ii 4 SCIENCE PROGRESS 



slightly elevated temperatures, the double salts are unsuitable for 

 such determinations, and the rapid deliquescence of the simple 

 selenates also excludes them. But the simple rhombic sulphates 

 have been found suitable, and adequately large blocks were 

 eventually obtained out of perfect crystals. The result was to 

 show that the co-efficients of cubical expansion exhibit a pro- 

 gression corresponding to that of the atomic weights of the three 

 alkali metals. The differences in the expansions of the three 

 salts are only one and a half per cent, of the total expansion, 

 yet this is five times as great as the possible error of the 

 interference dilatometer, so that the results are absolutely 

 conclusive. 



Conclusion 



The outcome of the writer's investigations has thus been to 

 demonstrate that the exterior geometrical, internal structural, 

 optical and thermal properties of the crystals of each of these 

 series of isomorphous salts exhibit progressive variations which 

 follow the order of progression of the atomic weights of the 

 alkali metals which the different salts contain, and that similar 

 variations attend the replacement of sulphur by selenium in 

 the acid radicle present. Further, the applicability of this rule 

 to such widely different series as the rhombic anhydrous sul- 

 phates and selenates, and the monoclinic double sulphates and 

 selenates containing six molecules of water of crystallisation, 

 indicates its probable validity as a general law of nature. 

 This law has been defined in one of the writer's memoirs in 

 the following terms : " The difference in the nature of the 

 elements of the same family group, which is manifested in 

 their regularly varying atomic weights, is also expressed in a 

 similarly regular variation of the characters of the crystals of 

 an isomorphous series of salts of which these elements are 

 the interchangeable constituents." 



This law definitely allocates specific angles, structure, and 

 physical properties to every member of isomorphous series, 

 thus removing the only supposed exceptions to the original 

 conception of Haiiy, that difference of chemical composition 

 involves difference of crystalline form. At the same time, it 

 explains the real meaning of Mitscherlich's principle of iso- 

 morphism, and indicates the limited and exact sense in which 

 it is true. 



