Toumansky, for the historical treatise of Ulugh-bek. Judging by what is already known of the latter work, its discovery



1. On the discovery of the Toumansky MS. and its contents see Zapiski Vostochnago Otdeleniya (ZVO), X, 121-37.


2. Now we know that the year 377 H. is given not only in the Leiden MS., as stated in Brockelmann, GAL, i, 147, but, for instance, in Yāqūt’s Irshād (vi, 408). It is known that the same date is several times given in the treatise itself (cf. ZVO, iv, 402); for the completion of parts i and ii even the day is given (Saturday, ist of Sha'bān = 26.xi.987), though in isolated passages, apparently written by another person, later dates are given. Particularly characteristic are the words (p. 132) about the scholar Marzbānī, born in Jumādā II, 297 (II-III. 910): “And he is alive in these our times, in 377; and we beg of God for him health and continuation of life, from God’s clemency and bounty: and he died in 378, may God have mercy on him." The last words belong, evidently, not to the author (otherwise he would have deleted the previous words) but to another person. [?V.M.]


3. The work of the wazīr Maghribī who continued al-Nadim, see Irshād, vi, 467, has not reached us. On the wazīr Maghrībī see Brockelmann, i, 353; E. Zambur, Manuel de genealogie et de chronologie pour l’histoire de l’slam, Hanovre 1927, p. 15.





would have brought but little benefit to science. [1] But while engaged in his search, Gulpāyagānī found a document quite unknown until then and mentioned in none of the bibliographical surveys, which has proved to be of the greatest scientific importance.


The MS. does not contain the author's name, but the date of its composition is indicated : the author began his work in 372 H. (between and for Abul-ārith Muammad ibn Amad, prince of the province of Gūzgān or Gūzgānān (in Arabic garb Jūzjān, or Juzjānān), which lies in the north-western part of the present-day Afghanistan. It is quite natural that the author allows more space (ff. 20b-21a) to this province than would have been expected from its comparative unimportance, even though at that time Gūzgān was experiencing, under the Farīghūnid dynasty, a period of political and cultural prosperity.


On Gūzgān and the Farīghūnids Toumansky’s article [2] gives only a fragment from Rashid al-dīn's Jāmi‛ al-tawārīkh, almost literally copied, as is the whole of the corresponding part of this work, from 'Utbī's history in its Persian

(P 4)

| translation by Abul-Sharaf Jarbādhagānī. [3] In a note added to Toumansky's article I have mentioned a fact recorded by an author of the eleventh century — Gardīzī — that the amīr of Samarqand, Nū ibn-Manūr (A.D. 976-97), had accepted ties of relationship with the head of the Farīghūnids. This fact belongs to the beginning of the reign of Nū (who ascended the throne at the age of thirteen). The prince of Gūzgān in the Gardīzī text, as in that of our author, bears the name of Abul-ārith Muammad ibn-Amad (in 'Utbī: Amad ibn-Muammad). [4]


Both our author and other geographers of the tenth century describe the Gūzgān of the Farīghūnids as much more extensive than it had hitherto been. Both at the time of the Muslim conquest [5] and later, according to the geographers of the ninth century A.D., Fāryāb (on the site of the town of Daulatābād, or that of the village of Khayrābād) [6] was not reckoned as in Gūzgān, the western frontier of the latter, as attested by Ibn Khurdādhbih’s itinerary, [7] passing between



1. V. Barthold, Ulugh-bek, Petrograd, 1918, p. 113 and sq. (in Russian).

2. ZVO, X, 128 and sq.

3. On the original and translation see my Turkestan, ii, 20 and sq.; Engl. ed., GMS, p. 19 and sq. The text of the chapter on the Farīghūnids borrowed from the Persian translation of 'Utbī, Tehran, 1272, p. 305 and sq., is given by Rashid al-din with some unessential alterations and abbreviations. The Arabic original vUtbī-Maninī, ii, 101.

4. Text of Gardīzī, according to the Cambridge MS., King's College, 213, f. 104b: va bā amīr Abul-asan Farīghūn khwīshī kard tā bad-īshān pusht-i ū qawī gasht.

5. e.g Balādhuri, p. 406 below.

6. V. Barthold, Historico-geographical Survey of Iran (in Russian), St. Petersburg, 1903, p. 23. [See Map viii.]

7. BGA, vi, 3211.





Faryāb and Shabūrqān [1] (or Shubūrqān; [2] in our author: Ushbūrqān or Ashbūrqān, [3] now Shibarghan), at an equal distance (9 farsakhs) from both. According to Ya'qūbī, Faryāb was "the old town”; the second most important town at that time, and the residence of the Arab governor ('āmil) of Faryāb, was the town of Yahūdān (in Iakhrī and others: Yahūdiya; in our author: Jahūdhān, on the site of Maymana). But, on the other hand, the mountainous province Gurzivān, on the upper course of the river Āb-i Maymana, was considered a part of Gūzgān; there lived the local ruler (malik) of Gūzgān, whereas the Arab governors of Gūzgān resided in Anbār (according to our author: Anbīr), on the site of the present town Sari-i-Pul. In Iṣṭakhrī’s time (or that of his source, Balkhī) the situation had changed. It is not mentioned whether at that time Fāryāb was part of Gūzgān, but Yahūdiya was reckoned as belonging to it and as being even its principal town, while Anbār was the seat of the government (sulān); evidently at that time a distinction no longer existed between the Arab governor and the local ruler. As to our author, he names Jahūdhān as the residence of the “king of Gūzgān”, and Anbār as the capital of the province (qaaba, the same term is used by Bīrūnī, [4] in whose work we also find the form Anbīr).


Our author assigns to the rulers of Gūzgān the first place among the vassal princes (mulūk-i araf) of the Sāmānid kingdom, not only for their political importance, but also for their “love of science”. At that time the sway of the prince of Gūzgān reached to the north as far as the Amū-daryā and to the south was recognized by all the chiefs (mihtar) of the mountain provinces of Gharchistān and Ghūr. A part of Gharchistān was called “Gharchistān of Gūzgān” and was administered directly by the prince of Gūzgān, whose frontier towns were those of ālaqān, on the site of Qal'a-Walī, [5] and Rabā-i Karvān on the upper Harīrūd. [6] On the upper Murghāb Gūzgān had a common frontier with that of the prince of Bust (on the Hilmand). It may be concluded therefrom that nearly the whole of the province of Ghūr owed allegiance to the prince of Bust. But in his description of Ghūr (f. 21b) the author calls the ruler

(P 5)

| of this province, the Ghūrshāh, vassal of the amīr of Gūzgān. To the latter tribute was



1. Such vocalization in Iṣṭakhrī, BGA, i, 270; in Marco Polo, Sapurgan, with several variants.

2. Vocalization, BGA, vi, 3211; vii, 28710; distorted vocalization in Yāqūt, iii, 25410 and 3054, though Yāqūt had visited the place.

3. BGA, ii, 32117 and 322­7: Ashbūrqan.

4. Thus in the work of A.-Z. Validov [Validi], Al-Bīrūnī āthāri, now in the press, p. 18: in the Berlin MS. (on it see Ahlwardt, No. 5667), f. 123a, instead of qaaba stands qar.

5. On its situation see ZVO, xiv, 031.

6. Similarly in Iṣṭakhrī, BGA, i, 272; cf. ibid. 265 below. [V.i., p. 336.]





likewise paid by the nomad Arabs of the neighbouring steppes, who numbered 20,000, possessed herds of sheep and camels and were considered to be the richest of all the Arabs of Khorāsān.


The Farīghūnids called themselves descendants of the mythical Farīdūn, [1] but apparently there exists no information as to whence this dynasty sprang, when and how it gained its power, and whether or not it was related to the pre-Islamic rulers of Gūzgān, the Gūzgān-khudāts. [2] The name of the dynasty had some relation to a locality in the extreme north of the province; Maqdisī [3] mentions a Rabat Afrīghūn, one day’s march from Andkhoy and two from Karkī. According to Narshakhī, [4] Amad ibn-Farīghūn was already amīr of Gūzgān in the last years of the ninth century, at the time when the relations between the affārid 'Amr ibn-Laith and the Sāmānid Ismā'īl ibn-Amad were broken off. Since Iṣṭakhrī [5] mentions an Abul ārith ibn-Farīghūn, apparently the same Abul ārith Muammad ibn-Amad who was a contemporary of our author, this ruler must have lived unusually long. [6] It is probable that the name of this prince was not yet recorded in Balkhī’s original work, as it is mentioned not in the chapter on Khorāsān, but in that on Fārs, a chapter which, according to de Goeje, [7] belongs without doubt to Iṣṭakhrī and not to Balkhī, though Iṣṭakhrī wrote it a long time before his work was brought out, not later than in 933, i.e. half a century before the appearance of the Ḥudūd al-’ālam. Iṣṭakhrī mentions also a secretary or minister (kātib) of the amīr of Gūzgān, Ja'far ibn-Sahl ibn-Marzubān, of the family of Marzubān ibn-Zādiya, who was a native of Shīrāz. This Ja'far was still alive at the time when Ibn auqal composed his chapter on Khorāsān, i.e. at the end of the nine hundred and sixties; [8] Ibn auqal [9] was acquainted with him and speaks of



1. There is no foundation for reading Afrīghūn instead of Afrīdhūn, as Toumansky proposes, ZVO, X, 130.

2. I. Marquart (Markwart), Ērānšahr, p. 80.

3. BGA, iii, 347.

4. Ed. Schefer, P. 85.

5. BGA, i, 1482.

6. The year of his death apparently is not mentioned anywhere. He was still alive in 999 at the time of the conquest of Khorāsān by Mahmūd ('Utbī-Manīni, i, 316); the account of the battle of Charkhiyān (4 January 1008; cf. my Turkestan, ii, 287) names, as the ruler of Gūzgān, his son and successor Abū-Nar ('Utbi-Manīni, ii, 84), who died in 401 (1010-11). Contrary to Markov, Invent. Catal. of Muslim Coins of the Hermitage (in Russian), SPb. 1896, p. 178 and sq., and Zambaur, Manuel, p. 205, the Farighūnids never possessed Balkh and did not strike coins. The names and dates given by Zambaur do not in the least correspond to reality and represent a step backwards in comparison with Sachau's article to which Zambaur refers.

7. ZDMG, XXV, 50.

8. As the Sāmānid amīr contemporary with himself Ibn auqal names Manūr ibn-Nūh (961-76), BGA, ii, 34111. In 358 (9689) this author was on the Gurgān (ibid., p. 28210), in the same year “for the last time" in Mosul, ibid., p. 1463, and apparently returned no more to the east.

9. Ibid., p. 208.





the rare unanimity with which the qualities of the Gūzgān minister were extolled by his contemporaries. About all other statesmen, alongside with favourable reports, unfavourable ones might be heard or read; but Ibn-auqal never encountered any one who had an unfavourable opinion of Ja'far ibn-Sahl. Every one who visited Khorāsān during the previous fifty years was indebted to him for some kindness; those who could not visit him personally were not excepted, as they received letters and presents from him. On his lands he built rabats and assigned revenues of his estates for their maintenance; in every rabat and village he kept cows, to the number of one hundred or more, in order to provide milk for the refreshment of passing travellers. In no respect had he his equal in Khorāsān. It is very probable that Ja'far ibn-Sahl patronized Ibn-auqal's work.


Whether the author of the Ḥudūd al-’ālam made any travels himself does not appear from his work. He speaks only of borrowing information from books, though he names

(P 6)

| none of his Muslim sources. As Toumansky [1] remarks, “nowhere does he name his sources, except for Ptolemy, and even him, probably, only as a rhetorical figure”. This remark does not entirely correspond to the facts, for besides Ptolemy, Aristotle is named (f. 2a), and his “Meteorologica” (al-Āthār al-'ulwiya) cited. The same passage (about the ocean encircling the earth) is quoted by al-Kharaqī, an author of the beginning of the twelfth century. [2] Ptolemy, as a matter of fact, is cited twice (4a and 5a), not in the chapters consecrated to separate provinces, but in the general part, viz. in the chapter on islands. There are mentioned thirteen islands and two mountains projecting into the Indian Ocean, and it is added that these two mountains are found in Ptolemy’s books; but in Ptolemy's Geographica there is nothing on which this information could be founded. According to our author Ptolemy enumerated twenty-five islands in “the Western Ocean” (Ptolemy’s δυτικὸς Ὠκεανός). These names are given and the majority are really borrowed from Ptolemy, beginning with the six “islands of the Blest” (αἱ τῶν Μακάρων νῆσοι, Ptolemy, iv, 6, 34, in Arabic authors generally al-Khālidāt, in our author al-Khāliya, and in Battānī al-Khālliyāt). [3] From Ptolemy was derived the information



1. ZVO, X, 132.

2. Text in Nallino, Al-Battānī sive Albatemi opus astronomicum, pars i, Mediolani, 1903, p. 175.

3. Battānī-Nallino, i, 17, note 2. This is not the only case of coincidence of Battānū’s text with that of our author. According to Battānī, ibid., p. 18, note 5, and our author (f. 4b), near India and Ceylon there were fifty-nine islands; according to Ibn-Rusta, BGA, vii. 8415, and Khar, in agreement with the text of Ptolemy (vii, 4, 1-13), the islands were nineteen. In Nallino's opinion Battānī read which stood in his list; this mistake was evidently made by the source common to Battānī and our author.





about the “isles of Britannia”, of which, according to our author and to Arab geographers, [1] there were twelve (this number is not in Ptolemy). Concerning Britannia, as well as the “Isles of the Blest” (Canary Islands), our author gives information which, apparently, does not exist in other sources: he says that in the “Isles of the Blest” there are “gold mines; once a year people from the Sūdān and from towns of Sūs al-'Aqā make their way there and bring away gold from those mines; no one can live there on account of the intense heat”. Britannia is called (f. 37b) “the storehouse of goods from Byzantium (Rūm) and Spain (Andalus)”. Yet among the names of the twenty-five islands there are some that do not occur in Ptolemy: by mistake the author places Rhodes and Arwad in the Western Ocean; as to the legendary “Isle of Men” and “Isle of Women”, their mention at this place is, no doubt, due to the fact that the legend of the Amazons was in Islamic times localized in the Baltic sea, [2] perhaps owing to a linguistic misunderstanding. The references of the author, like those of many other Muslim geographers, [3] are, evidently, not to the original text of Ptolemy, but to the readaptation of his work by the Arabs; but there is nothing “rhetorical” about these references.





The history of Arabian geographical science has been very insufiiciently investigated. [4] In the Encyclopaedia of Islam, which is not quite consistent

(P 7)

| in the choice of the catch-words (cf. Adab, al-Djabr, on the one hand, and on the other Astrology, Astronomy), where we might have expected to find an article on this subject, nothing is to be found either under Djaghrāfiya, or Geography. In Brockelmann's Geschichte der Arabischen Literatur there are sections consecrated to geographical literature but, as has already been pointed out by its reviewer, [5] the insufficiency of Brockelmann's book [6] is apparent, particularly as regards this topic.



