The historians of Alexander the Great make no mention of the Aral, nor does any allusion to it occur in the pages of any classical writer until early in the second part of the fourth century of the Christian era, when Ammianus Marcellinus speaks of two rivers that rushed down headlong from the mountains, and, descending into the plains, commingled their waters and formed the Oxia Palus. In Dr Smith's Dictionary of Greek and Roman Geography, however, it is suggested that the Oxia Palus is identical with the Karakul Lake, which is now the termination of the Kohik, or river of Samarkand. Both Herodotus and Strabo appear to have had some knowledge of a series of lagoons watered by the overflowing of the Jaxartes, which was supposed by most early writers to empty itself into the Caspian some 80 parasangs from the mouth of the Oxus. The Aral, when first recognized as something more than a region of pools and swamps, was regarded as the extreme eastern portion of the Caspian, which was thought to communicate, by a long narrow strait, with the Northern Ocean.

Sir Henry Rawlinson rejects Humboldt's theory-adopted by Sir R. Murchison-of the Aralo-Caspian Sea that once filled the entire concavity of Turan, extending to the Black Sea, the Northern Ocean, and the Balkash Lake. Briefly, Sir Henry maintains that from B. c. 600 to A. D. 500, that is, for a period of 1100 years, both the Oxus and the Jaxartes emptied themselves into the Caspian, and that the Aral did not then exist as an inland sea. Even so late as A. D. 570, when Zemarchus, the Byzantine envoy, was returning to the west from the encampment of the Khakhan at the foot of the Ak-tagh, or White

an undoubted fact, and can be proved by the division of the large Lake of Alakul into three smaller basins, the separation of Lake Chelkar-Tenghiz, which receives the water of the Irghiz, f. on the bay of the Aral Sea, SaryCheganak, with which it was at one time united, and many other instances of a similar kind which have occurred in the steppe.'

Mountains, to the north of Samarkand, the Aral was only entitled to the appellation of a reedy marsh, though some thirty years afterwards the Oxus may have ceased to fill its western branch, and have kept to its direct northern channel. About that time the Sea of Kardar, or south-western portion of the Aibugir Lake, which had hitherto been fed by the Urghunj branch, became dried up, and exposed to view the ruins of a treasure city submerged in a remote age. According to Persian traditions these ruins were successfully excavated and rifled for a period of twelve years, and are placed by General Abbott, under the name of Berrasin Gelmaz, in a small island, though Admiral Boutakof fixes the locality of 'Barsa Kilmesh' in a salt marsh a little to the west of the Aibugir Lake.

For the next 600 years the Aral is described by the Arab geographers in terms that might be applied to it at the present day, though between the ninth and twelfth centuries three successive capital cities, situated at the apex of the Delta, were destroyed by sudden floods. The course of the river, too, was constantly shifting, according to the greater or smaller quantity of water diverted from the main channel for purposes of irrigation. In A. D. 1221,-Sir Henry continues,-Octai or Okkadai Khan, the son of Chinghiz Khan, when besieging Urghunj, broke down the dam that regulated the flow of the waters, and directed the full force of the stream against the walls of the town, which, being built of clay, were speedily undermined and swept away; about three years later the Oxus again forced its way to the Caspian, and the desiccation of the Aral commenced at the same time.

The elder Poli travelled from the Volga to Bokhara about 1260 A. D., by a route which must have taken them across the Aral, but that sea is not once mentioned by Marco Polo. Again, in 1330, Hamdullah Mustowfi describes the Amou as flowing from Hazarasp, a town about 40 miles south-east of the

The historians of Alexander the Great make no mention of the Aral, nor does any allusion to it occur in the pages of any classical writer until early in the second part of the fourth century of the Christian era, when Ammianus Marcellinus speaks of two rivers that rushed down headlong from the mountains, and, descending into the plains, commingled their waters and formed the Oxia Palus. In Dr Smith's Dictionary of Greek and Roman Geography, however, it is suggested that the Oxia Palus is identical with the Karakul Lake, which is now the termination of the Kohik, or river of Samarkand. Both Herodotus and Strabo appear to have had some knowledge of a series of lagoons watered by the overflowing of the Jaxartes, which was supposed by most early writers to empty itself into the Caspian some 80 parasangs from the mouth of the Oxus. The Aral, when first recognized as something more than a region of pools and swamps, was regarded as the extreme eastern portion of the Caspian, which was thought to communicate, by a long narrow strait, with the Northern Ocean.

