1902.] on the Constitution of Crystals. 147 



lithium light is now very much smaller than it was at the ordinary- 

 temperature ; for red hydrogen light it is still smaller, and for sodium 

 light the crossing has now actually arrived, while for thallium green, 

 hydrogen greenish-blue, and the blue light for which crossing occurred 

 at the ordinary temperature, the axes are separated at increasing angles 

 in the perpendicular plane. 



Our last experiment will illustrate the effect of change of tempe- 

 rature in the case of caesium selenate. The behaviour of this salt is 

 so similar to that of the long-known case of gypsum, that, as I can 

 obtain a very much larger crystal of this beautiful mineral, suitable 

 for projection purposes, I shall use it to demonstrate the phenomena. 

 You see at the ordinary temperature nothing whatever beyond the 

 fact that some light gets through to the screen. The optic axes are 

 at present separated in the horizontal plane to such an extent that 

 they are well outside the field. We are now warming the crystal, and 

 you see colour making its appearance at the sides; now the axes 

 themselves, surrounded by their coloured rings, are appearing. They 

 approach the centre, they now coalesce to form the uniaxial cross and 

 spectrum circles. They now separate in the vertical plane, and we 

 remove the source of heat, the axes still continuing to separate verti- 

 cally until the crystal and the metal frame in which it is being heated 

 have taken up the same temperature. Now the motion stops, and 

 the phenomena repeat themselves in the inverse order, once more 

 coming to a cross and circles, and again separating in the horizontal 

 plane, until finally they disappear on the margin of the field. 



These beautiful ca&es of crossed axial plane dispersion, you will 

 remember, are entirely due to the operation of the rule which we 

 found to govern the progress of the double refraction, in accordance 

 with the progress in the atomic weight of the alkali metal, so that 

 these very exceptional cases are in reality strong proofs of the main 

 generalisation derived from this work. 



It has thus been amply demonstrated to you that the chief result 

 to which these researches have led is, that the members of every 

 series of salts known to the chemist, which differ by containing 

 different elements of the same family group, exhibit perfectly regular 

 variations in their exterior morphology and in their interior physical 

 properties ; and that these variations follow the order of the differ- 

 ences between the atomic weights of those interchangeable elements. 

 Thus, it has been shown that the crystallographical properties of the 

 elements are in line with all their other properties, chemical and 

 physical, in exhibiting the same progressive character which is so 

 conveniently expressed by their atomic weights. 



We do not yet know why a particular series of salts chooses the 

 specific type of symmetry which is common to its members, but we 

 have reasonable ground for hope that further work in the direction 

 indicated, together with a successful development of the interesting 

 mathematical and geometrical researches now being conducted by 



h 2 



