The name Oxus is that by which the river is mentioned in the writings of the ancient Greek historians. In the older traditions of the Parsi books it is named the Veh-rιid, in some form of which originates the classical name which we find it most convenient to use, and also it may be presumed the names of various territories on the banks of its upper waters, such as Wakhan, Wakhsh, and Washgird, which are no doubt identical in formation, if not in application, with the classical Oxiani, Oxii, and Oxi-Petra. The classical names have long ceased to be known to the inhabitants of the country. In early Mohammedan history the river was usually styled Al-Nahr, whence the title Ma warâ Ί Nahr, or “beyond the river,” which came to be bestowed on a province of Persia lying to the north of the Oxus, and which in modern use has been rendered Transoxiana. In subsequent Mohammedan writings Al-Nahr gives place to Jaihùn, corresponding to the Gihon of the Mosaic garden of Eden. And now the river is known by Asiatics as the Amù Daria, a name of which the origin is uncertain.

In the most remote ages to which written history carries us, the regions on both sides of the Oxus were subject to the Persian monarchy. Of their populations Herodotus mentions the Bactrians, Chorasmians, Sogdians, and Sacae as contributing their contingents to the armies of the great King Darius. The Oxus figures in Persian romantic history as the limit between Iran and Turan, but the substratum of settled population to the north as well as the south was probably of Iranian lineage. The valley is connected with many early Magian traditions, according to which Zoroaster dwelt at Balkh, where, in the 7tlι

century B.C., his proselytizing efforts first came into operation. Buddhism eventually spread widely over the Oxus countries, and almost entirely displaced the religion of Zoroaster in its very cradle. The Chinese traveller Hwen Tsang, who passed through the country in 630-644 a.d., found Termedh, Khùlm, Balkh, and above all Bamian, amply provided with monasteries, stirpas, and colossal images, which are the striking characteristics of prevalent Buddhism; even the Pamir highlands had their monasteries.

Christianity penetrated to Khorâsân and Bactria at an early date; episcopal sees are said to have existed at Merv and Samarkand in the 4th and 5th centuries, and Cosmas (c. 545) testifies to the spread of Christianity among the Bactrians and Huns.

Bactria was long a province of the empire which Alexander the Great left to his successors, but the Greek historians give very little information of the Oxus basin and its inhabitants. About 250 b.c. Theodotus, the “governor of the thousand cities of Bactria,” declared himself king, simultaneously with the revolt of Arsaces which laid the foundation of the Parthian monarchy. The Graeco-Bactrian dominion was overwhelmed entirely about 126 b.c. by the Yuéchi, a numerous people of Tibet who had been driven westwards from their settlements on the borders of China by the Hiongnu, the Huns of Deguignes. From the Yuéchi arose, about the Christian era, the great Indo-Scythian dominion which extended across the Hindi! Kιish southwards, over Afghanistan and Sind. The history of the next five centuries is a blank. In 571 the Haiâthalah of the Oxus, who are supposed to be descendants of the Yuéchi, were shattered by an invasion of the Turkish khakan; and in the following century the Chinese pilgrim Hwen Tsang found the former empire of the Haiâthalah broken up into a great number of small states, all acknowledging the supremacy of the Turkish khakan, and several having names identical with those which still exist. The whole group of states he calls Tukhâra, by which name in the form Tokhâristan, or by that of Haiâthalah, the country continued for centuries to be known to the Mohammedans. At the time of his pilgrimage Chinese influence had passed into Tokhâristan and Transoxiana. Yezdegird, the last of the Sasanian kings of Bokhara, who diedin 651, when defeated and hard pressed by the Saracens, invoked the aid of China; the Chinese emperor, Taitsung, issued an edict organizing the whole country from Ferghana to the borders of Persia into three Chinese administrative districts, with 126 military cantonments, an organization which, however, probably only existed on paper.

In 711-12 Mohammedan troops were conducted by Kotaiba, the governor of Khorâsân, into the province of Khwârizm (Khiva), after subjugating which they advanced on Bokhara and Samarkand, the ancient Sogdiana, and are said to have even reached Ferghana and Kashgar, but no occuption then ensued. In 1016-25 the government of Khwârizm was bestowed by Sultan Mahmud of Ghazni upon Altιintash, one of his most distinguished generals.

