CHEMISTRY. 211 



reagent. Had oxidizing agencies been active to any great extent below the 

 surface the pyrite must have been decomposed, and the inference from the 

 facts is strong that such lias not been the case. 



The formation of pyrite might conceivably either take place immedi- 

 ately at the expense of the bisilicates or be formed from secondary minerals; 

 but the ferruginous silicates, chlorite and epidote, are frequently deposited 

 in veins and patches quite free from pyrite, and nothing has been observed 

 in their association with pyrite to indicate an epigenetic connection. It is 

 therefore more probable that p3'rite resulted immediately from the action of 

 hydrosuiphuric acid and similar compounds on the bisilicates. This action 

 could not possibly be unaccompanied by the formation of other alteration 

 products, for the whole stochioraetric relations of the bisilicates would be 

 changed by the abstraction of iron. Since hydrosuiphuric acid is a pow- 

 erful reducing agent, it is a pnori probable that the accompanying products 

 would contain little ferric oxide, and as the bases in the bisilicates are fully 

 saturated with silicon a separation of silicic acid is indicated. 



Formation of chlorite. — The chloHte, which, as has been seen in Chapter III., 

 nearly always results from the decomposition of the ferro-magnesian sili- 

 cates, is of uncertain species, but it is neither clinochlore nor pennine, and 

 answers well to Werner's chlorite (the ripidolite of G. Rose). This mineral 

 has approximately the composition of a semisilicate, and contains little or 

 no ferric oxide. It is also accompanied in a great proportion of cases by 

 secondary quartz, and often also by calcite. The occurrence of this last 

 mineral shows that carbonic acid, as well as hydrogen sulphide, must have 

 been present during the decomposition of the rocks, and probably from the 

 commencement, for chlorite contains no calcium; and had hydrosuiphuric 

 acid alone acted on the bisilicates a calcium silicate must have resulted in 

 the first instance. Of such a preliminary change, however, there is no 

 trace, although, as has been seen in Chapter III., it appears possible to fol- 

 low the course of decomposition mineralogicall}- from its incipient stages. 

 Calcite, however, is not usually prominent among the decomposition pro- 

 ducts of tlie bisilicates in specimens collected under ground, unquestionably 

 owing to its great solubility. 



Circumstances favoring the formation of epidote. That chloritC Or cllloi'ltic minerals 



