244 LEADING AMERICAN MEN OF SCIENCE 



specific gravity; and he proposed an original Latin nomenclature 

 similar to that used in botany and zoology. In the third edition 

 (1850), the " natural classification " was abandoned, with the frank 

 statement that it was "false to nature in its most essential points;" 

 and Dana's own Latin nomenclature was not even mentioned in 

 the synonymy. It had become obvious that the primary basis of 

 mineralogical classification must be found in chemistry, while, 

 within the groups established on chemical grounds, subdivisions 

 must be based largely on crystalline form. It was further recog- 

 nized that the more comprehensive groups must be founded not 

 on the metallic, or electropositive, constituents, but rather on the 

 non-metallic, or electronegative, constituents, and very largely 

 on the type of the chemical formula. For instance, the iron- 

 holding minerals, or the copper-holding minerals, would form an 

 utterly heterogeneous group; but the sulphides, or the carbonates, 

 or the silicates, would form a relatively homogeneous and consist- 

 ent group. Among the sulphides, again, the monosulphides and 

 the disulphides would form rational subdivisions. A system of 

 classification based primarily on chemical principles as now under- 

 stood, with subdivisions characterized by identity or similarity of 

 crystalline form, recognizes all the true relations which were 

 expressed in the so-called " natural system," and which the earlier 

 forms of chemical classification conspicuously failed to recognize. 

 In Part VI of his third edition, Dana outlined the "chemico- 

 crystallographic classification" a classification whose general 

 plan has been almost universally adopted, though the progress of 

 mineral chemistry has made possible great improvement in the 

 details. 1 This classification was proposed tentatively in the third 

 edition, a merely provisional classification being used in the ar- 

 rangement of the descriptive part of the book. The new classifica- 

 tion was definitively adopted in the fourth edition (1858). 



The fourth edition introduced a remarkably elegant system of 

 symbols for crystalline forms. The treatment of crystallography, 

 in this, as in the previous editions, followed in general the method 



1 A classification essentially similar to Dana's was proposed two years 

 later by Gustav Rose, in his Krystallo-chemisches Mineral-system. 



