1839.] Lieut. Irwin's Memoir oj Afghanistan, 797 



springs which are found in them do not generally give out water. 

 It is therefore plain that the periods of such streams as may originate 

 in them must be the same as those of the rains and snows of the 

 country ; such are often of the greatest importance to the husbandry 

 of a little neighbourhood, but their fame does not pass beyond those 

 bounds. The Swan and Huro alone of this class are deserving of 

 mention. They seem to have no periods distinct from the rains in 

 the country, but their springs are sufficient to preserve them running 

 streams at all seasons until they gain the Indus, whereas most others 

 lose themselves, or are expended on the fields, in all seasons but the 

 rainy, and some do not in any season reach the sea or a river. 



79. We every day hear of mountains so lofty as to be covered 

 with never melting snow* The expression construed in strictness 

 would lead to an erroneous conclusion, for, that ice or snow can 

 only remain unmelted which lies in a place whose temperature is 

 never above the freezing point, and few such can be found within 

 the habitable climates. Snow gradually disappears even during 

 a hard frost. Part it is true, is carried off by evaporation, but 

 part also is melted by the heat of the earth. The rivers of Swit- 

 zerland rise from under glaciers of solid ice. As the inferior snows 

 are gradually melted away, part of the upper also deprived of this 

 support, either gradually slide down the declivities, or fall in ava- 

 lanches, themselves to be melted in lower and warmer regions. The 

 snow and ice are therefore perennial only because they are sup- 

 plied from time to time as fast as they are consumed. It is also evi- 

 dent from the same principles, that the fall of snow in winter must 

 in all cases have some tendency to augment the streams, since part 

 is forthwith melted by the heat of the earth. But where these 

 streams originate in hills of considerable altitudes, a far greater part 

 is as it were stored up for a warmer season, and according to the 

 degree of that altitude, and the cold consequent upon it, the season 

 of its melting is later or earlier. While the snows of the low hills 

 are rapidly melting by the warmth and the rains of March, it is at 

 the same time snowing on the high mountains, whose previous stores 

 are as yet unaffected by the weather. The increasing heat at length 

 dissolves them in the order of their altitude, the highest of all melting 

 at midsummer. It is therefore evident that as far as depends on the 

 melting of the snow, streams rising in low hills must be highest in the 

 spring, and streams rising in high hills in the summer ; and the periods 

 of the streams would thus be an index of the altitude of their sources. 

 But when a river is fed by the snows of both high and low hills, we 



