384 Beniarks on Glacierx. 



called crystals by Hugi. We may easily convince ourselves of 

 this by breaking a portion, or infusing a coloured liquid which 

 penetrates into the fissures separating the fragments, and 

 allows us to distinguish their form and size. It is easy to per- 

 ceive that their size diminishes in proportion as we ascend 

 either from the bottom of the glacier towards the surface, or 

 from its lower to its upper part or origin. Here they may be 

 seen reduced to mere granules, so that the ice, losing more 

 and more its transparency and compactness, insensibly passes 

 (nearly at a uniform elevation among the Alps) into the state 

 of a coarse snow which is known to the mountaineers by the 

 tevxafirn or haut neve. A glacier is, therefore, a spongy mass, 

 continually imbibing atmospheric waters, as well as those pro- ' 

 duced by the melting of its surface, and which infiltrate into 

 the capillary fissures which the ice presents throughout its 

 whole thickness, and particularly at the portion nearest the 

 siu'face where it is less compact. The temperature of this 

 water being always near the freezing point, it is converted 

 into ice by the least sinking of the temperature, and tends to 

 dilate the glacier in every direction. But as it is restrained 

 on two sides by the flanks of the valley, and above by the 

 weight of the superior masses, the whole act of dilatation, aided 

 besides by that of gravitation, tends to urge it down the de- 

 clivity to the only side which offers a free passage. This ex- 

 planation being once admitted, it follows that the more fre- 

 quently the alternations of freezing and melting take place, or 

 the variations of the temperature are above and near zero, 

 the more rapid will be the advance of the glacier subjected to 

 them. Thus it happens that winter, when the entire mass is 

 frozen in an equal manner, is the season when it is in a state 

 of rest. 



The progress of the glacier is not uniform throughout the 

 whole thickness of the mass ; but if we suppose it divided into 

 beds parallel to its surface, each of these beds or layers will 

 advance with greater rapidity in proportion as it is nearer the 

 surface, or, in other words, as it is more exposed to the in- 

 fluence of atmospheric changes. It will be perceived that this 

 difference in quickness will become more obvious in the upper 

 beds, because there must be added to the quickness, i)roper to 



