ii2 SCIENCE PROGRESS 



further apart, but that the resulting totally different structure 

 happens by pure coincidence to be capable of exhibiting similar 

 planes arranged according to the same order of symmetry as the 

 metallic salts. The optical similarity of the rubidium and 

 ammonium salts, presently to be referred to, would appear, 

 however, to negative the idea of mere chance determining the 

 development of similar faces. We know so little yet as to 

 the actual character of the atom in the molecule, and its 

 state of rest or motion, that the writer has always regarded it as 

 a sphere of operations rather than as an immobile entity. Even 

 if the packing is as close as it is held to be by Barlow, it will 

 only probably be so as regards these spheres of operation or 

 motion, within which there may at any given moment be a 

 large amount of free space compared with that occupied by 

 the atomic matter. 



Before leaving this subject of the structure of crystals, 

 however, it may be remarked that the experimental evidence 

 of intermolecular and interatomic spaces brings the work of 

 the writer into close touch with the important work of Prof. 

 J. J. Thomson, whose discovery of the complex nature of 

 the atom, and its constitution out of corpuscles of negative 

 electricity (the so-called electrons, immensely smaller than the 

 atom itself), shows that we have now a third kind of inter- 

 space to consider — namely, the interelectronic. Moreover, the" 

 discovery of the corpuscular structure of the atom renders it 

 the more probable that atoms may take up definitely orientated 

 positions in the molecule, and consequently enhances the value 

 of the teaching of directional changes in topic parameters as 

 indicative of such positions. 



The results of the optical investigations may now be briefly 

 referred to. The optical properties of both series may be repre- 

 sented by an ellipsoid with three unequal rectangular axes. 

 We may employ either the ellipsoid of Fresnel, or that of 

 Fletcher termed the " indicatrix," the two being polar reci- 

 procals and exhibiting the same facts. The indicatrix is 

 preferable, because the relative lengths of its axes are imme- 

 diately expressed by the refractive indices of the crystal along 

 those directions. In the rhombic system the positions of the 

 axes of the ellipsoid are fixed, being identical with the directions 

 of the crystallographic axes. In the monoclinic series the 

 ellipsoid is free to rotate about the rectangular symmetry axis, 



