METHOD OF ASCERTAINING PRIMARY FORMS. 77 



(ft.) If the crystal does not occur under one of the primary forms, 

 but under one of those new figures which we have seen, in . 51, 

 are capable of being derived from the primary forms by symmetrical 

 modifications, in these cases, (from our knowledge of the relations 

 of the various new figures, as they have been called, to the prima- 

 ry solids,) we shall be able, in general, to say, that a crystal of 

 this sort belongs to some one of two or three of the primary forms, 

 and in a few cases we are sure of the identical form from which it 

 comes. Thus, if the crystal is a trapezohedron, we know it can 

 come only from the Cube, the regular Octahedron, or the rhombic 

 Dodecahedron. If a pentagonal dodecahedron, it must have for its 



dron. The different varieties of octahedron are also distinguishable from 

 each other, by the angles at which their several planes respectively meet, 

 as well as by the S3'mmetry of their modifications. (. 50.) 



The different species of parallelopipeds, are the most difficult forms 

 among crystals to be identified. With regard to a crystal of this denom- 

 ination, we should first observe whether it possesses a series of planes 

 whose edges are parallel to each other. If we observe such a serie?, 

 we should then hold the crystal so as to bring the parallel edges into a 

 vertical direction. When thus situated, we should observe whether 

 there be any plane at right angles to the vertical planes. 



It sometimes happens that agreeably to the foregoing directions, the 

 crystal will possess two sets of vertical planes, according as one or the 

 other of two sets of parallel edges are placed upright. In such a case, 

 we should endeavor to ascertain whether the planes belonging to one set 

 are not so symmetrically arranged with respect to those of the other as 

 to possess the character of modifications of the terminal edges of the pre- 

 dominant form ; if this should be the case, we should not make them the 

 vertical planes. 



If there be a series of vertical planes, and a horizontal plane, we should 

 observe whether any of the vertical planes are at'right angles to each 

 other, and whether there be any oblique planes lying between some of 

 the vertical planes and the horizontal plane. We should remark the 

 equality or inequality of the angle at which any of the vertical or ob^ 

 lique planes incline on the several adjacent planes. 



We will suppose a parallelepiped with oblique planes, situated be- 

 tween the vertical and horizontal planes: if the inclination of the former 

 planes is uniform to the adjacent vertical and horizontal ones, it belongs 

 to the rigbt square Prism, provided the vertical planes are perpendicular 



7* 