1. BGA, vii, 8513. Ibn Rusta; Battānī-Nallino, i, 18.

2. Kunik-Rosen, Izvestiya al-Bekri, &c., i, 80; Peschel-Ruge, Geschichte der Erdkunde, p. 90; Nallino, Al-uwārizmī, p. 50.

3. Nallino, l.c., p. 52.

4. In the broadly planned Introduction to the History of Sciences (G. Sarton, Carnegie Institution of Washington. Publication No. 376, 1927; cf. a review by E. J. Holmyard in JRAS, 1929, 209 and sq.) much space is allotted to geographical science “from Hecataeus to Bīrūnī”. It is stated there that Greek knowledge was transmitted by the Christian school to other Oriental Christians — Syrians and Arabs — and finally to the Muslims; that from the second half of the seventh century to the end of the eleventh century Arabic was the principal language of science and progress, and that in the twelfth century "the intellectual supremacy of the Muslims had already come to an end” (p. 18), which is hardly true.

5. [Seybold, Edrisiana, I.] ZDMG, lxiii, 596.

6. [See now its Supplement. V.M.]





The learned critic, writing in 1908, was of opinion that the best survey of Arabian geographical literature was that of Reinaud, published in 1848. [1] The work of Abū-Ja'far Muhammad ibn-Mūsā al-Khuwārizmī, which in the first half of the ninth century laid the first foundations of Arabian geographical science, became accessible in a printed edition only in 1926. [2] But as early as 1895 this work had been the subject of a classical study by c. A. Nallino, [3] who later took up again numerous questions concerning Khuwārizmī and Arabian geographical science in general, in his extensive Latin work on the astronomer Battānī (d. in A.D. 929). [4]


It is a well-established fact that Arabian geography, like Arabian astronomy, was founded on Ptolemy. In the Middle Ages Ptolemy was studied only in the East, at first in the Christian East, later in the Muslim East, whereas in Western Europe until the fifteenth century he remained quite forgotten. [5] From Ptolemy’s astronomical work was also borrowed the historical canon, that is, the chronology of reigns from the eighth century B.C. to the second century A.D., which was adopted by the Christian world jointly with the chronology of Eusebius (fourth century A.D.), in which history begins with Abraham and the kings [6] contemporary with him and with his descendants. Some efforts, not always successful, were made in Muslim literature to localize ancient geographical traditions: thus Bīrūnī in his Canon (eleventh century) tried to identify the classical Ilion with the Syrian Tripoli. [7] There is no literal rendering of Ptolemy’s text in Arabic; from the outset Muslim scholars treated this text much more independently than at a later date did the West-European scholars.



1. Geographie d’Aboulféda, t. i, Introduction générale à la géographie des orientaux.

2. Das Kitab ūrat al-ar des Abū Ğa’far Muammad ibn Mūsā al-uwūrizmī, herausg. von Hans V. Mžik, Bibliothek arabischer Historiker und Geographen, III. Band.

3. C. A. Nallino, Al-uwūrizmī e il suo rifacimento della geograpa di Tolomeo, Reale Acad. d. Eincei, Serie V, Alemorie della classe di scienze morali, &c., vol. ii, parte 1a.

4. C. A. Nallino, Al-Battānī sive Albatenii opus astronomicum, Public. del Reale Osservatorio di Brera in Milano, No. xl, parte i-iii. The third part (Arabic text) appeared in 1899, the first in 1903, the second in 1907.

5. K. Wright, The Geographical Lore of the Time of the Crusades, New York, 1925, Amer. Geogr. Soc., Research Series, No. 15, pp. 10, 19.

6. See the text of Ptolemy's Canon, e.g. C. Wachsmuth, Einleitung in das Studium der alten Geschichte, Leipzig, 1895, p. 305 sq. On Eusebius, ibid., p. 163 sq. According both to Eusebius and Bīrūnī, Chronologie, ed. Sachau, p. 85, the list of kings begins with Bel, father of Ninus.

7. Berlin MS. Ahlw. 5667, f. 34a:

Chronologie, p. 86.





Already in Muammad Khuwārizmī’s ūrat al-ar we find a new version of Ptolemy, partly corrected and completed, partly distorted. According to Nallino [1] Khuwārizmī’s rifacimento is a work the like of which no European nation could have produced at the dawn of its scientific activity. Yet this early independence of Muslim scholars had its negative side as well. There was no firm and definite startingpoint for scientific thought and no possibility of discriminating between facts borrowed from different sources; even in the tenth century, geographers did not know what was authentic in Ptolemy and what had been added by Muslim authors.

(P 8)

| The exact date of M. Khuwārizmī’s work is unknown. The scanty biographical information about him has lately been summarized by Wiedemann. [2] The appellation al-Qutrubbulī indicates that he was associated not only with Khorāsān, but also with the locality on the Tigris whither, perhaps, already his ancestors had migrated; the appellation al-Majūsī shows that his ancestors were not Christians but Zoroastrians; this may partly account for the fact that he was more influenced by Indian and Persian traditions than by Greek ones. Besides astronomical and mathematical treatises (it is well known that "algorithm” is a distorted form of al- Khuwārizmī’s name), he compiled an historical work, references to which are found in so early an historian as Amad ibn-Abī-āhir ayfūr; [3] in abarī the earliest reference to Khuwārizmī’s work deals with the death of the caliph Mahdī (A.D. 785), [4] while the last reference is made under 210 (A.D. 825-6). [5] Khuwārizmī took part in the scientific activity which flourished in the reign of the caliplt Ma'mūn (a.d. 813-33); his geographical work is surely connected with the map drawn up for Ma'mūn, which was regarded as a joint production; [6] later, under the influence of the legend of the Septuaginta, &c., it was said that seventy scholars took part in this work. [7] Consequently in Nallino’s [8] opinion Khuwārizmī’s work undoubtedly was composed under Ma'mūn (A.D. 813-33). Meanwhile, Nallino determines tentatively the terminus post quem and the terminus ante quem. Among the towns of the third climate the insignificant village of Qiman [9] in upper Egypt is named, which could have become known in consequence of a victory of the Government troops over the rebels in 201 (A.D. 816-17); [10] if so, Khuwārizmī wrote not earlier than 201 H.



1. Al-uwārizmī, p. 53.

2. Enc. of Islam.

3. Kitāb Baghdād, ed. Keller, p. 349.

4. abarī, iii, 55112.

5. Ibid., p. 108513.

6. BGA, viii, 3314.

7. Reinaud, Introd., p. xlv, note 2. Nallino, Al-uwārizmī, p. 13.

8. Nallino, ibid., p. 22.

9. In Mžik's edition. No. 153.

10. A more exact date in Kindī (GMS, xix, 1666): Jumādā I, 201 (25 Nov.-24 Dec. 816).





As the latest date, 210 (A.D. 826-7) is proposed, but no explicit reason for its adoption is advanced. In reality the work of Khuwārizmī in its present form cannot be placed in the reign of Ma'mūn, as it mentions the new capital Suria-man-ra'ā (Sāmarra), [1] the construction of which began in 211 (A.D. 836) [2] under the caliph Mu'taim (A.D. 833-42). The terminus post quem must therefore be advanced by twenty years; as terminus ante quem could be taken the date of Khuwārizmī’s death, if that date were known to us; the last time Khuwārizmī’s name seems to be mentioned is on the occasion of the caliph Wāthiq’s death in A.D. 847. [3]


In Khuwārizmī’s treatise we meet along with geographical names of the Muslim period a great number of ancient names; later these names rapidly begin to disappear; Yāqūt in his dictionary says, with reference to geographical names occurring in pre-Muslim authors, that “owing to the length of time” [4] they have mostly become unintelligible. It is interesting to note the efforts of Khuwārizmī to connect the ancient names with those of his time. Germany is called land of the Slavs; the two Sarmatias are respectively identified with the land of the Danube Bulgars and that of the Alans; both Scythias, respectively, with the land of the Turks in general and that of the Turks of the extreme east, the Toghuzghuz; Serika, with īnistān, i.e. China. [5] The last example shows that for comparison with Greek terms Persian geographical names were utilized as well.

(P 9)

| For the exact title of Ptolemy's book Γεωγραφική Ὑφήγησις, “Geography” or in the Arabic version Jaghrāfiyā was substituted; this word was generally translated as “image of the earth” (ūrat al-ar), [6] and here probably lies the explanation of the title of Khuwārizmī’s book. The author of the Fihrist [7] knew that Ptolemy's work consisted of eight books or sections (in Greek βιβλίον, in Arabic maqāla). The first translation, an unsatisfactory one, was made for a younger contemporary of Khuwārizmī, Abū-Yūsuf Ya'qūb al-Kindī, tutor and familiar of Amad, son of the caliph Mu'taim. The death of Kindī is given as 260 (a.d. 873-4). [8] It is very probable [9] that Kindī utilized this translation for his own geographical work, “Description of the inhabited part of the earth” (Rasm al-ma'mūr min al-ar), mentioned



1. In Mžik's edition. No. 301.

2. abarī, iii, 1180; BGA, viii, 357a.

3. abarī, iii, 136416.

4. Li-taāwuli-l-zamān, Yāqūt, i, 78.

5. Mžik's edition, p. 105 (Nos. 1593, 1596, 1600, 1601, 1602).

6. e.g. Yāqūt, i, 76, also H. Khalīfa, ii, 601. In Mas'ūdī, BGA, viii, 3313, the translation is qa’ al-ar (the crossing of the earth).

7. Fihrist, p. 268; ZDMG, 1, 213.

8. Thus according to Nallino's Arabic work, 'Ilm al-falak, p. 115; Tj. de Boer, Enc. of Islam, ii, p. 1095, says only that he was still alive in 256 (a.d. 870).

9. Thus Brockelmann, i, 225.





by Masudī. [1] A pupil of Kindī, Amad ibn-Muhammad ibn-al-ayyib Sarakhsī (d. in a.d. 899), [2] was also author of a geographical work; the title “Book of Routes and Kingdoms” (Kitāb al-masālik wal-mamālik), [3] frequently occurring in Arabic geographical literature, is also often applied to this work. At the same time an improved translation of Ptolemy was made by Abul-asan Thābit ibn Qurra (a.d. 836-901), a native of pagan arrān and a great admirer of his native pagan culture. By a similar disposition Nallino [4] explains the tendency of Battānī, who also belonged to the pagan (ābian) milieu of Harrān, to revert in some cases from Khuwārizmī to Ptolemy, though it constituted a step backwards (un vero regresso).


From the geographical works of such mathematicians and astronomers as Khuwārizmī, Kindī, Thābit ibn-Qurra, and Battānī the “Books of Routes and Kingdoms” greatly differed in that much more space was allotted in them to political and economical than to mathematical and physical geography. These works contained not only a list of provinces into which the world, and principally the Muslim world, as known to Arabic science, was divided, but also information on towns, commercial routes, articles of export from particular provinces and towns, See. According to the Fihrist [5] the author of the first work on “Routes and Kingdoms” was Abul-'Abbās Ja'far ibn-Amad al-Marwazī ; his work remained unfinished; after his death in Ahwāz his books were taken to Baghdād and there sold in 274 (a.d. 887-8). These data might lead to the belief that the composition of Marwazī’s work belongs approximately to the same time, and this renders doubtful Marwazī’s priority'. Another work of the same title was also written by a ninth-century' geographer, Abul-Qāsim 'Ubaydullāh ibn-'Abdillāh ibn-Khurdādhbih. This work is likewise mentioned in the Fihrist, with no historical details except that the author was a familiar of the caliph Mu’tamid (a.d. 870-92). [6] Ibn Khurdādhbih dedicated his work to some member of the 'Abbāsid dynasty, whom he addressed in the second person without giving his name. The question of the dates of Ibn Khurdādhbih’s life and work is treated in detail in de Goeje’s Preface to the edition of the text. According to de Goeje [7] Ibn Khurdādhbih originally wrote his work in 232 (a.d. 846-7), i.e. in the reign of the caliph Wāthiq (a.d. 842-7), and rewrote it in 272 (a.d. 885-6), under the caliph Mu'tamid. If the first date is exact, the “Book of Routes and Kingdoms”



1. BGA, viii, 2519.

2. On the author Brockelmann, i, 210; Fihrist, p. 261.

3. H. Khalīfa, i, 509 (No. 11870).

4. Al-uwārizmī, p. 24.

5. Fihrist, 150; GMS, vi, 2, p. 400.

6. Fihrist, 149.

7. BGA, vi, p. xx.





by Ibn Khurdādhbih appeared in its first version long before the work of Marwazī, and the mistake of al-Nadīm

(P 10)

| must be explained by the fact that only the second version was known to him, as is shown by his words relating to the caliph Mu'tamid. De Goeje’s opinion was opposed by Marquart, who [1] sought to prove that there was only one edition of Ibn Khurdādhbih’s work, terminated not earlier than 272. Marquart attributes a decisive importance to the fact that already in that version of Ibn Khurdādhbih’s work which de Goeje considers as the earlier one are mentioned the Toghuzghuz, as the Arabs usually called the Uyghurs, this information having been borrowed by Ibn Khurdādhbih from the traveller Tamīm ibn Bar al-Muṭṭawwi'ī, who visited the Toghuzghuz in a region which the Uyghurs, according to Chinese sources, conquered only in a.d. 866. Yet in another passage [2] Marquart himself quotes (though with a wrong interpretation) the text of Jāiz, who died in 255 (a.d. 868-9), in which the Toghuzghuz are mentioned not as new-comers, but as old inhabitants of the same locality, in the neighbourhood of the country of the Kharlukhs (Qarluq). Evidently the word Toghuzghuz, as is only to be expected from its etymological origin (toquz-oghuz “nine Oghuz”), did not always designate the Uyghurs; the Arabs apparently transferred to the Uyghurs the name of the previous inhabitants of the locality conquered by them. [3] Therefore, in order to refute de Goeje’s opinion on the two versions, other proofs ought to be adduced. It is doubtful, for instance, whether Ibn Khurdādhbih could speak about the caliph Wāthiq in his lifetime without using the traditional formulae accepted in such cases; but this argument would have significance only in case the full and not the abbreviated version of Ibn Khurdādhbih had reached us.


