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

§ 53. V.n.nd.r.

Marquart, Streifzüge, passim; Moravcsik, Zur Geschichte der Onoguren, v.s., § 22.

The natural sequence of the three closely connected chapters would be: § 22 (Majgharī), § 53 (V.n.nd.r), § 46 (Mirvāt). The subject is of considerable difficulty and the following points must be examined:

THE SEATS OF THE V.N.ND.R. Our peoples V.n.nd.r (§ 53) and Mirvāt (§ 46) have direct parallels only in Gardīzī's N.nd.r and M.rdāt. In both the Ḥ.-'Ā. and Gardīzī the V.n.nd.r / N.nd.r are the immediate neighbours of the Majgharī though the latter's habitat is conceived differently: our author places them near the Urals, whereas Gardīzī describes the Southern Magyars as living in the region of great rivers in the north-western corner of the Black Sea. Gardīzī's views on the Magyar territory are supported not only by I.R. and Bakrī but by the consensus of Byzantine and Western European sources as well. Therefore in discussing the location of the V.n.nd.r / N.nd.r territory contiguous on that of the Majgharī we have to depend chiefly on Gardīzī and disregard our author's theoretical constructions. [3] Such is the conclusion arrived at after a long series of attempts to co-ordinate our data with those of Gardīzī until it became evident that our author's starting-point was based on an error.

According to Gardīzī the N.nd.r lived between the river separating them from the Majgharī and the mountain from which another river flowed down and behind which lived the M.rdat. The reading of the Oxford MS. according to which the mountain stood above the N.nd.r. would suggest

3. Gardīzī simply describes the facts and our author forces them into a geographical scheme. His error arises the moment that he tries to dispose his materials in map form.

466   Commentary   § 53

that it stretched in a northern direction. The river from the eastern (or northern) bank of which the Majgharī could see the N.nd.r on the opposite bank is most probably the Danube, or alternatively its northern affluent Sereth mentioned in Const. Porph.'s description of Atelkuzu (v.s., § 22). Consequently the N.nd.r lived west of the last mentioned river, or south of the Danube, with the Transylvanian Carpathians standing "above" them. Gardīzī adds that the N.nd.r lived in the direction (bar janb "on the side") of the Saqlāb. As stated in § 43 the latter term may refer to the western Slavs (or even to the Macedonian Slavs, § 42, 17.).

Our author, in spite of his cartographical error, preserves the original disposition of the peoples with regard to one another, but this goes only as far as the original triad Majgharī - V.n.nd.r - Mirvāt  is concerned. In § 46, north [east?] of the Mirvāt  are named "some of the Inner Bulghār and [!] the V.n.nd.r mountains". As the Inner Bulghār belong definitely to the Iṣṭ. < Balkhī tradition which does not know the V.n.nd.r, this combination may be disregarded as the author's own guess. See diagram on p. 440.


HARKAVI'S AND MARQUART'S VIEWS. In the Hebrew document quoted below Harkavi, as early as 1875, explained the name V.n.nt.r by that of the Bulgarian  but it was a long time before the parallel names in Ḥ.-'Ā. and Gardīzī became known. [1] When Marquart first studied Gardīzī's passage, Streifzüge, 172, he was led astray by the fact that Bakrī also mentions a pair of the Majgharī's neighbours. Having very ingeniously located the latter in the western Caucasus Marquart was less happily inspired in identifying them with the two peoples found in Gardīzī. He overlooked the fact that Bakrī (see notes to § 50, 4.) speaks of their south-eastern neighbours, while Gardīzī has in view the later Atelkuzu territory and its south-western neighbours. The identification of Gardīzī's  and  with Bakrī's  and  has often been taken for granted, but after the publication of the Ḥ.-'Ā., where the two series of names are separated, no place for doubt could remain as to its inconsistency. [2] Twenty-three years after the publication of the Streifzüge, Marquart dropped en passant a hint for a new identification of the V.n.nd.r with a promise to develop the subject. His sudden demise (4.ii.1930) prevented him from carrying out this intention and his note buried, as if intentionally, at an unexpected place does not seem to have attracted the notice which it merits. In his Arktische Länder (1924) Marquart, among other things, studies the disappearance of the sound g in old Bulgarian and Turkish and gives as an example the name of the Turcoman tribe Salur < Salghur. As another instance of the same phenomenon he quotes (p. 275) "den bulgarischen Hordennamen  (Nikephoros);  Ulughundur (Ibn al-Kalbī, + um 820, bei Jāqūt); Ołxontor (Anania

