XIIL] ORIGIN OF CRYSTALLINE ROCKS. 309 



which constitutes dolomite. The experiment of Marignac, who 

 thought to form dolomite by substituting a solution of chloride 

 of magnesium for the sulphate, I found to yield similar results, 

 the greater part of the magnesian carbonate produced passing at 

 once into the insoluble condition, without combining with the 

 excess of carbonate of lime present. The process for the pro- 

 duction of the double carbonate described by Charles Deville, 

 namely, the action of vapors of anhydrous magnesian chloride 

 on heated carbonate of lime, in accordance with Von Buch's 

 strange theory of dolomitization, I have not thought necessary 

 to submit to the test of experiment, since the conditions re- 

 quired are scarcely conceivable in nature. Multiplied geognos- 

 tical observations show that the notion of the epigenic pro- 

 duction of dolomite from limestone is untenable, although its 

 re-solution and deposition in veins, cavities, or pores in other 

 rocks is a phenomenon of frequent occurrence. 



The dolomites or magnesian limestones may be conveniently 

 considered in two classes : first, those which are found with 

 gypsums at various geological horizons ; and, second, the more 

 abundant and widely distributed rocks of the same kind, which 

 are not associated with deposits of gypsum. The production 

 of the first class is dependent upon the decomposition of sul- 

 phate of magnesia by solutions of bicarbonate of lime, while 

 those of the second class owe their origin to the decomposition 

 of magnesian chloride or sulphate by solutions of alkaline bi- 

 carbonates. In both cases, however, the bicarbonate of mag- 

 nesia, which the carbonated waters generally contain, contributes 

 a more or less important part to the generation of the magnesian 

 sediments. The carbonated alkaline waters of deep-seated 

 springs often contain, as is well known, besides the bicarbonates 

 of soda, lime and magnesia, compounds of iron, manganese, 

 and many of the rarer metals in solution ; and thus the metal- 

 liferous character of many of the dolomites of the second class 

 is explained. The simultaneous occurrence of alkaline silicates 

 in such mineral waters would give rise, as already pointed out, 

 to the production of insoluble silicates of magnesia, and thus 

 the frequent association of such silicates with dolomites and 