Unlike the work of Marwazī, forgotten at an early date, that of Ibn Khurdādhbih obtained a wide circulation and was utilized by many scholars, among whom was undoubtedly, though perhaps not at first hand, our author. The problem of what exactly was borrowed by later authors and from which of Ibn Khurdādhbih’s works it was borrowed, is somewhat obscured by the fact that the “Book of Routes and Kingdoms” has reached us, as de Goeje has proved, [4] only in an abridged form. In quotations from Ibn Khurdādhbih by other authors a more complete text is sometimes found than in the two



1. J. Marquart, Osteuropäische und ostasiatische Streifzüge, Leipzig, 1903, p. 390. Cf. Westberg in Journ. Ministry of Public Instruction (JMNP), New Series, xiii, 1908, p. 374 (in Russian).

2. Streifzüge, p. 91.

3. Cf. for instance, my Report on a Scientific Mission to Central Asia, 1893-1894, SPb. 1897 (in Russian), p. 33 and sq.

4. BGA, vi, p. xv and sq.





known MSS. of Ibn Khurdādhbih’s work; many statements of such authors as Ibn al-Faqīh, [1] Ya'qūbī, Ibn Rusta, [2] and others are founded on Ibn Khurdādhbih. Besides the “Book of Routes and Kingdoms” Ibn Khurdādhbih wrote several other treatises, of which the nearest to the “Book of Routes and Kingdoms” as regards subject-matter was, judging by the title, the “Book of the Genealogies of the Persians and of their Colonies”, and it is possible that some of the references to Ibn Khurdādhbih belong to this latter work. Another question to be elucidated is whether some of the authors could have utilized directly the same sources as those from which Ibn Khurdādhbih’s data were derived.


Ibn Khurdādhbih says in his Preface that he translated Ptolemy’s Geography [3] from a foreign language (it is not said whether from Greek or Syriac); this translation is not mentioned in Arabic literature. According to Nallino, [4] the translation of Ptolemy’s Geography was made by Ibn Khurdādhbih for his own use and was not put into circulation. It is remarkable that even this author, who calls himself a translator of Ptolemy, attributes to Ptolemy the statement, which does not occur in the Greek original, namely, that in his times there were 4,200 towns altogether. [5] In Ptolemy there is no such estimate of towns.

(P 11)

| In spite of his study of Ptolemy, Ibn Khurdādhbih wrote his book on a totally different plan. The astronomical divisions are entirely put aside; the principal part is devoted to “itineraries”, i.e. the description of routes connecting provinces and towns, with an exact indication of distances. Mas’ūdī, [6] with some contempt, calls geography, as understood by Ibn Khurdādhbih, a science for couriers and letter-carriers (to a Russian these words may recall the well-known words of Mme Prostakov in Fonvizin’s comedy Nedorost). Actually, hardly any one will deny that the “Books of Routes and Kingdoms” form precisely the most precious part of Arabic geographical literature. Thanks to them we know the topography of the Muslim Near East of the ninth to tenth centuries much better than that of the ancient world. It is a matter of regret that among documents of ancient literature such compositions as Isidore of Charax’ Σταθμοὶ Παρθικοί, and to a certain extent the “peripli” of the Black Sea and the Indian Ocean, occupy such an isolated place, though Ptolemy would have probably classed them with “chorography”, which he opposes to scientific geography.



1. BGA, v.            2. BGA, vii.            3. BGA, vi, 39.

4. Al-uwārizmī p. 7.

5. BGA, vi, 54.

6. Prairies d'or, ii, 70 and sq.; BGA, vi, p. xii.





We have seen that a translation of Ptolemy had also been in the hands of a contemporary of Ibn Khurdādhbih, Kindī, who wrote a geographical treatise under a different title, indicating a closer relation to mathematical geography; but a pupil of Kindī, Amad Sarakhsī, wrote, like Ibn Khurdādhbih, a book of routes and kingdoms. Sarakhsī, as his appellation denotes, was a native of Khorāsān, but his life and work, as far as it is known, were connected only with Baghdād, where he perished in 899, seemingly a victim to court intrigue. [1] Another disciple of Kindī, who came to 'Irāq from the eastern provinces, Abū Zayd Amad ibn Sahl al-Balkhī, returned to his birthplace, where he lived for many years "(he died in 934) and where he wrote his geographical work, which had a great influence on later geographers, among whom was our author.





De Goeje devoted a detailed article [2] to the question of the geographical work of Abū Zayd Balkhī and of its relation to those works of Iṣṭakhrī and Ibn-auqal which have reached us. In it he quotes biographical data on Balkhī found in the biographical dictionary of afadī, Al-Wāfī bil-wafāyāt. [3] It is clear now that afadī borrowed this information from Yāqūt, [4] who, in his turn, found it in the book on Abū Zayd, which was composed by Abū Sahl Amad ibn-'Ubaydillāh ibn-Amad, “client of the Commander of the Faithful”. As to Abū Sahl, he utilized an earlier biography of Balkhī, the author of which was Abū Muammad asan ibn-Muammad al-Wazīrī who personally knew Abū Zayd Balkhī and had studied under him. [5]


The most important addition to de Goeje’s data is Yāqūt’s testimony according to which Balkhī died (in Dhul-qa'da 322/October 934) at the age of 87 or 88; he was therefore born about 235 (a.d. 849-50). [6] His geographical work, composed, as may be gathered from de Goeje, [7] in 308 or 309 (a.d. 920 or a little later), was therefore written by him

(P 12)

| in his late old age. His journey to 'Irāq, mentioned by de Goeje (Yāqūt [8] says that he went there on foot with a caravan of pilgrims), belongs to his early youth, as is confirmed by the fact



1. The version of the Fihrist, p. 261, according to which the wazir Qāsim fraudulently added the name of Sarakhsī to the list, confirmed by the caliph, of persons condemned to death, is in contradiction with that of a familiar of the caliph Mu'taid, Ibn amdūn, recorded by Yāqūt, GMS, vi, 1, p. 159, according to which the caliph deliberately sent Sarakhsī to his death as an heretic who had tried to lead astray the caliph himself.

2. ZDMG, xxv, 42-58.

3. Brockelmann, ii, 32.

4. GMS, vi, 1, pp. 141-52.

5. Ibid., pp. 143, 144, and 147.            6. Ibid., p. 141.

7. ZDMG, xxv, 49.

8. GMS, vi, 1, p. 145.





that he studied under Kindī, who died soon after 870 (see above). Balkhī spent eight years in 'Irāq and while there visited the neighbouring countries. He acquired broad and many-sided knowledge and when, by way of Herat, he returned to his native Balkh, his learning won him great fame. [1] The eight years spent in 'Irāq do not cover, in all probability, the whole of the time of his travels; seeing that nothing is said about his life in Balkh before the accession to the throne of the Sāmānid Nar II (a.d. 914-43), one may conclude that he only returned to his birthplace in his old age. To the first years of the reign of Nar II belong, in all probability, Balkhī’s comments, quoted in the Fihrist, [2] about his relations with the general usayn ibn-'Alī al-Marwazī (or Marwarrūdhī) and also with the wazīr Abū-'Abdillāh Muammad ibn-Amad Jayhānī. Balkhī received from usayn and his brother Muammad Su'lūk [3] regular material assistance, but forfeited this subsidy in consequence of having composed a religious treatise, which later was highly appreciated in orthodox circles. Yāqūt [4] quotes an opinion according to which Balkhī’s work was ranked with the most useful, from the Muslim point of view, that had ever been written. (usayn was an Ismā'īlī heretic; Balkhī, too, in his youth held Shī'ite views, which he later abandoned.) The wazīr Jayhānī used to send to Balkhī presents of female slaves, but later deprived him of this attention because of Balkhī’s treatise on sacrifices (al-Qarābin wal-Dhabā'i), which he disliked. The wazīr Jayhānī was suspected of dualism, and some peculiarities of his personal life were connected, in the minds of the people, with his religious opinions: he would not touch a man otherwise than through cloth or paper, and could not suffer the presence of cats. [5]


So far as is known, Balkhī was employed in the service of the State only during the short administration in Khorāsān of the eminent dihqān of Marv, Amad ibn-Sahl (a.d. 918-19), who was held to be a descendant of the Persian kings. [6] Amad was at the head of the Sāmānid troops who quelled the revolt of usayn Marwarrūdhī, and took the latter prisoner. Subsequently, while in Nīshāpur, Amad abandoned the cause of the Sāmānids and was obliged to retreat to Marv where he was defeated and taken prisoner, and later died in the prison of Bukhārā. Amad ibn-Sahl came from a family



1. GMS, vi, 1, p. 147.

2 Fihrist, 138, quoted in Yāqūt, l.c., 141 and sq. In the Fihrist Abū-’Alī stands, by mistake, instead of Abū-'Abdillāh.

3. On him de Goeje, ZDMG, xxv, 54, note 1.

4. GMS, vi, 1, p. 149 below.

5. On this GMS, vi, 6, p. 293, according to Sallāmī’s History of the Rulers of Khorāsān, now lost. On this cf. my Turkestan, ii (Engl. ed., p. 10).

6. On him my Turkestan (Russian ed.), i, 6 and sq., ii, 251 and sq. (Engl. ed. p. 240).





of zealous Iranian patriots; his brothers fell victims to the national fanaticism (ta'aṣṣub) of the Arabs; there lived with Amad in Marv a certain Sarv, to whom Firdausī refers when recounting the lays of Rustam. [1] At that time Abū Zayd tried to keep outside of the national disputes about the relative superiority of Arabs and Persians, as well as outside of the religious discussion of the relative merits of ‘Alī and the other companions of the Prophet. [2] Whatever his own national origin, and whatever his native language, he, as a scholar, spoke the literary Arabic, and in the same language, though with no great success, did the amīr Amad ibn-Sahl [3] try to communicate with him when he arrived in Balkh (there is no other information on Amad ibn-Sahl’s stay in Balkh). When Balkhī declined the office of wazīr offered him by Amad ibn-Sahl, there was appointed to this position a friend and countryman of his, Abul-Qāsim 'Abdullāh ibn-Amad ibn-Mamud Ka'bī, who also wrote treatises of a religious nature, though even farther removed from orthodoxy. In Sam'ānī he

(P 13)

| is called head of the Mu'tazilites. [4] Balkhī took a post as secretary under Ka'bī with an allowance of 500 dīnārs a month. Abul-Qāsim was entitled to a sum of 1,000 dīnārs, but he himself gave orders to the cashier to pay him 900, and to increase Balkhī’s salary to 600, on the express understanding that Balkhī should receive his salary in good coin, while all questionable coins were to be put down to his own account. [5] At that happy time Balkhī, thanks to the generosity both of the amīr and the wazīr, acquired some property in his native village of Shāmistiyān, on the Gharbangī canal (one of the twelve canals irrigating the environs of Balkh), and this property was inherited by his descendants.


After the fall of Amad ibn-Sahl, Balkhī, apparently, lived as a private person on his own lands. Without indication of date [6] it is reported that a Sāmānid amīr (probably Nar) invited him to come to Bukhārā, and that Balkhī declined the invitation, giving as his reason that he was frightened by the violence of the current and the width of the Amū-daryā. Other persons of high rank, with whom Balkhī kept up a correspondence, were the amīrs of Chaghāniyān (later viceroys of Khorāsān), Abū-Bakr Muammad and his son Abū-'Alī Amad, [7] but he seems not to have met them in person.



1. ZVO, xxii, 280.

2. GMS, vi, 1, p. 148.

3. Ibid., p. 150.

4. GMS, xx, p. 485. Ka'bī, who died in 319 (a.d. 931) also wrote historical works; cf. Turkestan, p. 11.

5. GMS, vi, 1, p. 147.

6. Only in Maqdisī, BGA, iii, 4. De Goeje, ZDMG, xxv, 55, refers to Maqdisī and afadī, but the reference of Yāqūt, GMS, vi, 1, p. 152, to Maqdisī shows that afadī borrowed this information through Yāqūt from the same Maqdisī.

7. GMS, vi, 1, p. 143.





The number of Balkhī’s compositions, according to his grandson, [1] was sixty. The geographical treatise of Balkhī, which in all probability (reports are somewhat contradictory) [2] bore the title uwar-al-aqālim (“Images of Climes”), is not expressly mentioned among them. The contents of Iṣṭakhrī’s work, founded, as is known, on that of Balkhī, make one suppose that the title referred not to the division of the habitable world into seven climes from south to north, but to climes as geographical divisions, representing independent entities. Of such climes Iṣṭakhrī enumerates twenty, and the same number appeared in Balkhī. [3] As a matter of fact in the list [4] of Balkhī’s works there are mentioned some titles referring to geographical contents. Such are, for instance, the “Book of the Heavens and the Universe” and a “Commentary on Images” (tafsīr al-uwar). It is possible that by the latter title is meant the geographical work of Balkhī which, according to Maqdisī, was only a very short commentary on Balkhī’s maps. [5]


Already in those times the question of the authorship of the work, which now forms the first volume of the Bibliotheca GeographoRūm ArabicoRūm, was not quite clear. Maqdisī [6] saw only three copies of this work, one—in Rayy, [7] another—in Nīshāpūr, and the third—in Bukhārā. In the first case, the authorship [of the maps ? yunsab ilā Abī Zayd bil-ashkāl. V.M.] was attributed to Balkhī; in the second (in the MS. itself there was no author’s name), to Abū Bakr Muammad ibn-al-Marzubān al-Muawwalī al-Karkhī, who died in 309 (a.d. 921-2); in the third, to Abū-Isaq Ibrahīm ibn-Muammad al-Fārisī al-Iṣṭakhrī, who was named in the MS. itself. Maqdisī considers the last to be the most probable, as he had seen several persons who had known Iṣṭakhrī and witnessed the composition of his work; one of these witnesses was Abū-Nar al-arbī, mutasib of the town of Bukhara. [8] The putative authorship of Karkhī is mentioned



1. Ibid., p. 150.

2. ZDMG, xxv, 37.

3. According to Maqdisī, BGA, iii, 410.

4. Given by Yāqūt (GMS, vi, 1, p. 142 and sq.) more fully than in the printed edition of the Fihrist, p. 138.

5. [In an additional note Professor Barthold gives expression to the view that, in the printed editions, Fihrist, 13821, Yāqūt, GMS, vi, 1, p. 14217, a full stop may have wrongly cut into two the unique title Kitāb tafsīr uwar kitāb al-samā’ wal-ālam li Abī Ja’far al-Khāzin “Book of Interpretation of the Maps of Abū Ja'far al-Khāzin’s Book

of the Heavens and the Universe”. He then proceeds: If this title refers to the geographical work of Balkhī, this could lead to the conclusion that to Balkhī belonged only the explanation of the maps, but not the maps themselves. The astronomer Abū-Ja'far al-Khāzin is often mentioned in Arabic literature, e.g. in Bīrūnī, see Sachau’s Index to his edition of the Chronology.]

