194 CHEMISTRY OF THE EARTH. 



certain portion of alkali and silica. The soda-feldspars, being more easily 

 decomposed and disintegrated by atmospheric inflnences, are broken np 

 by mechanical agencies more readily than the potash-feldsi>ar. The sa me 

 is true of silicates like hornblende and pjToxene, which are less hard than 

 the feldspars. From the mechanical and chemical disintegration of 

 ordinary crystalline rocks, which consist chiefly of these various minerals, 

 together with quartz, there will result a coarse sandy sediment, in which 

 quartz with more or less orthoclase will prevail, while the finer mud 

 will contain only the more minutely divided particles of these, together 

 with partially decomposed soda-feldspar, clay, and the comminnted 

 hornblende and pyroxene. 



§ 24. This process is evidently one which must go on in the wear- 

 ing away of rocks by aqueous agency, and explains the fact that while 

 qnartz, or an excess of combined silica, is, for the most part, wanting in 

 rocks which contain a large portion of alumina, it is generally abundant 

 in those rocks in which potash- feldsi^ar predominates. The coarser and 

 more silicious sediments are readily permeable to infiltrating waters, 

 which gradually remove from them the soda, lime, and magnesia which 

 they still contain, and, if organic matters intervene, the oxide of iron, 

 leaving at last little more th?in silica, alumina, and potash, the elements 

 of granite, trachyte, gneiss, and mica-schist. On the other hand, the finer 

 sediments, whose origin, simultaneous with the coarser, we have just 

 explained, resisting the penetration of waters, will retain all their soda, 

 lime, magnesia, and iron-oxide, and containing an excess of ahunina, 

 with a small amount of silica, may, by their metamorphism, give rise to 

 basic lime and soda-feldspars, and to pyroxene and hornblende — the 

 elements of diorites and dolerites. 



§ 25. The disintegration of alkaliferous rocks, however, frequently 

 takes place under such conditions as to be more mechanical than chemi- 

 cal, and it may often happen that sediments still retaining a considerable 

 amount of combined soda become mingled with carbonates of lime and 

 magnesia. The reaction which then goes on between the liberated 

 alkaline silicate and these earthy carbonates gradually effects the con- 

 version of these into silicates, while the alkali is eliminated in the form 

 of soluble carbonate of soda, giving rise to alkaline mineral waters, whicb, 

 as I have shown, are abundantly generated in sediments where feldspathic 

 matters and earthy carbonates are intermingled. It is only from rocks 

 destitute of these carbonates that silicated alkaline waters can issue. 



§ 2G. A decomiiositiou more exclusively chemical is observed particu 

 larly among the crystalline schists of tropical and semi-tropical regioiss, 

 Avhere a process of disintegration often destroys the cohesion of the 

 rocks to a considerable depth. This change, which has been but imper- 

 fectly studied, is probably dei)endent in great part on the action of tlie 

 soluble products of vegetable decomposition, aided by the elevated tem- 

 perature. It, however, requires careful investigation ; and a considera- 

 tion of the causes which have induced it, and the extent to which it may 

 have in former periods i)revailed on the earth's surface, is of gTeat geolo- 

 gical imjiortance, since the immense erosion of which geognosy a£t"ords 

 us evidence, and w^hich seems so difficult to explain if we conceive the 

 rocks to have been as hard as we now find them in many regions, becomes 

 more easily intelligible if we suppose the cohesion of the crystalline rocks 

 to have been previously much weakened by decay. 



§ 27. The oi)eration of the mechanical and chemical agencies which ]>re- 

 side over the disintegration of x)re-existing rocks naturally divides the 

 insoluble products into two types, approaching in chemical composition, 

 as we have shown, to granites, gneiss, and mica-schist, on the one hautl^ 



