[ 178 ] 



XXII. On a Connexion between Crystalline Form and Chemical 

 Constitution, with some inferences therefrom. By James D. 

 Dana*. 



AMONG oxides, the protoxides, like the metallic elements, 

 are characteristically isometricf in crystallization. The 

 sesquioxides are as characteristically hexagonal, this being the 

 form of the sesquioxides of iron, aluminium, and chromium. The 

 deutoxides are typically tetragonal, as seen in the deutoxide of 

 tin (tin-ore) and of titanium (rutile and anatase). There are 

 other forms among protoxides, sesquioxides, and deutoxides : 

 for example, ZnO is hexagonal; TiO 2 in Brookite^ andMnO 2 in 

 pyrolusite are orthorhombic; but these cases, as the following 

 observations make apparent, may be regarded as a consequence 

 of polymerism — a principle that has been recognized by others 

 as underlying dimorphism. 



* Reprinted from a separate impression, communicated by the Author, 

 from Silliman's American Journal for July 1867. 



f I propose to employ in the forthcoming edition of my ' Mineralogy ' 

 the terms Isometric, Tetragonal (having a square base), and Orthorhombic 

 (erect on a rhombic base), in place respectively of Monometric, Dimetric, 

 and Trimetric. Monometric describes a line better than a cube ; the hexa- 

 gonal prism is as much dimetric as the square prism ; and the oblique 

 prisms are as truly trimetric as the right rhombic. It is very desirable that 

 the technical terms of science should be uniform over the world, as far as 

 possible, and that authors should be willing to yield their own usage for 

 the sake of uniformity, The terms adopted appear to be the best that have 

 been proposed, and have already extensive use in Europe. Isometric is 

 Hausmann's term ; tetragonal and hexagonal, with rhombic, are employed 

 by Naumaun. Mohs's terms pyramidal for the tetragonal system, and 

 prismatic for the orthorhombic, are exceedingly bad, as there are pyramids 

 among isometric, orthorhombic, and hexagonal forms, as well as the tetra- 

 gonal; and prisms in all the systems excepting the isometric. 



There is additional reason for our proposed change, in the natural rela- 

 tions of the systems of crystallization; for the similarity in the names 

 monometric, dimetric, trimetric (the latter two the monodimetrische and 

 trimetrische of Hausmann) implies a fundamental relation in the forms, 

 while the true classification is as follows: — (1) Isometric, including the 

 isometric system, peculiar in the absence of double refraction or polariza- 

 tion ; (2) Isodiametric (from 'iaos, equal, and diameter), including the te- 

 tragonal and hexagonal forms (alike named from the shape of the base), 

 characterized by equal transverse axes or diameters, and uniaxial polariza- 

 tion; and (3) Anisometric (from civ ig os, unequal, &c), including the re- 

 maining systems, and distinct in having the axes or diameters all unequal, 

 and biaxial polarization. 



Monoclinic, Diclinic, Triclinic (from Naumann) I would retain, as they 

 express admirably the relations of the systems. Clinorhombic is often used 

 for the monoclinic system, and is well enough. But clinorhomboidal for 

 the triclinic would not be desirable, as the French commonly use the word 

 rhomboidal where others use rhombic; and the diclinic could have no cor- 

 responding name, unless it be clinorect angular, which would be very ob- 

 jectionable. 



