168 THE PRINCIPLES INVOLVED IN EOCK-WEATHEEING 



when carbonated water was used, a distinct loss of iron and 

 magnesia throngli solution. Lepidolite, voigtite, vermiculite, and 

 pyroselerite were similarly acted upon, the iron and magnesia 

 being removed in the form of carbonates. The fact was noted 

 -that whenever anhydrous micas, or lower hydrated niicas, 

 become hydrated, they always at the same time increase in bulk. ' ' 

 This fact he regarded as accounting for the rapid weathering of 

 micaceous sandstones. 



(4) Solution. — The solvent action of water is perhaps the 

 most important of its immediate effects, though there are many 

 incidental chemical changes set in operation which, in the end, 

 are of equal or even greater significance. It is the solvent action 

 only that concerns us here. 



As long ago as 1848 the Eogers brothers showed^ that pure 

 water partially decomposed nearly all the ordinary silicate 

 minerals which form any appreciable part of our rocks. The 

 action of carbonated water was recognizable in less than ten 

 minutes, but pure water required a much longer time before 

 its effect was sufficient for a qualitative determination. So pro- 

 nounced was the action of carbonated water that the presence of 

 lime, magnesia, and the alkalies could be recognized in a single 

 drop of the filtrate from the liquid in which the powdered min- 

 erals were digested. By digestion for forty-eight hours they 

 obtained from hornblende, actinolite, epidote, chlorite, serpen- 

 tine, feldspar, etc., a quantity of lime, magnesia, oxide of iron, 

 alumina, silica, and alkalies amounting to from 0.4% to 1% of 

 the whole mass. The lime, magnesia, and alkalies were ob- 

 tained in the form of carbonates; the iron of the horn- 

 blende, epidote, etc., passing from the state of carbonate to that 

 of peroxide during the evaporation of the solutions. Forty 

 grains of finely pulverized hornblende, digested for forty- 

 eight hours in carbonated water at a temperature of 60°, with 

 repeated agitation, yielded — silica, 0.08%; oxide of iron, 

 0.095%; lime, 0.13%, and magnesia, 0.095%, with traces of 

 manganese. Commenting on these results, Bischof remarks^ 

 that *'by repeating this treatment 112 times with fresh carbon- 

 ated water, a perfect solution might be effected in 224 days. 

 If now," he says, ^^40 grains of hornblende, unpowdered, in 

 which, according to the above assumption, the surface is only 



^ Am. Jour, of Science, Yol. V, 1848. 



* Chemical and Physical Geology, Vol. I, p. 61. 



