MANUAL OF NATURAL HISTORY. 593 



Eegarding the classification of minerals, great 

 diversity of opinion has prevailed at various periods. 

 Formerly they were, like most other natural objects, 

 arranged almost entirely according to their external 

 characteristics and crystalline affinities. But of late 

 years, in this, as in other branches of natural sci- 

 ence, internal constitution and arrangement have 

 been allowed gradually to assume a more important 

 part, until it has been admitted by most modern 

 writers, that a truly natural system of classification 

 must be entirely based on chemical composition. 

 That this is the most correct view, according, at 

 least, to the present amount of our knowledge, can 

 scarcely be doubted, as by trusting to form only, 

 substances of very opposite qualities and structure 

 may be placed side by side ; and yet there certainly 

 is a considerable connection between external and 

 internal properties, but exceptional cases are very 

 numerous, and are very puzzling to the systematic 

 mineralogist. The first general approach to an ar- 

 rangement of this nature, included salts under the 

 heads of their acids instead of their bases, thus bring- 

 ing together all sulphates, nitrates, carbonates, &c, 

 and binary compounds were placed according to 

 their electro-negative element, making groups of 

 sulphurets, phosphurets, &c. To a strictly chemi- 

 cal system there are, at present, several obstacles, 

 some arising from the phenomena of isomorphism, 

 others from the difficulty of selecting the essential 

 element in compound minerals, or those containing 

 several bases, but more especially from the complex 