Sir Henry Rawlinson rejects Humboldt's theory-adopted by Sir R. Murchison—of the Aralo-Caspian Sea that once filled the entire concavity of Turan, extending to the Black Sea, the Northern Ocean, and the Balkash Lake. Briefly, Sir Henry maintains that from B. c. 600 to A. D. 500, that is, for a period of 1100 years, both the Oxus and the Jaxartes emptied themselves into the Caspian, and that the Aral did not then exist as an inland sea. Even so late as A. D. 570, when Zemarchus, the Byzantine envoy, was returning to the west from the encampment of the Khakhan at the foot of the Ak-tagh, or White

an undoubted fact, and can be proved by the division of the large Lake of Alakul into three smaller basins, the separation of Lake Chelkar-Tenghiz, which receives the water of the Irghiz, f.om the bay of the Aral Sea, SaryCheganak, with which it was at one time united, and many other instances of a similar kind which have occurred in the steppe.'

Mountains, to the north of Samarkand, the Aral was only entitled to the appellation of a reedy marsh, though some thirty years afterwards the Oxus may have ceased to fill its western branch, and have kept to its direct northern channel. About that time the Sea of Kardar, or south-western portion of the Aibugir Lake, which had hitherto been fed by the Urghunj branch, became dried up, and exposed to view the ruins of a treasure city submerged in a remote age. According to Persian traditions these ruins were successfully excavated and rifled for a period of twelve years, and are placed by General Abbott, under the name of Berrasin Gelmaz, in a small island, though Admiral Boutakof fixes the locality of 'Barsa Kilmesh' in a salt marsh a little to the west of the Aibugir Lake.

For the next 600 years the Aral is described by the Arab geographers in terms that might be applied to it at the present day, though between the ninth and twelfth centuries three successive capital cities, situated at the apex of the Delta, were destroyed by sudden floods. The course of the river, too, was constantly shifting, according to the greater or smaller quantity of water diverted from the main channel for purposes of irrigation. In A. D. 1221,-Sir Henry continues,-Octai or Okkadai Khan, the son of Chinghiz Khan, when besieging Urghunj, broke down the dam that regulated the flow of the waters, and directed the full force of the stream against the walls of the town, which, being built of clay, were speedily undermined and swept away; about three years later the Oxus again forced its way to the Caspian, and the desiccation of the Aral commenced at the same time.

The elder Poli travelled from the Volga to Bokhara about 1260 A. D., by a route which must have taken them across the Aral, but that sea is not once mentioned by Marco Polo. Again, in 1330, Hamdullah Mustowfi describes the Amou as flowing from Hazarasp, a town about 40 miles south-east of the

modern town of Khiva, by the Muslim Pass and Kurlawa to Akricheh on the Caspian, near the mouth of the Attrek,-traces of which course were seen by General Abbott in 1840. The same geographer remarks that, owing to the divergence of the Amou in the previous century, the level of the Caspian was sensibly raised, and that the post of Aboskun at the mouth of the Attrek was consequently overwhelmed. During the whole of the 14th century the entire volume of the Amou was poured into the Caspian, while the Jaxartes, or Syr, lost itself in the sands of the desert, but early in the 15th century an anonymous writer, whom Sir Henry Rawlinson suspects to have been Shah Rokh's minister,—and whose manuscript he obtained at Herat, -speaks of the Aral as being dried up, through the drainage into the Caspian of the waters of the Jyhoon and Syhoon, the names given by Mohammedan geographers and historians to the Oxus and Jaxartes.

So far back as the middle of the 13th century the Franciscan Friar, William de Rubruquis, states that the Jaxartes, after creating numerous swamps, was lost in the desert, and about the year 1340 Pegoletti advises travellers bound for Tatary to leave Urghunj to their right, and to strike straight across from Saraichik on the Yaik, or Ural river, to Otrar on the Jaxartes, a route that would traverse the bed of the Aral: nor does that sea appear in the Catalan map of 1375. In short, the existence of the Aral Sea depends upon its two great tributaries, the Amou and the Syr. When the former is deflected, the bed of the sea contracts, and the Syr, being no longer able to force its way to the receding shores, becomes absorbed in reedy marshes.

There can be no doubt that in ancient times the main branch of the Oxus disembogued itself into the Caspian, for both Strabo and Pliny tell us how the merchandise of India was conveyed across the mountains to a stream that flowed into the Oxus, how it descended that river to the Hyrcanian or Caspian