latitude 39o. The Caspian forms its western boundary; and a line, from the Caspian to the Sea of Arâl, and thence again to the Alâ-tagh, or Ming Bulâk Mountains, which run north of the Sirr, or Jaxartes, as far as Tashkend, completes its northern frontier.

That part of Usbek Tûrkistân which lies south of the Asfera Mountains, may be divided into the countries south of the Amu, or Oxus, and those to the north of that river.

The divisions to the south of the Amu, including also those that extend to both its banks, or which are contained between its branches, are four: 1. Badakhshân; 2. Balkh; 3 Khwârizm; and, 4. The Deserts of the Turkomans.

The divisions to the north of the Amu are five: 1. Khutlân; 2. Karatigîn; 3. Hissâr, or Cheghâniân; 4. Kesh, or Sheher Sebz, including Karshi and Khozâr; and, 5. The Vale of Soghd, in which are the celebrated cities of Samarkand and Bokhara.

The countries lying along the Vale of the Sirr, or Jaxartes, may be considered as being six in number: 1. Ferghâna, now called Kokân and Nemengân; 2. Tashkend; 3. Uratippa, or Ushrushna; 4. Ghaz, or the Arâl Desert; to which may perhaps be added, 5. Ilak, extending between Uratippa and Tashkend; and, lastly, The district of Türkistan Proper.

I. DIVISIONS SOUTH OF THE AMU.

It may be convenient, in reviewing the different divisions of Usbek Tûrkistân, to follow the course of the two great rivers, as they proceed from the hills to the Sea of Arâl.

It will not be necessary to say much of the southern divisions, as they are, in general, sufficiently well known.

1. BADAKHSHAN.

Badakhshan is the first district to the south of the Amu. In the age of Baber, it was considered as being bounded on the south by Kafferistân, on the east by Upper Kâshghar, on the north by Khutlân, and on the west by Kundez and Anderâb. It is chiefly mountainous, and appears to be formed by the course of two considerable rivers, that unite to form the Amu. That river of the two which has the longest course and the greatest body of water, is the Penj, called also the Hammû,* which appears to be the Harat of the Arabian geographers. It has lately been ascertained to rise in the high grounds east of the Belût-tagh range, issuing from under the snow of the lofty mountains of Pûshtekhar, and working its way by the lower grounds of Shughnan and Derwâz. The second river, which is called the Kokcha, or Badakhshân river, is inferior in magnitude and length of course to the first, rising to the south of it, in that high mountainous ridge of Belût-tagh, which separates Badakhshân from Chitrâl, and Mr Elphinstone's MSS.

* Hence probably the name of Amu.

the course of the Kâshkâr or Cheghanserâi river; and on the north, divided from the course of the Penj, by a chain of lofty hills which intervene, and form the ridge of the opposite valleys. Badakhshân Proper lies along the Kokcha river, though the dominion of the King of Badakhshân generally embraced all the country south of the Penj. The country north of the Penj✶ belonged to Khutlân. The mountainous tracts near its source still called Wakhân, and by Marco Polo, Vochan, are probably part of the Wakhsh † of the Oriental geographers. Besides the two great valleys which run along the river, through all the extent of the country, there are numerous others which wind among the hills, particularly on the south, towards Kafferistân, and which transmit several streams of considerable size to the larger rivers. The Penj and Kokcha unite just below the Badakhshân territory.

The soil in the valleys is fertile, and the country has always been famous for producing precious stones, especially rubies and turquoises. It was visited in the 13th century by Marco Polo, whose account of this and the neighbouring provinces is far more correct than has been generally supposed. It belonged to Baber in the latter period of his life, but was not the scene of any of his more eminent exploits. He mentions, that its native king claimed descent from Sekander, or Alexander the Great; a claim which is continued down to the present day. The family may, perhaps, be descended from the Grecian dynasty of Bactriana, which subsisted so long unconnected with the empire of Alexander's successors.

2. BALKH.

The country between Badakhshân and the desert of Khwârizm, on the east and west, and the Hindûkûsh hills and the Amu, on the south and north, which, following Mr Elphinstone, I include under the general name of Balkh,‡ comprehends a variety of districts that, at the present day, are under several different governments. They are chiefly valleys formed by rivers that descend from the Hindûkûsh hills, and which, after forming glens and dales, frequently of considerable extent and fertility, discharge themselves into the Amu. The principal districts mentioned by Baber, are Anderâb, Talikân, Kûndez, and Khûlm, to the east; Balkh, in the centre, in a plain below the Dera Gez, or Valley of Gez, and Shiberghân, Andekhûd, and Meimana, to the west. The eastern districts are generally level and fertile towards the mouth of their different rivers; but the valleys become narrower, and contract into glens as they are followed towards the sources of their parent streams on the Hindûkûsh. The country round Balkh is level and rather sandy. The Dehâs, or Balkhâb, as it ap

* See Ebn Haukal.

+ See Abulfeda, Ap. Geograph. Min. Græc. vol. III.

