222 EOCK DISINTEGRATION AND DECOMPOSITION 



at all affected by the ordinary atmospheric agents of solution. 

 Bischof goes so far as to say that the silicate of alumina is not 

 at all affected by carbonic acid, but the researches of Mtiller, to 

 which reference has been made, and our own investigations, tend 

 to disprove this. Dana states^ that in the decomposition of 

 basalt, on the island of Tahiti, the olivine is the earliest to give 

 way, becoming first iridescent and finally falling away to a soft, 

 pulverulent, ochreous yellow or brown powder. The compact 

 base of the rock yielded next, the augite holding out until the 

 last. Those silicates which are least liable to atmospheric de- 

 composition are, as is to be expected, those which have resulted 

 from the alteration of less stable silicates, as serpentine from 

 olivine, epidote from hornblenvde, or kaolin from feldspar, etc. 

 A few silieaj;es like tourmaline and zircon, or garnet, or oxides 

 like rutile and magnetite, or the salts of rarer earths like mona- 

 zite and zircon, are scarcely at all affected by any of the ordi- 

 nary agents of decomposition, but remain in the form of residual 

 sands in the beds of streams, from whence the lighter, more 

 decomposed material is removed by erosion. 



In the weathering of potash-feldspar rocks carrying black 

 mica, the latter mineral is as a rule the first to give way, and at 

 times almost wholly disappears. With basic rocks, on the other 

 hcind, the dark mica is one of the most enduring of the constitu- 

 ents, and in the residual sands may be found in surprisingly 

 large proportions. 



In the kaolini/ed gneisses of northern Delaware, the biotite, 

 as a rule, is in an advanced stage of decomposition, while the 

 small amount of primary muscovite is still fresh and intact, 

 retaining all its original lustre and elasticity. 



Among the feldspars the potash varieties are far more re- 

 fractory than the soda-lime, or plagioclase varieties. This is 

 shown not merely by our own investigations, but by those of 

 others as well. Roth shows^ from analyses of fresh and weathered 

 phonolites, nepheline basalts, and dolorites, that the loss of soda 

 is almost invariably greater than that of potash. 



In the coarse, pegmatitic dikes of Delaware County, Penn- 

 sylvania, the microcline masses, as mined for pottery purposes, 

 are beautifully fresh and translucent, while the associated oligo- 

 clase is snow-white through a splitting up along cleavage lines 



lOp. cit., p. 298. 



^Op. cit., 3d ed., Sd Heft. 



