strong ice that loaded carts were driven across from one bank to the other. At the commencement of the 15th century the Portuguese ambassador, Ruy Gonzalez de Clavijo, alludes with reverence to the Viadme, as he calls it, and describes it as one of the rivers that descended from Paradise. It attains, he says, a league in width, and traverses a flat country with great and wonderful force, until, at last, it reaches its goal in the Sea of Baku-one of the many names of the Caspian. Its muddy waters, he continues, are lowest in the winter season, when its sources in the mountains are congealed, but in April the melting snows begin to fill its broad, deep bed, and for the next four months it is a noble river.

In 1558, Anthony Jenkinson, the shrewd and adventurous representative of the London Muscovy Company,' mentions, that on the 5th October he struck' a gulf of the Caspianthough it is far more likely to have been the Aibugir Gulf of the Sea of Aral-where the water was fresh and sweet. 'Note,' he continues, that in times past there did fal into this gulf the great river Oxus, which hath his springs in the mountains of Paroponisus in India, and now cometh not so far, but falleth into another river called Ardock, which runneth toward the North, and consumeth himself in the grounde, passing undergrounde above 500 miles, and then issueth out againe, and falleth into lake of Kithay.' A little further on he attributes the desiccation of the Oxus to the numerous canals of irrigation it had to supply, and complains that the water that serveth all the countrey is drawen by ditches out of the river Oxus, unto the great destruction of the said river, for which cause it falleth not into the Caspian Sea as it hath done in times past, and in short time all the land is like to be destroied, and to become a wilderness for want of water when the river of Oxus shall faile.' Six or seven weeks later, he says, he crossed the great and rapid river Ardock, a branch of the Oxus, running 1000 miles to the

northward, for half of that distance passing under-ground, then issuing from beneath the earth's surface and terminating in 'the Lake of Kitay.'

It is not a little singular that Jenkinson should make no allusion to the Aral as the recipient of the river he calls the Ardock, which could be no other than the main channel of the Oxus, and which some later writers designate as the Khesel or Kizil, that is, the Red River. In Pinkerton's Collection the Khesel is said to take its rise in the mountains to the northeast of Sogdiana, and to fall into the Aral 50 or 60 miles after receiving the Amou. The writer has here evidently confounded the source of the Amou with that of the Syr, but he goes on to explain how the Khesel, which formerly fell into the Caspian at St Peter's Bay, was in 1719 diverted from that channel by the Tatars, in the hope of destroying by thirst the expedition commanded by the gallant but ill-fated Prince Beckovitch. The Dutch Orientalist Bentinck, in his notes to the 'Histoire Généalogique des Tatares'-published in 1720-describes the Oxus as bifurcating about 40 leagues from its embouchure; the left arm turning off to the Caspian, while the right arm, which 80 years previously flowed under the walls of Urghunj, then fell into the Kizil, to the ruin of that once flourishing city, and so passed on to the Aral. Jenkinson's confusion is no doubt attributable to his own ignorance of Persian and Toorkee, and consequent dependence on the intelligence and good faith of his interpreter.

That the Oxus did at one time, perhaps at several times, empty itself into the Caspian is a fact that cannot be disputed. The unfortunate Conolly proceeded for some distance up the deserted bed, whose width he estimated at 2000 paces, and General, then Captain Mouravief, met with verdure, reeds, and pools, along the deep broad channel excavated by the mighty river, which seems to have finally divided into two branches

inclosing the present peninsula of Dardji. According to the latter writer the Syr formerly effected a junction with the Amou, but an earthquake, which is said by the Khivese to have happened in the 14th century, separated the two rivers and gave them each the course they now hold. This earthquake is a very questionable event, though it is mentioned also by Baron Meyendorf as an item of Khivese faith, but he adds that others assert that the south-western channel was dammed up in 1670 to check the ravages of the Kuzzaks.

