no SCIENCE PROGRESS 



to y}r. The above rules governing the relations of the members 

 of the series are, if anything, more elegantly exhibited on this 

 assumption than if the parameters are merely calculated for the 

 directions of the three rectangular axes of the rhombic system, 

 but the facts shown on both suppositions are identical. The 

 monoclinic double salts do not exhibit any pseudo-hexagonal 

 tendencies, and the directions of the topic parameters are here 

 those of the three crystallographic axes. 



As regards the very interesting question of the internal 

 structure of the molecule, and of the orientation of the individual 

 points representing the various atoms, which are all averaged 

 into the single point representing the molecule, there is some 

 evidence already available. In the case of the simple sulphates 

 and selenates, when one alkali metal is replaced by another of 

 higher atomic weight, the separation of the molecular points 

 increases to a far greater extent along the vertical axis than 

 along the three pseudo-hexagonal horizontal axes, the smaller 

 increase along the ^ direction being very nearly the same as 

 that along the pair of axes %. On the other hand, when sulphur 

 is replaced by selenium, while there is also an increase of 

 separation in every direction, and the amounts along the three 

 lateral directions are again nearly equal, that along the vertical 

 axis is less than the latter. Thus, when the metal is changed 

 there is a preponderating change of dimensions of the molecule 

 in the direction of the vertical axis, and when the dominant 

 negative element, sulphur, is replaced by its family analogue, 

 selenium, the change is chiefly along the lateral axes. This 

 would appear to suggest that the two atoms of the alkali metal, 

 or their spheres of influence, are situated along the vertical axis, 

 while the sulphur atom or its sphere of influence occupies the 

 centre. The graphic formula, 



K 



I 

 o 



I 



o=s=o 



I 



o 



I 



K 



although only representing the molecule in one plane, will at 

 once show how likely this is. The actual arrangement in space 

 is probably that of a structure consisting of three regular point 



