IX.] CHEMISTRY OF NATURAL WATERS. 97 



soil, the result of these affinities is an almost complete elimina- 

 tion of the ammonia and potash from infiltrating waters. 



7. That the replacement of one base by another in this 

 way is not complete is shown moreover by the experiments 

 of Liebig, Deherain, and others, who have observed that a 

 solution of gypsum removes from soils a certain amount of 

 potash- salt, which was insoluble in pure water. In this way 

 gypseous waters may also acquire portions of sulphate of soda 

 from silicates. 



It is not certain that all the above reactions observed for 

 chabazite are applicable without modification to the double 

 hydro-aluminous silicates of sedimentary strata. Were such 

 the case, important changes might, in certain conditions, be 

 effected in the composition of saline waters. Thus in presence 

 of a great amount of a hydrous silicate of lime and alumina, 

 solutions of chloride of sodium might acquire a considerable 

 amount of chloride of calcium ; but it is probable that these 

 reactions, however important they may be in relation to the 

 soil, and to surface-waters with their feeble saline impregna- 

 tion, have at present but little influence on the composition 

 of the stronger saline waters. It is however not impossible 

 that the action of the ancient sea- waters, holding a large amount 

 of chloride of calcium, upon the hydrated and half-decomposed 

 feldspars which constituted the clays of the period, may have 

 given rise to those double silicates which formed the lime- soda 

 feldspars so abundant in the Labrador series. 



8. The reactions just described assume an importance in 

 the case of waters impregnated with soluble matters from vege- 

 table decay ; and in this event, another and not less important 

 class of phenomena intervenes, due to the deoxidizing power 

 of the dissolved organic matter. By the action of this upon 

 the insoluble peroxide of iron set free from the decomposition 

 of ferruginous minerals and disseminated in the sediments, pro- 

 toxide of iron is formed, which is soluble both in carbonic acid 

 and in the excess of the organic (acid) matter. By this means 

 not only are great quantities of iron dissolved, but masses of 

 sediments are sometimes entirely deprived of iron-oxide, and 

 5 G 