Tokhâristan in general formed a part successively of the empires of the Sasanian dynasty of Bokhara (terminated 999 a.d.), of the Ghaznavi dynasty, of the Seljukian princes of Persia and of Khorâsân, of the Ghori or Shansabanya kings, and of the sultans of Khwârizm. The last dynasty ended with Sultan Jalal-ud-din, during whose reign (1221-31) a division of the Moghul army of Jenghiz Khan first invaded Khwârizm, while the khan himself was besieging Bamian; Jalal-ud-din, deserted by most of his troops, retired to Ghazni, where he was pursued by Jenghiz Khan, and again retreating towards Hindustan was overtaken and driven across the Indus.

The commencement of the 16th century was marked by the rise of the Uzbek rule in Turkestan. The Uzbeks were no one race, but an aggregation of fragments from Turks, Mongols, and all the great tribes constituting the hosts of Jenghiz and Batu. They held Kι!ndιiz, Balkh, Khwârizm, and Khorâsân, and for a time Badakhshan also; but Badakhshan was soon won by the emperor Baber, and in 152'9 was bestowed on his cousin Suliman, who by 1555 had established his rule over much of the region between the Oxus and the Hindi! Kιish. The Moghul emperors of India occasionally interfered in these provinces, notably Shah Jehan in 1646; but, finding the difficulty of maintaining so distant a frontier, they abandoned it to the Uzbek princes. About 1765 the wazir of Ahmed Shah Abdali of Cabul invaded Badakhshân, and from that time until now the domination of the countries on the south bank of the Oxus from Wakhan to Balkh has been a matter of frequent struggles between Afghans and Uzbeks.

The Uzbek rule in Turkestan has during the last twenty years been rapidly dwindling before the growth of Russian power. In 1863 Russia invaded the Khokand territory, taking in rapid succession the cities of Turkestan, Chemkend and Tashkend. In 1866 Khojend was taken, the power of Khokand was completely crushed, a portion was incorporated in the new Russian province of Turkestan, while the remainder was left to be administered by a native chief almost as a Russian feudatory; the same year the Bokharians were defeated at Irdjar. In 1867 an army assembled by the amir of Bokhara was attacked and dispersed by the Russians, λvho in 1868 entered Samarkand, and became virtually rulers of Bokhara. In 1873 Khiva was invaded, and as much of the khanate as lay on the right bank of the Oxus was incorporated into the Russian empire, a portion being afterwards made over to Bokhara. Russia acquired the right of the free navigation of the Oxus throughout its entire course, on the borders of both Khiva and Bokhara. The administration of the whole of the states on the right bank of the Oxus, down to the Russian boundary line at Ichka Yar, is now in the hands of Bokhara, including Karategin—which the Russians have transferred to it from Khokand—and Darwâz at the entrance to the Pamir highlands. At the present time the states on the left bank of the Oxus, from its sources in the Panjah river down to the town and ferry of Khwaja Saleh, are mainly subject to Afghanistan; from Khwaja Saleh to the frontiers of Khiva and Russia at Ichka Yar the left bank of the Oxus is subject to Bokhara; from the same point the Afghan boundary is supposed to stretch across the Dasht-i-chul plains of the Turkomans, above Maimana, to Sarakhs, where it meets the Persian frontier.

The regions in which the Oxus takes its birth, and through which it passes until it becomes lost in the Sea of Aral, may be divided into upper, middle, and lower: the upper is constituted by the highlands between the Thiân Shân and the Hindù Küsh ranges, and the middle by the plains and uplands which are situated in the broad valley between the western prolongations of the same ranges; the lower lies in the plains of western Turkestan. Descriptions of the chief provinces and states in the middle and lower regions will be found under Afghan Turkestan (vol. i. p. 241), including the eastern khanates of Kùndûz, Khülm, Balkh, and Akcha, and the Chahâr Wilâyat, or Four Domains, viz., the western khanates of Sir-i-pül, Shibrghân, Andkhüi, and Maimâna; also under Badakhshan, Karategin, Hissar, Bokhara, and Khiva; accounts have also been already given of Bactria, Balkh, and Bamian. Here we shall only treat of the highland regions of the Oxus, and the river itself in its downward course to the Sea of Aral, postponing all other matter to the article Turkestan (see also the map of Turkestan).

