168 FLORA INDICA. 



Indian side^ separate the great rivers which flow towards the 

 plains of India^ and which^ successively uniting in their courses 

 through the plains, ultimately discharge their waters into the 

 Indus and Brahmaputra, from which they are at first sepa- 

 rated by the whole width of the chain of the Himalaya. The 

 great rivers from west to east are in succession — the Jelam, 

 the Chenab, the Ravi, the Beas, the Satlej, the Jumna, the 

 Ganges, the Gogra, the Gandak, the Cosi, the Tista, the ]Mo- 

 nas, and the Suhansiri; all of these are separated by chains 

 at first of great elevation, but which terminate at last abruptly 

 in the plains of India. Some of these chains are now well ex- 

 plored, but others, especially those in Nipal and Bhotan, arc 

 still very imperfectly known. They vary a good deal in direc- 

 tion, some being almost perpendicular to the main axis, fl'hile 

 others form with it a very acute angle. They all ramify very 

 much, giving off chains of the third order, separating the tri- 

 butaries of the great rivers. 



The length of the chain of the Himalaya, from the Indus to 

 the Brahmaputra, may be estimated at about 1400 miles, while 

 its width varies from 200 to 100 miles. ]Most of the lofty 

 peaks with which we are acquainted are situated on the se- 

 condary chains, but the mean height of the main axis is pro- 

 bably greater. The elevation of the secondary chains dimi- 

 nishes, on the whole, as they approach their termination in the 

 plains, though with a certain degree of irregularity. In length 

 these vary considerably, according to their direction, but we 

 must refer to the map for details of their structure and ar- 

 rangement. It will be seen that their ramifications are innu- 

 merable ; their flanks are in general steep, and separate deep 

 valleys. Open plains are rare, but occur occasionally at all 

 elevations, and there are a few inconsiderable lakes. The 

 mean slope of the Himalaya from the plains to the axis is not 

 more than 1 in 25, and the mean slope of the ridges of the 

 scondary chains, which are usually very oblique, and always 

 sinuous, must be considerably less. It is important to keep 

 in view these numbers, which serve to correct the erroneous 



