94 SCIENCE PROGRESS 



arranged in orderly succession below each other thus producing 

 eight vertical columns. Now it is the members of these vertical 

 columns, the family groups of very similar elements, which give 

 rise to the isomorphous series of Mitscherlich, for they are 

 capable of replacing each other in their compounds without 

 altering the symmetry of the crystals. 



It was at this stage, in the year 1891, that the writer com- 

 menced an organised attempt to bring some order into the 

 domain of chemical crystallography, by an attack on certain 

 series of isomorphous salts of definitely ascertained chemical 

 relationship. The results of five other crystallographical in- 

 vestigations had previously been published, the last of which, 

 concerning the effect of the addition of the CH 3 group in an 

 organic homologous series, suggested the mode of tackling 

 the greater problems to be solved. The four objects placed 

 in the forefront were : (1) to obtain definite information as to 

 the nature of the relationship between chemical composition and 

 crystalline form ; (2) to settle the question whether or not each 

 chemical substance has its own specific crystalline form or (if 

 polymorphous) forms ; (3) if the answer to (2) is in the 

 affirmative, and the members of isomorphous series produce 

 crystals which, while similar, are not identical, to ascertain the 

 law which governs their differences ; (4) to extend the investiga- 

 tion to the whole of the physical properties of the crystals, in 

 order to discover the relationships of the various salts with 

 respect to those properties, and to determine whether or not 

 they are in line with the structural relationships. 



It was decided that the series which offered the best chance 

 of obtaining definite results were those "containing the three 

 alkali metals — potassium, rubidium, and caesium — which not only 

 belong to the first of the vertical groups of family elements in 

 the periodic table, but to the same section (even series) of the 

 group. They are thus related in the most intimate manner 

 possible, and are, moreover, the most electro-positive of the 

 metals. Their atomic weights are separated by considerable 

 intervals, and the atomic weight of rubidium (84-9) is almost 

 exactly the mean between the atomic weights of potassium 

 (38*85) and caesium (131-9). It was to be expected, therefore, 

 that the crystallographical differences between the members of 

 isomorphous series would be here found at a maximum. It was 

 also decided to include the analogous salts of ammonium in the 