6. BGA, iii, 5a.

7. On the library of the minister Ismā'il ibn-Abbād in that town see ibid., p. 391.

8. Ibid., p. 136,2.





again [1] in the chapter on Sind; but in the references and quotations Maqdisī names only Balkhī and Iṣṭakhrī. According

(P 14)

| to de Goeje [2] all the quotations in which Balkhī is named correspond entirely to Iṣṭakhrī’s text. Nevertheless de Goeje thinks it possible that Maqdisī might have had in his hands, besides the text of Iṣṭakhrī, that of Balkhī, [3] but that Yāqūt, on the other hand, was in possession of a single book, and that quoting from this he referred principally to Iṣṭakhrī, but sometimes to Balkhī as well, “as though following a definite system”. [4] This last guess is hardly supported by the facts: Yāqūt refers to Balkhī without mentioning Iṣṭakhrī only once, with regard to the distance between Jedda and 'Aden; [5] the corresponding words are of course to be found also in Iṣṭakhrī. [6] In all the other cases Iṣṭakhrī alone is quoted, e.g. with regard to the distance between aramut and 'Aden. [7] Consistency, which de Goeje vainly seeks in Yāqūt, can be discovered only in Maqdisī: with regard to three out of the twenty climes mentioned, viz. the last three: Khorāsān, Sīstān, and Mā-warā’ al-nahr, Balkhī is preferentially quoted; while in three others, Fārs, Kirmān, and Sind, preference is given to Iṣṭakhrī.


In de Goeje’s [8] opinion the work of Iṣṭakhrī represents a second and greatly enlarged edition of Balkhī’s work, compiled between 318 and 321 (a.d. 930-3), i.e. in Balkhī’s lifetime. In Russian works [9] the date 340 (a.d. 951) is often attributed to Iṣṭakhrī’s work, but according to de Goeje [10] this was the date of the MS. which was the basis of most of the copies circulating in the East; at that date the work, composed twenty years earlier, was published. De Goeje places Iṣṭakhrī’s meeting with Ibn auqal at the same date. [11] The meeting is confirmed by Ibn auqal himself, who, with Iṣṭakhrī’s consent, undertook the revision of his work. [12] Unfortunately, Ibn auqal does not say a word as to when and where this meeting took place, and only mentions [13] that by that time he had already compiled a map of Ādharbayjān [14] and Mesopotamia.


Ibn auqal intended to give at the end of his work a full synopsis



1. Ibid., p. 47510; cf. the interpretation of the text, ibid., p. 5a, in fine, as against ZDMG, xxv, 48. Grammatically, however, the previous interpretation seems more natural.

2. ZDMG, xxv, 47.

3. Ibid., p. 52.            4. Ibid., p. 46.

5. Yāqūt, ii, 41,7.

6. BGA, i, 27 above.

7. Yāqūt, ii, 28519; Iṣṭakhrī, 273.

8. ZDMG, xxv, 50.

9. Cf. e.g. Toumansky’s article, ZVO, x. 127.

10. ZDMG, xxv, 51 and sq.

11. Ibid., p. 48 and 51 (below): in the one place: “vermutlich nicht später als 340”; in the other: “schwerlich früher”, though the same date is meant.

12. Whether Iṣṭakhrī in his time had received a similar consent from Balkhī, and whether he had met him at all, is not known.

13. BGA, ii, 2360.

14. In the final edition of Ibn auqal’s work, Ādharbayjān, as in Iṣṭakhrī, is represented on the same map as Armenia and Arrān. [Cf. our § 35.]





of his travels, [1] but never carried out his intention; the only definite date given is that of his departure as a young man from Baghdād (Thursday, Ramaān 7, 331, i.e. in May, a.d. 943); [2] otherwise it is merely said that he visited certain towns in certain years. The year of the termination of his work is held to be 367 (a.d. 977-8). [3] During such a lapse of time Ibn auqal could evidently visit

(P 15)

| the same towns several times; thus in 358 (a.d. 968-9) he was in Mosul for the last time. [4] He wrote his work as a subject of the Fāimid caliph, and apparently spent in the West the years preceding the completion of his book, since in 361 (a.d. 971-2) he was in Sicily. [5] This may account for the fact that his work did not acquire, in the Eastern parts of the Muslim world, the same fame as that of his predecessor. Only the work of Iṣṭakhrī was translated into Persian; the manuscript which Sir W. Ouseley took for a copy of the translation of Ibn auqal and edited as such was found to be an abridged version of Iṣṭakhrī’s book. [6] The acquisition by the library' of Shāhrukh in the fifteenth century of a copy of the Arabic original of Iṣṭakhrī gave an impulse to the composition in Persian of the geographical work of Hāfiz-i Abru. [7]



1. BGA, ii, 2365-6.

2. Ibid., 518. It is apparently not quite exact, as the day of the week does not correspond to the date. If instead of khalauna one reads baqīna, the date would be 1 June, 943,but such a supposition would be untenable. Ibn auqal adds that on the same day the amdānidir al-daula left Baghdād, fleeing from the Turks. Nāir al-daula became chief amīr of Baghdād (amīr al-umarā) on Sha'bān 1, 330 (a.d. 21. iv. 942); the same date in Ibn-Miskawaih, Eclipse of the Abbasid Caliphate, ii, 28, and in Ibn al-Athīr, viii, 286. His rule came to an end as a result of the battle mentioned by Ibn auqal; it had lasted, according to Ibn-Miskawaih (Eclipse, ii, 41), 13 months and 3 days; according to Ibn al-Athīr, 13 months and 5 days, which brings us in any case to the first days of Ramadan 331; it is possible that Thursday, Ramadān 3, is meant (a.d. 11.v.943). In Zambaur (Manuel, p. 9) the day of Ramaān 7, 331, is given as the date of the passing of the power from Nāsir al-Daula to the Turk Tuzūn; but, according to Ibn al-Athīr (viii, 298), Tuzūn received from the caliph the title of amīr al-umarā only on Ramaān 25.


3. In de Goeje’s opinion, BGA, iv, p. v, this date results from the mention, BGA, ii, 2015, of the khuba, read “last year”, by the Ziyādid ruler of Yemen, (Lane-Poole, The Muhammadan Dynasties, Russian transl., p. 72; Zambaur, p. 115), in the name of the Fāimid caliph. But de Goeje fails to say at this place (and apparently at any other) in what source he found the change of the khuba in Yemen in 366. In Ibn al-Athīr there is no information on the subject. Sir J. W. Redhouse in his Introduction to Khazraji’s History of Yemen (GMS, iii, 1, p. 11) says that when in 377 (a.d. 987) “the Hiwāliyy ruler of San'ā” took Zabīd, he suppressed the khuba in the name of the Fāimids. Reinaud (Geographie d'Aboulféda, Introd., p. lxxxiii), without proofs, places the termination of Ibn auqal’s work in 366 (a.d. 976) [evidently after Uylenbroek, Specimen, p. 15?].


4. BGA, ii, 1463; see above.

5. Ibid., 2214-16.

6. Rieu, Pers. MSS., p. 416.

7. Al-Muaffariya (a volume dedicated to Baron Victor Rosen by his pupils), pp. 3, 13, and 18





It is beyond doubt that our author had before him a copy of the work of Balkhī or of Iṣṭakhrī. This is particularly evident in the chapters devoted to the western provinces; for instance, the words of our author (f. 34b) on Malaya correspond perfectly with Iṣṭakhrī’s text, p. 62:



The text of Iṣṭakhrī is sometimes not very exactly rendered; our author calls Mārida “the greatest town of Andalus” (36b), while Iṣṭakhrī, p. 43, speaks of it as “(one) of the greatest towns of Andalus”. A quotation from Iṣṭakhrī, p. 68 and sq., not exactly understood, accounts for what our author says of Athens. South and west of Constantinople Iṣṭakhrī distinguishes the Athenian and the Roman shores, but the words “Athens” (Athīnās) and “Rome” (Rūmiya) remain to him names of towns; of Athens, as a town, it is said that there was “the residence of the wisdom of the Greeks” (yūnāni-yūn). According to our author (f. 37b), “the Athenian coast” included the entire sea-coast from the strait of Constantinople (Bosphorus) to Andalus (Spain); he knows “Athīnās” only as the name of a locality where in ancient times there stood a town Yūnāniyān, and, as he says, “all the wise men and philosophers rose from this region (iyat) of Athīnās”.


Historical facts are likewise now and then borrowed from Iṣṭakhrī (or Balkhī). In the chapter on mountains (f. 7b), as in Iṣṭakhrī’s account of the Arabian peninsula, a mountain is mentioned, the summit of which occupied an area of 20 farsakhs in circumference, where there existed cultivated fields and running water, and, also as in Iṣṭakhrī, it is said that the locality was conquered by the Qarmaian Muammad ibn-al-Fal. According to our author this event took place “in ancient times”, which is not very accurate, in that it refers to an event of circa 300 H.; [1] but perhaps the words andar qadīm, used also two lines above with regard to the ancient capital of the Yemen kings, were repeated by a clerical error. Some passages in our author more nearly resemble Ibn auqal than Iṣṭakhrī. In the chapter on Irāq (f. 31b) Qar ibn-Hubayra is called the largest town between Baghdād and Kūfa; these words

(P 16)

| do not figure in BGA, i, 85, but they exist in BGA, ii, 166. Of course one cannot conclude from this



1. The Qarmaian Muammad ibn-al-Fal, apparently, the brother of the Qarmatian 'Adī ibn-al-Fal, who sacked Zabīd according to Lane-Poole, The Muammadan Dynasties, 1894, p. 90, shortly after 292/904, and according to Zambaur, Manuel, p. 115, in 303 H.





that our author utilized Ibn auqal’s original; in this case, as in many others, Nöldeke [1] is right in saying that Ibn auqal’s relation to Iṣṭakhrī cannot be determined by a simple comparison of the two texts of BGA, i and BGA, ii. The missing words of Iṣṭakhrī’s text, as published by de Goeje, are to be found in the abridged version (Gotha MS.) and in the Persian translation edited by Ouseley; consequently they were undoubtedly in Iṣṭakhrī.


The terms of our author’s description (26a) of three Muslim colonies on the lower reaches of the Sı̊r-daryā [2] corresponds almost literally to Ibn auqal’s text, p. 393. In BGA, i, these colonies are not mentioned; but it is sufficient to compare Iṣṭakhrī’s text in de Goeje’s edition, p. 333, with Yāqūt’s [3] quotation from Iṣṭakhrī, to be convinced that the course of the Sı̊r-daryā was described in Iṣṭakhrī with much more detail than in the de Goeje edition. [4]


Did our author have before him Balkhī’s work in its primitive form, or in Iṣṭakhrī’s version? Some passages apparently show the influence of those chapters of BGA, i, which are principally attributed to Iṣṭakhrī, e.g. the chapters on Sind and especially the description of Manūra (f. 26a), cf. Iṣṭakhrī’s text, p. 173. But this passage may also have stood in Balkhī. The dependence on Balkhī-Iṣṭakhrī is still more noticeable in the chapters of the Ḥudūd al-’ālam dealing with Khorāsān and Transoxiana which, in the original, belong undoubtedly to Balkhī: ride the passages on the Herat mosque and the number of people who spend their time there (f. 19b, cf. Iṣṭakhrī, p. 265); the description of Būshang (ibid., cf. Iṣṭakhrī, p. 270); the account of the river Murghāb, which crosses the village Diza (f. 20a, cf. Iṣṭakhrī, p. 270); the account of the three Buttam (or Butmān) mountain chains (f. 23b, cf. Iṣṭakhrī, p. 333); the account of the outposts at Osh (f. 24a, cf. Iṣṭakhrī, p. 333); and the account of Khatlām or Khaylām as being the birthplace of the amīr Nar (ibid., cf. Iṣṭakhrī, p. 334). In two instances, namely in the accounts of the market in Marsmanda (f. 23b, cf. Ibn auqal, p. 3845) and of sixty villages near Sokh (f. 24a, Ibn auqal, p. 3961), our author’s words can be compared only with the text of BGA, ii, because in BGA, i, the corresponding passages of the Balkhī-Iṣṭakhrī text have



1. ZDMG, lvi, 433.

2. Barthold, Turkestan, ii, 179, Engl. transl., p. 178; Barthold, History of Irrigation in Turkestan, SPb. 1914 (in Russian), p. 149.

3. Yāqūt, ii, 404 and sq. The text in Yāqūt, as de Goeje points out in a footnote, BGA, ii, 393c, is clearly corrupt, especially 4055 where instead of fa yamtaddu 'alā al-atrāk al-Ghuzziya one must read fa yamtaddu ilā al-qaryat al-adītha.

4. De Goeje himself says that in BGA, i, he gives only the abridged text of the description of Transoxiana “während der eigentliche Text des Iṣṭakhri bei Ibn auqal und in den Anmerkungen dazu zu finden ist”.





not come down to us. Apparently, among the passages of the Balkhī-Iṣṭakhrī text, that have distinctly influenced our author, there is none that from a chronological point of view could belong to Iṣṭakhrī alone. allāj, who was executed in 309 (a.d. 922), is mentioned by both our author (f. 28a) and Iṣṭakhrī (p. 148 and sq.), but he would hardly have been named by Balkhī. Yet it is possible that our author had another source in this case, as allāj is mentioned by Iṣṭakhrī elsewhere than in the description of allāj’s native town al-Bayā.


Moreover, the question whether or not Balkhī’s version has been preserved in Arabic MSS. along with that of Iṣṭakhrī would now require a fresh consideration. De Goeje has proved very convincingly [1] that the MSS. that were at his disposal, inclusive of the Berlin MS. (which Brockelmann [2] in spite of de Goeje still ascribes to Balkhī),

(P 17)

| all contained Iṣṭakhrī’s version. But since then certain new MSS. attributed to Balkhī have been discovered; viz. the MS. acquired in Egypt by Amad Zakī bey [3] and the “Balīkodex mit schönen Karten”, acquired in Baghdād for the Hamburg library: [4] the necessary evidence as to the extent to which their texts differ from that of BGA [5] is still to be given.





A geographical work, under the same current title of “Book of Routes and Kingdoms”, was written by the Sāmānid wazīr Abū Abdillāh Muammad ibn-Amad Jayhānī, who is mentioned in Balkhī’s biography. References to this work are often met with, but the work itself seems to have completely disappeared. Among the geographers of the tenth century who utilized it are Ibn auqa [6] and Maqdisī. It can be seen from the latter’s [7] comments that Ibn Khurdādhbih’s work formed the basis of that of Jayhānī. Occasionally the same MS., if it did not contain an indication of the author’s name, was attributed by some to Ibn Khurdādhbih, and by others to Jayhānī. But it can be gathered from Maqdisī that Jayhānī,



1. ZDMG, xxv, 42-58. The final conclusion, p. 57, is that both MSS., taken as the basis of the edition, viz. the Bologna one (on which see V. Rosen, Remarques, &c., Rome, 1885, p. 94) and the Berlin one (in printed editions respectively A and B), transmit if not entirely, at least in its greater part, the work of Iṣṭakhrī.

