Mr. J. F. Daniell on Crystallization. 33 



upon its edges and angles. But this rhombohedron, unlike 

 that of calcareous spar, is not only divisible in directions 

 parallel to its six faces, but may be split into two tetrahedrons 

 and one octohedron, the four solid angles again, if the two 

 tetrahedrons maybe split off, and two octohedrons will remain ; 

 and the octohedron may be divided into six smaller octohe- 

 drons and eight tetrahedrons. Thus the whole mass may be 

 resolved into tetrahedra and octohedra ; no one of which can 

 we conceive so small as not to be again divisible in a similar 

 manner. 



Which, then, of these two solids is entitled to be considered 

 the primitive form of the crystal ? Neither of them can fill space 

 without leaving vacuities ; and we can scarcely conceive either 

 of them forming an arrangement sufficiently stable to consti- 

 tute the basis of a permanent crystal. 



They may both be symmetrically arranged, so as to afford 

 to the eye the external forms of the series of secondary 

 crystals, which may be geometrically calculated from their 

 various decrements ; but they must be conceived to attract one 

 another by their edges only ; and the tetrahedral arrangements 

 will be regularly interspersed with octohedral, and the octohe- 

 dral with tetrahedral cavities. The following figures, which I 

 never yet saw represented in works upon crystallography, 

 exhibit the construction of the tetrahedron, the octohedron, 

 and the cube upon each of these hypotheses. 



VOL. II. Aua, 1831, 



