4) There is a striking difference in the altitude of climatic snowline, which affects the main feeders of the rivers, between the Tien Shan and the Kunlun mountains which are bordering the north and the south of the Tarim basin. In the southern exposures of the Tien Shan the snowline is approximately 4, 000 m and gradually decreases its height until about 3, 500 m in the east part of the mountains, while in the Kunlun it generally stands at 5, 000-5, 500 m and at the great glacier-clad watershed it rises up to nearly 6, 000 m. The change of these altitude is simplified and shown in Fig. 4 (a-b), which represents the climatic profile along the longitudinal zone of 81°-83° E. It generally decreases its height from south to north, however, the inclination grows steeper in the Kunlun mountains and thereafter, while nearing the Tien Shan it becomes gentle. In other words, the rapid decrease in the height of the snowline occurs between the latitudes of 35°-40° N.
Fig. 4 (c-d) is the approximate profile of the snowline during the last Ice Age compiled from the data of F. Machatschek (1912, 1913, 1944), K. Hermes (1955) and H. von Wissmann (1959, 1960, 1961), which is about parallel to the present one. However, the depression was greater in the Kunlun mountains. With the consequence of this, coupled by the development of extensive plateau-like gentle topography of the high altitude as well as the mightier body of the Kunlun itself, can be accounted for the more extensive snowfield during the last Ice Age than in the Tien Shan (Fig. 5). This exemplifies the greater fluctuation in the size of snowfield in the Kunlun mountains than in the Tien Shan when changes of the snowline occurred. And this might be true of the smaller changes of the snowline during historical times, which might have resulted in the greater fluctuation of water of the rivers flowing down the Kunlun mountains.
5) The retreating of glaciers in the Tien Shan and the Kunlun mountains in the last several decades has been observed by many scientists and explorers, which might have generally kept pace with the fluctuations of glaciers of the northern mid-latitude during the last over a hundred years. These fluctuations must have caused the changing of sea-level, as investigated and analysed by S. Thorarinsson (1940) and the eustatic curve for the last 2, 000 years (Fig. 6) was worked by R. W. Fairbridge (1961, 1963). And “historical records over the last two millennia indicate one of the longest sun-spot lows in Roman times, when sea level seems to have dropped to about 2 metres below the present level” (Fairbridge, 1963) is worth noting, because that period roughly corresponded to the Han (Former and Later) dynasties in the history of China, when the silk trade with Rome flourished, and in the Taklamakan desert several ruins such as Niya, Kara-dong, Miran and part of Endere, have been found which were flourishing settlements during the Han period.
To what extent the curve of Fairbridge would apply to the following historical times of this region is a problem, although it seems probable that the extension and the shrinkage of rivers occurred after the Han period, particularly in the southern border of the basin according to the glacial fluctuations of the Kunlun mountains.
6) One way to solve the problem is to examine the Chinese dynastic Annals and other historical records concerning this region and assume the geographical background, particularly dealing with the volume of waters during each dynasty. The weak point is that there were several blank periods, which means the absence of the reliable records caused by the temporary or complete withdrawal of Chinese control from the Tarim basin.