2. GAL, i, 229.

3. Ign. Kratchkovsky, Abū Hanīfa ad-Dīnawarī, Leiden, 1912, p. 24.

4. C. Seybold in ZDMG, lxvii, 541.

5. In the summer of 1929 when the present work had already gone to press I had the opportunity of examining the Hamburg MS. Like the Berlin MS., it proved to be the work of Iṣṭakhrī, not of Balkhī. It also contains the famous story of the author’s stay in Samarqand (BGA, i, 318), which could not belong to Balkhī, who, according to the direct evidence of Maqdisī (BGA, iii, 414), never crossed the Oxus.

6. BGA, ii, 2362, with an unfavourable mention both of Jayhānī’s work and of that of Ibn Khurdādhbih.

7. Cf. my Turkestan, p. 12 and sq.





besides written sources, utilized oral information; he assembled foreigners and bade them speak of their native lands and of the roads leading thereto. Thus, according to Maqdisī, [1] it was a 140 days’ journey from Tūnkat [2] to the principal town of China, “as Jayhānī was told by the ambassadors, and he mentioned this in his book and clearly expressed it in his statement”.


Unfortunately this itinerary has not come down to us either through Maqdisī or any other author; I have not met with quotations from it. But one might suppose a priori that the great number of geographical names belonging to Central Asia and found in our author shows the latter’s dependence on the itinerary given by Jayhānī. It is somewhat difficult to determine the extent of such a dependence, seeing that our author does not give any itineraries; but many of the geographical names of the Ḥudūd al-’ālam are also quoted by an author of the eleventh century, Gardīzī, [3] who gives the distances between the towns and the itineraries, i.e. precisely the information missing in the Ḥudūd al-’ālam. Gardīzī states that he borrowed these data from Ibn Khurdādhbih, Jayhānī, and a third anonymous work under the title Tawāu al-dunyā. [4] At one place [5] Gardīzī’s expressions literally coincide with the quotation from Jayhānī found in Bīrūnī, [6] though referring not to the route to China, but to that from Khotan to Tibet. According to Bīrūnī, Jayhānī said that


“the Chinese in ancient times built a bridge from the summit of one mountain to the summit of another, on the way from Khotan into the province of the Tibetan [7] Khāqān; whoever crosses this bridge enters the locality where the air impedes respiration and renders the tongue heavy; many of

(P 18)


those who pass there die from this, but many recover as well. The Tibetans call it Mountain of Hell.”


The same passage (of course in Persian translation) is found almost verbatim in Gardīzī, where the building of the bridge is attributed to the people of Khotan, which perhaps can be explained by a mistake of the copyist . Evidently mountain sickness is



1. BGA, iii, 346 (345b).

2. South of Tashkent, on the river Āhangarān (in Russian: Angren). But it is possible that instead of Tunkat one should read Navīkat, the town of Navākat or Navīkat being the starting-point of several routes to China, V. Barthold, Report, p. 114, note 1.

3. V. Barthold, Report, See., pp. 78-126.

4. Ibid., p. 103 (text) and p. 126 (transl.). [Barthold translates Tawāu by “Insignificance”, “Frailty”, which is rather a strange name for a geographical treatise. It is more probable that the book bore the name *Rab' al-dunyā, “Habitable part of the World”, as indicated by a variant, see M. Nāzim’s ed. of Zayn al-akhbār, Berlin, 1928, p. 4. V.M.]

5. Ibid., p. 88 (text) and p. 112 (transl.).

6. Chronologie, ed. Sachau, 271.

7. Instead of byt, read Tbbt.





meant here, which even now hampers traffic along the high passes leading from Eastern Turkestan into India. [1]


It is hard to say how far such a specific dependence of Gardīzī on Jayhānī confirms a similar dependence of the Ḥudūd al-’ālam on Jayhānī, for in the passage on Tibet [2] there is less resemblance between the text of the Toumansky MS. and Gardīzī than in such other passages as those on the Toghuzghuz and China. Gardīzī mentions none of the Tibetan towns, while the Ḥudūd al-’ālam names a whole series of them and even attributes to Tibet several towns placed by Gardīzī on the way from Kāshghar to Khotan, [3] though the town of Khotan itself (as in Gardīzī) is described in the chapter on China (f. 14a); moreover, Khotan is placed on the frontier between China and Tibet, and the title “Chief of Turks and Tibetans” [4] is attributed to the Khotan ruler.


A passage at the beginning of the chapter on Tibet in the Toumansky MS. (on the involuntary gaiety felt by every one entering Tibet) is clearly borrowed from Ibn Khurdādhbih. [5] The data on the Tibetan towns are apparently derived from various sources; two names, Lhāsā and Krsāng, [6] designating, it seems, the same town, are given as names of two different towns. As in all compilations, such examples are fairly frequent in the Ḥudūd al-’ālam, which not only refers to the Burās and the Barādhās as two distinct nations, [7] but also, in the chapter on Khūzistān (f. 28b), separately mentions Rāmhur (?), i.e. Rāmhurmuz (as in Balkhī-Iṣṭakhrī, [8] place of the assassination of Mānī), and Rām-Urmuzd, a large and rich commercial town on the frontier between Fārs and Khūzistān, though it is evident that the latter is only a more correct and fuller Persian form of the first name. In the chapter on Khūzistān the name of another large town is given in the Persian pronunciation, Vandūshāvur, instead of in the Arabic, Junday-Sābūr; the spelling of the Toumansky MS. approximates very closely to the pronunciation Vandēw-Shāpūr proposed by Nöldeke. [9] In the data borrowed from Balkhī-Iṣṭakhrī the author sometimes substitutes a local Persian form for the literary



1. Kornilov, Kashgariya (in Russian), Tashkent, 1903, p. 349.

2. Cf. Doklady Akademii Nauk, Series B, 1924, p. 73 and sq.

3. V. Barthold, Report, p. 94 (text) and p. 119 (transl.); Tadrūf and Rastūya, mentioned there, figure in the Toumansky MS. among the towns which formerly belonged to China, and “now” belong to Tibet.

4. Cf. the title 'Aīm al-Khotan in a twelfth-century author, Turkestan, i, 20a.

5. BGA, vi, 1709.

6. Muammad aydar (sixteenth century) gives Arsāng, cf. Ta’rīkh-i Rashīdī, transl. by E. Denison Ross, London, 1895, pp. 136 and 411.

7. [But see p. 44, note.]

8. BGA, i, 93. [See my notes. V.M.]

9. Nöldeke, Geschichte der Perser und Araber, p. 42.





Arabic one; thus in the passage on Herat (f. 19b), evidently borrowed from Balkhī-Iṣṭakhrī (p. 265), he writes Harī instead of Harāt.


The similarities of the Ḥudūd al-’ālam and Gardīzī, and the dependence of both on a common source, are perhaps most obvious in the chapters on China and the Toghuzghuz. Many names of towns, and among them the Persian names Baghshūr and Sangalākh, [1] occur in both authors as names of localities between Turfan and Khami and between Sha-chou and Su-chou. Gardīzī gives itineraries which are not to be found in the Ḥudūd al-’ālam, but the sequence of names in both clearly shows that the itineraries were also known to the latter, who in accordance with the general plan of his composition enumerates the towns

(P 19)

| from east to west, while Gardīzī does it in reversed order. [2] There is no complete correspondence between the text of the anonym and that of Gardīzī; the former (f. 14a) mentions a “stone tower” (burj-i sangīn), which seems to be the only trace of influence of Ptolemy’s geography in this locality (i, 12, 9, λίθινος πύργος, cf. Qudāma, [3] burj al-ijāra); in Gardīzī no such name occurs. On the other hand, the Ḥudūd al-’ālam contains no mention of Qamul or Khami, which is found in Gardīzī. [4] From this we may conclude that at this place Gardīzī reflects a later stage of geographical knowledge; and it is possible that here our author depends on Ibn Khurdādhbih, and Gardīzī on Jayhānī. In any case our anonym’s information cannot be up to his own epoch, or even to that of Jayhānī. Particularly characteristic is the description of the town of Kan-chou (Khāmchu, f. 13b and sq.; same in Gardīzī):


“Half of it is owned by the Chinese, half by the Tibetans; a perpetual war goes on between them; they are idol-worshippers; their government [5] is on behalf of the Tibetan khāqān.”


Such could have been the situation in the times of Ibn Khurdādhbih, or in those of him whose work was his source, the traveller Tamīm ibn-Bar al-Muṭṭawwi'ī; [6] but during the whole of the tenth century Kan-chou was an Uyg’nur principality. [7]


In no greater degree does our anonym reflect the situation in India in the tenth century. The original source of Ibn Khurdādhbih and



1. Ḥudūd al-’ālam, f. 13b and f. 14a; Gardīzī in Barthold, Report, p. 92 (text) and p. 117 (transl.).

2. Cf. e.g. the itinerary in Gardīzī, text p. 91, transl. p. 116, and the order in which the towns are enumerated in the Ḥudūd al-’ālam, f. 17a.

3. BGA, vi, 26415, cf. J. Marquart, Ērānšahr, p. 316. Text in Bīrūnī, India, ed. Sachau, p. 14913; cf. Ērānšahr, p. 155. In Muammad uwārizmī (ed. Mžik, No. 865): burj ijāra.

4. Report, p. 92 (text) and p. 117 (transl.). [But see note to § 12, 9.]

5. [Barthold translates: sulān, but in the tenth century this word meant more probably “government”. V.M.]

6. Yāqūt, i, 840, above; Barthold, Report, p. 34.

7. Cf. Bretschneider, Medieval Researches, i, 241 and sq.





other early Arabian geographers was in this case provided by the work of the traveller Abū 'Abdillāh Muammad ibn-Ishaq, who lived two years in Khmer (Qimār), i.e. Cambodia; [1] he it was who originated the passage (f. 14b) on the strict forbiddance of adultery in that country, [2] as well as the story of the woman who ruled Orissa. The name Orissa occurs in two forms: Ūrsfīn [3] and Ūrshfīn (in the story of the queen). [4] The work of Balkhī-Iṣṭakhrī was utilized for instance in the passage dealing with the Arab rulers of Multān and with the town Bābī (in Iṣṭakhrī, Bāniya). According to Iṣṭakhrī [5] the khuba in Multān was read in the name of the caliph; according to Ibn auqal [6] in the name of the 'Abbāsids whom Ibn auqal, writing in the kingdom of the Fāimids, did not recognize as caliphs; according to Maqdisī, in the name of a Fāimid. [7] Our anonym (f. 15a) says that the khuba was read in the name of Mu'izzī (bar Mu'izzī); but it is not clear whether we have to do here with a clerical error, the possibilities being Mu'izz (the Fāimid caliph who ruled from 953 till 975), and “Mu'izzī” which might designate the son and successor of Mu'izz, the caliph 'Azīz. The acceptance of either of these hypotheses [but see my translation and note, V.M.] would bring us to the conclusion that the anonym, perhaps from oral sources, knew of the Shī'ite coup d'etat in Multān which evidently took place after Ibn auqal and before Maqdisī. It is known that Multān remained in the hands of the heretics till its conquest by Mamūd the Ghaznavid in 1006; [8] the epitomizer of Ibn auqal, who wrote in the twelfth century, gratuitously supposed, in order to explain Mamūd’s expedition, that Multān, after Ibn auqal, had again passed for a certain time into the hands of the Hindus.


In the chapters on Central Asia and China there are no indications of events that could have taken place a short time before the work was composed. It is possible that here, too, as in many other instances, the use of different sources made the author mention the same localities under different names. The town Panchul (Bnjūl), Wen-su of the Chinese sources, was situated probably on the site of the present-day

(P 20)

| Uch-Turfan, [9] as confirmed by the Chinese source in which it was stated that this town bore the name of Yü-chou. Gardīzī



1. BGA, vii, 132.

2. Ibid., and vi, 66 and sq.

3. Together with Smndr, as in Ibn Khurdādhbih, p. 642, who gives Ūrn-shīn. [But see my note, p. 243. V.M.]

4. As in Ibn Rusta, p. 13413.

5. BGA, i, 175.

6. Ibid, ii, 2304.            7. Ibid, iii, 4853.

8. On this 'Utbī-Manīnī, ii, 72; Elliot, History of India, ii, 441.

9. E. Chavannes, Documents sur les T’ou-Kiue (Turcs) occidentaux, SPb. 1903, p. 9, placed Wen-su on the site of Aqsu, but later (in M. A. Stein’s Ancient Khotan, p. 544) adopted the opinion that Wen-su was Uch-Turfan. [See my notes, pp. 294-7, V.M.]





uses the name Bnchūl, [1] but not Ūj; Mamūd Kāshgharī (eleventh century) [2] gives Ūj but not Bnchūl, whereas our author (18a) names Bnchūl and Ūj separately, with the additional remark, absent in other sources, that Bnchūl “now” belongs to the Khirkhiz. This detail can hardly pertain to the times of the author, since it can only reflect the situation at the time of the Qirghiz empire, at the end of the ninth and the beginning of the tenth centuries, [3] but could have been incorporated in the works both of Ibn Khurdādhbih and Jayhānī.


In the chapter on lakes are mentioned side by side (f. 3b) the lake of Tuz-kul, from which seven tribes of the Qarluq procured salt, 10 farsakhs long and 8 farsakhs wide, in the country of the Khallukh (Qarluq), and the lake of Issik-kul, lying between the possessions of the Chigil and the Toghuzghuz, which was 30 farsakhs long and 20 farsakhs wide, and on the shore of which was situated the town of Barskhān. In spite of the different estimate of the size of the lakes, [4] it is very probable that the first name Tuz-kul also designates Issik-kul. [5] Nallino, [6] on the strength of a quotation in Kharaqī, an author of the twelfth century, has shown that Issik-kul was mentioned by Jayhānī; the same quotation from Jayhānī, unnoticed by Nallino, exists in Yāqūt, II, 224, where some copyist substituted for the name of Issik-kul that of the port Abaskūn on the Caspian.


In the few cases where the author mentions events of his own times these events pertain to the history of the Muslim world. In the year 372, in which the author completed his work, there died the Būyid shāhānshāh Fanākhusrau (in our author Panākhusrau) who bore the title 'Aud al-daula (d. 8 Shawwāl 372 = 26 March 983); [7] his massacre of the Baluches, an event mentioned by our author (f. 26b) and by Maqdisī after him, [8] probably belongs to the end of his reign. [9]



1. V. Barthold, Report, p. 91 (text) and p. 116 (transl.).

2. Mamūd al-Kāshgharī, Constantinople ed., i, 38; he several times (i, 335, 381; ii, 121) cites words from the dialect of its inhabitants.