1. In his translation of § 52 (in annex to Markov's work) Toumansky illustrates V.n.nd.r by  found in Ibn al-Athīr, i, 243 ( < Mas'ūdī, Murūj, ii, 58-64). On other similar hints cf. now Kokovtsov, o.c., 92.

2. Cf. Barthold's Preface, p. 43. Marquart, Streifzüge, 172, 517, knew only a stray quotation from the Ḥ.-'Ā through Westberg's Beiträge, p. 215.

§ 53   V.n.nd.r.   467

Širakac'i, VII. Jahrh.) > W(u)ł(u)ndur Bułkar (Ps. Moses Chorenac'i, letztes Drittel des IX. Jahrh.), Wunundur (Ḥudūd al-'Ālam, Ende des X. Jahrh.), bereits mit prothetischem w vor labialem Vokal, wie im čuwaschischen;  Wulundur (al-Mas'ūdī, 943-4 n. Chr.) = magy. Nándor Fejérvár = Belgrad."

The exact references of this cryptic passage are: Nicephori Archiepiscopi Constantinopolitani Opuscula, ed. de Boor, Lipsiae 1880, p. 24; Yāqūt,ī, 404: Japhet's sons: Yūnān, al-Ṣaqlab,  (sic), [1] Burjān, Jurzān, Fārs, Rūm; Géographie de Moïse de Corène [attributed sometimes to A. Shirakats'i], ed. by Soukry, Venice 1881, p. 25, transl. p. 34 (Marquart's translation in Streifzüge, 57); Moses of Khoren, History, book ii, ch. 6. The reference to the Ḥ.-'Ā. evidently hails from Westberg's Beiträge. Mas'ūdī mentions  both in the Murūj, ii, 58-64, and in the Tanbīh, 180, 183 (see in detail Streifzüge, 60-74).

Marquart thinks that Onoghundur belongs to the type of names formed with the Turkish suffix -dur (Bayandur, Mongoldur). The forms attested in the sources would then suggest for our V.n.nd.r the reading *Vunundur. [Gardīzī's N.nd.r can hardly be compared directly with the Magyar form Nándor; most probably the initial v taken for the conjunction va was dropped by the scribe in the same way as we find in our text Khān instead of Vakhān, cf. also Mas'ūdi's  with initial w.]


THE ONOGHUNDUR. The people called Onoghundur were a Bulgarian tribe (cf. § 51) which "from the sixties of the fifth century down to the end of the seventh century" lived north of the Caucasus, to the east of the Azov sea in the Kuban region. Their great ruler Kobrat () organized them into a powerful state but after his death (circa A.D. 642) the advance of the Khazars split the Bulgar kingdom; a part of the tribes under Bayan (said to be Kobrat's son) remained in their former seats as Khazar subjects, whereas another of Kobrat's sons Asparukh travelled westwards and after having crossed the Danube (A.D. 679) conquered the territory of the present Bulgaria. Const. Porph., De thematibus, p. 48, says that since that time the name of the Bulgar has become known for "previously they were called ". [2] The centre of Asparukh's kingdom was in the strong locality  surrounded on one side by marshes and on the other by very high rocks. Jireček, Geschichte d. Bulgaren, 1876, p. 129, read the name  < Slavonic glŭ "angle, corner" and identified it with the southern part of Bessarabia known under the Turkish name Bujaq which also means "corner". [However the situation of  better suits some place in Dobruja.]

Considerably later, in the second half of the ninth century, the Onoghundurs who had stayed in the old seats and became mixed with the Magyars [3]

1. Not in the general index of Wüstenfeld's edition.

2. Marquart, Chronologie, pp. 89-96, Streifzüge, pp. 126, 505; Bury, A History of the later Roman Empire, 1889, ii, 333; Moravcsik, o.c., pp. 65, 71—2, 89.