This is the ancient Bactria, a term probably taken from its old Persian name of Bakhterzemîn, or Eastern country, which is given it as late as the Institutes of Taimur. Khorasan is sometimes made to include this, as well as the whole country below the hills, as high up as Badakhshân on the one side, and round their ridges to Kandahâr on the other. See Ebn Haukal, Baber's Memoirs, &c. The name of Khorasan may be derived either from its being the country east of Persia, or that west of Bakhter Zemin; as, by an odd singularity, Khawer, in the ancient Persian, is used to signify either east or west. The first certainly seems to be the more probable.

proaches that city, after leaving the Dera Gez, diminishes in size till it nearly disappears in the barren plain; and the western districts are ill watered, and indicate, by their sandy soil, the approach to the desert.

3. KHWARIZM.*

Baber never visited this country, which lies near the mouth of the Amu or Oxus ; and being surrounded on all sides by desert, may be considered as an island formed in the waste by the Amu; by innumerable branches and cuts from which, the whole country is enriched. Its geography is very defective and erroneous, though considerable materials exist for correcting it. The Amu, soon after it passes the cultivated country of Urgenj, meets the sandy desert, in which it is nearly swallowed up, so that the river is of no great volume when it reaches the sea of Arâl.

4. DESERT OF THE TURKOMANS.†

This desert, which extends from Khwârizm and the borders of Balkh to the Caspian, and from the limits of the Persian Khorasân to the sea of Arâl, and the country of the Kerghis, is inhabited by wandering Turkomans, some of whom own submission to the Chiefs of Khwârizm, or Urgenj, and others to the Persians; while a considerable portion of them yield scarcely even a nominal submission to either.

II. DIVISIONS NORTH OF THE AMU.

It has already been remarked, that these divisions are bounded on the east by the Belût-tagh mountains, which extend northward from the Hindûkûsh to the Asfera mountains, are very lofty and precipitous, and bear snow on their summits the greater part of the year, some of them without intermission. They are probably very broken and abrupt, as no pass is known to cross them, except from Badakhshân. And it is remarkable, that, in consequence of the height and abruptness of the moun

* The Chorasmin of the Persarum Syntaxis, (see Geograph. Gr. Minor, vol. III. p. 5,) is, I presume, the two Khwarizms; and indeed it includes places both in Khwarizm and Balkh. The Greek translator, to express our B, always uses M. as in Bokhara, Balkh, Tibet (TEMIIT), &c. The Latin translator does not seem to have understood some of the names; thus, MAYPANA, Mayrana, is Mâweralnaher; ZAOYL, Saul, is Zabul, or Zablestân; KOIZTAN, Koistan, is Kohistan; xapatay, Chamatan, forte Chamultan, is Hamadân; Sirazin is Shirâz ; Sistarin, Shûster; Artuel, Ardebil; Giaz is Chach, &c. &c. + The term Turkoman, James de Vitri derives from Turci et Comani, by an etymology, says Gibbon, which few critics will adopt. Yet, as we find the Turkomans pushing in on both sides of the Caspian, by Azerbaejan and the desert of Khwarizm, in both instances advancing out of the country called Comania, by the earliest travellers, from the wandering tribe of Komans, who inhabited it, there seems to be no good reason why they might not have received their name from being designated as Tûrk-Koman, or Koman-Turks, to distinguish them from the numerous tribes of the same race. See the travels of Carpini and Rubriquis in Hakluyt and Bergeron. The Cuban derives its name from this tribe of Comani or Cobani.

tains which inclose the country that has been denominated Usbek Turkistân on the east, there appear in all ages to have been only two passes across them for caravans and armies, both of which are gained by following the course of the two great rivers the Amu and the Sirr, to which the country appears to owe many of its most obvious features. The one of these grand passes leads through Badakhshân, and is the route taken by the caravan of Kâbul, and frequently by that of Samarkand and Bokhâra, on its road to Khoten and Kâshghar, This was the road followed by Marco Polo, in the thirteenth century, and more recently by De Goes,* the last European who is known to have crossed these mountains. The second pass, which ascends by the sources of the Sirr, lies in the hills that separate Ferghâna from Kâshghar, to the eastward of Ush. This is the road by which the ambassadors of Sharokh returned from China. Some inroads of Taimur's generals, by this pass, are recorded; and the caravan of Kâshghar seems to have taken this road in going between that city and Samarkand in the time of Baber's father, as it does at the present day. The route pursued by the caravan of Tâshkend, in its way to Kâshghar and China, is not quite clear; but, in some instances, it seems to have gone up the right bank of the Sirr; and after passing the Julgeh Ahengerân, or Blacksmiths' Dale, to have crossed the range of hills that encloses Ferghâna on the west, near Ahsi; to have proceeded on thence to Uzkend, and from that place, by the same pass, as the caravan of Samarkand. There is, however, reason to imagine, that the caravan of Tâshkend frequently kept a more northerly course, skirting the Ala-tagh hills that enclose Ferghâna on the north and east; and that after rounding them, and passing near Almâligh, it proceeded straight to Kâshghar. These are the only routes by which Eastern Turkistânt appears to have been reached from the west; and an attention to this fact will explain several difficulties in the earlier historians and travellers. If the supposed route to the north of the Ala-tagh hills was really one of those followed by the caravan of Tâshkend, it will perhaps explain a difficulty stated by Major Rennell, in his Memoir of a Map of Hindostan. After mentioning that Kâshghar was 25 days' journey from Samarkand, he observes, that one account differs so much from the rest, that he will draw no conclusion from it. It is one that makes 27 journeys from Tâshkend to Kâshghar, “although Tâshkend is supposed to be five journeys nearer to it than Samarkand is." If the Tâshkend route led round the hills to the north of Ferghâna, whence the traveller had to return southward towards Kâshghar, the itinerary in question will not be so inconsistent with the others as it might at first seem to be.

