678 Dr. A. E. H. Tutton [March 14, 



in the polarized light, which renders them more visible. Again, if 

 we crystalhze a substance from solution, say potassium bichromate, it 

 entirely depends on whether the degree of supersaturation is slight, 

 the " metastable " condition of Miers and Ostwald, or excessive, the 

 " labile " condition, as to what kind of crystal we obtain. From the 

 metastable solution perfect little single crystals are started into slow 

 growth, by the advent of germ-crystals of the same or an isomor- 

 phous suljstance from the air ; while the labile solution spontaneously 

 and rapidly crystallizes in the beautiful feathery forms illustrated on 

 the screen. Ammonium chloride (Fig. 1) and metallic silver afford us 

 even more beautiful screen pictures of arborescent crystallizations, and 

 nothing can exceed the beauty of snow crystals, an example of rapid 

 crystallization of water-vapour. For goniometry these labile forms 

 are useless, but, nevertheless, they teach us much concerning the 

 structure of crystals. For in them the skeleton, or inner framework 

 and plan of architecture of the crystal is revealed. 



It is hard to realize the clearly proved fact that our atmosphere 

 teems with excessively minute crystals — for they possess the complete 

 organization of a crystal — often not exceeding the one thousand 

 millionth of a milligramme in weight, and capable, whenever they 

 fall into a quiescent slightly super-saturated solution of a crystalline 

 substance of like structure, of calling forth its power of crystallizing. 

 In order to be able to exercise this remarkable power, however, the 

 germ ciystal must be isostructural in a very strict sense, if not 

 identical with the substance which is set crystallizing. Xot only 

 must its symmetry be similar, but the dimensions of its structural 

 units — the " bricks " of the crystal edifice, so to speak — must be all 

 but identical. The conditions are, indeed, similar to those required 

 for the facile formation of parallel growths of one crystallized sub- 

 stance on another, so admirably investigated by Barker. 



Perhaps the most striking cases for the purpose of illustration are 

 those of the rhombic alkaline sulphates, selenates, perchlorates, chro- 

 mates, or other of the well crystallized salts of the alkali metals potas- 

 sium, rubidium, and cresium, and of the base ammonium which is so 

 extraordinarily capable of replacing them. In any such group of salts 

 the periodic law of Newlands and Mendeleeff is most beautifully illus- 

 trated by the regular progression of all the properties of the crystals 

 of the three metallic salts, corresponding to the progression in the 

 atomic weights of the metals, the salt of rubidium, the metal' of 

 intermediate atomic weight, having invariably intermediate properties, 

 both morphological and optical. A slide showing the gradual slight 

 change in the prism angle of the crystals of the sulphates will make 

 the point clear, the slope of the prism face of rubidium sulphate 

 being intermediate between the greater slope of that of the potassium 

 salt and the lesser slope of that of the caesium salt. The variation of 

 the position of the optical ellipsoid in the three monoclinic double 

 sulphates of the 6H2O series containing these three salts is also illus- 

 trated by a slide in which the ellipsoid can be rotated through the three 