3. The name Onoghundur (Onoghur, &c.) may be responsible for the western designation of the Magyars as Hungar-. See Munkácsi and Németh quoted by Moravcsik, o.c., 81, note 3.

468 Commentary § 53

began their westward trek which finally brought them into the present-day Hungary, cf. Moravcsik, o.c., 89.

THE SOURCE OF THE Ḥ.-'Ā. AND GARDĪZĪ. If Our two Muslim sources have preserved the name of the Onoghundur it remains to be seen to which of the two migrations the item can be assigned. It does not look probable that the original name of the Danubian Bulgars, not recorded in the earlier Muslim sources, should have suddenly emerged at a later time. [1] Both in the Ḥ.-'Ā. and Gardīzī the V.n.nd.r / N.nd.r appear not as an abstract symbol but as a tribe in flesh and blood. As shown in the notes to § 42, 17. our item on the "Christianized Slavs" is due to some later source of circa A.D. 900 when the Magyars sat in Atelkuzu and it is most likely that the additional details on the Magyars's neighbours (§§ 46 and 53) found in the Ḥ.-'Ā. and Gardīzī belong to the same source (Hārūn b. Yaḥyā?). If so, the special information of our two sources must refer to the second lot of Onoghundur pushed on by the Magyar migration. [2] Neither the Ḥ.-'Ā. nor Gardīzī mentions any enmity between the V.n.nd.r and Magyars. The qualification of the V.n.nd.r in our source as cowards (badh-dil) may be due to a wrong interpretation of the word tarsā (which means both "Christian" ' and "coward"). [3] In Gardīzī[4] the N.nd.r are definitely called Christians (tarsā) and Rūmī, i.e. "Byzantine", very possibly with a reference to their religion. In the list of bishoprics dating from the middle of the eighth century a bishop of the Onoghurs () is mentioned under the metropolitan of Crimean Gothia (), cf. Moravcsik, o.c., 64. The Onoghurs in question were certainly those who still remained to the north-east of the Black Sea and therefore could be controlled from the Crimea. [5] The rest of our author's characteristics may be only a development of his initial mistake about tarsā. [6]

1. The Khazar king's letter (v.i.) refers to the events of A.D. 679, but this detail may point to the literary origin of the passage.

2. Unless the name V.n.nd.r < Onoghundur refers to some special Bulghar territory, such as the original  occupied by Asparukh ?

3. Was then the original source on Eastern Europe, or the text in which it was available, in Persian? The absence of underground canals (kārīz) in the M.rdāt country, mentioned in Gardīzī, could hardly strike any one except an Iranian. Cf. also the strange transcription of the name  (§51)  (§ 52). These facts still await an explanation. Mas'ūdī, Murūj, ii, 59, says that dissensions among the W.l.nd.rī tribes arose in connexion with the presence among them of a Muslim merchant from Ardābīl. Consequently Persian traders penetrated into the southern Russian steppes and could be the source of information for their coreligionists.

4. [And also in our § 22.]

5. Were it not for the name *Vunundur one might consider as the Magyars' neighbours the Rumanian Vlachs, see Kunik in Izvestiya al-Bakri, ii, 16, and Niederle, Manuel, Map.

6. The Danube Bulgars were baptized under King Boris in A.D. 864. If indeed our data refer to them ( = Burjān = Inner Bulghār = Bulghari), their weakness in comparison with the Magyars could be explained by the fact that the latter were moving westwards and their forced energy (under the Pecheneg impact) could be mistaken for strength.