About the middle of the 18th century Captain Woodroofe, while engaged in surveying the Caspian, under Captain Elton's orders, by desire of Nadir Shah, was informed that it is now a hundred years since the Oxus emptied itself into the upper end of this (Balkan) Bay,' and he adds that, the river drying up in many places through the intense summer heat, the Toorkomans had imagined that by closing the mouth they would retain the water in the channel. The result, however, proved contrary to their expectations, for the river, having no longer any current to clear away the sand that was continually being blown into it, became choked and gradually silted up altogether. This version is adopted also by M. Khanikof, who mentions, in addition, the construction of a second dam to the eastward of Kunya or Köhne Urghunj, which effectually turned the Amou to the north.

General Abbott's theory is probably as near the truth as need be desired. The natural course of the Oxus, he says, would be to the Aral, but at some remote period it must have encountered an obstacle that deflected its waters through the easier country lying to the westward. At the same time it never ceased battering the barrier which opposed its proper course. until at length a breach was effected, when it rushed straight onwards to the Aral. General Abbott is further of opinion that many centuries ago the Syr and the Amou met a

little to the north of Urghunj, and that while a small portion of the out-pourings of the former river found its way to the Aral swamp, the main volume rolled on with its sister stream to the Caspian. By degrees, however, the Syr worked out a straight channel for itself to the north-west, while the Amou, relieved from the shock that sent it off to the south-west, also excavated a way to the northward, and thus in the end both rivers separately disembogued themselves into the Aral, as in ancient times they had conjointly deposited their waters in the bed of the Caspian. The old channel of the Amou is still the natural drain of the Kara-koum desert, so that, as already observed, it may not be impossible that the sound of subterranean waters may be heard at Kara Goombuz.

The Oxus, or the Amou, takes its rise in the Pameer Steppe, the loftiest table-land in the world, and known to the people of Central Asia as the Bam-i-Dunya, or the flat, or terraced, roof of the world. The northern branch, issuing from Lake Sir-iKul, has been generally accepted as the father of the stream, though the southern branch issuing from the Pameer Kul appears to be somewhat the longer of the two. The northern branch passes by the specific name of the Panja, and on quitting the lake is barely ankle deep, though 15 feet wide, and running at the rate of 3 miles per hour over a smooth bed. It tends at first in a south-west direction as far as Hissar, where it effects a junction with the Dara Sarhad, the southern branch that flows from the Pameer Kul. From this point it pursues a westerly course to Ishkashm, where it turns suddenly to the north, inclosing in the angle thus formed the once celebrated ruby mines of Badakhshan. Bending north-west-by-north, and then towards the south, it forms nearly a semi-circle, receiving three affluents-the Shakh-dara, the Bartang, and the Surk-ab, or Red River-until it reaches the point where it is joined by the Kokcha, nearly due west from the mouth of the Shakh

inclosing the present peninsula of Dardji. According to the latter writer the Syr formerly effected a junction with the Amou, but an earthquake, which is said by the Khivese to have happened in the 14th century, separated the two rivers and gave them each the course they now hold. This earthquake is a very questionable event, though it is mentioned also by Baron Meyendorf as an item of Khivese faith, but he adds that others assert that the south-western channel was dammed up in 1670 to check the ravages of the Kuzzaks.

About the middle of the 18th century Captain Woodroofe, while engaged in surveying the Caspian, under Captain Elton's orders, by desire of Nadir Shah, was informed that it is now a hundred years since the Oxus emptied itself into the upper end of this (Balkan) Bay,' and he adds that, the river drying up in many places through the intense summer heat, the Toorkomans had imagined that by closing the mouth they would retain the water in the channel. The result, however, proved contrary to their expectations, for the river, having no longer any current to clear away the sand that was continually being blown into it, became choked and gradually silted up altogether. This version is adopted also by M. Khanikof, who mentions, in addition, the construction of a second dam to the eastward of Kunya or Köhne Urghunj, which effectually turned the Amou to the north.

General Abbott's theory is probably as near the truth as need be desired. The natural course of the Oxus, he says, would be to the Aral, but at some remote period it must have encountered an obstacle that deflected its waters through the easier country lying to the westward. At the same time it never ceased battering the barrier which opposed its proper course. until at length a breach was effected, when it rushed straight onwards to the Aral. General Abbott is further of opinion that many centuries ago the Syr and the Amou met a