SPITI VALLEY, &c. 259 



is owing the fluctuation in level of the lakes in Tartary, in defiance 

 of increasing cold. The lake of Mdnsarovara celebrated in Hindu my- 

 thology for giving efflux to several rivers in opposite directions, (a me- 

 taphorical figure to indicate the point of their divergence) was not admitted, 

 upon Moorcroft's assertion, to be land-locked, from ideas of the feeble- 

 ness of evaporation at that great height then unknown and unsuspected :* 

 and though the lake does appear to have an outlet in the Satlg, 

 this does not alter the question in regard to basins (inferior it is 

 true to Mdnsarovara, but under similar circumstances) having been 

 found wholly inclosed; and Moqrcroft was right as to the fact, though 

 his reviewers could not reconcile it with their preconceived opinions. 

 Chamorertl (which is probably fifty miles in circuit) has no passage 

 outward, though it is fed by streams which have a broad channel, 

 and ran with great volume in their season, t Evaporation by an atmos- 

 phere which from its extreme rarity and dryness, greedily drinks up 

 moisture, is here amply sufficient to graduate the marginal limit of those 

 lofty reservoirs to the extent of four or five feet, which was the maximum 



* The table land of Thibet was estimated by European theorists, at eight thousand feet 

 above the sea, though Captain Turner had shewn tlie unprecedented rigors of tlie climate even iu 

 so low a latitude, and Mooucropt's Narrative had given us a sufHcicntly f'riglufiil idea of midsummer 

 in that country. 



f This lake occurs in Riipshu at an elevation of fifteen thousand feet. It is a long sheet of 

 blue water with a var3 ing breadth. My route took mc by its margin for a whole day's journey, and 

 1 encamped at its eastern extremity where the shore was of turf. No water mark appeared above 

 five feet, and as I was here in the end of September, that may be considered as the limit of fluctuation, 

 a circumstance which was assumed by theorists in regard to Mdnsarovara as proving the reverse of 

 what Mr. Moorcroft asserted, or that there must be a drain from the waters of the lake. C/iumorrril 

 has likewise no efflux, though several streams pour the li(iuilied snow of the neighbouring mountains 

 into its basin. Evaporation in this dry air is fully suUicient to preserve the balance, and it is more 

 surprising that any water should rcin.iin at all, than that no outward conununication should exist. 

 The northern margin ol" the l ike is hemmed in li\' a mas-i of mountain which shouts up in a nearly 

 mural precipice of baro rock to a laij^lu of twi niy thousand t'cef and upwards. The snow rested 

 close to the summit, hut in vast bodies, having a chtf of several hundred feet, and but for its da/zling 

 whiteness might have been confounded with tlie rock itself". It had ccascil to melt. In winter the 

 lake freezes, and remains ii.xcd for several months, the snow then accumulates upon the ice and 



