32 Mr. J. F. Daniell on Crystallization. 



angles of the primitive forms. There is no difficulty in con- 

 ceiving a compound cube made up of a large number of small 

 cubes ; but if we place, upon each of the six faces of the cube 

 so formed, layers of cubic particles, decreasing each by a row 

 of particles parallel to the edges, till a pyramid is constructed 

 nipon each, terminating in a single particle, the figure becomes 

 converted into a dodecahedron with twelve equal rhombic 

 sides. If the decrement take place upon the angles, instead 

 of the edges, of the original cube, the figure is converted into 

 an octohedron. 



By decrements of more than one row of particles, and by 

 intermediate and mixed decrements, it may be shown that an 

 almost infinite variety of secondary forms may be constructed ; 

 any or all of which may be assumed by the substance to 

 which the primitive form belongs. Figures of these various 

 decrements are now so common in elementary works of science, 

 that there is no occasion to present them here. So far our ob- 

 servations upon the mechanical properties of such bodies will 

 go hand in hand with the hypothesis. Parallelopipedons, or 

 six-sided figures, of the nature of the rhombohedron and cube, 

 might attract one another by their similar sides, would form 

 stable combinations, fill all the spaces which they occupy, 

 and would yield to mechanical division only in the direction of 

 their joints. 



But there is a class of substances, which, affording the same 

 series of secondary crystals as sulphuret of lead, yields to me- 

 chanical division in very different directions ; and affords pri- 

 mitive forms of a very different character. This class of crys- 

 tals is well represented by fluor spar. 



If we apply the edge of a knife with a little dexterity to 

 a cube of Jiuor spar, we shall find that its eight solid 

 corners may be removed, and that the new-formed planes will 

 coincide with those of a regular octohedron. We may go on 

 separating slices from any of these faces, all of which may be 

 split into acute rhombohedrons. Had observation rested here, 

 no difficulty would have occurred in applying the hypothesis, 

 the acute rhombohedron would have been the primitive 

 form and the cube, the octohedron, and all the modifications 

 of this series might readily have been produced by decrements 



