630 ADDITIONS. 



3. Classification of Minerals. 



In the Philosophy of the Inductive Sciences, B. vui. C. iii., I have 

 treated of the Application of the Natural-history Method of Classifi- 

 cation to Mineralogy, and have spoken of the Systems of this kind 

 which have been proposed. I have there especially discussed the sys- 

 tem proposed in the treatise of M. Necker, Le Regne Mineral ramene 

 aux Methodes d'Histoire Naturelle (Paris, 1835). More recently have 

 been published M. Beudant's Cours elementaire d'Histoire Naturelle, 

 Mineralogie (Paris, 1841) ; and M. A. Dufresnoy's Traite de Mingra- 

 logie (Paris, 1845). Both these works are so far governed by mere 

 chemical views that they lapse into the inconveniences and defects 

 which are avoided in the best systems of German mineralogists. 



The last mineral system of Berzelius has been developed by M. 

 Rarnmelsberg (Nurnberg, 1847). It is in principle such as we have 

 described it in the history. 



M. Nordenskiold's system (3rd Ed. 1849,) has been criticised by G. 

 Rose, who observes that it removes the defects of the system of Ber- 

 zelius only in part. He himself proposes what he calls a " Krystallo- 

 Chemisches System," in which the crystalline form determines the 

 genus and the chemical composition the species. His classes are 



1. Simple Substances. 



2. Combinations of Sulphur, Selenium, Titanium, Arsenic, Antimony. 



3. Chlorides, Fluorides, Bromides, Iodides. 



4. Combinations with Oxygen. 



We have already said that for us, all chemical compounds are mine- 

 rals^ in so far that they are included in our classifications. The 

 propriety of this mode of dealing with the subject is confirmed by our 

 finding that there is really no tenable distinction between native mine- 

 rals and the products of the laboratory. A great number of eminent 

 chemists have been employed in producing, by artificial means, crystals 

 which had before been known cnly as native products. 



