191:5] on Great Advance in Crystallography (58 1 



the similar atoms of that element, while if it be a compound we have 

 a structure composed of atoms of as many kinds as there are chemical 

 elements present combined in the substance, and in the same relative 

 proportion as is expressed by the chemical formula of the substance. 

 In the case of a compound, moreover, the structure may also be 

 considered to be that of the molecules of the substance, for they or 

 a simple arrantjement of a small number (group) of them form the 

 grosser units of the structure, whilst the atoms are the ultimate units. 



Suppose we now represent this molecular or polymolecular grosser 

 structural unit by a point, and that such point be analogously situated 

 within each unit. The essence of crystal structure then is that these 

 points are so arranged in space that if they are joined by imaginary 

 lines the latter form a " space-lattice " (German, " Raumgitter "), each 

 unit cell of which may be conceived to be the " brick " already alluded 

 to, and the domicile of the chemical molecule or group of molecules 

 (indeed, it is immaterial whether the points are considered as placed 

 at the corners or in the centres of the cells) or, in the case of an 

 elementary substance, of a group of similar atoms. We may, there- 

 fore, define a crystal as follows : — 



" A crystal of any definite chemical substance consists of a homo- 

 geneous arrangement of grosser units of matter, each consisting of 

 one chemical molecule or a small group of molecules of the substance, 

 and the kind of arrangement is such that these grosser units are 

 all identically (sameways, parallelwise) orientated, and that their 

 analogously chosen representative points, one from each such grosser 

 unit, form a space-lattice (Raumgitter)." 



There are fourteen kinds of space-lattices, slides of several of 

 which are exhibited on the screen. Three possess full cubic symmetry, 

 two are tetragonal, four are endowed with rhombic symmetry, and 

 two are monoclinic ; while triclinic, trigonal, and hexagonal crystals 

 have each one space-lattice coi'responding to their type of symmetry. 

 In every case it is the full (holohedral) symmetry of the system which 

 is present, no space -lattice possessing merely the lower degree of 

 symmetry corresponding to one of the so-called hemihedral or 

 tetartohedral classes of the system in question. 



Now in the solid crystal, not only are the grosser units arranged 

 so that their representative points are repeated in space with extra- 

 ordinary accuracy of position, with production of unit cells or 

 " bricks " of al)Solutely identical dimensions throughout the crystal, 

 but the shapes of the grosser units themselves are identically similar 

 and identically similarly orientated in space. Suppose, however, that 

 the force of crystallization, the directive molecular force concerned in 

 bringing the molecules together in this regular order of marshalling. 

 is only adequate just to attain this marshalling of the grosser units 

 into a space-lattice formation, without being able to fix the units 

 aliout their own centres of gravity, a certain amount of wobbling 

 about the latter being still permitted. We might, in such circum- 



2 z 2 



