1849.] On the Physical Geography of the Himalaya. 771 



indeed it is not easy to perceive how they are impelled, the plateau of 

 Tibet sloping on their right to Digarchi, and seeming to invite the 

 streams that way. There is however of course a water-shed, though 

 by no means a palpable one ; and we know by the signal instances of 

 the vast rivers of South America and those of north-eastern Europe, 

 how inconspicuous sometimes are the most important water-sheds of 

 the globe. The sources and courses of the feeders of the Tishta will 

 shortly be fully illustrated by Dr. Hooker, my enterprising and ac- 

 complished guest, to whom I am indebted for the above information re- 

 lative to the Lachen pass and its vicinity, and whose promised map of 

 Sikim, which state is the political equivalent for the basin of the 

 Tishta, will leave nothing to be desired further on that head. 



But the Himalaya must necessarily be contemplated in its breadth as 

 well as its length ; and we have therefore still to consider what regional 

 divisions belong to these mountains in relation to their breadth, or the 

 distance between the ghat line of the snows and the plains of India. 



The Himalayan mountains extend from the great bend of the Indus, 

 to the great bend of the Brahmaputra, or from Gilgit to Brahma Kuud, 

 between which their length is 1800 miles. Their mean breadth is 

 about 90 miles ; the maximum, about 110, and the minimum, 70 miles. 

 The mean breadth of 90 miles may be most conveniently divided into 

 three equal portions, each of which will therefore have 30 miles of ex- 

 tent. These transverse climatic divisions must be, of course, more or 

 less arbitrary, and a microscopic vision would be disposed to increase 

 them considerably beyond three, with reference to geological, to botani- 

 cal, or to zoological, phcenomena. But, upon comparing Capt. Herbert's 

 distribution of geological phcenomena with my own of zoological, and 

 Dr. Hooker's of botanical, I am satisfied that three are enough. These 

 regions I have already* denominated the lower, the middle and the up- 

 per. They extend from the external margin of the Tarai to the ghat 

 line of the snows. The lower region may be conveniently divided into 

 — I. the sand-stone range with its contained Dhuns or Maris ; — II. the 

 Bhaver or Saul forest ; — III. the Tarai. The other two regions require 

 no subdivisions. The following appear to be those demarcations by 

 height which most fitly indicate the three regions. 



* J. A. S. for December 1847, and June 1848. 



5 G 2 



