166 THE PEINCIPLES INVOLVED IN EOCK-WEATHERING- 



these the iron also passes over to the hydrated sesquioxide state, 

 as is indicated by the general discoloration, the rock becoming 

 first streaked and stained, and finally uniformly ochreous. The 

 more common minerals thus attacked are the ferruginous car- 

 bonates of lime and magnesia, and silicates of the mica, amphi- 

 bole, and pyroxene groups. As the oxidation progresses, the min- 

 erals become gradually decomposed and fall away mto unrecog- 

 nizable forms. The red and yellow colors of soils are due to 

 the iron oxides contained by them. In some cases, the mineral 

 magnetite, a mixture of proto- and sesqui-oxides, undergoes 

 further oxidation and also loses its individuality. 



(2) Deoxidation is a less common feature than oxidation. 

 Water, carrying small quantities of organic acids, may take 

 away a portion of the combined oxygen of a sesquioxide, con- 

 verting it once more into the protoxide state. The local bleach- 

 ing of certain ferruginous sands and sandstones is due to this 

 action and to a partial removal of the ferriferons salt in solution. 

 Through a similar process of deoxidation, ferrous sulphates may 

 be converted into sulphides, a process which undoubtedly takes 

 place in marine muds protected from atmospheric action. 



(3) Hydration^ commonly accompanies oxidation, and, indeed, 

 is an almost constant accompaniment of rock decomposition, as 

 may be observed in comparing the total percentages of water in 

 fresh and decomposed minerals and rocks, as given in the 

 analyses. 



This assumption, provided it be not accompanied by a loss of 

 constituents, either by solution or erosion, must be attended by 

 an increase m bulk, such as may be quite appreciable. The 

 Comte de la Hure, as quoted by Branner,* has expressed the 

 opinion that some of the hills of Brazil have actually increased 

 in height through this means. The present writer has calcu- 

 lated that the transition of a granitic rock into arable soil, pro- 

 vided the same took place without loss of material, must be 

 attended by an increase in bulk amounting to 88%. This ex- 

 pansion cannot, however, be attributed wholly to hydration. 



Hydration as a factor in rock disintegration is, in the writer's 



^ This word is liere used in a more compreliensive sense than is customary, 

 and would include, in part at least, the hydrolysis of recent writers. See 

 Bell and Cameron, Bull. 30, Bureau of Soils, U. S. Dept. of Agriculture, 

 1905. 



2 Op. cit., p. 284. 



