44 SCIENCE PROGRESS 



this view, for their origin is now explained ^ by local uplift of 

 the beds of the rivers which originally traversed the region. 

 Hence, if the rate of elevation exceed that of the erosion of the 

 river, the stream would be unable to keep its channel open and 

 consequently a lake would be formed. In other words, " local 

 elevation has enhanced the erosive power of the river below 

 and diminished it above the line of uplift." However, in 

 some cases at any rate, a river has been able to keep pace with 

 the local uplift in the plateau-region. In particular, this 

 occurrence was noticed by Dr. Hedin ^ in the case of the Bog- 

 tsangpo ; and he remarks, " one is often astonished at the whim 

 of the stream in turning sharply to cut through a rocky crest, 

 whereas it would seem much easier to flow on along the open 

 longitudinal valley." This is a phenomenon which I observed ^ 

 on more than one occasion on the Armenian plateau, e.g. at Gop 

 and at Sheep in the valley of the Murad or Eastern Euphrates. 

 The discovery of volcanoes, only recently extinct, in the interior 

 of Tibet, near the Dupleix range, on the thirty-fourth parallel, 

 also lends colour to this view of the land having been fractured 

 into blocks, which have been subjected to differential movement. 

 As a corollary to the explanation which I offer of the Trans- 

 Himalayan system, it follows that the natural continuation of 

 the parallel ranges of the block lies now sunk beneath the 

 Brahmaputra valley, at the base of the great fault-scarp, to 

 which the river flows in parallel alignment. Accordingly this 

 valley must be of the nature of a rift-valley or sunken trench, 

 especially since the opposite (southern) wall of the valley 

 lies parallel to the northern wall and in like manner pos- 

 sesses an average height of 23,000 feet. This deduction again 

 receives some substantiation even from the scanty geological 

 data which (until Dr. Sven Hedin's scientific results are 

 published, i.e. in two or three years' time) we at present possess 

 concerning the Upper Brahmaputra valley and its continuation 

 westwards in the Nari-Khorsum (15,000 ft.) or upper Sutlej 

 valley (Hundes). The sacred lakes Mansaro-war and Rakas-tal 

 lie centrally in a glacial trough in this W.N.W. — E.S.E. rift-valley, 

 which occupies the site of a relative depression between 

 uptilted mountain blocks. It is filled not only by quite recent 



• R. D. Oldham, Records, Geol. Survey of India, 1888, xxi. p. 156. 



^ Op. cit. i. p. 207. 



' A Treatise on the Geology of Armenia, pp. 181, 192. 