§ 53   V.n.nd.r.   469

MAS'ŪDI'S "W.L.ND.R". An entirely independent use of the term [1] is found in Mas'ūdi's well-known report on the incursion into the Byzantine Empire of the nomads called W.l.nd.rī in (or after) 320/932. [2] In the Murūj (written in 332/943), i, 262, ii, 58-64, Mas'ūdī calls the invaders "Turkish peoples" and enumerates their four tribes, namely, B.jnī, [3] Bajghurt (= evidently Magyar), Pecheneg (the most valiant of the four), and Nūkarda (still obscure). In the Tanbīh, 180, 182, Mas'ūdī refers to the incursion "of the Burghar and the Turkish tribes" and under the latter mentions the same four names. The reasons of this association of tribes are not quite apparent and it is possible that information belonging to different epochs has been telescoped in Mas'ūdi's version. As regards the date, the invasion seems to correspond best to that of the  (i.e. in Byzantine terminology: Magyars) recorded under 934! However, Mas'ūdī presents the four tribes as living in the neighbourhood of the Khazars and Alāns, [4] which after the events of 889 (v.s., § 22) could be true only with regard to the Pechenegs. The kings of the four tribes appear as independent chiefs and only by the consent of his three colleagues is the king of the Pechenegs invested with the supreme command on the day of battle. Mas'ūdī says that the tribes were called  "after the town of  situated in the extreme frontier region of the Rūm towards the east" and adds that the cavalry dispatched by the Emperor against the invaders reached this frontier post in 8 days. The exact situation of W.l.nd.r has been a matter of much speculation. Some scholars looked for it even in the Caucasus and in the Crimea, but Marquart, Streifzüge, 499-500, with some probability identified it with the fortress of  which lay in the neighbourhood of Burgas and was mentioned in the delimitation treaty of 864 concluded between the Emperor and the Bulgarian King Boris. [5] Jireček, o.c., 499, already suspected in W.l.nd.r a Bulgarian (non-Slavonic) name corresponding to some different official term (Debeltos?). Mas'ūdī must have got it from some oral source. Already in his innumerable "Zusätze" in Streifzüge, 500, Marquart wondered whether "Walandar" has not preserved the name of the "Unughundur-Bulgars" and in his Arktische Länder (1924) he finally adopted this point of view. The fortress, of which the name must consequently be restored as *Vulundur, could have received this name either

1. The form W.l.nd.r peculiar to Mas'ūdī results from the dissimilation n.n>l.n. Cf. the Armenian form Vłəndur.

2. See Marquart, Streifzüge, 60-74, 499-500, 527.

3. Contrary to Marquart, o.c., 67,  mentioned alongside with Pecheneg can hardly be identical with the latter. Perhaps it is only a metathesis of äpni, as one of the Oghuz clans is called in Kāshgharī, i, 57; on their later history see M. F. Köprülü-zade, Oguz etnolojisine dayir, pp. 24-7 (v.s., § 18). However, cf. infra the Khazar king's letter. [Rashīd al-dīn, ed. Bérézine, vii, 7, among the Oghuz tribes issued from Kök-khan mentions separately  and ]

4. Marquart, o.c., 74: "verblasste Erinnerungen".

5. See now V. Zlatarski, Istoriya na Bŭlgarskata dŭržava, Sofia, 1927, i. 25: the frontier left Develt to the Byzantine Empire.

470   Commentary   § 53

from some colony of Onoghundurs with whom the Greeks were in relations since the times of Kobrat, Streifzüge, 529, or because it was directed against the Vulundur (in Arabic one might say: 'alā thaghr al-Wulundur), and consequently Mas'ūdi's term *Wulunduriya (referring to all the four, or even five different tribes), most probably has to be taken in the sense of "the coalition attacking on the *Vulundur front". [1] Whatever the explanation of the raid, [2] the survival of the name *Vulundur in Mas'ūdī is a firmly established fact interesting as a parallel to our *Vunundur.


THE KHAZAR KING'S LETTER. Among the parallels to the name V.n.nd.r it remains for us to consider V.n.nt.r found in the Hebrew letter supposed to have been sent by the Khazar king Joseph in answer to that of Chasdai ben Shafrut, an agent to the Cordovan caliph 'Abd al-Raḥmān (A.D. 912-61). The year 961 is the terminus ante quem of Chasdai's original letter and the king's reply must have followed it within a not too long period. As has been recently discovered (1924), the existence of King Joseph's letter was known already to Yahuda ben Barzillai (lived towards A.D. 1100) who wondered "whether it was genuine or not". The question is complicated by the existence of two versions of the document: [3] the one (A) in a shorter form was published in Constantinople in 1577 (this text is very close to the Christ Church College MS. 193); the other (B) in a more complete form came to light only towards 1873 among the manuscripts collected by Firkovich. This fact, in view of this collector's suspect practices, was not in favour of a blind acceptance of the contents of this particular version.

