wherry: classes of crystals 483 



stance. When, on the other hand, external conditions prevent 

 equiUbrium from bemg attained, the symmetjy of the structure- 

 unit may conceivably find expression in the development of the 

 faces, and, as a matter of fact, a tetrahedral or hextetrahedral 

 habit is too frequently exhibited by diamond to be considered an 

 accident.^ Etch-figures bring out partial symmetries when they 

 are produced under conditions where attainment of equilibrium 

 is delayed, and in diamond they are actually hemihedral at 

 first, and become holohedral in the later stages of development. 

 The electric polarity, however, should be determined by the 

 structure as a whole, which holds in this case. Rotation of the 

 plane of polarized light, as it could not occur in a tetrahedral 

 substance in any case, has no bearing on the matter. The 

 features of diamond thus agree with its assignment to two dif- 

 ferent classes at the same time, and its crystallization may be 

 stated in the following manner: 



System, cubic; structure, holohedral; structure-unit, tetra- 

 hedral. 



In sylvite the habit and the absence of optical rotatory power 

 correspond to a holohedral structure, which is found to be pres- 

 ent by the X-ray examination. In this case electric polarity 

 could not occur, as no polar axes are present under any interpre- 

 tation. But on the basis of the gyrohedral symmetry shown by 

 the etch-figures this substance is commonly assigned to that 

 class. As pointed out by the Braggs a very slight distortion, 

 which might have sufficient influence on the growth of the crystal 

 or development of etch-figures to give rise to the gyrohedral sym- 

 metry observed, would not affect the X-ray spectra to a recog- 

 nizable extent ; the structure as a whole might then be essentially 

 holohedral, while the structure-units are not. Such a relation 

 was in fact predicted by Barlow and Pope*' some years before 

 the development of the X-ray methods. They pointed out that 

 if, in a close-packed cubic assemblage of atoms, alternate ones 

 are of slightly different sizes, gyrohedral symmetrj'- would neces- 



* This word is not to be taken too literally; it is merely a convenient de- 

 scriptive term for phenomena, the causes of which do not at a given time appear 

 worthy of extended investigation. 



« Trans. Chem. Soc. 91: 1179-1187. 1907. 



