part 3] INDUS, BBAHMAPUTBA, AND GANGES. 157 



which the Gangetic drainage reached the sea, or the risers of the 

 Punjab flowed to the Indus, as produced by erosion or capture in 

 the ordinary sense. The gaps were broad and the alluvium, though 

 doubtless shallow as compared with the depths elsewhere, still 

 reached some hundreds of feet. He considered that these gaps 

 were more probably due to earth-movements, by which the surface 

 was depressed below the level of deposition ; once this began, and a 

 continuous alluvial plain was established, further changes in the 

 courses of the rivers would result from the balance between erosion, 

 deposition, and subsidence. He did not attach importance to the 

 argument from the general westward trend of the valleys in the 

 Himalayan region, because the secondary structure of the range 

 was that of a series of secondary ranges crossing the main axis 

 obliquely and tailing off in echelon westwards. Consequently, 

 the general direction of the valleys may equally well have been 

 determined by structural conditions as by the direction of the 

 main drainage-channel. In other respects he was in general 

 agreement with the Author's conclusions, although the exact 

 sequence of the changes could not be determined with certainty. 



The historical evidence of changes in the drainage-system was 

 of considerable interest; but, in appealing to the ancient Hindu 

 writings, it must be remembered that one school of Sanskrit 

 scholars held that the composition of the Yedas antedated the 

 Aryan invasion of India, that the prototypes of the rivers men- 

 tioned in them must be sought for in the Oxus valley, that the 

 Saraswati of the Veclas Avas the Helmund, and that the names 

 were subsequently applied to the rivers of the Punjab. In the 

 Mahabharata, which was certainly composed in India, the Saraswati 

 has dried up, and the legend of an underground course to the 

 Ganges is mentioned, showing that this legend is at least a couple 

 of thousand years old. If this interpretation is correct, no 

 argument can be based on the descriptions contained in the Yedas; 

 but that derived from the nomenclature of the Jumna certainly 

 suggests that the transfer of this river from the Indus to the 

 Gangetic system was subsequent to the introduction of the 

 existing language into India. The Author was. however, in error 

 in attributing this suggestion to the speaker; it is. in reality, of 

 older date, and should be credited to the late H. P. Medlicott, 

 although never published by him (so far as the speaker knew). The 

 dry bed of the Lost River of the Indian Desert seems to have 

 carried a considerable body of water at a much later date : for the 

 chronicles of the earliest Arab invasions of India record some 

 incidents which it is difficult to understand, unless this channel 



carried a Large body of water so late as the commencement of the 

 1 1th century. Later chronicles equally show that it no longer 



Mowed in the 13th century, so that the date of the final drying-up 

 of this river can be established within fairly narrow limit-: 

 but it is not possible to decide whether the water was supplied by 

 the Sutlej or by the .lainna. 



