384 M. Berzelius on Isomorphism as [May, 



have species, subspecies, and varieties. What I have said of 

 garnet and chabasie applies equally to pyroxene, amphibole, 

 mica, &c. 



But these ideas cannot be applied to general systematic clas- 

 sification, without producing a deviation from the usual method. 

 Certain general forniulye of chemical composition do not present 

 the same crystalline form ; for instance, felspar and albite have 

 the same formula, but not the same crystalhne form, and must 

 consequently be considered as more distinct species than two 

 garnets or amphiboles of different composition. 



I shall now endeavour to show that the difficulties may, in 

 great measure, be removed, by means of a change in the che- 

 mical classification. In a former essay, I demonstrated that the 

 products of the mineral kingdom are best arranged in the order 

 of the electro-chemical relations of their elements, and that they 

 may be placed according to their most electro-positive, or 

 electro-negative principle. Each of these methods has its 

 advantages, and may be equally well employed. In ray essay 

 on Chemical Mineralogy, I have grouped the families according 

 to the electro-positive elements, because most of them, in their 

 combinations with the electro-negative elements, impress on 

 them particular characters, which are preserved, more or less 

 perfectly, in all minerals in which they exist; such are lead, 

 copper, cobalt, nickel, iron, barytes, &c. ; and since these com- 

 pounds are often the objects of labours, whose end it is to 

 extract from them a like electro-positive substance, it appeared 

 to me that the facility of applying the science to practice, which 

 results from the circumstance that the combinations of these 

 metals form separate classes, would equal the advantages of the 

 other mode of classification, which, however, is not to be 

 despised, and in which all the metallic sulphurets, for instance, 

 as well as all the silicates, are arranged together. At that time 

 the difficulties arising from the changes between isomorphous 

 bodies were not foreseen. On considering the modifications 

 which that circumstance, now fully established, must introduce 

 in systematic arrangement, it is immediately evident, that where 

 isomorphous changes most frequently occur, classification, if 

 not impossible, must at least be more difficult. The excellent 

 labours of M. Mitscherlich show that these substitutions may 

 take place between electro-negative, as well as electro-positive 

 bodies, without changing the figure of the crystals ; but in the 

 combinations presented by the mineral kingdom, fiequent 

 changes occur between the most common positive bodies ; 

 whilst among the negative, they have only been discovered 

 between the phosphoric and ersenic acids, bodies of rarer 

 occurrence. If negative isomorphous bodies, with sulphur, or 

 silica, were as frequently met with in the mineral kingdom, 

 either classification would present the same difficulties : there 