For a right understanding of the highland region, notice must be taken of its position relatively to the two great longitudinal systems of mountains, the Thiân Shân and the Indian Caucasus, and their respective prolongations east and west, which form such a prominent feature in the physical geography of the continent of Asia. These mountain systems include between them a belt of tablelands of varying breadth, and generally of considerable altitude. The forces of nature by which both the mountains and the intermediate table-lands were primarily evolved from the earth’s crust .appear to have acted concurrently over the entire region, but with greatest elevating effect along the northern edge of the Caucasus; for, though the highest peaks of the Hindù Küsh and the Himalayan ranges are more frequently met with on spurs some distance to the south than on the northern waterparting, the elevated masses are here of greatest magnitude; here there are mountains whose peaks rise to great altitudes above the sea-level, but which’ are comparatively insignificant differentially, the visible height above the surrounding table-lands being rarely more than a third, and often less than a tenth, of the height above the sea; and here there are passes across great ranges of which the level is barely distinguishable from that of the surrounding table-lands, so that the traveller may cross a great waterparting without being aware of it, a tussock of grass deciding the course of the waters, whether towards the frontiers of China or of Europe or towards the Indian Ocean.

The elevated mass which forms a bridge between the Thiân Shân and the Hindu Caucasus, in the quarter where they approach each other most closely, constitutes the governing geographical and political feature of these regions, and gives birth to all the principal sources of the Oxus. A happy instinct has led the inhabitants to call it the Bam-i-dünia, or Roof of the World; modern European geographers have called it the “heart of Asia,” the “central boss of Asia.” It is the Tsungling of Chinese writers, the northern Imaus of Ptolemy, the Mountain Parnassus of Aristotle, “the greatest of all that exist toward the winter sunrise.” The geographical indications of the Puranas, considered in any but a fabulous light, point to it as Méru, the scene of the primeval Aryan paradise. Old Parsi traditions point to it as the origin and nucleus of the Aryan migrations. And it is here that the Mohammedan invaders are shown, by their identification of the great rivers with the Gihon and Pison of the Mosaic narrative, to have believed that the terrestrial paradise, the cradle of the human race, was situated.

Few regions can present claims to interest and just curiosity so strong and various as this one. Its past history is interwoven with that of all the great Asiatic conquerors, and its position on the rapidly narrowing borderland between the British and the Russian dominions gives it additional interest at the present time. But its geography is most intricate and complicated, and has long been a fruitful subject of controversy. The region is intersected with mountain ridges and depressed river beds which are alike difficult to cross; its altitude is unfavourable for the growth of cereals, and it mostly lies buried in snow for half the year; it is, moreover, sparsely inhabited, and does not produce sufficient food for the requirements of the inhabitants. It interposes a formidable barrier between eastern and western Turkestan across the ancient highway from Europe to China; and, though this barrier has been repeatedly crossed, the extant narratives of the journeys and descriptions of the routes present only occasional glimmerings of truth amidst a mass of error and confusion, and are at times barely available for sober inquiry; genuine facts of observation have been so mixed up with erroneous information that it has become impossible to reconcile conflicting statements or separate the true from the false. Thus within the last quarter of a century maps have been published by eminent geographers in Eflgland and Germany in which the great cities of eastern Turkestan are placed 3° to 4°, or over 200 miles, too far to the west, and the limits of the “heart of Asia” are materially narrowed.

The interest attaching to the region has even led to the fabrication of spurious documents which have darkened the mist already enveloping it, and have betrayed eminent geographers into error and confusion.

While geography remained under the spell of these mischievous fictions, research was impeded, and an insurmountable obstacle placed in the way of the true delineation of the region; doubt was even thrown on the accuracy of the work of genuine explorers. But within the last decade the mist in which the “Roof of the World” had so long been enveloped has been largely dispelled by the labours of Russian and British officers, and also by natives of India trained to geographical exploration and employed in connexion with the operations of the Great Trigonometrical Survey of India. In some parts there is still much doubt and uncertainty, but enough is now known to furnish the geographical student with a fairly accurate idea of the general course of the rivers and configuration of the tablelands and mountains.

Two systems of rivers give birth to the sources of the Oxus, one to the north rising in and around the Alai plateau, the other to the south rising in the Pamir plateaus, of which there are several. The two systems are divided by a great chain of mountains known locally as the Kizil-yart range, but called by Fedchenko (looking from the north) the Trans-Alai range, and by recent Russian surveyors the Peter the Great range; it lies from east to west on the southern border of the Alai plateau, and throws out spurs westwards to Darwâz; its medium height above the sea-level is 18,000 or 19,000 feet, with occasional peaks rising to 25,000 feet. Of the Oxianian affluents to its north and west the principal are the Wâkhsh or Siirkh-âb ( = the Kizil-su = the Red River), rising in the Alai, and the Mùksù and Khing-ab rivers, which join the Wâkhsh in the district of Karategin.

