265 



sphere, lay down limits which they cannot pass ; 

 it is the scorified lava of the Malpays, the pow- 



infinite number of these curves may be traced through the 

 small number of points which are known to us under the lati- 

 tudes of 0°, 20°, 45°, 62°, and 71° north, calculation is a very 

 imperfect substitute for observation. Without advancing any 

 thing very positive, we may say, that it is probable in 28° 

 17 1 the limit of the snows is above 1900 toises. From the 

 equator, where the snows begin at 2460 toises, that is near 

 the height of Mont Blanc, to the twentieth of latitude, con- 

 sequently to the limits of the torrid zone, the snows descend 

 only a hundred toises ; now ought we to admit, that eight 

 degrees farther, and in a climate which still bears almost the 

 character of a climate of the tropics, this line already lowers 

 four hundred toises ? Supposing even a lowering in arith- 

 metical progression from the twentieth to the forty-fifth de- 

 gree of latitude, a supposition which is contrary to known 

 facts (Rec. d'Obs. astron., vol. i, p. 134), the perpetual snows 

 would not begin under the parallel of the Peak but at the 

 height of 2050 toises above the level of the Ocean, conse- 

 quently 550 toises higher than on the Pyrenees and in Swit- 

 zerland. This result is supported also by other consider- 

 ations. The mean temperature of the stratum of air, with 

 which the snows are in contact during the summer, is, on the 

 Alps, a few degrees below the point of congelation, and under 

 the equator, a few degrees above it {I. c.p. 137). Admitting 

 that, at 28 degrees and a half, this temperature is 0, we find 

 according to the law of the decrement of heat, reckoning 98 

 toises to each centesimal degree, that the snows^ought to exist 

 at the height of 2058 toises above a plain, the mean temper- 

 ature of which is 21 degrees, and consequently equal to that 

 of the coasts of Teneriffe. This number is almost identical 

 with that deduced from the hypothesis of a diminution in 

 arithmetical progression. One of the high tops of the Sierra 