3. V. Barthold, A Sketch of the History of the Semirechye, p. 19; Barthold, The Kirghiz, Frunze [= Pishpek], 1927, p. 19.

4. Gardīzī attributes to Issik-kul a still greater size, viz. of 7 days’ journey, see Barthold, Report, p. 89 (text), p. 114 (transl.).

5. Cf. Report on Capt. If. Unkovsky’s Embassy (1722-4), ed. Veselovsky, pp. 187 and 193, and the map appended to it on which Issik-kul is called Tuskel ( < Tuz-köl.) [But see my note. V.M.]

6. Battānī, i, 172 and 175. [But Battānī, p. 169, quotes al-Jayhānī, wa ghayra-hu min al-'ulamā. V.M.]

7. Ibn al-Athīr, ix, 13; Zambaur, Manuel, 202.

8. BGA, iii, 4895.

9. Ibn auqal (BGA, ii, 22112) speaks only of the victory which 'Aud al-daula, with the help of the Balūches, won over the Kūfich; one must suppose that the rupture with the Balūches occurred later. [It is doubtful that 'Aud al-daula assumed the title of shāhānshāh. As to the crushing defeat of the Balūches by 'Aud al-daula, it took place in 361 January 972, see Ibn Misakawaih, The Eclipse, ii, 299-301. V.M.]





In the description of the town of Qum (f. 29a) it is said that the secretary (dabīr) Bul-Fal, the son of 'Amīd, was a native of that place. The person here meant is the famous Būyid minister Abul-Fal ibn al-'Amīd, [1] who died in Hamadān on the night of Thursday, [2] 6 afar 360 (8 December 970). Quite singular is the mention by our author (33a), in the chapter on Ādharbayjān, Armenia, and Arrān, of the large village of Mubārakī which was situated at the gates of Barda'a where “the camp of the Russians (Rūs) stood, at the time when they seized Barda'a”, and where they were afterwards besieged, an event which, as is known, happened in 332 (a.d. 943-4), [3] forty years before the composition of the Ḥudūd al-’ālam.


The mention of the Russian raid is perhaps corroborative of a fact which I have pointed out elsewhere, [4] viz. that the Caspian provinces are described by our anonym with particular detail. Here he gives us a whole series of details

(P 21)

| which one would vainly look for elsewhere. The same remark applies in part to his description of the southeastern shore of the Caspian; especially worthy of attention is the fact, apparently not mentioned in other sources, that two languages were spoken in Astarābād (f. 29b). However, it is evident that our author knows the eastern Caspian shore less than the western one, for in the description of the former several notable inaccuracies occur. In two places (f. 11a and f. 29b) the river Hirand is mentioned as rising in the mountains of ūs, traversing the confines of Ustuvā and Jarmukān, [5] flowing between the two parts of the town of Gurgān, [6] then directing itself towards the town of Abaskūn, and finally emptying itself into the Khazar (Caspian) Sea. One sees that the upper course of the Atrak has been confounded with the lower course of the Gurgān, to form one river. (It is remarkable that the river Atrak, in spite of the fact that its waters irrigated the town of Dihistān and its environs, [7] is not mentioned by the tenth-century geographers.) If, therefore, the anonym’s data upon the Caspian provinces were borrowed from one source, it is probable that this source was composed not in the eastern, but in the western part of the Caspian region.


With less geographical detail are described the provinces of modern



1. Cf. his biography by Amedroz (from Ibn-Miskawaih) in Der Islam, iii, 323 and sq.

2. Ibid., p. 346; in the translation, p. 339, by mistake Wednesday. The correct translation (night of Thursday) in The Eclipse of the ’Abbasid Khaliphate, v, 293.

3. Cf. Yakubovsky in Vizantiiskiy Vremennik, xxiv, 63-92; the Ḥudūd al-’ālam quoted, ibid., p. 91.

4. Izv. Kavkaz. Instituta, vi, 63 and sq.

5. J.rm.kān, f. 11a and f. 19b, BGA, iii, 300 below and 320 J.rmūkān; on its site, ibid., p. 352, where J.rm.qān is placed at three stages from Nasā.

6. Cf. BGA, ii, 273; iii, 358 and sq.

7. V. Barthold, The History of Irrigation in Turkestan, p. 32.





Afghanistan, more closely related to the author, but the fact is interesting that the Afghāns (Afghānān, f. 16a) are mentioned as a people; until now 'Utbī [1] was considered the oldest author mentioning this ethnographical term (al-Afghāniya). Particularly little information is given on that part of Afghānistān where, at that time (since a.d. 977) ruled Sabuktagin, founder of the Ghaznavid dynasty, later so powerful. [2] Apparently, the text here has been somewhat corrupted by the copyists. The name of the town of Ghazna occurs several times in its usual form “Ghaznīn”, but it is not impossible that to this same town may pertain the words about the rich commercial town Ghazaq, which at first belonged to India but later passed under the sway of Islam and formed the frontier between the possessions of the Muslims and those of the Infidels (f. 22a).





From the Preface translated by Toumansky [3] Russian readers could form an idea of the author’s conception of his task. Not quite clear is the title chosen, Kitāb udud al-ālam min al-mashriq ilā al-maghrib, which in Toumansky’s translation is rendered “The Book of the Frontiers (or Limits) of the World from East to West”. [4] The second variant of the translation (“the limits”) is apparently the more correct, though in Toumansky’s mind it was perhaps connected with the peculiarity of the work in which “for each province the frontiers are given first of all”. The word udūd in Arabic geographical literature means not so much “frontiers”, in the sense of frontier-line, as “limits”, in the sense of the total extent of a territory. In Ibn Khurdādhbih’s words, [5] Ptolemy abāna al-udūd, which de Goeje translates “a donné une bonne description”. [6] However, in the description of two provinces Khorāsān and Transoxiana, our author uses the word udūd in some special and not very clear sense. Separately from the description of the provinces themselves are described their udūd, and of the Sāmānids, the rulers of the whole country, it is said (f. 19a):


“In the whole of Khorāsān are their lieutenants, while on the frontiers (andar add-hā) of Khorāsān there are kings, who are called margraves (mulūk-i arāf).” [7]


If the author

(P 22)

| meant by this that in the chapter on Khorāsān would be described the provinces under the immediate rule of the Sāmānids, and that



1. Enc. of Islam, under Afghānistān (M. L. Dames). Cf. also in 'Utbī the chapter on the Afghāns, 'Utbī-Maninī, ii, 300. [But cf. p. 349, n.2.]

2. Barthold, Turkestan, ii, 274, Engl. ed., p. 261.

3. ZVO, x, 127.

4. Ibid., p. 125, “frontiers”; p. 128, “limits”. [I translate “regions”. V.M.]

5. BGA, vi, 39.

6. Ibid., p. 1.

7. [Barthold translates andar add-hā “within the limits”, but this expression means: “on the frontiers”; on the other hand he renders mulūk-i arāf by udelniye praviteli (“vassal rulers”) whereas I prefer the term margraves as better expressing the conception of “princes of the periphery (arāf)” of the original. V.M.]





in that on the “limits [read: ‘marches’, V.M.] of Khorāsān” would be described the vassal principalities, then he did not adhere consistently to this distinction. Enough to say that the possessions of the amīr of Gūzgān, the most important of the vassal rulers, are included in Khorāsān proper (f. 20b), and not in “the limits [read: “marches’, V.M.] of Khorāsān”.


In spite of the relatively insignificant size of the Ḥudūd al-’ālam, as compared with the works of the Arabic geographers of the tenth century, it was meant to contain all data “that became known until then” on the countries and kingdoms of the world, i.e. all that could be learnt from books or from the words of learned men. [1] Such a claim, expressed in the Preface, is repeated in the text in the passage where the author passes from the physico-geographical description of the inhabited world to that of separate kingdoms and towns, with the reservation that “all the particulars of the world may be known to none, save God” (f. 13b). In various other passages the same assurance is expressed as to the fullness of the information given. At the end of the chapter on freshwater lakes (f. 4a) it is said:


“These are the lakes that are known and on which books give information; besides these, there are numerous small lakes, of which one is in the mountains of Gūzgān in Mānishān, near Bistarāb; [2] its length is one farsakh, its width half a farsakh. There are similar lakes in the mountains of ūs and in the mountains of abaristān; but these lakes are not known and are not ancient; or it happens [read: va yā vaqt buvadh, V.M.] that they dry up so that there remains no water in them; therefore we have not mentioned them.”


The same reservation is further made where swamps (batīa-hā) are described. The chapter on islands ends with the words (f. 5b):


“There is no other reputed and inhabited island in the whole world, besides those that we have mentioned; we have represented (on the Map) [3] all these seas, gulfs, and islands, as they are and at their respective places.”


At the end of the chapter on deserts and sands (f. 13a) it is said:


“In the limits of the Muslim world the large and known deserts and sands are those which we have mentioned; in the lands of the Infidels, except (those) of the Turks, they are also such as we have mentioned, and God knows best of all and from Him is assistance.”



1. To the not very clear words of the Preface: andar gird-i akīmān [I read: yādkird-i akīmān, “memories of the sages”, V.M.] correspond on f. 13b the words: ba-akhbār-hā shanīdan.

2. On the district Mānishān see also the description of Gūzgān, f. 20b, the townāb is not mentioned there.

3. Toumansky, ZVO, x, 128, had already noticed that the text mentions the Map which is absent in our copy.





In other words, the author admits the possibility of not having enumerated all the deserts and sands of the country of the Turks (where they are most numerous); but for the rest his list seems to him absolutely complete.


With the tendency towards completeness is connected a tendency towards numerical exactitude; the author tries to give the precise number of seas, salt and freshwater lakes, islands, countries into which the inhabited part of the world is divided, &c. As far as it is possible to judge by the Arabic geographical works that have come down to us, the author is largely independent in his geographical generalizations and terminology. The conception of the seven seas, as developed by our author (Eastern Ocean, Western Ocean, Indian Ocean, Mediterranean, Caspian, Black Sea, Aral Sea, f. 2b and sq.) does not apparently exist anywhere else. The author applies the term Green Sea (daryā-yi akhar or daryā-yi sabz, in Arabic al-bar al-akhar) to the Eastern Ocean, and the term Great Sea (al-bar al-a'am) to the Indian Ocean, while in Muammad ibn-Mūsā al-Khuwārizmī [1] both terms are applied to the Indian Ocean (the Great Sea — al-bar al-kabīr), and the Caspian Sea [2] is called Khuwārizmian. [3]

(P 23)

| The Black Sea is called “Sea of the Georgians” (daryā-yi Gurziyān), a term which does not seem to occur anywhere else. But even in our MS. the Georgians are not [4] mentioned among the people living around the Black Sea; in another passage, that dealing with the description of Byzantium, the Black Sea bears the name of daryā-yi Gurz (f. 37b) and the same form Gurz is given in certain Muslim sources as the name of the town of Kerch in Crimea; this has induced Westberg [5] to suppose that our author gives the name “Sea of Kerch” to the Azov Sea; but in reality the Black Sea is meant here as it is mentioned in the neighbourhood of Thrace. Nor does the variant daryā-yi Gurziyān support this supposition, though Westberg at another place endeavours to explain the name of Kerch by that of a people called Garsh [6]



1. Ed. Mžik, p. 74.

2. Ibid., p. 80.

3. Cf. Enc. of Islam, i, under Bar al-Khazar. As stated there, this early terminology of the Arab geographers may account for the Russian designation of the Caspian: Khvalinskoye, or Khvalimskoye more.

4. [But see § 42, 13. V.M.]

5. Izvestiya Akad. Nauk, 1899, p. 214.

6. “Die Garschen”, ibid., p. 309, but the reference to the Russian translation by Patkanov, p. 29, is wrong, and I have in general failed to find such a passage in the [so-called] Geography of Moses of Khoren. [As a matter of fact Patkanov translated first the abridged version of the Armenian geography, ascribed by him to Anania Shirakats'i. The complete text, edited by A. Soukry, Venice, 1881, p. 25, transl. p. 35, mentions the Garsh, whom Marquart, Streifzüge, p. 171, identifies with the Kashak, or Circassians. Moreover, see p. 401, note 1, and p.446, note 2. V.M.]





who lived, according to the [so-called, V.M.] “Geography of Moses of Khoren”, between the country of the Bulgars and the Black Sea.


Quite as original seems to be the author’s conception of the division of the inhabited world into “parts of the world” and separate “countries”. Like all Arab geographers, he accepts the division of the world into three parts, Asia, Europe, and Libya. Of course, the first place by extent belonged to Asia, and the term Āsiyat al-kubrā (f. 13a) entirely corresponds to Ptolemy’s expression ἡ μεγάλη Ἀσία (beginning of books v and vii, also viii, 3). In the author’s opinion Asia occupies two-thirds of the inhabited world, Europe one-quarter, and Libya one-twelfth. The belief that the area of Asia is twice as great as that of the other parts of the world put together occurs in other Arabic authors, notably in Bīrūnī, [1] but in any other author we should vainly look for the opinion that Europe is three times as large as Africa. To the division of the world into parts, borrowed from the Greeks, the author lends as little importance as do the other Muslim geographers, and in the survey of separate provinces he does not approach the question whether they are situated in Asia or in another part of the world. Our author counts fifty-one countries (nāhiyat) in all, of which five are situated south of the Equator, one (the Sūdān) is astride it and forty-five lie north of it. The number of the provinces is very near to that given by Khuwārizmī, viz. fifty-six, but the names of the provinces in Khuwārizmī [2] are entirely different, and many of them are borrowed from Ptolemy, which is not the case with our author. The provinces situated to the south of the Equator are enumerated in the usual order from east to west; the first to be named is Zāba (but f. 2b and f. 39a, as well as in Arab geographers, Zābaj); further on come Zangistān (country of the Zanj or Negroes, actual Zanzibār), abasha (Abyssinia), the country of Buja (or Baja, a people of Hamitic descent, still existing, and divided into several branches), [3] and Nubia. In the description of countries situated to the south of the Equator (f. 39a) the order is somewhat different: Zangistān, Zābaj, abasha, Buja, and Nubia. The author places the country of Zābaj, as he does also in the case of Zangistān, to the south of the Equator (f. 39a). The geographical term “Zābaj” is not very distinctly used by Arab geographers, who sometimes confuse the names Jāba (Java) and



1. Cf. quotation in Yāqūt, i, 63. It is remarkable that on the other hand Ibn Khurdādhbih, p. 155, entirely ignores Asia and divides the world into four parts: Europe, Libya, Ethiopia, and Scythia.