1. KHUTLAN.

The two districts of Khutlân and Karatigîn, which stretch along the Belût-tagh mountains, are more inaccessible and less known than most of the others. The name of Khutl,

* See Kircher's China Illustrata, p. 62, folio; and Astley's Collection of Voyages, vol. IV. p. 643, quarto.

+ The country very absurdly called Little Buchariâ.

or Khutlân, does not appear to be known at the present day; but it was applied in the time of Baber, and as far back as the age of Ebn Haukal,* to the country lying between the upper branch of the Amu, called Hârât, or Penj, which divided it from Badakhshân on the south; the Wâkhshâb or Sûrkhrûd, which separated it from Cheghânian or Hissâr on the west; the hill country of Karatigîn on the north; and the Belût-tagh on the east. Khutlân is broken in all quarters by hills. Its few valleys are said to be narrow, and overhung with lofty mountains. The glens of Shughnân and Derwâz, which lie near the source of the Penj, are fertile. The country of Wakhsh, which is always joined with Khutlân by the earlier geographers, probably extended between Khutlân and Karatigîn, or may have included Karatigîn itself. Its name is still to be found not only in the uncertain district of Wakhîka, but in the country of Wakhân, the Vochan of Marco Polo, which lies above Badakhshân, near the source of the Penj, close upon Pûshtekhar. The name Wakhshâb, anciently given to the river which divided Cheghâniân from Khutlân, is said, by Ebn Haukal, to be derived from that of the country of Wakhsh, where it originates. It ran by Weishgird, the ancient capital of the country, and joined the Amu above Kobâdiân. On this river was the Pul-e-Sengîn, or Stone-bridge, so often mentioned in the history of Taimur Beg. While some circumstances seem to point out the river which joins the Amu above Kobâdiân, opposite to Kundez, others certainly accord much better with the Sûrkh-âb, or the river of Karatigîn, which has a course of upwards of 160 miles before it falls into the Amu. The Wakhi language still remains in many districts in the hills of Badakhshân and Khutlân; and it is not improbable, that the Wakhi or Wakhshi race were the most ancient inhabitants of this hilly region. Many of the rivers that flow into the Amu in the earlier part of its course, descend from the hill-country of Khutlân. It is said to have been the seat of a splendid dynasty, before the Musulman con

* See Ouseley's translation of Ebn Haukal's geography, pp. 232 and 239.-The geographical position of certain places seems to require a departure from the reading in the text of that work. In the description of the boundaries of Mâweralnaher, we may perhaps read p. 232, "On the north, Mâweralnaher is bounded by Tûrkestân, which, enclosing Ferghâna, extends round Khotl, on the river Heriât, (Penj or Hârât.) To the south, Mâweralnaher begins from Badakhshân, and extends along the river Jihoon down to the sea of Kharezm.”—Again, p. 238, "Advancing in one direction from the Jihoon, we have the territory of Soghd, Samarkand, Siroushteh, Chaj, and Ferghâna; and, in another direction, from the borders of Samarkand, that of Kesh, Cheghanian and Khotlân; whence we have the river Jihoon from Termed, and Cobadiân, down as far as Kharezm. Farab, Sinta, Tarjaz, and Ailak, are reckoned to belong to Chaj," &c. This is true of them, though not of the preceding districts specified in the text. In p. 240,"The desert extends all round Kharezm. On one side it is bounded by Ghazneh, that is, the western side; the eastern and southern sides are bounded by Khorasan and Mâweralnaher. Below Kharezm, there is no town on the Jihoon until you come to the lake." In p. 241, Ghizni is probably Ghaz or Ghaznah; and for Kab, the sense requires Kat or Kath, the ancient capital. Debzek and Deirek, pp. 262 and 263, should be Dizak, the modern Jizzekh. "The mountains of Ashehreh" should probably be "the mountains of Aspereh." Rud-i-Haas, p. 224, must be Rud-i-Dehás. The running title (or Surkhi) is sometimes included in the text, of which instances occur in p. 275, p. 279, &c. Having no copy of the original, these changes are merely conjectural; but they seem to be necessary for the sense, and the alteration of a single letter, or of the points, is sufficient to produce most of them. These observations are made solely to justify the sense in which I have read Ebn Haukal. They cannot affect the acknowleged merit of the learned translator, who followed his text.

In regard to the era of the work, as it stands, is not the mention of the palace of Kharczm-Shah (p. 241) rather suspicious? Perhaps, however, the palace was only of a king of Kharezm.

e