On Crystallography. \0\ 



of view that Llnnceus was led astray in his methodical dis- 

 tribution of crystals. Roine de I'lsle, by employing it with 

 more art and justice, avoided the breaches of natural con- 

 nections which disfigure the system of the Swedish natu- 

 ralist. But there was something arbitrary in the choice of 

 the forms which De I'lsle regarded as primitive, by consult- 

 ing only the external aspect of crystals, without regarding 

 their structure. Bergman, who had so successfullv seized 

 the nucleus of carbonated lime, by mechanically dividing 

 the metastatic crystal, was not equally fortunate with respect 

 to several other crystals, and in particular with respect to 

 the dodecahedron variety of the same substance which vv'as 

 then called tele de clou (noil head). It would result from 

 the explanation which he gives of the structure of this cry- 

 stal, that its nucleus should have angles totally different 

 from those of the true ; and Bergman has even been obliged 

 to suppose that the planes which he calls fundamental were 

 truncated in the present case, which presented a new excep- 

 tion to the principle on which his system was founded*. 



The mechanical division of minerals, which is the only 

 method of ascertaining their true primitive form, proves 

 that this form is invariable while we operate upon the same 

 substance, however diversified or dissimilar the forms of the 

 crystals belonging to this substance may be. Two or three 

 examples will serve to place this truth in its proper light. 



Take a regular hexahedral prism of carbonated lime (PI. T. 

 figs. 1 and 2). If you try to divde it parallel to the edges 

 which form the contours of the bases, you will find that 

 three of these edees taken alternately in the upper part, for 

 instance, the edges If, cd, b m, may be referred to this 

 division : and in order to succeed in the same way with re- 

 spect to the inferior base, we must choose, not the edges 

 I' f\ (f d', U m , which correspond with the preceding, but 

 the intermediate edges d' f , h' c', V m' . 



The six sections will uncover an equal number of trape- 

 ziums. Three of the latter are represented upon fig. 2, viz. 

 the two which intercept the edges If, cd, and are designated 

 by pp 00, aa kh, and that which intercepts the lower edge 

 d'f, and vvhich is marked by the letters nn it. 



Each of these trapeziums will have a lustre and polish, 

 from which we may easily ascertain that it coincides with 

 one of the natural joints of which the prism is the assem- 

 blage. You may attempt in vain to divide the prism iu any 

 other direction. But if you continue the division parallel 



• See tlic remarks on this lientl in my Essay towards a Theory of the 

 Struciure of Crystals.— Sec Pljilosopljical Maj^azine, vol. i. p. 35 el seq. 



G 3 - tQ 