2. Ed. Mžik, pp. 101-5, Nos. 1548-1603.

3. Cf. articles “'Abābde”, “Bedja”, and “Bishārin” in Enc. of Islam.





Zābaj; [1] but in any case the term Zābaj always refers to the Malay coast or archipelago. [2] The data of the Ḥudūd al-’ālam on the islands of the Indian Ocean are borrowed from Ibn Khurdādhbih. Besides the island of Jāba there is also mentioned “the continental Jāba” (Jāba-yi khushk, f. 6b), corresponding probably to

(P 24)

| “the kingdom of Jāba the Indian” of Ibn Khurdādhbih. [3] A certain influence of Balkhī is also felt in that our author, similarly to Iṣṭakhrī, p. 11, places the country of the Zanj opposite Fārs and Kirmān, evidently on the assumption that the African coast extended much farther to the east than it does in reality. [4] But in this part of his work the author seems to have utilized sources unknown to us. Thus in Abyssinia are mentioned the following towns: Rāsun, on the seashore, residence of the king; Savār, where the army is stationed; and Rīn, the residence of the commander-in-chief. In other works we find entirely different names. [5] The folio containing the description of the countries of Buja and Nubia has been considerably damaged.


The order of enumeration of the forty-five lands situated to the north of the Equator is somewhat different in the general enumeration (f. 13a) to the order followed in the description itself (see the table of contents, f. 1b). [6] In the disposition of the chapters in the text the principle of movement from east to west is observed more scrupulously than in the general introduction (f. 13a), but without complete consistency. Thus India is described before Tibet, though in the text it is said that to the east of India are situated China and Tibet, and to the east of Tibet only China. After Tibet are described the countries of the Turkish peoples; after the Toghuzghuz (the western neighbours of the Tibetans) follow their western neighbours, the people Yaghmā; [7] after them the author passes to the north and speaks of the Khirkhīz, who, in his opinion, lived in the east towards China and the Eastern Ocean; [8] then again he passes to the south



1. BGA, vi, 46, note 2; also Enc. of Islam, ii, under “Java”. [See my note, p. 473. V.M.]

2. According to Bīrūnī, India, ed. Sachau, p. 103 above, the islands Zābaj are nearer to China than to India.

3. BGA, vi, 665: mamlakat Jābat al-Hindī [referring to the maritime Jāba. Y.M.]

4. Our author places Zanj as well opposite Sind, and so does Iṣṭakhrī; at another place Iṣṭakhrī, p. 36, places Zanj opposite some parts of Hind. [This seems to be a misunderstanding, as Iṣṭakhrī, p. 36, refers to the Indian Ocean and not to the ar al-Zanj. V.M.]

5. [See my note, p. 474. V.M.]

6. [I omit here the enumeration, which will be found at its place in my translation. V.M.]

7. F. 17a, Yghmiyā, but 13a, 17b, and 18a correctly Yghmā. The chapter on this people mentions Kāshghar, though at the same time it is said that the town is situated on the frontier between the possessions of the Yaghmā, the Tibetans, the Khirkhiz, and the Chinese.

8. Iṣṭakhrī, p. 9 below, also speaks of the Ocean (al-bar al-muīt) as the frontier of the Khirkhīz.





and describes the Khallukh (Qarluq) whose province on the east adjoined Tibet [1] and the limits of the Yaghmā and the Toghuzghuz; then the Chigil, who had separated themselves [? v.i. § 16] from the Khallukh, whose country on the east [?] and south adjoined the limits of the Khallukh and whose western neighbours were the Tukhsī. [2] Again passing to the north the author speaks of the Klmāk, living to the west of the Khirkhīz and to the north of the Irtish, and of the Ghūz. In the chapter on the Ghūz it is said that to the east and south of their province is situated the Ghūz desert [3] and the towns of Transoxiana; the Kīmāk are not mentioned in this connexion, but in their special chapter their peaceful relations as well as their wars

(P 25)

| with the Ghūz are mentioned. [4] After this come the Turkish Pecheneg, living to the west of the Ghūz, and the Khifjākh (Qipchaq), of whom it is said that they adjoin the Pecheneg on the south, and the northern desert on all the other sides. Elsewhere it is said that the Khifjākh separated from the Kīmāk, from which one might deduce that these latter had once been the eastern neighbours of the Khifjākh. After the Khifjākh is mentioned only one “Turkish” (according to the ideas of the Arab geographers!) people, the Magyars (Majghari). Nothing is said of the frontier between them and the Khifjākh, although it is mentioned that to their east were mountains, to their south lived a Christian people called Vanandar, and to their west and north lay the country of the Rūs. The text presents some resemblances to that of Ibn Rusta, [5] and likewise to that of Gardīzī [6] (e.g., the number of horsemen and the mention of the great extent of the Magyar country; Gardīzī estimates both its length and width as 100 farsakhs, while according to the Ḥudūd al-’ālam it was 150 farsakhs in length and 100 in breadth). All three texts are probably derived from the same source (perhaps the work of Ibn Khurdādhbih) which has been most fully utilized by Gardīzī.


After having spoken of the Magyars and their struggle with their neighbours, the author feels confident that he has finished with all



1. [See p. 256, note 2. V.M.]

2. In the translation of the text of Gardīzī (Report, p. 125) and in the Sketch of the History of the Semirechye, 7. 15, I wrote “Takhsī”; but in the MS. of Mamūd of Kāshghar, judging by the printed edition (i, 28, 85, 342; ii, 243), everywhere stands Tukhsī. [In the Ḥudūd al-’ālam: Tukhs, probably formed from Tukhsiyān, on a false malogy with Ghūziyān < Ghūz. V.M.]

3. The expression biyābān-i Ghūz corresponds to the expression mafāzat al-Ghuzziya, in Balkhī-Iṣṭakhrī (BGA, i, 217 and sq.).

4. According to Iṣṭakhrī, p. 222, the frontier between the countries of the Kimāk and the Ghuzz was formed by the river Itil (Ithil), by which is probably meant the lower course of the Kama (cf. my article “Ghūz” in the Enc. of Islam).

5. e.g. the mention of 20,000 Magyar horsemen, BGA, vii, 142.

6. Barthold, Report, text, p. 98, transl. p. 121 and sq.





the Turks: “now I shall enumerate all the lands of Islam, and then the rest of the lands of the Infidels which are situated in the west.” However, we shall see that in spite of this intention, the author, after the description of the Muslim provinces, comes back to such peoples as lived even farther to the east than the Magyars.





The description of the Muslim world forms, naturally, the greater part of the description of countries (17 1/2 out of 26 folios), yet even this proportion shows that the Ḥudūd al-’ālam allots to the non-Muslim world a greater space than do the Arab geographers. In the description of the Muslim countries the general order, from east to west, is again often disturbed by transitions from south to north. From Khorāsān and its frontier provinces [“marches”, V.M.], among which figure Sīstān and the provinces along the Hilmand, the author passes to the north, to Transoxiana and its frontier provinces. No special chapter is devoted to the desert Karaskūh [read: Kargas-kūh, V.M.], i.e. the “Khorasan desert” of Balkhī-Iṣṭakhrī. [1] Then follows the description of the southern provinces: Sind, Kirmān, Fārs, Khūzistān. From Khūzistān again a transition is made to the north, to the Jibāl and Daylamān (plural of Daylam). The latter comprises all the provinces along the southern and south-eastern shore of the Caspian, including the province Kūmish (Qūmis of the Arab geographers), with Bistām, Damghān, and Simnān. Rayy, with Khwār and Qazvīn, is included in the Jibāl, and not, as in Balkhī-Iṣṭakhrī, [2] in Daylam. Rayy is called “the residence of the king of the Jibāl” (f. 29a). Not until after this digression does a description of 'Irāq, lying to the west of Khūzistān, follow; then again comes the description of northern provinces: Jazīra, Ādharbayjān, Armenia, and Arrān. As in Balkhī-Iṣṭakhrī, the description of the last three provinces is united in a single chapter, in the following order: Armenia, Arrān, Ādharbayjān, [3] though one would have expected to see Arrān before Armenia. The other provinces of the Muslim world are: Arabia, Syria, Egypt, Maghrib, and Spain (Andalus).


Khorāsān, Transoxiana, and their frontier provinces are described with more detail than the other parts of the Muslim world, because the author’s materials on

(P 26)

| them were the most detailed. But he shows no such partiality to the eastern provinces as would be detrimental to the western ones, and there are no eastern provinces or towns among those to which, in some respects, an exceptional place is ascribed in the whole of the Muslim world. Khorāsān is placed near



1. BGA, i, 227.            2. Ibid., p. 207.            3. Ibid., p. 180 [But v.i., p. 142].





the centre of the inhabited world (f. 19a) but 'Irāq near the centre of the world in general. 'Irāq was the most prosperous [1] province of the Muslim world, Baghdād the most prosperous town, and Wāsit the most pleasant town in 'Irāq (f. 31a). The most pleasant countries in Islam were Ādharbayjān, Armenia, and Arrān (f. 32b), a statement which was probably borrowed from the same source as that from which the data on the Caspian provinces in general were derived. The noblest town of the world is Mekka, the birth-place of the Prophet and the House of God. Mekka was built by Adam; its construction was completed by Abraham; from Adam’s time God has loved this house (the Ka'ba, f. 33b). The first town built after the Flood was an'ā in Yemen (f. 34a). The chief town of the province of 'Omān, [2] oār (the name of this town has been for some reason replaced by the name of the province), is the storehouse for goods from all over the world; there is no other town where merchants are richer; all the merchandise from east, west, south, and north is brought to this town and from here re-exported. In Khūzistān prosperity [3] is greater than in any of the neighbouring provinces (f. 28b). Egypt is the richest country of the Muslim world; Fustāt (Cairo) the richest town of the world (f. 35a-b). The Egyptian pyramids were built by Hermes even before the Flood. The length, width, and height of each of the two large pyramids equalled 400 arash (a comparison of this passage with the text of Ibn Khurdādhbih [4] shows that the Persian arash corresponds to the Arabic dhira). Sometimes the author mentions an edifice as having an exceptional significance for the whole world, whereas in his source (Balkhī-Iṣṭakhrī) this remark is made only with regard to the Muslim world. In the passages on the Christian church in Edessa (Ruhā, f. 32a), and on the large bridge over the Euphrates (f. 34b), the words



1. The Persian term ābādhān does not easily lend itself to translation; it points to a state of inhabitedness and prosperity, as contrasting with the state of desolation, though without reference to a large number of inhabitants; cf. f. 34b, on two towns of the Mesopotamian frontier zone: ābādhān va kamardum. [I translate ābādhān by the neutral term “prosperous”. See on all these terms Index E. V.M.]

2. In the MS. everywhere wrongly Ommān instead of 'Omān.

3. In the text bisyār-ni'mat-tar. In the Ḥudūd al-’ālam the terms ni'mat and khwāsta are often juxtaposed but not as synonyms. That ni'mat and khwāsta are not one and the same thing, may be seen from the fact that in a country there may be little ni'mat and much khwāsta (f. 16b) and contrariwise (f. 37a). Apparently the word ni'mat refers to the general level of prosperity and wealth, and khwāsta, to the separate sources of prosperity or income, as for instance cattle: cf. f. 16b above, on the inhabitants of a Tibetan province: “their khwāsta are sheep.” [See Index E. V.M.]

4. BGA, vi, 1593- Therefrom, too, are borrowed the words about the inscription [but not the wording of it. V.M.].





andar hama jihān correspond to Iṣṭakhrī’s [1] fil-Islām or fī bilād al-Islām.


Everywhere careful attention is paid to what goods are exported from a given place and what localities have a particular importance in trade. Such details will undoubtedly complete in many respects what we already know from Arabic geographical literature on the various branches of industry in the Muslim world. These data might form the subject of a special treatise but unfortunately the interesting terms will not always be found in dictionaries. [2]


Often occurs the expression “place of merchants” (jāy-i bāzar-gānān [present-day pronunciation bāzurgān, V.M.] or jāyagāh-i bāzar-ganān); [3] thus are called whole provinces, e.g. Transoxiana (f. 22b), and separate towns; only in the chapters on Mesopotamia, Syria, Egypt, and Spain merchants are not mentioned, though with regard to Syria it is said that to it are brought all the goods imported from Maghrib, Egypt, Byzantium, and Spain (f. 34b). Single localities and towns are mentioned as

(P 27)

| gates (dar) into, or as store-places (bārgāh and bārkadha) of, some particular country. The situation of a town or of a locality on a given route is also mentioned, but only in the chapters on the eastern provinces; of the pilgrim routes to Mekka, the only one that merits the author’s attention is that of the Khorāsān pilgrims (f. 29a). [4]


Samarqand was a resort of merchants from all over the world (f. 23a), as were also Isfījāb (f. 24b where for the word jāy is substituted the word ma'dan), and the port of Abaskūn (f. 29b) on the Caspian Sea at the mouth of the Gurgān. The name of “Gates of Turkistān” is given to the whole of Transoxiana (f. 22b), to Khorāsān (f. 19a), and separately to Farghāna (f. 23b) and to the town of Gurgānj (f. 25b) in Khuwārizm. The capital of Khuwārizm, Kāth (spelt: Kāzh) was “the gate to the Ghūz Turks (read: Turkān instead of Turkistān) and the store-place of the Turks, Turkistān, Transoxania, and the Khazars”. About the province of Isfījāb it is said that “whatever is produced in any place of Turkistān is brought here” (f. 24b); the town of abrān, or aurān, was “the place of the Ghūz merchants” (ibid.). The situation of Karmīna, Dabusiya, and Rabinjan on the way (from Bukhārā) to Samarqand is specially mentioned (ff. 22b-23a). The small town Bāsand [5] in Chaghāniyān was



1. p. 62 (the bridge) and p. 76 (the church).

2. [See Index D.]

3. Or sometimes bāzargānān bisyār, meaning that at a given place there are numerous merchants.

4. The Khorāsān pilgrims went via Baghdād; this may account for the mention of Qādisiya “on the way of the pilgrims’’ (f. 31b).

5. On it see my Turkestan, p. 76; English ed., p. 74.





“the place of paupers, though with abundant riches”; paupers were also the inhabitants of the town of Chaghāniyān (ibid.). Other “places of paupers” were the small town of Sakalkand or Iskalkand [1] in the mountains of Hindūkūsh (f. 21b) and, as now, the Zarafshān mountains (f. 23b). The expressions “highway” (shāhrāh) and “highway of caravans” are used only with reference to the road from Marwarrud to Balkh, through Faryāb and Shapūrqān (or Ushpūrqān, f. 21a). [2] The store-house of Balkh was, however, the principal town of Gūzgān, Anbīr (or Anbār, ibid.), situated away from the above-mentioned road. The store-houses of India were Balkh (ibid.) and Lamghān, i.e. Laghmān (f. 16a); the gates to India were Bust on the Hilmand (f. 22a) and Parvān near the Hindūkūsh (f. 22b). To Andarāb near the Hindūkūsh was brought the silver from the mines of Panjhīr and Jāriyāba and here dirhams were coined from it (f. 21b). On the frontier of Vakhān there was a village that was called “the gate of Tibet” where was a Muslim customs and guard-post (f. 25b). Several towns of Vakhān are enumerated, and as the last place in the limits of Transoxiana is named the large village Samarqandāq (“Little Samarqand”), where live Hindus, Tibetans, Vakhanians (Vakhī), and Muslims (ibid.); here, too, is described Bolor (Kāfiristān), a locality not mentioned by the Arab geographers.