The passage containing the name V.n.nt.r is found only in version B. The Khazar king says that his ancestors fought against "many nations" whom they expelled and whose country they occupied. Then comes the additional paragraph: "In the country in which I live lived formerly the V.n.nt.r. Our Khazar ancestors warred against them. The V.n.nt.r were more numerous, as numerous as the sea sand, but they could not resist the Khazars. They left their country...." After this the two versions agree in saying that the enemies were driven beyond the great river Rūnā (A. ) or Dūnā (B. ), and "until the present day they are situated on the river Rūnā/Dūnā, near Kushtantiniya/Kustandina [i.e. Constantinople] and the Khazars have occupied their country". [4]

1. Unless the coalition was formed on some special territory, v.s. = Bujaq.

2. In his final "Zusatz", o.c., 528, Marquart writes: "was es mit der Erstürmung der Festung Walandar für eine Bewandtnis hat, lasst sich bei dem völligen Schweigen der Chronisten . . . auch jetzt noch nicht erkennen, so viel ist aber nun mehr klar, dass die Walandarhorden eigentlich die Bulgaren (B.rgh.r) und ihre damaligen Verbündeten, die Pečenegen, sind…." [Cf. C. A. Macartney in Byz.-Neugr. Jahrb., 1930, pp. 159-70.]

3. See Prof. P. K. Kokovtsov, Yevreysko-khazarskaya perepiska v X veke, ed. by the Academy of the U.S.S.R., 1932, which gives the originals of all the documents bearing on the correspondence with the Khazar king with translation and a very valuable commentary. The third document discovered lately in Cambridge does not concern us here.

4. Kokovtsov's transl., pp. 75 and 92.

§§ 53-5   V.n.nd.r.   471

In a later passage the king gives an account of the Khazar boundaries and, immediately after a very detailed enumeration of the localities belonging to the Crimea [Firkovich's home!], the frontier is said to turn northwards to the country of Batsra ( most probably  *Bačna referring to Bajnī or Bajnā whom Mas'ūdī associates with the Pechenegs, v.s., p. 469, n. 3). The (inhabitants) of this country lived near the river V.zg (A. spells Y.zg, very probably *Uzu = Dniepr) and wandered in the steppe down to the limits of the H.gry'īm (A. Hyndy'īm), i.e. evidently Hungarians * Consequently the lands of a (Turkish) tribe and those of the Magyars stretched to the west of the Khazar and separated the latter from the Danube. The writer clearly refers to the expulsion of the V.n.nt.r beyond the Danube as a remote past (events of A.D. 679), whereas the account of the Khazar frontiers presupposes the arrival of the Pechenegs in the second half of the ninth century. The form W.n.nt.r has a striking resemblance to our V.n.nd.r, and on the other hand considerably differs from the forms attested in Greek and Armenian sources. Numerous names in version B seem to have been borrowed from Muslim geographers [1] and the question arises whether such is not the case of W.n.nt.r as well. The interpolator could not possibly know the Ḥ.-'Ā. or Gardīzī [which in Europe have come to light at a very recent date] but could he not have seen their common source? The text of the Khazar letter as it stands, if confronted with our two Persian authors, would confirm the interpretation of our Rūtā/Dūbā as Danube [2] and, on the other hand, suggest the identity of our V.n.nd.r with the Danubian Bulgars. However, the origin of the Hebrew interpolation remains obscure and the clever interpolator may have read his own sense into his source. Therefore in our own explanation of the Muslim texts we have to go principally by their internal evidence.

1. The most striking example is the Arabic form  *Ṣlawiyūn, Kokovtsov, o.c. 98-9.

2. See, however, supra, p. 217.

[Back to Index]