In the Tian-Shan mountains there are remains of two glaciations. The earlier was the greater, and the glaciers descended well below 10,000 feet. This glaciation was followed by a long interval, when the erosion of the rivers converted the U-shaped glacial valleys into V-shaped gorges. A second glaciation descended to a level of 10,000 feet, and again developed U-valleys to this level; the end-moraines of these glaciers are young and fresh-looking. In the Altai range there were also two glacial periods. In the older and greater the snow-line was depressed by 3000 feet. The glaciers attained a length of twelve miles and descended to a level of only 3000 feet above the sea. The second glaciation was less extensive. So far we have been dealing with small mountain glaciers only. But in north-eastern Siberia we find a different state of affairs. The Stanovoi and Verkhoiansk mountains were heavily glaciated, and during the first glaciation were probably the centre of an actual ice-sheet similar to that of Scandinavia. The ice descended the valleys of the rivers Yana, Indijirka and Kolyma and covered the New Siberian Islands, which were at that time connected with the mainland. The upper valley of the Lena was also heavily glaciated by an ice-sheet moving southward, probably from the Patom highlands. When this glaciation drew to a close the source of supply among the mountains ceased, and the ice on the lowlands and in the lower parts of the river valleys was left stranded as “dead” ice. When the mountains became free of ice, the re-born rivers carried great quantities of moraine-clay and other debris with them, and flooding the ice-surface over wide areas deposited their load above the ice. In course of time the remains of the ice-sheet were deeply covered by a layer of earth and stones, which prevented the ice from melting and preserved it to the present day. This is the probable origin of the well-known “fossil ice” of Siberia. Other theories have been put forward, such as the freezing of ground water during the winter, but none are satisfactory, and that given here was generally adopted by Russian geologists.