Sind was not a rich province, but there were many merchants in it and several of its towns carried on sea-trade (f. 26a).


From Khorāsān is mentioned the road to Rayy through Bahman-ābād and Mazīnān and the road to Gurgān through Jājarm; Jājarm was the store-house of Gurgān (f. 19b). Several industrial and commercial towns are mentioned in the Caspian provinces; an interesting description is given of Pirīm (or Firīm), principal town of the mountain province Qārin. [3] In the detailed and precise description of the Caspian provinces is to be found a striking absurdity: the words of the Qor’ān (xviii, 78), concerning the ruler who seized every ship by force, are applied to the continental town of Ahar in Ādharbaijān ff. 32b-33a). The legend, as in Iṣṭakhrī, [4] refers to the dynasty of the Julandids, but Iṣṭakhrī has in mind not the principal branch of the dynasty, which ruled in 'Omān, [5]

(P 28)

| but the “family of 'Umāra”,



1. The pronunciation in Yāqūt, i, 254, ind iii, 108; cf. BGA, i, 275d.

2. Narshakhī, ed. Schefer, p. 1125, uses the expression shāhrāh with regard to the Bukhārā-Samarqand road.

3. Cf. my Historico-geographical sketch of Iran, SPb. 1901, p. 155. [See the translation of f. 30a-b. V.M.]

4. BGA, i, 140.

5. See Salīl ibn Razīk, History of the Imams and Seyyids of Oman, transl. by G. P. Badger, 1. 1871, quoted by Barthold in the Russian translation of Lane-Poole’s Muammadan Dynasties, p. 284; E. de Zambaur, Manuel, p. 125. [See my explanation in the Notes. V.M.]





which possessed the district on the coast of Fārs, on the Kirmān frontier. This district is also mentioned by our author (f. 27b) who calls it “a place of fishermen and a haunt of merchants”. It is incomprehensible why the words on the Julandids and the Qor’ānic legend should have been transferred from one place to the other, to which they evidently can have no reference. Another evident and incomprehensible mistake occurs at the end of the chapter on 'Irāq (f. 13b), where it is said that the construction of the small towns of Karkh and Dūr (in the text by mistake Dwn), near Sāmarrā was begun by Mu'tasim (833-42), and completed by Ma’mūn (a.d. 813-33).


Store-places were: in Kirmān, Hurmuz (f. 26b); in Fārs, Sīrāf, and Māhīrūbān (or Māhrūbān, f. 27b); [1] in Khūzistān, Sus (f. 28b, here and in other passages — Shūsh). Consequently in Kirmān and Fārs such significance was attributed to sea-side towns, and in Khūzistān to a continental one. Yemen (f. 24a) is described in much greater detail than by Iṣṭakhrī: enough to say that the town of Zabīd, which according to our author occupied the second place after San'ā, is not mentioned at all [2] in Iṣṭakhrī.


The westernmost town [“land”, V.M.] of the inhabited world was “Sūs-the-Distant” (Sūs al-aqā, f. 36b). Innumerable quantities of gold were found there; the people by their customs hardly resembled men; foreigners seldom penetrated there. The chief object of export from the Berber country were panthers (or leopards, palang); the Berbers hunted them and brought their pelts for sale into Muslim towns.





From Spain the author again passes to the non-Muslim provinces, first of all to Byzantium, information on which is mostly borrowed from Ibn Khurdādhbih. The following remark is curious (f. 37a): “In Rūm (Byzantium) there are ancient towns; formerly there were many towns, now there are few.” A series of misunderstandings is explained partly by the careless rendering of the Arabic text, and partly by the fact that the author, as usual, did not discriminate between information borrowed from various sources. Ibn-Khurdādhbih [3] says that Thrace (Trāqiya) is situated “beyond Constantinople, on the side of the country of the Burjāns” (Danube Bulgars); in our author the name B.rqiya, i.e. Trāqiya, is given to the town [4] of the Burjāns.



1. On this town see BGA, i, 29c; Yāqūt, iv, 699.

2. According to Maqdisī, Zabīd was “the Baghdād of Yemen” and only by its size was inferior to San'ā, which was in a state of decadence, BGA, iii, 84 and 86.

3. BGA, vi, 109, above.

4. [“Land” ? V.MJ.





To the north-west of Rūm the author places the Bulgars (Bulgharī), not noticing that they are identical with the Burjāns whom he has just named. As distinct from the Bulgars are named, but not described, “the Slavs who have accepted Christianity”, who paid tribute to the emperor of Byzantium. [1] The country of Rūm extended to the Western Ocean, where on the south it bordered upon Spain. The Franks and the peoples living between the Franks and Spain, are described with less detail than by Iṣṭakhrī. [2] Rome is included in the Frank country. As already explained, the information on Britain as a store-house of Rūm and Spain stands isolated. The deformation of Balkhī-Iṣṭakhrī’s report [3] on the Greeks and Athens has been noted above, p. 21.


The subsequent chapters, those concerning the Slavs and the Rūs (ff. 37b-38b), the text of which has been edited by Toumansky, give little that is new. [4]

(P 29)

| The chapter on the Rūs forms a characteristic example of the indiscriminate amalgamation of data pertaining to different periods; one finds the report probably derived from Ibn Khurdādhbih about a single Russian khāqān, side by side with that about three independent Russian towns, taken probably from Ibn Falān. The fact that our author places the said three towns on the same river Rūs seems to result from an arbitrary combination by him of his sources. By the river Rūs in the present case is probably meant not the Volga in its upper course, above its junction with the Kama, as supposed by Toumansky, [5] but the Don. [6] The next people after the Rūs are “the Inner Bulgars”, in whom Marquart [7] sees the Danube Bulgars, and Westberg [8] the Black Bulgars who, according to the Russian Chronicle and to Constantine Porphyrogennetos, occupied the Don and the Azov Sea. It would be hardly expedient to attempt to analyse these hypotheses, founded as they are on the evidently insufficient and fragmentary information which has come down to us, especially in view of the fact that the author has blended together data belonging to different periods and in spite of the scarcity of his



1. This passage has been edited by Toumansky, ZVO, x, 132; translation and notes, 134. Under the influence of the record on an aqueduct coming, BGA, vii, 12618, “from the town called Bolghar” (the aqueduct near the village Belgrad) the Arabs imagined a river flowing from Bulgaria through Thrace and falling into the Bosphorus.

2. BGA, i, 43.

3. Ibid., p. 70.

4. Solely to an unhappy conjecture of Toumansky is due a detail, not to be found in the text, viz. “that in one of their (Rūs) tribes there are Mirvats”, ZVO, x, 136, note 3. The word mrvut of the text (va andar gurūhī az īshān muruvvat-ast) is probably the translation of the Arabic rujla (BGA, vii, 14614).

5. ZVO, x, 137, note.

6. BGA, ii, 27b.

7. J. Marquart, Osteuropäische und ostasiatische Streifzüge, Leipzig, 1903, p. 517.

8. Journal of the Ministry of Public Instruction (JMNP), N.S., xiii, p. 387 and sq. (in Russian).





information, has tried, with illusory exactitude, to fix the geographical situation of the countries and towns which he enumerates. There are seemingly no contradictions in his system, but this system can hardly have ever corresponded to the actual facts.


The country of the Sarīr (i.e. that of the Avars), which on the west was bounded by Byzantium, was in the south conterminous with the Muslim possessions in the Caucasus, i.e. with Darband and southern Daghestan, [1] and not with Armenia, as our author takes it. To the north and west of the Sarīr lived the Alāns whose land adjoined Byzantium and not the Muslim possessions; to the north the Alān land extended to the Black (or Azov) Sea and to the possessions of the Khazarian Pechenegs. The latter, who were the northern neighbours of the Alāns, are the first people mentioned to the east, and not to the south, of the Black Sea; their eastern frontier was the “Khazar mountains”, i.e. the mountains which, in the author’s opinion, constituted the western frontier of the Khazar country; in his conception (f. 8a) the mountains stretched west of the Caspian Sea, between the possessions of the Sarīr and the Khazars, as far as the beginning of Alān territory, whence they followed a northern direction to the end of the Khazar country, then passed between the lands of the Khazarian Pechenegs, those of the Inner Bulgars and those of the Rūs (the text here is not quite correct), to the limits of the Slavs, then followed a northern direction passing through the middle of the Slav possessions and skirting the Slav town of Khurdāb [2] until they reached the end of the Slav country. To the west of the (probably the same) mountains, to the north of the Black Sea and to the north-west of the Khazarian Pechenegs lived the Mirvāt (the Khazarian Pechenegs were for them partly eastern, partly southern neighbours); on the west, too, the Mirvāt adjoined the Black Sea. To the north-west of the Mirvāt and also to the north of the Black Sea, lived the Inner Bulgars whose land on the north reached the “Russian mountains”. The westernmost country on the northern shore of the Black Sea was that of the Slavs, conterminous to the south with Byzantium. The eastern neighbours of the Slavs, besides the Inner Bulgars, were the Rūs; on the north and west, the Slav country marched with the ‘Uninhabited Deserts of the north’. With the same deserts was conterminous on the north the Rūs country, which to the south extended down to the Danube, on which stood the capital of the Slavs, Khurdāb. [3]



1. Cf. Enc. of Islam, Barthold, Daghestān.

2. On this town, cf. the equally unfounded hypotheses of Marquart, Streifzüge, p. 471, and Westberg, JMNP, l. c., p. 12.

3. [Cf. § 6, 45. V.M.]





On the other hand, on its north-east [1] the country of the Sarīr marched with that of the Khazars; the latter on their east had “a wall between the mountains and the sea, the sea and a part of the river Itil” (f. 38b). Among the Khazar lands are counted ūlās and Lūghar, but their

(P 30)

| situation is not indicated; according to Ibn Rusta [2] thus were called the peoples living on one of the outskirts of the Khazar country, near the high mountains which stretch to Tiflis. Marquart [3] endeavours to prove that Ibn Rusta’s ūlās correspond to the Nandars (or Vanandars) of both the Toumansky MS. and Gardīzī, and likewise that the Lūghar correspond to the Mirvāt, the name Lūghar itself being a distortion of Aughaz, or (with the article) al-Aughāz, by which are meant the Abkhāz. [4] Meanwhile he (Marquart) thinks that in the source common to the Ḥudūd al-’ālam and Gardizī the information about these peoples had been confused, and that the Nandar, who according to their description correspond to the Alāns, received the name of Abkhāz.

All this is, of course, more than questionable, and since the ūlās and Lūghar are mentioned in the Toumansky MS. separately from the Mirvāt and the Nandar, these theories become still less likely.


North of the Khazars, and west of the Volga, lived the Barādhās, and west of the latter the Vanandar. East of the Volga lived the Burās, [6] and north of the Burās and the Barādhās lived the Turkish Pechenegs, while north of the Vanandar were the Magyars. Gardīzī adds [5] that the Nandar (Vanandar) were separated from the Magyars by the Danube and that from their (northern ?) bank the Magyars could see the Nandar. On the bank of the river stood a mountain and from its slope a water descended; beyond the mountain lived the Mirdāt (Mirvāt); between their province and that of the Nandar there was a distance of ten days’ journey. The neighbours of the Burās [6] on the south-east were the Central-Asian [?] Ghūz. Of the Bulgars it is only said that they lived on the Itil, with no explanation about their neighbours.



1. The author speaks only of the north; to the west of the Khazar country were “mountains”.

2. BGA, vii, 1399.

3. Streifzüge, pp. 31, 176, and 496.

4. Westberg, l.c., p. 388, was also inclined to consider the Mirvāt as Abkhaz.

5. V. Barthold, Report, p. 98 (text) and p. 122 (transl.). To the Danube  (Dūnā) apparently refer the words of our author about the river, the name of which in our MS. can be read Rūtā or Rūthā, ZVO, x, 135, note 11. On the different names of the Danube in Arabic literature, cf. BGA, viii, 67, note o.

6. [Instead of Burās (§51) read: Bulghār, as suggested by Barthold in an additional note. V.i., p. 450.]





Of the Pechenegs and other peoples considered as Turks, inclusive of the Magyars, information is given elsewhere, as we have seen. The northern outskirts of the Inhabited World begin in the east with the land of the Qirghiz (Khirkhiz), whose neighbours on the west were the Kīmāk, who in their turn had as neighbours the Qipchaq (Khifjākh, f. 18b). But meanwhile it is mentioned (f. 19a) that the Qipchaq country on all sides, with the exception of the south, where it adjoined the Pechenegs country, marched with the northern desert. Of the Pechenegs it is said that their neighbours to the east were the Ghūz; to the south, the Burās [1] and the Barādhās; to the west, the Magyars and the Rūs; to the north the Danube formed their frontier, provided that this river is identical with the river Rutā, or Rūthā (f. 19a). [2] From this summary it may be seen that, in addition to the erroneous location of the Pechenegs too far to the north (on the middle course of the Volga where they have never been), all the information on their old and new territories is jumbled together, notwithstanding the fact that their migration is mentioned in the text (f. 38a, chapter 47). The southern neighbours of the Magyars were the Vanandar, their western and northern neighbours, the Rūs; of the eastern side it is only said that a mountain was found there. [3]


In spite of the uncertainty of these data, it does not seem superfluous, in view of the attention which this part of the Toumansky MS. has already attracted, to give a full translation of the corresponding chapters, from the place where Toumansky stopped (f. 38a), to the end of the section on the northern part of the inhabited world (f. 38b below), save for the information of the frontiers that has been separately treated above.


[We omit the translation of ff. 38a-38b which will be found at the proper place in the text. V. V. Barthold concludes his Preface as follows:]

(P 32)

| The translation of this short fragment shows that the pronunciation of several names could not be established. A great number of geographical names of which the reading remains unknown is the principal reason that has made me renounce the attempt to give a full translation of the MS., and limit myself to the present Preface in which I do not touch upon the outward aspect of the MS., as these details have already been given by Toumansky [see my Preface, V.M.].



1. [Read: Bulghār. V.M.]

2. [Very doubtful. V.M.]

3. For a detailed survey and explanation of the data quoted by Barthold on pp. 42-4, see my notes to §§ 46, 53, &c. V.M.]


[Previous] [Next]
[Back to Index]