VIEWS ON MINERAL SPECIES. 439 



as isolated species if the system be rigidly adhered to, since 

 they cannot be referred to any known group. 



Above all the chief difficulty remains with the silicates, 

 of which the anhydrous section is by Dana sub-divided into 

 bi-silicates, meta silicates, ortho-silicates and sub-silicates, 

 according to the ratio of the oxygen of the silica to that of 

 the bases. In reality it is extremely difficult to say what is 

 to be regarded as acid and what as base ; in the hydrous 

 section this is not even attempted, but the minerals are 

 classed under Zeolites, Mica Division, Serpentine Division, 

 Kaolin Division, etc. 



Now in these cases we clearly have a partial retention of 

 the old Natural method ; for the Micas, Chlorites, etc., are 

 practically grouped by their physical characters, the com- 

 position being too hopelessly complex to serve as the sole 

 basis of classification ; if a mineral is obviously allied to the 

 Micas it remains with them although the formula cannot be 

 brought into exact harmony with that of any Mica. 



This is only one instance among many which might 

 raise in the mind of an unbiassed reader the suspicion that 

 after all our system is to a certain extent designed to bring 

 together minerals which are declared by their external 

 characters to be allied and would therefore have been 

 classed together under the old Natural Classification ; and 

 when the system does not harmonise with* the natural 

 grouping it is the former which must give place to the latter. 



The Zeolites, the Micas, and the Chlorites are natural 

 groups like the Felspars or the Pyroxenes, although their 

 affinities cannot yet be fully explained like those of the 

 Garnets or of the Calcite group by a simple relation of 

 isomorphism. 



.MINERAL SPECIES. 



The success, however, which has attended the present 

 system in some of the better-defined groups, such as the 

 Garnets, has led to a conviction that all minerals will 

 ultimately be regarded as either simple compounds or as 

 isomorphous mixtures of simple compounds ; the five 

 Garnets mentioned above may be distinguished under the 



