OLDER PERIDOTITES OF SCOTLAND. 409 



silicates are among the first to be attacked by the solvent agents, and 

 the products formed by their decomposition are the first to be de- 

 posited within the negative crystals. 



To the same cause, the presence of water and other liquids under 

 pressure, we must assign the formation of a network of cavities 

 along fissures of the crystals of rock-forming minerals, these 

 cavities containing, in some cases, liquids with bubbles, and in 

 others solid substances which have crystallized within them. 



While these networks of cavities sometimes lie along actual 

 fissui'e-planes in the crystal, in many other cases they form bands 

 traversing the crystal where no actual rupture can be seen to have 

 taken place. These bands of cavities probably indicate portions of 

 the crystal which have been in a condition of intense strain, along 

 which, according to a well-known physical law, solvent agencies 

 operate with greater force than elsewhere. The same band of 

 enclosures (marking a plane of strain) is often found traversing a 

 number of adjoining crystals in a rock. 



The change by which certain felspars acquire their beautiful 

 play of colours is analogous to that by which the avanturine appear- 

 ance, or Schiller, is acquired ; but the former alteration appears to 

 be an ultramicroscopie one. It probably consists of the development 

 of thin plates of hydrous silica in a set of parallel planes within the 

 crystal. Like the Schiller structure, it is characteristic of minerals 

 which have formed parts of deep-seated rocks. 



The twin-lamellae found in most plagioclase felspars appear not 

 to be necessary and origiujl structures of the crystals, but to have 

 been developed in them by strain, like the similar twin-lamellae in 

 the rock-forming calcites. While some of these twin-lamellae are 

 probably produced by the stresses and strains set up during the 

 cooling of a crystal after its first formation, as was illustrated ex- 

 perimentally by Foerstner, others among them are clearly of long 

 subsequent date to the consolidation of the rock, and have been 

 developed by the mechanical forces which have affected the whole 

 rock-mass leading to the formation of cracks in the crystals which 

 compose it. 



By Schillerization the most striking mimicry of one mineral by 

 another may be produced. Thus the first stages of the Schillerization 

 of the monoclinic and rhombic pyroxenes are diallage and bronzite 

 respectively, minerals which have been constantly mistaken for one 

 another ; by a further change the same minerals may in turn pass 

 into pseudo-hypersthene and true hypersthene, minerals which 

 present the most striking similarity in their colour, lustre, and also 

 in their general aspect, when viewed in thin sections, and can only 

 be distinguished by their optical properties. 



All the minerals, whether in their normal form or in their Schil- 

 lerized condition, may be converted into their pseudomorphs ; and 

 this change is not always a molecular one only (paramorphism), but 

 is sometimes accompanied by hydration due to the action of water 

 penetrating from the surface or by other changes in their compo- 

 sition. Under such conditions augite is converted into hornblende, 