The system of southern affluents is, however, the most important of the two politically as well as geographically, comprising as it does the water-partings which define the boundaries between China, Afghanistan, and Bokhara, and all the rivers of what is generally known as the Pamir region. The name Pamir is suggested by Bournouf to have been derived from Ùpa-Mérü, meaning the lands “beyond the mountain of Meru”; a later and more probable suggestion, by Major Trotter, is that it is the Khirgiz equivalent of Bâm-i-dûnia. It means simply an elevated steppe or plateau. By the people of the country it is not applied, as European geographers apply it, to the entire region, which is one of mountains as well as table-lands, but to each of the plateaus with the addition of a distinctive designation. Thus there is the Pamir-Kalân (great), the Pamir-Khùrd (little), the Pamir-Alichûr, the Pamir-Khargoshi (of the hare), the Pamir-Sarez (of the water-parting), and the Pamir-Rangkùl, on which the Rangkιil lake is situated. There is also another, the Pamir-i-Shiva, which, though only recently brought prominently to the notice of European geographers, is of considerable magnitude, elevation, and importance; it lies in that part of Badakhshân which is enclosed to the north and east by the Panjah river, and to the south and west by a spur from the Hindi! Küsh range. This spur is an offshoot from the vicinity of the Tirich Mir peak (25,400 feet) north of Chitral; it lies between Faizabad and Ishkâshim, sinks to 10,900 feet at the Zebak pass, and then again, ascending to higher altitudes, trends to the north-west, and strikes the western spurs of the Kizil-yart range in the Darwâz district; it forms the water-parting between the Kokcha river of southern Badakhshân and the Panjah river. Though a spur from the main range, it is of itself an important range, and has some claim to be regarded as the western boundary of the Pamir table-lands, as it lies immediately over the Shiva Pamir; if the claim be admitted, the breadth of the elevated barrier between the plains of eastern and western Turkestan will be found to be about 250 miles, whereas geographers have hitherto accorded to the Pamir plateau a breadth of only 100 miles. The Panjah river flows downwards through the region where the spurs of this western bounding range meet those of the Kizil-yart range, passing between narrow and precipitous gorges which form a natural gateway to the highlands, though one which in many parts is barely accessible, or has to be quitted altogether for the easier mountain passes on either hand.

The most elevated portion of the highlands occurs on the north-east border, above the plains of Kashgar and Yarkand. Here a chain of mountains, interwoven with the Thiân Shân and the Kizil-yart ranges, trends to the east and south-east, and throws up peaks of great height, culminating in Tagharma (25,500 feet); viewed from the plains to the east, it seems to form part of a great chain— the Belût Tâgh of Humboldt—which connects the Thiân Shân range with the Hindi! Kιish; but it is broken through by rivers, and terminates over the plains of the Sarikoi district. The line of water-parting which constitutes the real connexion between the Thiân Shân and the Hindi! Kιish lies more to the west, in hills which, emanating from the Kizil-yart range, pass between the Rangkιil Pamir and the Kizil-yart plain, and then bending southwards strike an angle of the Hindi! Kιish range on the borders of the Sarikoi and Kanjιit districts; they are probably nowhere of any great altitude above the general level of the table-lands; but they are of importance in that they may be regarded as the natural boundary between the states of eastern Turkestan now subject to China, and those of western Turkestan subject to Afghanistan and Bokhara.

The best known river of the Pamir plateaus is the Panjah, which receives all the other rivers of this region before it enters the plains; above Kila Panjah it has two important affluents, one from the east rising in Kanjιit, and probably about 120 miles long, the other from the north-east rising in the lake· of the Great Pamir (Wood’s Lake Victoria), and about 80 miles long. From the point of junction to Kila-Bâr-Panjah is 140 miles; here the united waters of the Sochân and Shâkhdara rivers from the east are received; 33 miles lower down, near Kila Wâmar, the Bârtang river, also from the east, is received. The upper source of the Bârtang is the Ak-su (white water) river, which rises in the Oikιil or Gazkιil lake of Little Pamir, and, winding round the highlands, passes through the Sarez Pamir, where its name changes to the Murghâbi (water fowl), which lower down becomes Bârtang (narrow passage). The Aksù-Bârtang is probably the...